By Stephen Browning (Created: 2008-04-28 15:32)
Introduction to the 'Big+Little' picture
The way in which electricity is to be supplied is subject to radical change. Distributed and Renewable Generation, together with Demand Management, is being promoted to reduce the use of central fossil fired plant, increase efficiency in delivery of energy and reduce emissions.
However, this will only be achieved if all resources are properly monitored and controlled within a new framework for electricity supply management. Any electricity supply system is always in instantaneous Power balance; the wires hold no storage and electricity moves at the speed of light across the system. This group will try and look at the isses with continuous matching of generation to demand, the need for accurate prediction and the fact that time and power demand are crucial. The future system comprising Central Generation (Big) and Distributed Resources (Little) needs to work as a disaggregated but co-ordinated unit to make major improvements.
Links:
By Stephen Browning (Created: 2007-11-28 11:48)
Each power system is always in balance; generation = demand. On an AC system, this rule is maintanied in real time by the frequency, whose deviation from nominal (50 or 60 Hz) represents the difference between 'required demand' and 'delivered demand'.
The frequency has to be kept within strict limits to avoid system degradation; usually +/- 1% for normal operation. If the frequency deviates beyond 2%, automatic disconnection and measures are necessary to arrest the slide and avoid collapse.
When frequency changes, Synchronous Generators will instantaneously release or absorb inertial energy and some (resistive) demand will reduce. Extra geneneration (and increasingly demand) is set to provide additional response and backup for same so that any event, inclusing the loss of the largest infeed, can be compensated without an excessive frequency deviation.
We can try to show the deviation limits in terms of maximum excursion from 'generation requirement = required demand'. This assumes that demand is 40% frequency sensitive. On a large power system only resistive load will react. Motors and other inductive loads are not frquency sensitive. This shows that the mismatch of generation delivered to that required has to be tightly controlled.
So, we need to be able to predict both demand and generation, and ensure that the match is kept within tolerance, for all timescales from immediate out to planning. Also, the transport system must be secure - under both steady state conditions and for any credible fault, both transmission and distribution must not be overloaded, have unacceptable voltage excursions or be unstable.
To do this, it is necessary to predict and model the network loadings and voltage/stability conditions in detail.
Links:
By Stephen Browning (Created: 2007-11-28 14:21)
The demand is continually changing, thus generation has to be scheduled and dispatched to track it, plus provide adequate response and spare, which can react in the appropriate timescales to cater for innacuracies in demand prediction or unexpected generation output. Here are some examples of different weekday demand patterns in Great Britain (GB).
The metering from all generation sources +/- interconnector flows will of course summate to the demand as the system is in balance. The system operator will maintain contiunous and integrated metering for the main plant, transmission system and interconnection flows. Total and nodal demand histories are derived from this and stored.
Demand prediction is normally carried out from analysis of total historical demand data against weather for cardinal point periods (time of day) in the relevant groups (day of the week, period of the year) with forecast base patterns and weather then applied to predict the future demand profile.
Operation of coventional main generation is of course under the plant operator's control, with output commited and and dispatched through system operation and market mechanisms. The instructed profile is compared with the demand prediction and plant ordering and dispatch adjusted to match across all lead timescales. This diagram shows the business elements of the 'unbundled' industry in Great Britain.
The predicted loading profile on transmission is also derived by application of nodal demand and generation data, derived from nodal history and and instructed generation output and applied to the grid technical data. The system is then analysed to ensure it will be secure - loading, voltage and stability in the steady state and after fault.
For demand prediction to be accurate, the metering must not be 'distorted' by omission of large amounts of embedded generation meters from the generation summation.
Passive distribution systems are designed and customer connections analysed to ensure the system will be secure at the peak and trough conditions in each year. Thus the passive system is always sized to meet the maximum demand on it.
Links:
By Stephen Browning (Created: 2007-11-28 17:16)
Most thermal fossil fired generation is designed to be most efficient at full load. Large coal and oil units are typically 36% efficient at max ouput dropping to 32% at half load. CCGTs can be 55% efficient at maximum, but only 40% when at half load.
Max ramp rates are around 10MW/minute and other dynamic restrictions apply when operating thermal units - minimum stable generation, minimum run and shutdown times. In addition, each unit requires a miminum notice to synchronise and will consume start up heat to bring it on load. Both of these increase with the time the generator has been shut down.
The variations in the daily demand curve dictate that a number of generators start up for the plateau and peak periods of the day. Some demand rises are so fast (up to 3000MW/hhr in GB) that a number of units will be ramping simultaneously. At all times, some units are also part-loaded for response, reserve and spare duty, to cover unexpected demand or generation changes. Units have to be ordered far enough in advance that they will synchronise at the correct time
It is vital that the demand curve is accurately predicted and generation is reliably operated to avoid unneccessary part loading, allocation of excess reserve or ordering of generators that aren't actually needed in the event. Prediction, reliability and timing are the key to efficient operation.
The conventional power plant is designed to be controllable for instruction following. Thus, its output is predictable for the purpose of Generation-Demand matching. Even so, allowances have to be made to cover the risk of plant breakdown; response, reserve and spare output is carried to cover the anticipated level of generation shortfall and failure as against the instructed output.
Links:
By Stephen Browning (Created: 2007-12-10 11:59)
Renewable Generation replaces fossil fuel burn and consequent emissions. Distributed generation is more efficient at providing electricity near the point of consumption and multi-energy generation systems (heat, cooling, power) can provide that energy more efficiently than conventional methods, although still using fossil fuel.
The latest Distributed Generation at premises level comprises micro wind, photovoltaic and combined heat and power (sometimes with cooling) installations. Separate large wind generation is accomodated at higher distribution voltages although with careful rules for operation if the system becomes stressed.
The problem with any renewable generation is predictability and the fact that there is gross variation from day to day. Both irradiance (for PV) and wind speed are difficult to estimate at the lead times relevant to commiting main generation.
A quick summary of Generation types, 'drivers' and predictability:
Here is an example of wind predictability and the effect on the residual requirement for main generation in Great Britain. It shows the Day Ahead and 0500 forecasts, then the last forecast and actual trajectory for each time for which forecasts are prepared and actuals are recorded.
And here is an example of 3 days of micro PV outputs superimposed on an 'average' domestic load curve.
In general, CHP and CCHP systems are driven by the heating or cooling requirements of the premises, or the process for which the thermal energy is required. Temperature fluctuates less frequently than other weather variables and plant output is more in tune with demand, which increases with low and high temperatures.
Also, assuming the CHP runs continuously when the weather is cold, the domestic premises profile might appear as follows:
Note that both the PV and CHP systems may tend to export at times of low premises demand.
Within the current operational framework, it is assumed that more spare and reserve output on conventional generation will need to be carried to meet the increased level of uncertainty introduced by renewable and other distributed generation.
Links:
By Stephen Browning (Created: 2008-01-14 16:01)
Conventional distribution system management is based on supplying demand to customers on a discrete network connected to a transmission grid supply point. Some conventional, observable system instruction following generation is also accomodated at the higher distribution voltages, able to regulate active and reactive power export (and reactive import) to meet system matching and transmission and distribution security requirements. Such generation is carefully controlled to avoid Power Quality issues at adjacent customer premises.
Design of the network is carried out by simple analysis of maximum and minimum demand - Max Gen and Min Gen - Min demand conditions to determine system capcity and quality. Because the generation is controllable, output can be intertripped or limited if necessary at low demand periods to avoid the need for major reinforcements to accomodate excess export at such times. This simple analysis will cover all expected loading conditions with supply transformer tap changing and generator control maintaining a valid voltage profile.
The loading pattern is predominantly a power flow from the grid supply point, decreasing by distance from that supply point with the voltage profile behaving in a similar manner. On feeders with generation, control is exercised to ensure security and quality is maintained. This design method means connections are geared to maximum demand conditions without any provision at the lower (domestic) levels for customer action to reduce the Peak leadings which only occur a few times a year. As a result, the systems are heavily sized, which does increase the customer connection charge; a large proportion of the final delivered cost of electricity.
Links:
By Stephen Browning (Created: 2007-12-17 14:38)
What we would expect to see under the current development framework is an increased penetration of smaller 'fuel or customer requirement' driven, unobservable, distributed generation of different types at different levels.
At domestic level we have Micro CHP, PV or Wind technology. PV is expensive, due to local turbulence Wind gives low yield at roof height and CHP systems (boiler + heat recovery turbines) are stilll being commercialised. However, the penetration of Micro Generation is forecast to increase.
At commercial/industrial level we have an effective market for CHP and CCHP, albeit mainly based on fossil fuels. Use of PV and some mini wind is also being applied.
Larger stand alone Wind Generation parks are separately connected to Distribution feeders.
Uncertainity of Wind output has already been shown previously. As a further example, appllication of say 3m domestic CHP units in Great Britain, all working to heat requirement would have the following impact on a winter's day.
With a cold, uniform external temperature the CHP will run continuously day and night. This is not the most efficient way to reduce fossil generation output.
Commercial CHP only runs in the daytime period and should produce a better impact profile.
Distributed domestic Photovoltaic systems will produce maximum output during summer daylight hours while domestic premises demand is not at its maximum. This will cause the premises to export.
There have already been cases of resulting local high voltage causing the inverter to trip. As regards the National position, domestic PV output can contribute to reducing higher load levels but leaves an everning Peak. You need a lot of capacity to make a significant impact; 3 million 1.1 kWP panels against the Great Britain demand in this example.
On commercial premises, maximum PV output occurs during the building maximum demand period and is synergistic with any electrical cooling load.
The overall impact on the main system of large DG penetration would mean that generation output would have to be made observable, albeit aggregated over suitable groups; say by supply point and then by defined transmission area and Nationally. Local and aggregate prediction mechanisms will be required.
At the same time the customer 'attitude' to demand is changing. Energy use reduction and the development of energy efficient premises and processes is being progressed. Also non-time critical Electricity demand is being identified and aplliance operation co-ordinated for use as afficient short term reserve.
To get true electricity efficiency, the need to recognise the inefficiency introduced by large demand variations over time and the need for accurate prediction and operation is crucial.
Links:
By Stephen Browning (Created: 2008-01-15 18:49)
Although the overall distribution energy supplied at a grid supply point will decrease with distributed generation, the supply point and the individual feeders will experience variable power flow patterns depending on the amount, type and distribution and location of generation connected.
Renewable generation output will of course vary depending on weather (irradiance, wind speed,) while CHP and CCHP will run at a constant output depending on the heat requirement.
The result will be variations in flow patterns by weather, time of day, day of the week and time of the year, all of which will be hard to predict. Such variations will need to be managed to avoid voltage and stability excursions on the distribution system.
Customers are also actively trying to reduce their energy demand. In addition, Non-time-critical demand is being identified and proposed for use as a short term reserve.
In modern low energy and passive housing, the residual electrical demand will be cooking, lighting and entertainment plus the small ventilation system and heating load; demand will probably peak during darkness. At this level, premises CHP is inappropriate (low heating load), although communal heating/cooling CHP may be appropriate. Distributed generation for such premises will probably comprise PV or Microwind.
As regards Power Quality, more modern devices such as compact fluorescent lights and switched mode power supplies in electronic and entertainment equipment are introducing increased levels of harmonic 'pollution' at distribution level. The demand in low energy houses will comprise a higher percentage of such devices. DC-AC micro-generation inverters also introduce harmonic distortion into the supply.
Customers need to be made more aware of the impact of their demand and embedded generation at different times. Instantaneous delivery of electrical power matched to demand, not just energy over time, has to be securely managed to avoid interruption, overloading of circuits, voltage excursions and inefficient and unnecessary running of a main fossil fired plant. The resulting need for accurate and separate forecasting of generation and demand at this level needs to be made clear.
So, to maintain a secure active distribution network with its changing flow patterns, it is necessary to monitor embedded generation, demand and remote line flows and voltage levels to a greater extent than with a pure passive system. Levels of control also need to be exercised to maintain delivery and quality within prescribed limits. The correct level of control can also reduce peak flows and thus allow more efficient Network design without unnecessary excess capacity. This in turn leads to a cheaper but weaker and thus more volatile network, where both generation and demand need careful control and active power quality conditioning may need to be applied. The use of storage to buffer fluctuations may also be beneficial as an alternative to more capacity.
This leads to the conclusion that all 'Distributed Electricity Resources' (DER) on an active system, generation, demand and storage must be monitored and the appropriate level of control by 'trading' applied to ensure secure operation. The operator of the passive distribution network has to become more active - a distribution system operator.
Links:
By Stephen Browning (Created: 2008-01-16 19:43)
DDistributed resources need careful handling for Distributed Network Security and to ensure that fossil fuel burn is reduced most efficiently. Clever management of 'Distributed Electricty Resources' (DER) is the key. The new Active Network now carries sometimes unpredictable generation and some controllable demand. There is also the possibility that storage could be applied to facilitate flow management on the distribution system and avoid fluctuations and also reduce the system import at times of high demand when the most inefficient fossil fired plant has to start up and run.
We need to persuade customers that the 'Fit and Forget' approach to distributed generation isnt going to achieve the best results. At the same time we must remember that the customer's main activity is getting on with life (domestic), carrying out business (commercial) and manufacturing (industrial). They do not want to dedicate time or expensive resources to good power profile management; the process has to be automatic.
Let us look at the Customer Demand and Distributed Generation profiles in more detail
The domestic customer has a basic refidgeration demand, a smooth lighting and entertainment load which peaks morning and evening then a large but highly erratic heating appliance demand (e.g cooking, hair dryers) which puts large spikes onto the profile. A large laundry equipment heating load will appear when the machines are operated. Note that domestic distribution connections are rated at least 12kW. Although this historially would be to accomodate some direct heating load, coincident heavy cooking demand with other demands peaking still needs to be catered for. In addition, Eco house designs can include instantaneous electric water heating. This will cause new demand spikes at time of general peak demand as against tanked hot water storage systems using gas or off peak electricity as the energy source.
If the domestic customer adds some renewable generation, we would expect to see an 'erratic' generation pattern overlay for Wind (turbulence effect at low levels) and a more consistent generation pattern for PV, depending on cloud movements across the sun. This would probably lead to overall daytime export and morning/evening import. CHP systems would generate in blocks dependent on the outside temperature; however such technology is not appropriate for high efficiency houses with a low thermal and colling demands are supplied by heat recovery, heat pumps and solar thermal panels, plus heat stores. Overall there is a considerable level of 'unpredictability' at individual domestic premises level, both generation and demand, which limits the potential benefit of control.
Moving up to commercial level, assuming some heating load will be met by larger scale CCHP (20kWe)and with a day-night temperature variation on the building, we could get the following profile shape in Winter.
The demand is less erratic for a large commercial building but shows a large day-hight variation. The CHP would however cut in before and cut out after main occupancy times. On a premises basis predictability is better than domestic. Generation varies with temperature while demand shows a higher 'basic' level plus some light and temperature based variations. Again generation and demand need to be monitored separately to ensure records of each are accurate and some level of control could be applied.
For a commercial building with a large PV array we might get this profile in summer
The residual site import is reduced in the morning but comes back up in the afternoon before work finishes.
At industrial level, large CHP is geared to providing heat and electricity for major processes. The generation will normally operate when the process demand is applied. The sizing of such CHP will normally be limited so as not to exceed the heat or electrical demands to avoid unprofitable export under simple tariffs or production of unnecessary heat. The operation of the plant and the demand should be predictable against manufacturing process operation timetables.
The more predictable and controllable Generation and Demand is, the more scope there is for control to assist with system management by operating outside normal premises requirement. At individual domestic level where there seems limited scope for control, some 'non time critical' demand (e.g. Laundry) can be usefully set to operate at approprate times (low National Demand). Commercial and Industrial locations may be more suited to premises level control.
Links:
By Stephen Browning (Created: 2008-01-24 15:15)
The main issue with DER management will be monitoring, trading and control at all levels. Let us look at the overall objective again:
From the point of view of the market and the operator, there is a need to monitor by location and time what the demand is expected to be and what generation outputs are programmed, together with data on the ability to instruct changes to generation and demand power profile, with energy and notice restrictions as appropriate, plus reserve capability so that timely instructions can be made to ensure demand and generation match with adequate reserve and spare to cover the error margins.
All this with a framework of continually changing demand as in these examples of different Great Britain weekday profiles.
The objective is to both reduce and smooth the power output of fossil fired generation while making the residual requirement for such plant predictable. This will not only reduce the energy requirement, but also, when such a plant is required, ensure it runs at peak efficiency to avoid unnecessary fuel burn and emissions.
As we get more generation at distribution level and variable DER that can participate, a system matching a two-way communication system is required to monitor and also trade where feasible.
At DER level, RES generation normally operates at full achievable (albeit variable) output, except where distribution or transmission security and quality limitations apply. To do otherwise for system matching purposes is inappropriate as we are simply reducing 'free' output, which has zero pollution/emission effects.
It is appropriate to vary CCHP unit output for system matching, but the degree of action may be limited by the associated heat or cooling requirement.
Electrical storage can be employed to smooth out excursions in the import or export profile to assist system matching and where located appropriately, to avoid overloading or assist with maintaining voltage levels. However, this adds additional cost and some energy loss. For CCHP, heat stores can also be used to permit variation of plant electrical output and can be very efficient.
Domestic premises loads and RES generation, with inherrently fluctuating profiles may not be suitable for major particpation in power profile management, except for large time variable demand such as laundry.
Businesses and community CCHP systems are suitable for electricity or heat storage, to benefit the customer and the system. Distributed Generation must disconnect from the distribution system if supply is lost. Therefore, electricity storage at premises level can also be configured to provide UPS support and allow the premises generation to keep running.
So, with large DG penetration, we have the need to monitor and be able to exercise control, where available, over a large range of premises and devices below each supply point. We have premises with demand, generation and/or storage, individual generation sites (e.g. wind farms) and possibly system connected storage and power conditioning. This combination of premises, individual generating plants and devices, forms a microgrid.
A lot of individual data is required for distribution system (microgrid) security management, and the aggregated information by supply point is then required by the market and the system operator for demand-generation matching and to maintain transmission integrity.
From the customer's perspective, there needs to be a considerable change in their relationship with the electricity supply business to achieve a tariff benefit from DER control.
Links:
By Stephen Browning (Created: 2008-01-24 16:48)
The current relationship between the customer and the electricity supply buisness looks as follows. This diagram is based on the unbundled electricity supply structure in Great Britain (GB). With the exception of balance trading and settlement, all the business elements will be present in any electricity supply structure. This is the case even in counttries with full or partial vertical integration.
In Great Britain, each supplier party is responsible for trading by the half hour to ensure that a viable profile of generation is purchased to meet forecast demand. The resultant power profiles (by generating unit and supplier-demand grouping) are submitted to the system operator so that the overall match can be checked and adjusted and transmission security maintained. At 1.5 hours ahead, the operator takes over, making specific instruction to individual generation units.
Each customer deals with their supplier, who has to charge for both energy and use of system. Normally, this is on a per unit consumed basis by period billing, even though the system charges relate to capacity and maximum demand. Distribution security is maintained on the 'passive' model; the customer only contacts the distributor in cases of supply failure.
Demand falls into one of three types:-
The latter of these three is obviously being tackled vigorously as public awareness of energy costs rises. Use of power efficient light bulbs and recognition of the fact that empty rooms and inaminate objects are not frightened of the dark is being recognised; a change from the acceptance of 'passive energy waste' we have grown up with. However, more automatic systems may be necessary as 'manual' operation can be tedious and the requirement tends to be forgotton over time.
Tackling the non-time critical element is important to improve the operating profile of residual fossill fired plant. However, this has to be done in a predictable manner.
Under a new model, some customers are active participants in a short-term market system.
The classic commercial interface between the customer and the system has always been limited by the capability of the metering and logistics of obtaining meter readings. Historically, a simple electro-mechanical integrating energy meter was the only practical option. This was read at set intervals and the energy consumed charged at a pre-set tariff. A separate standing charge was levied to cover connection and use of system charges. Where electrical storage heating was appropriate, a second register and a simple clock switch was added to allow this load to be energised overnight at a lower tariff rate. Larger premises could justify some more sophisticated metering with such facilities as maximum power demand tariffs and alarms.
Modern data acquistion and storage technology, together with cheap communications, can make the customer to utility interface more dynamic. This has the potential to allow a wider range of premises to have demand and energy use monitored more frequently and to enable DER to have a more active role in generation to demand matching. However, suitable commercial mechanisms need to be developed to enable this effectively.
A variable tariff model that reacts to real time and short term predicted system conditions is one possibility, with price signals generated from the operator and the market respectively. However, this subjects the customer to uncertainty as regards future energy costs and makes budgeting difficult. Premises with generation will have justified the installation against a forward analysis of energy rates. For a large installation, the owner (industrial or commercial) will have secured a power purchase agreement to fix the value of the energy in their project plan.
Setting the price signals correctly is a tricky business. The objective here is to correct generation-demand mismatch and remove expensive and inefficient operation by main plant; smoothing and peak reduction. Marginal pricing mechanisms can show large swings and the apllication of raw data could give excessive inappropriate changes to power profile. Average prices will give the wrong message and may cause adverse behaviour to that required. The prices need to be set so that the customers deliver the level of power change required. Time staggering the application of price changes by customer groups (generic) can also give more precise results. Geographic control is also of course required to maintain transmission and distribution security under this model. The issue of differential treatment of customers, especially as regards charging due to transmssion/distribution congestion has to be carefully managed.
The second method is to enable trades in 'variations' from the expected power profile using incremental and decremental offers; this is similar to the way in which the Great Britain operator matching mechanism works. It should be a more accurate way of adjusting the generation-demand match. However, variation trading requires a pre-declaration of expected power profile with prices for increasing or reducing import-export. If changes to profile are instructed, the final premises metering needs to be compared with the declared profile. This facilitates calculation of energy charging/payment for the instructed difference and any penalties for non-delivery. Again, to be effective this process needs to be carried out in market and operator timescales; with small premises it is difficult to predict the power profile.
It is possible to consider other variants, such as capped/collared tariff pricing to reduce the level of price variations the customer sees. Also, as was made clear earlier, the process needs to be automatic; the customer does not want to be actively involed in power management on a continuous basis, unless there is a potential financial consequence, (say a short-term high price,) which could be avoided by simple manual action.
Links:
By Stephen Browning (Created: 2008-01-28 12:13)
Proper data communication between the active customer and the industry is vital for efficient operation of the system, to reduce and smooth out the operation of the remaining fossil-fired plant while maintaining adequate security of transmission and distribution systems.
The industry requires data for the following processes.
Market and System Operator - data to ensure accurate matching of Generation with Demand with adequate reserve capability.
System Operator - data to ensure the Transmission system is secure and stable (steady state and after credible fault) and that delivered Power quality is adequate.
Distribution operator - data to ensure maintenance of end user power quality and security of supply
This all requires accurate forward predictions and metered actuals for Generation and Demand Power, by time and by location.
At the lower levels, Distribution group loadings with feeder and voltage data is required, together with predictions of projected import/export and possible changes to same by participation.
For matching and transmission security, location aggregated data is appropriate; again both the intended trajectory and capability to alter same are required.
For extensive distributed resources, effective aggregation is of paramount importance.
The distributor needs an accurate view of his system conditions but not the full detail of each individual premises contribution. Any control action will probably be automatic at the lower levels, to alter active resources import/exports (demand, generation, storage) and any system compensation equipment fitted.
The market and operator require multi-megawatt aggregated data for demand and generation and variable resources capability. The market requires this information aggregated by suppplier for forward bi-lateral trading. The operator requires totals by grid supply point and overall.
Both operator and market need predictions of timescale and persistance information on variable resource capability - the lead time to activate a change and the duration that can be sustained. Various pilot initiatives are already being carried out for done on provision of short notice short term Demand management and backup Generation use to provide ancilliary (reserve) services to the operator.
The aggregation of Distributed Energy Resources (DER) forms a Virtual Power Pool (VPP). This can comprise all active elements (Generation, Demand Management, Storage and Reactive control).
A VPP can offer network services to maintain stability, security and power quality at local level.
VPP aggregation forms multi-MW blocks .
Supplier aggregated blocks can be used for short term energy trading in the market to meet the half hourly energy requirements. (Commercial VPP)
Separate location aggregation is used to provide services to the system operator. (Technical VPP).
Blocks of dispatchable power are used to maintain system demand-generation matching
Blocks for ancilliary service provision (power/reaction time/duration capability) can be used to provide response to cover unexpected changes in the generation-demand match at near real-time.
There are currently a number of experiments being carried out with VPPs and as noted above there are separate initiatives and mechanisms for provision of ancilliary services by DER management. It is imperative that the overall interface framework and the data content requirements are clearly defined for each purpose. There are different requirements to support local secuirty, provide ancillary services, dispatchable power and marketable energy. It is important that a single set of data from the premises level can be configured to meet each requirements by clever filtering and aggregation.
Links:
By Stephen Browning (Created: 2008-03-13 12:24)
Now, we need to look at the data flow in the opposite direction.
This will mainly comprise instructions and signals to change the intended Import-Export profile of DER premises in response to the data offered to do same.
The important thing to remember is that because the system is always in balance, everything effects everything else!
The reverse route comprises a series of disseminators in parallel with each aggregator unit. When the distributor, the system operator, or a market supplier accepts an 'offer' from a block of DER resource, the block instruction has to be disseminated back to the original premises and then to the individual equipment that will make the necessary response.
In addition, the instruction needs to be accomodated by the other parties. For example, if the system operator instructs a DER block power change by location, the resulting action will impact each supplier party who has a contract with one or more of the component premises. In Great Britain, the resultant energy change needs to be aggregated by the supplier by half an hour to avoid distorting the supplier's contracted energy within the settlement process. There are already mechanisms in place to do this for dispatch and ancilliary service provision on large units. In addition, the power change may only be achievable if the distribution operator takes (albeit automatic) action to maintain network stability.
It is important to note the change in role for the distribution operator who now starts to be a (partly automatic) distribution system operator. The concept of control of active DER resources is being tried out in various locations, to permit larger amounts of distributed generation to be connected as long as local security can be maintained by intertripping or other active output management schemes if fault conditions occur. In Great Britain, these initiatives are being configured within specific areas know as local dispatch zones (LDZs.)
The combined control of variable demand, storage and generation and aggregation/dissemination of instructions is being tried out within the virtual power pool concept as part of smart networks research.
For an instructed change, which will comprise power and duration, this is all fairly simple to manage. However, the easiest way to cause DERs to respond is by simple tariff price switching as with the ENEL (Italian) Telegstore system and other simple distributed demand switching methods. The impact of a price change on a group power and voltage profile and supplier energy can be more difficult to guage.
Links:
By Stephen Browning (Created: 2008-03-12 09:44)
To look at a communication strategy from the bottom up, we need to start at premises level.
There is a considerable move to increase the level of communications within domestic premises for various uses.
At domestic level entertainment, computing and security are driving intiatives for both wired and wireless connections internally. Communication based on digital internet protocol (IP) is increasingly being adopted, apart from simpler analogue signals and commands. IP packet addressing obviously allows more flexibility in handling signals and data between different devices on a single network. There are a number of initiatives supporting home automation and the US HAN project (Home Area Network) is concerned with configuration and coordination of energy management systems. There are a large number of companies offering different solutions to a number of areas of home automation.
This gives rise to a number of different communication systems and it is imperative that standards for data traffic are developed. This will allow a single communication backbone and facilitate interoperability of communication and control units with peripheral sensors and controllers from various manufacturers. A single 'network' needs to be configured from the wired and wireless elements.
The data content standards for energy management need to be defined carefully. Segregated information on power flow for different demand, generation and storage appliances and control signals to same need to be managed to ensure simple analysis of current/predicted states and the ability to vary.
This all leads us to facilitation of intelligent control at premises level, within the overall framework of electricity supply and able to react to current and predicted generation-demand conditions. This can only really be handled by automatic systems and communication through to industry.
High demand (industrial) users can already enter into various schemes with their suppliers to reduce their tariff rates in exchange for participation in demand reduction, but these have tended to be simplistic in the past.
At domestic level, the capability to vary premises import-export power profile needs to be analysed by device type and ability to vary output or input to determine the capability for control. What we are looking at here is the ability to 'time-shift' demand and possibly generation.
Lighting is time-critical and cannot really have its operating time period altered at domestic level. Likewise, instantaneous water heating, and in the main, cooking and entertainment are also fixed. It is interesting to speculate whether on-demand entertainment might alter time usage patterns, but that is unlikely. On weekdays, only the evening period is normally available to people for relaxation.
Fridges and freezers can have their duty cycles delayed to give some short duration demand reduction shift. However, it has to be remembered that that reconnection will cause a larger overall demand increase as more units simultaeously operate rather than the normal time diversity that would be expected. Control of refrigeration load is only really practical for short term ancilliary services provision. The achievable reduction will of course depend on the appliance demand cycle which is in turn related to the temperature at its location.
Laundry is a non-time critical load and has been an ideal target for domestic demand shifting initiatives (as in Italy.) The start time can be delayed by time or price signal, but once started, it is not efficient to interrupt operation of the appliances.
In hot climates air conditioning and air cooling are the most important loads to consider for time shifting. The peak demand will occur just after sundown, (combined lighting/cooling,) although some tests with price-varying thermostats have set the high price for a four hour afternoon block. The result of this is a large reduction at the start of the time block, then a gradual decay in the demand reduction over time. At the end of the period, there will be an increase above expected demand as delayed cooling comes back on. This results in a less than optimal reduction at the peak time with a sharper residual peak.
Here is a possible example of the effect of fixed period priced reduction in Great Britain. This is based on the average domestic load shape and includes the increased demand effect at the end of the period.
It requires a large number of households of this type to have a large cumulative impact on Great Britain's demand (see below.)
The main contribution the domestic sector can make is to shift the use of high energy non-time critical devices, primarily laundry to the off peak periods, which is already forming the focus of early smart metering applications. The use of dynamic pricing by sector, supplier group or geographical area alllows more precise control of the demand to be time shifted (more below) as against simple timed techniques.
Small-scale renewable generation needs to be allowed to operate at maximum level; to curtail output is a waste of free energy and an extra control complication. However, as we saw earlier, high levels of generation in the low demand daytime period (especially PV) can cause voltage rise and the generation will trip as required by the distribution operator. Some intelligent compensation may be needed, either in terms of optional demand or intelligent voltage control. Storage at individual premises level may be appropriate, but again adds cost and control complications. Microgrid level equipment may be more appropriate.
Commercial premises have a steady daytime demand, mainly lighting and office equipment. The size and scale of larger commercial premises with renewable generation may make storage and intelligent control effective at this level. Intelligent control of lighting at the ends of the working day will also help alleviate local and national demand peaks at these times, caused by the cumulative effect of commercial and domestic demand.
Let us say we have a large commercial premises with CHP. The following graph shows the premises Import-Export profile, CHP output, and the modified remises I-O profile without and with smoothing (storage). The storage removes export and the peak spikes of the remaining import, which thus alleviates strain on the local distribution system which will allow it to ccomodate more customers.
However, against the Great Britain demand profile, the simple smoothing at local level does not improve the load shape. In fact, it actually increases the level of the demand rise for the peak itself, as against the unsmoothed condition.
This all goes to illustrate that dynamic control is necessary to improve the overall demand profile as each sector has a different influence on the load shape. As such, effcient external communication is important.
The industrial sector can control the production loading to some degree, depending on the nature of the manufacturing process. Some trials are already in place as regards short term interruption of heavy electric (induction) heating loads to provide operator ancilliary services. Large scale changes to the timing of production runs will need carefully managed communication to co-ordinate.
The most critical area is handling the information on premises consumption and tariff rates and making the owner aware of critical periods, without overburdening and causing disinterest. Automatic monitoring of appliance power, storage and genertion states allows estimation of what changes to the forward import/export profile are possible.
Dynamic pricing can improve the load shape further by grading the level of reduction over time. Also, applying price changes on an area by area basis over time will also avoid gross over-reactions. From the customer perspective, predictive price information in advance is also vital. When high prices are forecast, the customer systems can take anticipatory compensating action both before and after the high price period. This will avoid violent changes to the overall load shape across price switches and prevent too much decay in price related demand reduction over the period of application.
Links:
By Stephen Browning (Created: 2008-03-19 11:00)
The Power profiles for different DER sectors, with and without generation, need to be considered carefully.
As we mentioned earlier, domestic premises' level of electrical demand is highly erratic. The relative level of short term variability could increase in modern high efficiency houses, even though the overall energy comsumption will decrease. To this could be added generation of an 'erratic' nature such as wind (turbulence at low height) and PV under changing light conditions (fast moving broken cloud).
PV under a clear sky or constant cloud level will give a smoother, but variable profile, rising to a peak then dropping again as the sun traverses. CHP will give periods of constant output depending on the outside temperature and thermal cycling requirement. However, in more efficient homes CHP will not be needed; CHP schemes on some pilot, low energy domestic developments can be seriously underutilised. Lower heating requirements, a move to instantaneous electric water heating (new high errratic demand) and heat pumps obviate the need for fossil fired systems. Against this, note that heat pumps may not be viable in high density developments!
On working days, PV tends to operate before the preak demand and thermal (rather than synoptic) wind can drop when darkness falls. As we noted before and above, demand in high efficiency dwellings will be dominated by lighting, entertainment, cooking and water heating while occupied; thus the weekday peak demand always occurs in the evening.
3m panels opeating on a bright day in Great Britain (GB) will actually shift the peak time to the evening.
Commercial premises have a more steady load during the working day period, comprising lighting, water heating, cooling and office equipment plus some (relatively) minor cooking load. Again, application of renewable generation is subject to the same observations as above; wind will be erratic but PV output will synergise with the highest demand level. Commercial space (especially high rise) in dense urban areas will again not be suitable for heat pump installations and natural cooling, due to lack of open ground and density of occupation. Thus, CHP for both heating and to drive cooling may be appropriate, as illustrated in the previous article (13).
For electricity generation alone, PV is also being increasingly installed on modern commercial buildings. The resulting profile for the premises can look as follows.
and when scaled up on a GB basis, again the peak gets shifted to the evening
Industrial demand is highly 'bespoke' and driven by the requirements of the individual production preocesses. It is usually more controllable within time periods and notice limits. Some large demand can be interruptible at short notice while other processes can have their schedules adjusted with some notice, but are 'uninterruptible' when in progress. Where heat and electricity are used by processes, fossil-fired CHP has been found be efficient and cost reducing. Renewables will make some reduction.
We need to consider what level of control is appropriate at premises level. As we said before Demand falls into one of three types:
The latter of these three is obviously being tackled vigorously as public awareness of energy costs rises. Use of power efficient light bulbs and recognition of the fact that empty rooms and inanimate objects are not frightened of the dark (turn lights off in unccoupied areas) is being recognised; a change from the acceptance of 'passive energy waste' we have grown up with.
However, manual actions are relatively time consuming and tend to be forgotten after a while; automatic monitoring and management is the key to ensuring continued reduction.
Tackling the non-time critical demand is more complex to handle; remember that predictability is vital - power, time and location. There are considerable gains to be made by smoothing and reducing the peak demands on - fossil fired generation. However, poorly controlled load movement can give rise to worse demand shapes as was experienced in the early days of fixed time off-peak domestic electrical heating. The remnants of this can still be seen as an artificial trough around midnight on the Great Britain (GB) Spring demand profile below. There are also examples of 'bad shaping' in the previous article (13) on intelligent buildings.
As we have already said, careful, ramped application of dynamic electricity prices (export and import) by time and group (supplier, geographical and/or sector) can influence premises Import/Export by changes to Demand, Generation and Storage. This should be able to effect compensation for unpredictable renewable plant output and produce a more efficient load profile for the remaining fossill plant which needs to run, rather than simple timer or advance time block pricing methods. Having said this it is important to recognise that forecasts of prices by time are important for effective control of customer DER resources.
However, any level of 'change' to the demand (and distributed generation/storage) profile gives rise to issues of predictability for the market and system/distribution operator. We will explore this in more detail later.
Links:
By Stephen Browning (Created: 2008-03-26 19:01)
As we have already stated, data traffic for distributed resource management will require the use of high speed aggregation and dissemination mechanisms between the customer and the commercial and operator sections of the industry.
Within premises we are seeing increased levels of data traffic and external interfaces - computing and entertainment in the domestic sector and business traffic in the commercial sector. The commercial sector also has buildings energy and facilities management systems while the industrial sector has large process management applications.
Digital audio/video, business and process management applications are all data intensive and are mainly comminucated by IP protocol packets. Power management data for future power systems is reasonably sparse and should not impose a great extra data burden at premises level, although some specialised equipment will be necessary. The main issue here is to define the data framework most applicable to each premises type and how that can be aggregated and disseminated at the higher levels.
Monitoring is important at device level for large premises demands, generation and storage with non time critical elements being managed directly. However, it is certainly not necessary to monitor every lamp bulb separately; presence and environmental sensors on a zone basis are already available to detect usage and control/override lighting, heating/cooling levels and appliance operation as appropriate. The individual demand for each large appliance and the other more general loads from zones should be monitored.
From the premises, simple data for demands, generation output and storage condition (kwH capacity and charge level) and any programmed activity are required. For controllable elements, timescales are required. Refrigeration can be interrupted for short periods at short notice while laundry loads can be timed in advance, but must normally run the cycle uninterrupted once started.
Renewable generation should not be interrupted except to maintain network stability but storage can to programmed at short or long notice.
The next level in the control sequence is the microgrid.
The premises controllers interface to a microgrid controller, which monitors import/export and exercises control over premises variable components. This system ensures real time and lead timescale secure, stable operation of the microgrid within power quality and any commercially applied limits. It also facilitates management of 'power variation' data from individual premises (generation, demand, storage) and instructions resulting from the acceptance of these offers by the market or operators (distribution and system). This offer/acceptance process again requires analysis of the microgrid integrity as a result of the instructions.
Premises data, comprising generation and demand power, storage power, energy and offers to change the same need to be aggregated in total for the microgrid (technical aggregation) and also by supplier (commercial aggregation) to support market activity. Any variation instructions will be on an aggregated basis for the microgrid and have to be disseminated back to the individual premises.
So we come to the operators and the market. The suppliers will use any offers to vary within forward market timescales and may operate trades to increase or decrease their total contracted energy in half hour blocks. The distribution operator systems will aggregate the data for the microgrids by supply point to give totals for the system operator. The system operator may use offers in the short term matching mechanism and for ancilliary service purposes.
All resulting instructions will be disseminated back by supply point, microgrid and then customer premises with operational instructions re-aggregated by supplier and market instructions aggregated by microgrid and supply point. This ensures supplier contracted energy is correct within settlement and that security, stability and power quality is maintained. In the case of ‘trigger’ instructions (ancilliary services activation or intertrip/restriction in case of fault), any execution of the associated action must be recorded for commercial and technical evaluation of the resultant power and energy change.
Links:
By Stephen Browning (Created: 2008-04-09 12:34)
We now need to explore the requirements for the content of different types of messages used between the customer premises and the industry.
We have already said that an IP protocol communications structure is appropriate for the transport of this data. Because of the number of businesses involved in the data chain and the suppliers of equipment to support it, a set of standards for the structure of messages needs to be defined. This not only affects equipment in the electricity management chain, but also those control systems, such as home automation or buildings management, within which energy management is incorporated.
The data structures for the main utility communications between supply and generation and the market and the operator are well defined. These carry data on energy (market timescale) and power/response profiles (operator timescale) with trades (market) and instructions (operator) to adjust same so that demand matches generation to an acceptable tolerance in real time. However, although the same principal data requirements also exist for communication with Distributed Energy Resources (DER), different data structures are appropriate. It will necessary for the aggregation and dissemination tools to handle any translations required.
The first thing to look at is the metering data streams and their uses on the Great Britain (GB) Power System.
The system operator uses continuous spot power metering of all the critical circuits. This covers all transmission circuits, supergrid supply transformers, main generators and interconnections. Because the power system is always in balance, summating the generation output gives the demand less that embedded generation which does not have operational metering. The operator uses the raw and calculated data in real time to monitor generation output versus instruction, also total and supply point demands with the latter used in on line system security analysis. Spot demand history is also used as a basis for demand shape prediction.
These meters give high accuracy half hour (hhr) energy and average power on the critical tariff circuits, which comprise the supergrid supply transformers, licensed main and transmission connected generators, interconnections, and low voltage circuits connecting each geographical distribution group. The operator uses this metering as a basis for his main demand databanks, by supply point and total, used for forecasting and to support off line security analysis. The tariff and settlement systems use the data to calculate generator energy delivery by generating company and total supplied energy within each distribution group. All this data is collected daily and held on a half hour basis. The generation company metering is compared with their contracted energy in the market plus or minus energy changes resulting from operator instructions and ancillary service delivery. The resulting difference (imbalance) between contracted+instructed and metered generation is charged or credited using prices derived from the market and system operator power matching actions
Half hour Import/Export metering is installed at all premises with maximum demand in excess of 100kw and smaller users who voluntarily request it. There are approximately 100000 such meters in GB from which data is collected approximately once per month.
he remaining 25 million GB meters are simple integrating energy units - mostly Import only although Export registers are included at premises with microgeneration. Readings are collected intermittently for billing purposes.
Half hour historic demand for every premises is calculated. Each non half hour metered site has a base profile set with a 24 hour curve for each day type (Summer/Winter/Spring/Autumn Weekday/Weekend) assigned, depending on the type of demand (domestic with or without overnight storage heating or commercial by load factor). For each half hour the individual profiled premises demands are summated within each distribution group and then ratioed to meet the total group demand from the high level meter summations less the total demand at all half hourly metered sites in the group. The resultant individual metered demands are then aggregated by supplier, by group and GB total, to give that supplier's total half hourly demand met. That demand is then compared with the supplier's contracted energy purchases for the half hour and the difference (imbalance) is charged or credited using prices derived from the market and system operator power matching actions. The supplier metered demands are adjusted as low level non half hour meters are read.
Note that the imbalance charges/credits are the incentive to ensure that the suppliers and generators contract energy accurately on a half hour basis. The market operates down to 1.5 hours ahead in GB with the operator only having sole control from real time up to that boundary.
There is fundamental issue as regards microgeneration in non half hour metered premises. The profiling mechanism assumes that all premises of one type will have an 'average' demand which matches the day type profile. However, microgeneration does not follow the same driver patterns as demand and it is impossible to produce a profile.
In addition, some incentive schemes for renewable generation, including Renewable Option Certificates in GB, need measurement of gross generated energy, not just net premises import/export.
Some Premises are participating in demand management activity, either within market timescales (scheduling of non time critical blocks of demand) or in operator timescales (short notice, short duration interruption of demand). These will also not follow the preset profile for the period of the demand management action.
A prerequisite to enable the execution of premises power management actions, except where triggered solely by frequency, is the establishment of a two way communication mechanism with the industry systems. As part of the intelligent premises control facilities, discrete metering of premises Distributed Energy (DER) elements; generation, demand and storage is required on a time period basis. This is needed to quantify and reward customer action, as instructed or requested by new tariff and trading mechanisms.
Although the number of meter points and the use of continuous metering appear extensive, it is appropriate for real time automatic monitoring of discrete resources and monitoring of control action responses to external signals. Indeed, home automation already makes more extensive use of sensors and controls than this metering system requires.
The enhanced metering, aggregated to allow simple half hour recording of DER elements, can then be signaled to external systems to give a more accurate record of the premises import/export, demand, storage and generation elements. This will assist with system operator, market and settlement processes and enable the microgrid control systems and the distribution operator to monitor states and capabilities of distributed resources connected to the lower voltage network, as flow patterns become more volatile.
Links:
By Stephen Browning (Created: 2008-04-16 12:42)
There are a number of commercial and technical mechanisms by which the variable elements of distributed energy resources (DER) can participate in the generation to demand matching process.
The timescales for participation are:
There are four basic mechanisms by which DER can participate.
Now let us look in more detail at the Great British market and operator functions.
In the market timescales, Supplier parties purchase energy and Generator parties sell energy, by account, for each half hour time slot. Some of these bi-lateral trades will be executed well in advance; further adjustment trades are made in shorter timescales as supplier account demand forecasts become more accurate.
Generators and Suppliers will process their traded positions into continuous power profiles for matching mechanism units. These represent individual generating units (gensets) and each supplier’s total demand within each distributor group. The power profiles are submitted to the operator, together with matching mechanism prices, limit data (availability) and dynamic data (ramp rates, minimum stable generation etc) for each unit
The operator initially uses this data to check that the synchronised gensets will be adequate to meet his evaluations of the requirements for demand, response/reserve and transmission security for each time period. He can execute limited scope unilateral trades to improve the position or use ancillary service contracts to place extra gensets on standby and later commit them to run.
Every half hour, market trading ceases for the third half hour ahead which is then added to the matching mechanism window. Within this window only the operator can instruct changes and only to the power profile and response settings of matching mechanism units, taking into account their limit, price and dynamic data. Within the window the limit and dynamic data can be revised by the owners as a result of a physical change (e.g. unit trips), but not the prices or the traded profile.
All trades, operator instructions and response/reserve delivery volumes are submitted to settlement. The sum of the energy of all contract actions (market trades, operator unit trades and unit response delivery) by party account by half hour is calculated and compared with the metered and meter derived energy for that half hour. The difference is party imbalance energy which is charged or credited at prices derived from operator instructions within and market trades for that half hour.
Let us look trading in Market timescales. Each Supplier party will build up a 'stack' of trades for each half hour during the period leading up to that time. Some of that trading will be in the form of seasonal contracts brokered months in advance.
The supplier party will build up trades over time, first for the block four hour period. He will then refine his position by half hour, as market closure approaches for his estimate of the demand for that period improves.
It is quite difficult for the supplier to make an accurate estimate of demand. Customers can change supplier after each agreed minimum contract period expires. This means that the supplier’s historic demand records are not a reliable basis for forecasting, unless he can maintain a stable customer base.
In practice the supplier may also sell energy back into the market if he has early contracts which summate to exceed his later estimates of demand.
The system operator will have received notification of the power profiles (PN in the next diagram) for each matching unit.
He has also received banded prices for increasing (Offer) and decreasing (Bid) power output/export or decreasing (Offer) and increasing (Bid) power import. These are shown as BOD +n ranges for output increase and BOD –n for output decrease in the next diagram.
Using his own estimation of demand and derived transmission flow limits the operator will determine where total generation does not match demand and the transmission system is not secure.
Using the price data and dispatching advice, he will then instruct closed blocks of changes to unit outputs from their power profiles (BOA in the next diagram) by ‘accepting’ the price offers or bids for such changes. The instructions will of course obey the submitted limit and dynamic data. MEL = Availability and SEL = Minimum Generation level in the next diagram.
To avoid the possibility of having to ‘reverse’ instructions, i.e. taking offers to increase overall output to cover a shortfall and then bids if total generation later exceeds demand, the duration of each instruction is kept small. Further instructions are used to extend the period of if necessary.
The operator will also instruct gensets and contracted customer sites to provide response and reserve ancillary services – primary response (instantaneous), secondary response (30 seconds) and five minute reserve. Anything beyond that timescale is normally covered by instruction.
Distribution operators will also have agreements in place, mainly for automatic action in the case of system disturbance. One standard example is that all small premises with generation must disconnect that plant from the mains in the case of supply failure, to avoid back energisation and bad power quality on the isolated section. On the other hand the ability of generation to ride through a fault, where the supply only suffers a transient disturbance, is important for larger distributed generation.
Links:
By Stephen Browning (Created: 2008-05-18 16:40)
Both the system and distribution operators need to forecast and monitor the transmission and distribution flows in terms of security and stability. They therefore need a reliable base for predicting demand and generation, both totals for matching and by location for security analysis. The system operator also needs adequate resources allocated to response and reserve duty. Both operators will also have plant allocated for emergency action, (e.g. intertripping), in parts of the system where this is required post fault to avoid breaching circuit thermal limits or local system voltage/stability limits. Introducing such generation control schemes has permitted extra plant, especially renewables, to be installed in parts of the system where they would not be accommodated under conventional (passive continuous operation) security criteria.
Active Distributed Energy Resources can thus be employed by the operators for matching, ancillary services and active local security management.
The matching mechanism is designed to allow submission of matching mechanism unit profiles and data to the system operator who can then make adjustments to those profiles as necessary. Demand units, which include all Distributed Energy Resources (DER), are by default configured by supplier within distribution groups. Individual large active customers, or aggregations of active groups of smaller customers, can be configured as separate demand units to enable participation. As DER premises are part of the system operator’s estimate of demand, only the variation of premises import-export should be modeled as part of the matching process.
All the major decisions which affect genset running profiles, principally commitment, are made in market timescales. Some operator intervention is permitted to ensure adequate plant margins and system security.
The matching mechanism takes banded prices called Bid-Offer data ranges (BOD). The range indexes have the same sense across all Generation and Demand units.
BOD positive ranges represent increasing genset output, increasing DER export or decreasing DER import as against the submitted unit profile. BOD negative ranges represent decreasing genset output, decreasing DER export or increasing DER import as against the submitted profile. Each range has a Bid Price, which represents the price the customer will pay for increasing his demand or reducing his export and an Offer price which the customer will receive for reducing his demand or increasing his export. The are shown as BOD +n ranges and BOD –n ranges on the next diagram.
The operator outturn sell and buy prices are also shown. These are the averages of Offers accepted (Buy) or Bids accepted (Sell) within the period, or the market closing marginal price where no offers or no bids have been made. Note that these are outturn prices and are not known in advance.
Lets say that in this case the customer has decided he can offer one range to decrease demand (BOD +1 Offer), the laundry demand for 4 hours from 1800. He therefore makes an Offer to increase demand for 4 hours from 1400 (BOD-1 Bid) which represents shifting the laundry run to the afternoon.
The operator can only ‘accepts’ offers and bids up to the third half hour ahead. If he accepts DER ranges then the customer’s profile and his use will start to change and the remaining future data will no longer be valid.
For example, let’s say the customer quotes a Bid price (pay to increase demand) on his BOD-1 range of 5p/kwh (£50/Mwh) between 1400 and 1700. If the operator has surplus output to sell across this period (starting at £30/MWh), then he will accept the bid in stages, say one half hour at a time. However when we get to 1700, the operator has less output to sell and the price rises; thus he will not take the Bid for that half hour. However the customer’s extra demand is contiguous (laundry) and cannot be interrupted…
Now let us look at the matching demand reduction over the period 1800 to 2300. Here the customer would be looking to reduce demand (BOD +2 Offer) and be paid at or above the £50/Mwh he paid earlier, say £55/Mwh. Up to 2100, the operator Buy price is above that level. But, for last two hours the operator buy price dips below that level and the offer will not be taken. The Bid and Offer acceptances are shown as BOA on the next diagram.
So the problem here is predictability, this time from the point of view of the customer. The matching mechanism is only really designed for short term adjustments to output on synchronised generation, not really for commitment for a period of running. As we said above such ‘commitment’ is really the province of the market although, as you will see from the previous article, attempting to trade DER in that arena also encounters problems.
So, once again let us look at using dynamic price messages to influence DER operation. In this case, however, we have an extra complication. The primary tariff is between the customer and his supplier. Thus, any contract to provide price related power change services to the operator must differentiate between the energy attributed to such operator initiated action and the import or export to be paid or credited at the normal premises tariff with the primary supplier. To make such distinction is not easy and the primary premises supplier may not wish to be involved in such DER activity. However, because the energy attributed to operator request is separated, administration of operator contracts with a number of premises can be managed by a unique agency holding a supplier licence. Also the primary supplier’s metered energy within settlement can be adjusted by any ‘operator action’ premises energy change to avoid consequential imbalance liabilities on that primary supplier.
Once again the price message has to be carefully constructed to give the desired result. The operator is working very much at the margin when instructing changes to plant profiles in the short term which translates to buying and selling energy. The buy and sell prices can be very different over the same period and also fluctuate significantly between periods. Also, at any one time, the marginal up and down price position will only be known to the operator for up to 3 half hours ahead, not as a complete profile over long periods.
It is again important to remember that the matching mechanism prices are energy only while the customer tariff consists of energy + system use charges. These elements again need to be kept segregated so that benefits and charges are correctly accounted for the customer, the operator, the ‘operator contract’ agency and the premises primary supplier!!!.
The classic model for priced DER participation is when the system is running on expensive plant. The operator will have high up (BOD +n) offer and high down (BOD -1) bid prices from conventional generation. If the operator has a selection of contract DER sites he can try and signal a price between the up offer and down bid to try and get a cheaper solution to achieving the match by influencing the customers to perform demand/import reductions or export increase. However, analysis of the likely customer reaction, the consequent change to the match position and the modified up and down prices is required. If the customers overreact, it could then cause the operator to reduce output on less expensive plant (more BOD-n Bids) than the DER price offered. The other point here is that the customer is being paid for reducing demand/import while also reducing his tariff payment to his primary supplier for the appropriate period. This all has to be carefully managed in respect of the overall tariff position.
The setting of the DER signal price is crucial to avoid excess, fast, demand reduction and then extra demand recovery at the end of the period. This can be handled by staggering the application of changing price signals across times and geographical groups. If this customer behaviour were to be reflected across say half the domestic premises in GB, then the result would be as follows.
Now, there is also a serious ethical regulatory, settlement and billing argument against such a practice of using priced DER management within matching mechanism timescales. Basically, the method violates the principle of the mechanism itself, namely that all power trading should be carried out within the mechanism framework, by unilateral instructions from the operator to the generator and supplier matching mechanism units. Although the operator can carry out ancillary service actions to secure the network and his ability to match demand and generation, he cannot use such facilities to carry out trading outside the mechanism. If he does so he is effectively bypassing the suppliers and trading direct with the customers, thus creating a dual tariff structure.
In a vertically integrated organization the problems do not arise in the same way. The Generator, Supplier and System Operator functions are combined, with a single primary tariff interface to the customer. Thus operator trading simply changes the primary customer tariff price for the period required.
These enable the Operator to send simple control signals to cause groups of customers to reduce or to increase demand/import, based on predetermined prices for the activity. The method can be employed by both system and distribution operators, to provide real time reserve and system security services. It is already configured to allow demand interruption at large participating premises.
Suppliers or supplier licensed agencies will act as the aggregators/disseminators to allow smaller sites to participate in this activity. Due to latency in the instruction process, for real time support this method can only be used to provide secondary response (30 sec to 5 minute) and reserve (5 minute delivery).
Anything beyond reserve requires dispatch instructions (BOA) within the matching mechanisms framework, subject to the rules stated above. However, the distribution operator (DSO) does not instruct plant within the matching mechanism. Therefore, when a DSO needs to exercise control over a prolonged period, some form of matching mechanism instrument is required.
The resultant energy from execution of ancillary control actions has to be calculated and included in the supply account contracted energy, for each supplier with participating sites. Such compensatory action is already carried out in respect of reserve delivery by generating units or large sites; the method simply needs to be incorporated in the aggregation/dissemination mechanism.
These involve instantaneous reaction to compensate for system problems, triggered by detection within the premises. As regards local stability, sites with distributed generation are already required to isolate the plant in the case of mains failure. As regards wider participation, it is possible for DER premises control systems to respond to frequency and voltage excursions. Contracts with the operator will normally relate to payments for providing the facility and setting it to operate. Energy delivery will again need to be evaluated and compensated to the relevant suppliers of sites where reaction has been executed.
Links:
By Hans De Keulenaer (Created: 2008-04-18 13:41)
When it comes to incorporating Distributed Energy Resource (DER) activity, the supplier will need to construct half hour energy blocks with prices, representing the aggregated ‘offers’ from, or contracts with, customers to vary premises export and import. These can then be compared with the sale and purchase of half-hourly energy blocks and prices being tendered in the market by other parties. Any trading with DER is simply between the supplier and its customers, to modify the supplier’s account demands and thus its purchasing requirements within the market. The trades do not enter the industry settlement process and customer actions will be reconciled dependent on the framework for participation.
All major decisions affecting genset running profiles, principally commitment, are made in market timescales. However, the market framework is the trading of energy by account by half hour. This is not precise enough to facilitate efficient merit order operation of main plant by minimising on load and startup fossil fuel burn against more accurate forecasts of total system and group demands. Thus, the market price messages may not promote the best use of DER to avoid unnecessary running of marginal plants.
The trading approach, comprising individual offers from the customer to increase or decrease import/export in market timescales, is difficult to handle for small premises. First, the system must know that a ‘tradeable’ demand, by definition non time-critical such as laundry, needs to run. This system next needs to schedule a running time period for the appliances, work out the half hourly energy demand and then calculate the worth of canceling the run. A single offer to represent the cancellation of the appliance run (reduction in demand) could be constructed, but putting an alternative bid to reschedule its operation cannot be time specific. All the customer really wants is to find the cheapest contiguous period to operate the appliance.
From the aggregation perspective, a large number of differently priced offers and bids, all covering different periods and each rather small, would be presented by such a mechanism. These would need to be aggregated by half hour and price band for the supplier to use. It is probable that the resulting trades from the supplier by half hour, when disseminated, would only request demand reduction for part the offer period for an individual customer site. This would indicate the appliance should be switched off and on over an elongated running period; not a practical option with laundry equipment.
In this example, the customer has planned to run laundry equipment (average demand 0.5kW), in two operations over 4 hours, from 1900 to 2300. He has offered to cancel the demand for an energy benefit of 5p/kWh (£50/MWh). The supplier will only accept this offer for the first and third half hours of the trade. Note that the customer benefit will be the saving at his normal tariff rate (say 5p/kwh) plus a payment from the supplier up to the offer price (2p/kwh).
What we also have to remember is that the market prices are energy only while the customer tariff consists of energy + system use charges. These elements need to be kept segregated for the customer-supplier interface to correctly identify the benefits. However, the opposite practice has sometimes been the case in the unbundled supply environment with standing charge elements to cover use of system being converted and added to the energy prices for simplicity. Thus, in the above example, the market energy price has been increased to reflect the DER price at the premises.
It could be considered that industrial premises with large scheduled process demands could construct offers and bids to ‘shift’ the process demand period. However, it is again likely that the aggregation/instruction/dissemination chain would again result in an impractical request for non-contiguous operation of the process plant.
The other problem with this method is proving that the customer site has delivered the instructed ‘trade’. In general customer demands are estimated as block aggregates. Unless the specific pre-trade demands for individual customer sites are submitted, they cannot then be compared with the final metering. This process is needed to determine the demand reduction which is to be credited at the specific trade price less customer normal tariff rate, as against the reduced metered demand charged at the customer’s normal tariff rate
The most practical methods for dynamic DER management involve the supplier signaling half-hourly prices in advance. Premises control systems can evaluate the tariff signals and determine when non time critical demand should run. Such systems should also warn the users of any critical high price periods so that ‘accidental’ unnecessary demand can be turned off.
Proving ‘delivery’ of customer Import-Export changes is of course not necessary with this method as the tariff simply applies to all Import or Export energy.
Once again, the market energy prices must be converted to the DER premises level by adding use of system elements.
In this case, the premises system has split the two laundry cycles over two periods, one in the afternoon and then a shorter and later run in the evening. It has also influenced the customer to reduce other demand around the peak time, (say, by delaying dinner) which is then recovered immediately afterwards.
If, say, this reaction was reflected over 2.5m households of a supplier with 5m households, the impact on the supplier’s energy would be as follows:
This shows that prices based on the raw market profile should not be used as dynamic tariff prices because they only reflect the marginal market position and the resultant customer reaction can be excessive. The consequential bilateral trading energy trading to cover such demand reaction would drive market prices in various directions to compensate.
The supplier will need to evaluate and forecast the likely impact of pricing in terms of the change to his half hourly account energy demands. Feedback from the premises control systems could also be employed for the purpose of analysing demand-price reaction. The resulting price profile issued would then be somewhere between the supplier’s marginal energy purchase rate and the customer energy tariff rate, with appropriate correction for use of system charges on the latter.
If the above action was reflected nationally, with say half of all households reacting in the same way, the result would look as follows:
As noted earlier, the issue of accurate demand recording is important to settlement and forecasting. Priced DER management obviously has a serious impact on demand history and conventional demand forecasting methods.
So, in summary, dynamic price setting should be carefully managed with regard to customer reaction. Staggered price changes can also be utilised with different groups of customers having different price profiles to avoid co-incident over-reaction.
Simple control signals can also be sent in advance to cause supplies to be reduced or to increase over a particular period, at a predetermined price or to avoid capacity charges. Such mechanisms for demand management have been used over many years to give large businesses advanced warnings of a peak occurring which will determine their chargeable maximum demand (use of system). Radio teleswitching has also been used to alter the operating time period for off-peak storage heating, on a day to day basis. This smooths the effect of the heating load switching on and off, against the different demand patterns encountered over the winter period and between weekdays and weekends.
Suppliers can also enter into other pre-arranged contracts with their customers, which reward the latter for changes to their Import-Export profile in response to signals sent in advance. As with the trading option above, this method has the issue of ‘proving’ delivery. The changed energy is credited at the special contract rate while the residual Import or Export is still charged or credited and the normal tariff rate for the premises.
These are not really applicable in market timescales, as they rely on independent action at the premises in response to abnormal conditions on the electrical supply (frequency, voltage, stability). These events can only be detected in real time; thus the management of automatic action falls within the operators’ domain.
Any market timescale changes to DER import/export need to be communicated back to the system and distribution operators, so that they can compensate their forecasts of generation, demand and system state within the matching and security processes.
Links: