The electricity supply

Transcription

1 Peacock Newborough 3/1/7 12:56 Page 74 Volume 3 - Number 4 - November 27 (74-78) Abstract The electricity supply industry in the UK will undergo substantial change over the next two decades principally caused by ageing infrastructure, increasing demand, the need to reduce CO 2 emissions and energy security. The revised network is likely to place a higher emphasis on the supply demand matching potential of load management and demand response programmes. Over the same time frame, it is likely that micro-chp systems based on engine and fuel cell technology will gain an increasing market share of the UK domestic boiler market. The potential exists therefore to develop micro-chp systems that will contribute to demand response. This paper explores the opportunities for modulating the load placed on the network Controlling Micro-CHP Systems to Modulate Electrical Load Profiles A.D. Peacock and M. Newborough Energy Academy, Heriot-Watt University, Edinburgh EH14 4AS, UK from groups of UK domestic dwellings via micro-chp system control. A control methodology is proposed which results in the daily load profiles placed on low voltage transformers to be smoothed and the commensurate load substantially reduced. This will reduce the requirement for part load operation of thermal generating plant and allow connection of additional renewable generation to Low Voltage distribution networks. The implication of this form of control on CO 2 emission savings is compared with current micro- CHP control methodologies based on heat demand in an individual dwelling. Keywords: Distribution networks, Electricity Supply Industry, Domestic sector, Micro-CHP 1. Introduction Much of the UK's existing electricity generation plant will reach the end of its useful life during the next two decades. Across this period, buildings-integrated micro-generation systems and distributed renewable power sources may be deployed to provide a significant proportion of the replacement generating capacity. Because of the intermittent supply characteristics of these types of generation, the net load profile placed on central thermal power plant will tend to become increasingly variable and so new supply/demand matching protocols will required. In this context, attention to the management of the demand side and in particular to the control of micro-generation embedded in LV networks may yield significant benefits to the electricity supply industry (ESI). Within the buildings integrated micro-generation subset, micro-chp (µchp) systems are being designed and developed to enter an existing marketplace, i.e. the domestic boiler market. Conventional wisdom dictates that the operation of µchp systems should be heat-led and at low penetrations, electricity generation will be almost incidental. The primary role of µchp will be to provide thermal comfort in individual dwellings. However, understanding the time-varying nature of electricity generation from µchp systems will become more important as the penetration levels within networks increase. Alternative control approaches to the heat-led strategy are feasible, especially if systems are packaged to include both a prime mover and a conventional boiler, i.e. a cogeneration and heating system. Systems of this configuration permit electricity production to be partially de-coupled in time phase from the thermal demand profiles of individual dwellings providing an additional degree of freedom with respect to their control. This work has taken as its springboard, an ESI perspective. The domestic sector is currently responsible for a large part of the variations in network demand seen on the national load profile for England and Wales. On a typical winter's day [1,2], the gradient changes on the total and domestic load profiles are nearly coincident, with the domestic sector contributing approximately 4.7 GW (38%) to the total rise in the morning and all of the 9.3GW rise associated with the evening peak. Gradient increases require fossil fuelled thermal power plant to be operating at part load in advance of a predicted time of change, in order to ensure sufficient generating capacity is always available to meet demand. For example, on January 25th 24, 23GW of coal plant and 18GW of CCGT plant was contracted to supply electricity to the grid [3]. At 5:am, prior to the morning gradient change, the electrical demand in England and Wales was 34GW and at this time 21% of the coal-fired and CCGT thermal plant was idling, 62% was part loaded and 17% was operating at rated load (Figure 1). The morning gradient rise ceased at 8:am with demand reaching a plateau of 47.9GW At this time 9% of coal-fired and CCGT thermal plant was idling, 36% was part loaded and 55% was operating at rated load. A similar pattern 74

2 Peacock Newborough 3/1/7 12:56 Page 75 A.D. Peacock and M. Newborough /ISESCO Science and Technology Vision - Volume 3, Number 4 (November 27) (74-78) : 1:3 3: 4:3 6: 7:3 9: 1:3 12: 13:3 15: 16:3 18: 19:3 21: 22:3 Thermal generating capacity (%) Demand (GW) Time of day (h) Figure 1: Operation of coal-fired and CCGT generating plant on January 25 th 24 in response to demand (England and Wales) of operation occurred during the afternoon plateau in anticipation of the evening gradient change. Idling and part load operation of coal and CCGT plants have negative implications for the carbon intensity of network electricity causing the carbon intensity in the period leading up to the morning gradient change to be higher than the average for the 24 hour period. Consequently, efforts to smooth load profiles in the domestic sector will generally have beneficial effects on the carbon intensity of network electricity. The aim of this investigation then is to consider the µchp approach from the perspective of a Low Voltage (LV) distribution transformer. In this context, for one envisaged method of control, this study aims to establish whether greater daily load factors (smoother load profiles) are achievable for LV networks containing high penetrations of µchp systems relative to (i) the base case of no µchp implementations and (ii) a conventional heat-led control strategy. The investigation is based on the operating performance and assumed constraints of present generation Stirling engine (SE) prime movers and it utilises a dataset of heat and electrical demand profiles for 5 dwellings. The work presented builds upon the authors previous work concerning micro-chp system simulation [4, 5, 6]; pertinent background material and a more comprehensive description of the modelling approach is contained in [4]. 2. Demand side The load seen at an LV transformer is described by considering the power demands of a 5 dwelling dataset. The dataset contains 1-minute heat and power demands and all performance parameters were computed on this time base. The aggregate demand is shown for a typical winter (January) day in Figure 2. The average heat and electrical requirements per dwelling on this day were 99.9kWh and 2.5kWh. The load factor of the electrical demand profile was 42.5%. The heat Thermal demand (kw) Electrical demand (kw) Figure 2: Aggregate heat and power demands from a group of 5 dwellings on a day in January 75

3 Peacock Newborough 3/1/7 12:56 Page 76 A.D. Peacock and M. Newborough /ISESCO Science and Technology Vision - Volume 3, Number 4 (November 27) (74-78) demand tends to lead the power demand, particularly in the morning as boilers commence operation prior to active occupation of the dwelling. 3. Micro-CHP System The principal characteristics of a prime mover that will determine its annual operating performance include electrical output, electrical efficiency, start-up response and effect of cycling on its lifetime. The delivered thermal output of the system is related to the prime mover's power output, electrical efficiency and the efficiency of heat recovery achieved. These operating conditions were defined for a generic SE µchp system. The definition of the start-up condition for the prime mover is likely to influence the veracity of results generated from any µchp modelling study [7]. Because the start-up condition for an SE prime mover is influenced by the period for which the prime mover has been idle, this investigation assumed both warm and cold start-up conditions (Table 1). TABLE 1: Start up conditions for a generic Stirling engine prime mover Time (min) % of rated load Cold start Warm start The cold start condition was applied if the µchp system had been inactive for >12 minutes. In both cases, an elapsed period of one minute after the activation signal was applied to allow for engine starting and frequency synchronisation purposes. Once full load was reached, no modulation of the electrical output of the µchp systems was assumed and operation occurred at full rated power. 4. Micro-CHP System Control Two control methods were evaluated; a) heat led control (HLC) - the presence of a thermal demand in the discrete dwelling to which the µchp system was deployed was the cause of activation and b) aggregated load control (ALC). A simple method was used whereby ALC took as its activation signal the aggregate load from the 5 dwellings according to the relationship shown in equation (1). N = (E-P a )/P oe (1) - E is the aggregate electrical demand of the dwelling dataset (kw) - P a is the aggregate load threshold above which a number of µchp systems within the group will be activated (kw) - Poe is the rated electrical power output of a single prime mover (kw) - N is the number of identically sized µchp prime movers to be activated once demand exceeds Pa (rounded up to the nearest integer). 5. Results At a high µchp penetration level of 76% (38 of the 5 dwellings) and Poe = 1.5kW the effects of a HLC strategy and the proposed ALC strategy on the resultant electrical load profile are quite distinct (Figure 3). The HLC strategy causes the load factor of the resultant electrical load profile to fall from 42.5% to 28.6% (Figure 3a). Reverse (export) flows were generated in the morning, prior to the morning gradient rise and, to a lesser extent in the late afternoon. With the ALC strategy, operation of the µchp systems was effectively curtailed prior to the morning gradient rise (which occurred between 6:41 and 9:2 on the January day) by setting Pa to the average aggregate demand during the night (i.e. approximately 26kW on the January day). For this prime mover size and penetration level, both the morning and evening gradients seen in the original demand can be eliminated. As a result the daily load factor increases from 42.5% to 48.5% (Figure 3b). A further property of aggregate load curves, the minimum load (L min ) can also be defined (Figure 4). The connection of embedded generation is only deemed to be advantageous as long as the quantity of generation connected to the LV distribution network (EC max ), minus the minimum load (L min ), is less than the peak load (L max ) (Figure 4) [8]. That is, when the net peak loading on a distribution transformer (whether import or export) is no greater than before, and is generally less. For the 5 dwelling dataset, Lmax was 1.6kW and Lmin was 17.2kW. EC max for this dataset was therefore 117.8kW or 2.4kW per dwelling. For µchp deployment as described (57kW) and HLC, L min fell to -2.1kW. Consequently a further 8.5kW of embedded generation would be permissible giving a total of 137.5kW or 2.8kW per household. With ALC, L min fell to 2.1kW and ECmax increased to 12.6kW meaning the total amount of embedded generation permissible was 159.7kW or 3.2kW per household. Thus, additional embedded generation connected to a LV distribution network containing high levels of µchp deployment can be seen to be a function of the control method employed. 6. CO 2 emission saving Using a simple emissions accounting methodology, the CO 2 emissions saved by deployment of the µchp systems described for this group of dwellings on the January day can be computed [4]. Assuming a carbon intensity of network electricity of.43kgco 2 /kwh and an emissions factor for natural gas of.19kgco 2 /kwh, the CO 2 emissions reduction on this day was 1.% and.7% for the HLC and ALC systems respectively. The reduced savings attributable to the ALC systems was as a result of reduced run times. However, this simple accounting method does not take account of any interaction between the µchp approach and the electricity network. 76

4 Peacock Newborough 3/1/7 12:56 Page 77 A.D. Peacock and M. Newborough /ISESCO Science and Technology Vision - Volume 3, Number 4 (November 27) (74-78) Figure 3: Effect of control method on the load profile placed on the network by 5 dwellings on the January day when n=76% and P oe = 1.5kW LV Transformer Load Export Import Figure 4: Definition of maximum capacity of embedded generation capacity (EC max ) that can be connected to the LV network 77

5 Peacock Newborough 3/1/7 12:56 Page 78 A.D. Peacock and M. Newborough /ISESCO Science and Technology Vision - Volume 3, Number 4 (November 27) (74-78) 7. Conclusions Micro-CHP has the potential to reduce the CO 2 emissions attributable to individual dwellings. As the concept moves towards mass market, it may also have the capacity to provide ancillary services to the ESI. A control method proposed here caused the load placed on the LV transformer to be substantially smoothed. This potentially could lead to reduced part load operation of network fossil fuel plants thereby contributing to the decarbonisation of electricity supply. Additionally, the control of µchp systems to ensure that their operation does not coincide with minimum load condition may allow additional embedded capacity to be connected to LV networks. Current economic and CO 2 emissions accounting models are insufficient in estimating the contribution of these additional benefits References [1] Electricity Association Sectoral Profiles, Electricity Association, 1998 [2] National Grid 7 Year Statement, 23, March 26 [3] Balancing Mechanism Reporting System (BM Reports), National Grid Company, February 26 [4] Peacock AD, Newborough M. Impact of micro-chp systems on domestic sector CO2 emissions. Applied Thermal Engineering 25, 25, pp [5] Peacock A.D., Newborough M., Impact of micro-combined heat-and-power systems on energy flows in the UK ESI, Energy, Article in press, 25 [6] Agar WR, Newborough M. Implementing micro-chp systems in the UK residential sector, J Inst Energy 1998, 71, pp [7] February 25 [8] Ingram S, et al. The impact of small scale embedded generation on the operating parameters of distribution networks, DTI, Report No K/EL/33/4/1,