Mutual trading strategy between customers and power generations based on load consuming patterns Junyong Liu, Youbo Liu Sichuan University
2 Outline Ⅰ Ⅱ Research Background Reviews on the development of electricity markets and trading strategy Problem Analysis Problem analysis of direct electricity purchase by large consumers Ⅲ Methodology Novel trading strategy for direct electricity purchase based on data analytics Ⅳ Trading Platform Design Trading platform for direct electricity purchase markets V Software Development Visual monitoring system for electricity markets
3 Reviews on the Development of Electricity Markets Research Backgroundd Advanced Development Worldwide Development in China Electricity Markets Electricity market reform Competitive market in generation side Competitive market in retail side Fair opening of power grid under government supervision Inter regional market transactions in Australia Settlement residues auction Nordic power system transactions Nord Pool Electricity market reform Competitive market in generation side Establish electric power trading center In process of establishing competitive market in retail side (No.9 Electricity market reform document in 25) Direct Electricity Purchase of Large Consumers Establish laws and regulations Empower customers to choose their power suppliers in accordance with the voltage level and power capacity Establish independent transmission and distribution price mechanism Establish surplus power handling mechanism Pilot projects including Price mechanism Cross subsidy Auxiliary service Transaction scheduling Trading strategy
4 Direct Electricity Trading Strategies Bilateral negotiation transaction model Research Backgroundd Large electricity customers and power generation companies meet directly and complete the transaction through bilateral consultation. It is the most commonly used model in pilot projects in China. Centralized bidding transaction model Large electricity customers and power generation companies proceed to bidding transaction in power trading center. It is applicable to facilitate the formation of transactions in a short period of time. Centralized matchmaking transaction model Large electricity customers and power generation companies proceed to transaction in power trading center based on the trading electricity curves. It is applicable for the situation which causes the minimal impacts on the original scheduling.
5 Problem Analysis Theoretical Research Most of the existing literatures in the field of direct electricity trading markets can be concluded into 3 categories In centralized matchmaking transaction model, optimization model is established based on the principle of high-low matching. Shortcoming: It is hard to establish a stable market between multiple individuals due to the frequent matchmaking process. The complexity of power system operation is also increased. In bilateral negotiation transaction model, game theory is applied to analyze the behavior characteristics of different trading individuals. Shortcoming: It is hard to obtain complete information between each other so that the system overall efficiency is hard to be optimized. Real option theory is used to establish the direct electricity trading model as well as calculate the trading price. Shortcoming: It has high requirement of electricity markets maturity. Thus, it is limited in practical applications especially in China.
6 Problem Analysis Practical situation Transition of market profits after No.7 electricity supervision document in China Transaction can be directly proceeded between large electricity customers and power generation companies. Large customers ask for lower electricity price and reducing the obligations of purchasing cross subsidies. Increase the operating costs and risks of electrical company Information asymmetry Power Generation Company Electric Company Large Customer Power Generation Company?? Electric Company Large Customer Difficult to optimal allocate system electricity Shortcoming of current trading model Opaque price Low market transaction efficiency Imperfect competition
7 A novel direct electricity trading strategy for large customers is proposed Principal-Agent Transaction Model Methodology Data Analysis Technologies Data distortion correction Data forecasting Data clustering Data pattern matching Principal A (Large customer) Common Agent (Electric Company) Principal B (Large customer) Principal C (Large customer) Achieve a win-win situation. Large customers want to ensure high electricity reliability and reduce electricity purchase prize. Electric company want to bring the direct purchase electricity into the overall power network optimization to make the system more stable, secure and efficient. Datasets Large customers Commission Electricity purchase prize Reliability requirement Electric company Network structure System Operation parameters Congestion condition Generation company Commission Electricity selling prize Auxiliary service capability
Electricity Consuming Scheme Formulation 8 Large Customers Electric Company Methodology Submit Mission Price Information from generation company Electricity Reliability requirement Data analysis Reward Complete Clear-price Trading electricity Operation Security check Real time scheduling Optimal allocation of system overall electricity Operation condition Balance Incentive Rewards and punishments based on the transaction Balance
9 Load Consuming Pattern Matching Technique First, the power generation output curve and large customer load curve is analyzed. Second, the most similar pairs of generation load are matched with each other. Third, transaction strategy is formulated with the objective of maximum the overall pattern matching performance in the whole power network. Correlation Analysis of Load/Generation Curve n r i i( k ) n k min min y o( k ) y i( k ) max max y o( k ) y i( k ) i k i k i( k ) y ( k ) y ( k ) max max y ( k ) y ( k ) o i o i i k Methodology Output curve of multiple generations Output curve of multiple large load customers 2 8 6 4 2 8 6 4 2 2 8 6 4 2 直购电厂 2 4 6 8 2 4 6 8 2 22 8 6 4 2 大用户 2 4 6 8 2 4 6 8 2 22 8 6 4 2 2 5 5 5 5
Load Consuming Pattern Matching Technique Effect on system power balance Methodology High correlation performance Low correlation performance P Generation Load P Generation Load t t Power network support Power network support P P Backup scheduling -ΔP Backup scheduling ΔP 2 Generation control ΔP t Generation control ΔP 2 t Effect on economical benefits Peak load periods ( 8:-: & 8:-2:) Reduce the peak-shifting pressure of power grid when large customer load increase sharply. Valley load periods ( 22: - 6: ) Reduce the shutdown risks of power generation company when large customer load decrease sharply.
Methodology Distortion correction of statistic load/generation data Distortion correction for similar typical days x N ni, xni, N n x x n, i ni, 3 i...96 N 2 2 i [ xn, i xni, ] N n x x x x,2 n, i ( n, i n, i )/ 2 ni, i Distortion correction for similar hours in typical days x 2 ' n, i xn, i j 5 j 2 x x x ' ' n, i n, i n, i i...96 x ( x x )/ n, i 2 n, i 2 n, i 2 2 Clustering of statistic load/generation data Mixed clustering algorithm based on K-means algorithm and self-organizing map algorithm Minimum distance connection weight Output layer Input layer V r min V r i a i j Topology iteration r ( t ) r ( t) ( t) f ( t)( V ( t) r ( t)) j j aj i j K-means clustering standard measure function based k on SOM output layer 2 E p m i p C i i
2 Load Consuming Pattern Matching Technique 8 6 4 2 Case study 8 6 4 2 6 4 2 8 6 4 2 Peak load period 8 6 4 2 8 6 4 2 4 2 Assessment indices 电厂 3 大用户 7 2 4 6 8 2 4 6 8 2 22 8 6 4 2 电厂 9 大用户 2 2 4 6 8 2 4 6 8 2 22 Case number Q System power balance requirement (unit: kw) R T System uncertainty factor (unit: RMB/kWh) F P ( C ) System real-time operation cost factor (unit: RMB/h) M F R Methodology R C f F M 353.69.4938.9945 7.38244 2-44.66.72422.8426-37.244 3-62.29.673752.88576-5.54 4-44.724 -.424.38377-59.59 5 49.55969.94785.22582.956 6 44.69.239723.9349 8.22486 7-24.7.8873.5738 -.3869 8-98.694 -.42.322654-64.94 9 2.3846.62978.89423.82286 247.53.8547.56 2.6625-7.7675.653.8883 -.5778 2-69.89.228474.88638-3.35 3-48.498 -.3656.356422-52.9279
3 Methodology Steps of Direct Electricity Purchasing Market Construction in Shanghai Orderly centralized matchmaking Step III Principal-Agent Platform Disordered bilateral negotiation Establish Regulations One-to-one Bilateral negotiation Passive security check Step I Matchmaking Trade Platform Many-to-many Multiple trading periods Generation-demand matching Many-to-many Load pattern matching Security check Comprehensive assessment Step II
Earnings curve Complete Electricity Trading Procedure + System scheduling System scheduling+ direct electricity trading Purchase and sale to the regular market (or) Trading Platform Design Large customers By sequence Power generations Pre Pair matching Transaction cancel (or) + Curtailment (or) 4 Trading Channel Transmission Capability Transaction curve P Operation Security check P P Capability Channel Transaction Non-conflicts Capability Transaction Channel t Serious negative impact Capability Transaction t Channel Curtailment needed t Direct electricity trading confirmation System operation scheduling confirmation
5 Case Study in Shanghai Load data distortion correction and clustering Trading Platform Design Input data: Statistic data of 48 large customers and 8 power generation companies in Shanghai. Generation output.5.5.5.5 5 5 2 5 5 2 5 5 2 5 5 2.5.5.5.2.5. 5 5 2 5 5 2 5 5 2 5 5 2 2 5 5 5 5 5 2 (A) Large customer load.5.5.5.5 5 5 2 5 5 2 5 5 2 5 5 2.5.5.5.5 5 5 2 5 5 2 5 5 2 5 5 2 5 5 5 2 (B) Clustering results.5.5 5 5 2 5 5 2
6 Case Study in Shanghai Enumeration of generation-load pattern matching Trading Platform Design Results of matching strategy 36, 37 (high correlation case) Strategy Number 36 37 Combination results of generation Combination results of large customers Correlation Degree (single match) 2/6/8 /3/4/5/6/7/8.988 /3/5/7 2/9.92 4.8896 Rest of electricity Rest of electricity.9494 2/6/8 /3/5/6/7/8/9.988 3/5/7 2/4/.96 /4+Rest of electricity Rest of electricity.8996 Rest of electricity means electricity that cannot be consumed by direct electricity transaction
7 Case Study in Shanghai Trading Platform Design Direct electricity purchase plan list based on the optimal-matching principle Plan Number Combination plan of generations Combination plan of large customers Integrated correlation degree (M) 24 2/6/8;/3/5/7;4 /3/4/5/6/8;2/7/9;.87547 36 2/6/8;/3/5/7;4 /3/4/5/6/7/8;2/9;.87435 37 2/6/8;3/5/7 /3/5/6/7/8/9;2/4/.8536 3 /2/6/8;5/7 /3/4/8/9/;2/5/6/7.84662 26 /2/6;5/7 /3/4/8/9/;2/5/6/7.8448 39 2/5/6;/7 2/3/4/5/8/9/;/6/7.83977 A n i n i e c e R i i r n i i e M r A i e i r Power grid dispatching department can obtain the power balance information in each time interval from the list above. Then, it can make scheduling plan on the basis of system operation status. Finally, the direct electricity transaction can be confirmed.
8 Trading Platform Design Case Study in Shanghai Comparison of different trading strategies Trading strategy Average power flow distribution Maximum load Minimum load Complete direct purchase electricity (MWh) Economic profits of electric company ( 4 RMB) Principal-Agent 63% 9% % 289 296.3 Centralized matchmaking Bilateral negotiation 57% % % 2465 34.6 66.5% % % 293 258.6 In terms of operation security, the presented principal-agent model provided the most uniform power flow distribution. The network is stable and secure with sufficient capacity margin. In terms of electric company economic profits, the result of principal-agent model is slight less than the centralized matchmaking model. That is because principal-agent model aims at maximizing pattern correlation instead of maximizing economic income. However, the volume of transactions is guaranteed, which has a positive impact on coordinating the profits of the whole market.
9 Software Development Functional structure diagram of power grid trading visualization software System realtime data visualization Power grid analysis visualization Application of GIS Technology Data integration, mining and intelligent analysis Data receiving platform for electricity markets SCADA/EMS WARMS DMIS Hydroelectric generation dispatching system Load forecasting
2 Software screenshot Software Development Power Generation distribution in Shanghai Load curve in different typical days Trade based on congestion analysis 3D map of power system stability using PMU
2
Conclusion 22 Based on the problems existing in direct electricity purchase markets in China, this research puts forward the corresponding solution and practical application, which has been successfully applied in Shanghai pilot projects. The main novelty of the research is proposing a new mutual trading strategy principal-agent transaction strategy based on the data analysis techniques. The most significant one, i.e. load consuming pattern matching technique, is able to economically optimize the allocation of power resources in the whole network. It also can reflect the real trade willingness of large customers so that the transparency and fairness of the electricity market are guaranteed. An active security check mechanism is considered in the electricity trading process, which ensures the high reliability performance for both power grid and large customers. The guarantee of secure and stable power network operation is positive for the power markets development.
23 Thanks for listening Please contact liujy@scu.edu.cn