A conditional value at risk based multi-consumer distributed transaction method
By adopting a multi-producer-consumer distributed trading method based on conditional value of risk, the problems of small controllable capacity of producers and consumers in the electricity market and the randomness of photovoltaic output are solved. This method achieves resource sharing and cost optimization, incentivizes more producers and consumers to participate, and improves the flexibility and economy of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- STATE GRID JIANGSU ECONOMIC RES INST
- Filing Date
- 2022-09-20
- Publication Date
- 2026-06-19
AI Technical Summary
In existing technologies, when producers and consumers participate in electricity market transactions, there are problems such as small controllable capacity, randomness of photovoltaic output leading to revenue risk, and inability to maintain system supply and demand balance. Furthermore, the measurement of uncertainty risk is not fully considered in distributed transactions.
We adopt a distributed transaction method for multiple prosumers based on conditional value of risk. By aggregating distributed resources within prosumers, we establish a model and introduce distributed transaction constraints. We use the ADMM algorithm to solve the model and combine it with generalized Nash equilibrium theory for cost allocation, thereby mitigating risks and incentivizing more prosumers to participate.
It enables resource sharing among producers and consumers, reduces transaction costs, avoids the randomness risk of photovoltaics, incentivizes more producers and consumers to participate in market transactions, ensures fair cost allocation, and improves system flexibility and economy.
Smart Images
Figure CN115587748B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of power system dispatching and optimization, and specifically relates to a multi-producer-consumer distributed trading method based on conditional risk value. Background Technology
[0002] With the massive integration of distributed renewable energy and the rapid growth of flexible loads such as user-side energy storage and air conditioning, end users are gradually evolving into "prosumers" with dual source and load characteristics, significantly improving system flexibility. However, prosumers still face several challenges in participating in market-based operations. First, as resource aggregators on the user side, prosumers have limited controllable capacity, sometimes even failing to meet basic market entry barriers. Second, prosumer aggregation units typically include resources such as rooftop photovoltaics, and the randomness of photovoltaic output means prosumers may face revenue risks when participating in market transactions. Finally, prosumers are mostly located at the end of the system architecture, resulting in a lack of ability to maintain system supply and demand balance and failing to fully leverage the flexibility of aggregated resources. Therefore, it is urgent to develop a mechanism for prosumers to participate in market transactions to maximize the value of their aggregated resource flexibility.
[0003] Traditional electricity market trading models include centralized and distributed trading. Compared to centralized trading, distributed trading offers decentralization and greater flexibility, making it suitable for dispersed producers and consumers to participate in market transactions. However, current research on producer-consumer participation in the electricity market focuses on distributed trading mechanisms and revenue distribution, with few studies considering the impact of internal uncertainties among producers and consumers on distributed trading. Further research is needed on measuring the uncertainty risk of producers and consumers in distributed trading. Summary of the Invention
[0004] To address the technical problems mentioned in the background, this invention proposes a multi-prosumer distributed transaction method based on conditional value of risk.
[0005] To achieve the above-mentioned technical objectives, the technical solution of the present invention is as follows:
[0006] A multi-prosumer distributed transaction method based on conditional value of risk includes the following steps:
[0007] (1) Aggregate distributed resources within the producer-consumer group to establish a producer-consumer model and a distribution network model, in which distributed transaction constraints are introduced;
[0008] (2) Set the confidence level and risk preference coefficient of the conditional value at risk model, and establish a multi-prosumer distributed transaction model based on conditional value at risk;
[0009] (3) Solve the distributed transaction model of multi-prosumer based on conditional value of risk using the ADMM algorithm;
[0010] (4) Establish a market clearing model to allocate distributed transaction costs among producers and consumers.
[0011] Furthermore, the specific process of step (1) is as follows:
[0012] (1.1) Assuming that the prosumer group includes photovoltaics, energy storage, central air conditioning, flexible loads, and fuel cells, and that photovoltaics use predicted output data, list the distributed transaction constraints for each prosumer group:
[0013] Energy storage constraints:
[0014]
[0015]
[0016]
[0017]
[0018]
[0019] In the formula: s represents the photovoltaic power output scenario; t represents the trading period; the photovoltaic power output varies under the time period t in each scenario, and the variables of internal energy storage, central air conditioning, flexible load and fuel cell of the producer and consumer are adjusted accordingly; and These represent the charging and discharging amounts of energy stored by consumer i during time period t in scenario s; P i c,max and P i d,max These are the maximum charging and discharging power of the energy stored in the consumer i, respectively; S i,s,t State of charge of energy stored by consumer i in scenario s during time period t; and These are the minimum and maximum energy storage capacities within producer-consumer i, respectively. and The charging and discharging efficiency of energy stored in the consumer's internal storage; The energy storage cost for consumer i during time period t in scenario s is the cost of battery wear and tear during charging and discharging. and These are the charging and discharging dissipation coefficients of the energy stored in producer i, respectively;
[0020] Central air conditioning constraints:
[0021]
[0022]
[0023]
[0024] In the formula: α i,t β is a parameter describing the building's cold storage characteristics and outdoor temperature for consumer i during time period t; i Parameters related to cold storage in building exterior walls and windows, and fresh air volume in air conditioning within the consumer's environment; γ i Parameters related to cold storage in building interior walls within the consumer's domain; σ represents the indoor temperature of consumer i during time period t in scenario s; i The energy efficiency ratio of central air conditioning refrigeration units in the production and consumption area; The cooling capacity of the central air conditioning system for consumer i during time period t in scenario s; T in,min and T in,max These are the minimum and maximum permissible indoor temperatures, respectively. The discomfort cost incurred by consumer i when adjusting the central air conditioning during time period t in scenario s; m is the user discomfort coefficient; T i ref The temperature is designed to be the most comfortable for consumers.
[0025] Flexible load constraints:
[0026]
[0027]
[0028]
[0029] In the formula: The flexible load value of consumer i in scenario s during time period t; The load baseline value for consumer i within time period t; and These represent the lower and upper limits of the adjustable flexible load for producer-consumer i within time period t; The discomfort cost incurred by consumer i in adjusting flexible load during time period t in scenario s; n is the user discomfort coefficient; fuel cell constraints:
[0030]
[0031]
[0032]
[0033] In the formula: P represents the power generation of the fuel cell for consumer i during time period t in scenario s; i min and P i max These represent the minimum and maximum output power of the fuel cell within producer-consumer i, respectively; and For the uphill and downhill ramp rates of fuel cells within the producer-consumer i; For consumer i, the cost of fuel cells in scenario s during time period t; The unit power generation cost of fuel cells within the producer-consumer i;
[0034] Distributed transaction constraints:
[0035]
[0036] In the formula: e i,t The energy that producer-consumer i trades with other producer-consumers during time period t;
[0037] Transaction constraints with power grid companies:
[0038]
[0039]
[0040] In the formula: The predicted photovoltaic output of producer-consumer i within time period t in scenario s; and These represent the electricity purchased and sold by producer-consumer i from the power grid company during time period t, respectively. For the transaction costs between producer-consumer i and the power grid company during time period t; λ t b and The prices at which electricity is purchased and sold from the power grid company within time period t;
[0041] (1.2) Introduce distribution network constraints to limit the transaction volume of producers and consumers, and ensure the safety of power grid operation:
[0042] Node power balance equations:
[0043]
[0044]
[0045] In the formula: P m,t Q represents the active power injected into node m during time period t. m,t P represents the reactive power injected into node m during time period t. mn,t and Q mn,t These represent the active power and reactive power of branch mn during time period t, respectively; P km,t and Q km,t Let F(m) represent the active power and reactive power of branch line km within time period t; F(m) represents the set of terminal nodes of the branch line with node m as the first endpoint; T(m) represents the set of first endpoint nodes of the branch line with node m as the last endpoint; r km For the branch circuit resistance (km); xkm For branch line km reactance; l km,t N is the square of the current amplitude on branch line km during time period t; S For the system node set;
[0046] Branch voltage drop equation:
[0047]
[0048] In the formula: V m,t and V n,t These are the squares of the voltage amplitudes at nodes m and n during time period t, respectively; r mn For branch mn, the resistance is x. mn For branch line mn reactance; l mn,t B is the square of the current amplitude on branch mn during time period t; S For the set of system branches;
[0049] Branch capacity equation:
[0050]
[0051] After applying second-order cone relaxation to the branch capacity equations, the power flow model becomes a second-order cone programming model:
[0052]
[0053] Node injection power:
[0054]
[0055]
[0056] In the formula: and These represent the active power and reactive power of the root node during time period t, respectively. The power factor of node m;
[0057] Line safety constraints:
[0058]
[0059]
[0060] In the formula: and These are the minimum and maximum values of the square of the voltage amplitude at node m, respectively. The maximum value of the square of the current amplitude on branch mn;
[0061] Under the supervision of distribution network operators, the network loss costs incurred by all producers and consumers during the transaction process are as follows:
[0062]
[0063] In the formula: C DSO For network loss costs; The penalty price for network loss during time period t.
[0064] Furthermore, the specific process of step (2) is as follows:
[0065] The conditional value at risk model is established as follows:
[0066]
[0067] z i,s ≥0
[0068]
[0069] In the formula: Conditional risk value for producer-consumer i costs; Value at risk for producer-consumer i; ξ is the confidence level; ρ s Let z be the probability of photovoltaic scenario s; i,s This represents the value at which the cost to consumer i exceeds the value at risk in scenario s;
[0070] The objective function of the multi-prosumer distributed transaction model based on conditional value at risk is:
[0071]
[0072] In the formula: C all The total cost of distributed transactions among multiple prosumers based on conditional risk value; the risk preference coefficient L represents the prosumer's attitude towards risk, and its value ranges from L∈[0,1]. The larger the value of L, the more risk-averse the prosumer is.
[0073] Furthermore, the specific process of using the ADMM algorithm to solve the multi-prosumer distributed transaction model based on conditional value of risk in step (3) is as follows:
[0074] (3.1) List the Lagrange equation for the objective function of the multi-prosumer distributed transaction model based on conditional value of risk:
[0075]
[0076] In the formula: λ i,t ω i,t σ i,t Let i be the dual variable of producer-consumer in time period t; As a penalty factor; Let i be an auxiliary variable for prosumer i during time period t, representing distributed transactions of prosumers i under the supervision of the distribution network operator and the electricity purchased and sold from the grid, respectively, satisfying the following constraints:
[0077]
[0078]
[0079]
[0080]
[0081] (3.2) The ADMM algorithm is used in GAMS software to iteratively update the producer-consumer, distribution network operator, and dual variables as follows:
[0082] (a) Update of the producer-consumer variable at the (k+1)th time:
[0083]
[0084] (b) The (k+1)th update of the distribution network operator variable:
[0085]
[0086] (c) The (k+1)th update of the dual variable:
[0087]
[0088]
[0089]
[0090] (3.3) Calculate the original residual and dual residual after each iteration:
[0091]
[0092] In the formula: s represents the distributed transaction volume of producer-consumer i in the (k+1)th iteration within time period t, and the original residuals with the grid's electricity purchase and sales volumes, respectively; i,t,k+1 , These are the distributed transaction volume of producer-consumer i in the (k+1)th iteration within time period t, and the dual residuals with the grid's electricity purchase and sales volumes, respectively.
[0093] (3.4) Determine whether the ADMM algorithm has converged based on the magnitude of the original residual and the dual residual. If it has converged, output the transaction decisions of each producer and consumer. If it has not converged, return to step (3.2) and continue iterating until it converges.
[0094] The iteration stopping condition is as follows:
[0095]
[0096] Where: ε pri and ε dual These are the tolerance upper limits for the original residual and the dual residual, respectively; if the above conditions are met, the algorithm converges.
[0097] Furthermore, step (4) establishes a market clearing model, and the specific process of allocating the distributed transaction costs of producers and consumers based on the generalized Nash equilibrium method is as follows:
[0098] (4.1) Allocate network transmission costs based on the transaction volumes of each prosumer obtained from the multi-prosumer distributed transaction model based on conditional value of risk:
[0099] When each producer-consumer only transacts with the power grid, the transaction costs for each producer-consumer are: The corresponding network transmission cost is When each producer-consumer transacts with the power grid and other producer-consumers, the transaction costs for each producer-consumer are: The corresponding network transmission cost is The increased network transmission costs associated with distributed transactions will be distributed among the prosumers / consumers according to the proportion of their transaction volume.
[0100]
[0101] Where: χ i The network transmission cost allocated to producer-consumer i;
[0102] Cost allocation constraints:
[0103]
[0104] Where: π i Distributed transaction costs reallocated for producer-consumer i;
[0105] (4.2) A market clearing model is established based on the generalized Nash equilibrium method to satisfy the distributed transaction cost constraints between producers and consumers:
[0106] Market liquidation model objective function:
[0107]
[0108] Where: μ i The market power of prosumers is determined by market operators and satisfies [the needs of consumers].
[0109] The constraints of the market liquidation model are:
[0110]
[0111] Taking the logarithm and negative of the objective function of the market clearing model, we obtain the following minimization problem:
[0112]
[0113]
[0114] In the formula: Cost savings for prosumer i after participating in distributed transactions;
[0115] (4.3) The market clearing model is solved using the Lagrange multiplier method to calculate the distributed transaction costs for each producer and consumer:
[0116] 1) The objective function of the market clearing model, the Lagrange equation, is:
[0117]
[0118] 2) Taking the first-order partial derivative of the above equation, we get:
[0119]
[0120] 3) and Substituting into the above equation, we get:
[0121]
[0122] 4) Calculate the distributed transaction cost for prosumers based on the cost savings and market power after participating in distributed transactions:
[0123]
[0124] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention proposes a multi-prosumer distributed transaction model, realizing resource sharing and reducing transaction costs for prosumers. It considers the internal photovoltaic randomness of prosumers, mitigates risk based on conditional value at risk theory, and enables prosumers to weigh risk and return. The ADMM algorithm is used to solve the multi-prosumer distributed transaction model, and distributed transaction costs are allocated based on generalized Nash equilibrium theory, achieving fair cost allocation and incentivizing more prosumers to participate in market transactions. Attached Figure Description
[0125] Figure 1 This is a flowchart of the method of the present invention;
[0126] Figure 2 This is a schematic diagram of the network topology in the embodiment;
[0127] Figure 3 This is a schematic diagram of the electricity purchase price from the power grid in the example;
[0128] Figure 4This is a schematic diagram of the load of Producer-Consumer 1 Scheme 1 in the embodiment;
[0129] Figure 5 This is a schematic diagram of the load of Producer-Consumer 1 Scheme 2 in the embodiment;
[0130] Figure 6 This is a schematic diagram illustrating producer-consumer costs and conditional value of risk under different risk preferences in the examples. Detailed Implementation
[0131] The following detailed description, in conjunction with the accompanying drawings, illustrates a specific implementation of the multi-prosumer distributed transaction method based on conditional value of risk according to the present invention.
[0132] This invention designs a multi-prosumer distributed transaction method based on conditional value of risk, such as... Figure 1 As shown, the steps are as follows:
[0133] Step 1: Aggregate distributed resources within the prosumer network, establish a prosumer model and a distribution network model, and introduce distributed transaction constraints;
[0134] Step 2: Set the confidence level and risk preference coefficient of the conditional value at risk model, and establish a multi-prosumer distributed trading model based on conditional value at risk;
[0135] Step 3: Solve the multi-prosumer distributed transaction model based on conditional value at risk using the ADMM algorithm;
[0136] Step 4: Establish a market clearing model and use generalized Nash equilibrium theory to allocate distributed transaction costs among prosumers and consumers.
[0137] This invention verifies the effectiveness of the proposed method using a simulation example of three prosumers connected to an IEEE 33-node system. Prosumer 1 includes a fuel cell, photovoltaic system, energy storage system, central air conditioning system, and flexible load. Prosumers 2 and 3 each include photovoltaic system, energy storage system, central air conditioning system, and flexible load. Assuming the prosumers are connected to nodes 4, 19, and 32 respectively, the network topology is as follows: Figure 2 As shown in Table 1. Parameters for fuel cells, energy storage, photovoltaics, and central air conditioning are listed below. The time-of-use tariff provided by the California Independent System Operator (CAISO) is used as the purchase price from the grid, such as... Figure 3 As shown, the electricity sales price is 0.5 times the electricity purchase price. The example demonstrates the optimization results from 1:00 to 24:00.
[0138] To verify the impact of distributed transactions on the transaction costs of each prosumer, this invention sets up two prosumer transaction schemes as follows:
[0139] Option 1: Multi-prosumer transactions with the power grid considering network constraints
[0140] Option 2: Considering network constraints, multiple prosumers transact with the power grid and other prosumers.
[0141] Prosumer energy management under the two schemes is as follows Figure 4 and Figure 5 As shown. Figure 4 and Figure 5 The comparison shows that the net load of Scheme 1 is lower than that of Scheme 2, with more significant load changes in fuel cells and energy storage. Under Scheme 2, fuel cell output is higher than under Scheme 1, and this increased output is sold to other prosumers to increase revenue. Under Scheme 1, energy storage charging and discharging fluctuates more significantly, constrained by the amount of stored energy, with excess energy released at the end of operation. In contrast, under Scheme 2, energy storage charging and discharging fluctuations are smaller, with prosumers trading with other prosumers before storing or releasing energy. Considering the costs associated with user comfort, prosumers slightly adjust their air conditioning and flexible loads, resulting in essentially the same air conditioning and flexible loads for both schemes. In conclusion, considering distributed trading is beneficial for unlocking the potential of load regulation, thereby increasing the net injected power of prosumer access nodes.
[0142] Table 2 compares the transaction costs under the two schemes. Since no distributed transaction costs are allocated, under Scheme 2, producer-consumer 1, acting as a producer, incurs higher costs; while producers-consumers 2 and 3, primarily acting as consumers, incur lower costs. Scheme 2 increases the net load of producers-consumers and the transmission power of lines, thus increasing network loss costs under the supervision of the distribution network operator. However, considering the overall lower social cost of Scheme 2, this demonstrates that the distributed transaction model can broaden the transaction channels for producers-consumers and improve the economics of their participation in transactions.
[0143] After solving the prosumer model, the distributed transaction costs need to be redistributed to maximize social benefits. This invention compares two methods: the generalized Nash equilibrium method and the Nash bargaining method. The distributed transaction costs for each prosumer under different allocation methods are shown in Table 3. Compared with Scheme 1 in Table 2, each prosumer saves costs under both allocation methods. Specifically, the generalized Nash equilibrium method calculates market power based on the shared energy among prosumers and allocates distributed transaction costs and network loss costs according to market power. Compared to Scheme 1, the more shared energy, the greater the cost savings, and the revenue per MWh (the revenue obtained for every 1 MWh of shared power) is the same, indicating that this method is fair in allocating revenue. In contrast, the Nash bargaining method has the same market power for all three prosumers. Distributed transaction costs and network loss costs are allocated equally based on market power. Compared to Scheme 1, the three prosumers save the same costs, but the calculated revenue per MWh differs. Prosumers with more shared energy receive less revenue per MWh, which discourages prosumers from actively participating in distributed transactions. Therefore, the choice of market forces has a significant impact on the distribution of profits, and a reasonable profit distribution method will encourage more producers and consumers to participate in distributed transactions.
[0144] Taking Scheme 2 as an example, the impact of the risk aversion value L on the total cost of prosumers in the conditional value-at-risk model is analyzed, and the results are as follows: Figure 6 As shown, as L increases, prosumers become more conservative, tending to increase total cost and decrease conditional risk value. Figure 6 This approach quantifies both returns and risks, allowing prosumers to determine their risk preferences based on their psychological expectations. Table 4 compares the total costs of the deterministic approach and the conditional value-at-risk (VAT) approach for prosumers. In day-ahead decision-making, the deterministic approach relies on predicted photovoltaic (PV) output, neglecting the uncertainty of PV output. The conditional VAT approach, however, considers PV output across multiple scenarios, resulting in lower day-ahead costs for the deterministic approach. However, when actual PV output falls below the predicted value, the deterministic approach requires prosumers to purchase the shortfall at higher prices, increasing intraday costs. The conditional VAT approach, by considering PV uncertainty in day-ahead decision-making, has lower intraday dispatch costs than the deterministic approach. Adding the day-ahead and intraday costs, the total cost of the conditional VAT approach is lower, demonstrating its economic viability.
[0145] Table 1 Prosumer Parameters
[0146]
[0147]
[0148] Table 2 Costs of each option / $
[0149]
[0150] Table 3. Distributed transaction costs for each producer-consumer under different allocation methods / $
[0151]
[0152] Table 4. Cost Comparison of Deterministic Model and Conditional Value at Risk Model / $
[0153]
[0154] The above embodiments are only for illustrating the advantages and features of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-substantial modifications, transformations and improvements made within the basic ideas and framework of the method proposed in the present invention shall fall within the scope of protection of the present invention.
Claims
1. A multi-prosumer distributed transaction method based on conditional value at risk, characterized in that, Includes the following steps: (1) Aggregate distributed resources within the prosumer to establish a prosumer model and a distribution network model, in which distributed transaction constraints are introduced; (2) Set the confidence level and risk preference coefficient of the conditional value at risk model, and establish a multi-prosumer distributed transaction model based on conditional value at risk; (3) Solve the multi-prosumer distributed transaction model based on conditional value at risk using the ADMM algorithm; (4) Establish a market clearing model to allocate distributed transaction costs among producers and consumers; The specific process of step (1) is as follows: (1.1) The producer-consumer group includes photovoltaic, energy storage, central air conditioning, flexible load and fuel cell. Photovoltaic uses predicted output data. The distributed transaction constraints of the producer-consumer group are listed respectively. (1.2) Introduce distribution network constraints to limit the transaction volume of producers and consumers, and ensure the safety of power grid operation: Node power balance equations: ； In the formula: For nodes m During the period t Internally injected active power; For nodes m During the period t Internally injected reactive power; and Branch roads mn During the period t Active and reactive power within; and Branch roads km During the period t Active and reactive power within; For nodes m The set of terminal nodes of a branch with a starting endpoint; For nodes m The set of the starting nodes of the branches of the terminal nodes; branch road km resistance; branch road km Reactance; branch road km During the period t The square of the internal current amplitude; For the system node set; Branch voltage drop equation: ； In the formula: and They are nodes m and n During the period t The square of the internal voltage amplitude; branch road mn resistance; branch road mn Reactance; branch road mn During the period t The square of the internal current amplitude; For the set of system branches; Branch capacity equation: ； After applying second-order cone relaxation to the branch capacity equations, the power flow model becomes a second-order cone programming model: ； Node injection power: ； In the formula: and The root node in the time period t Internal active power and reactive power; For nodes m Power factor; and Producers and consumers respectively i During the period t Electricity purchased and sold from the power grid company; Line safety constraints: ； In the formula: and They are nodes m The minimum and maximum values of the square of the voltage amplitude; branch road mn The maximum value of the square of the current amplitude; Under the supervision of distribution network operators, the network loss costs incurred by all producers and consumers during the transaction process are as follows: ； In the formula: For network loss costs; For the time period t Internal network loss penalty price; The specific process of step (2) is as follows: The conditional value at risk model is established as follows: ； In the formula: For producers and consumers i Conditional risk value of costs; For producers and consumers i Value at risk of costs; Confidence level; For photovoltaic scenarios The probability of; Indicates producer-consumer i In the scene s The cost exceeds the value at risk; For producers and consumers i In the scene s Time period t Internal energy storage cost; For producers and consumers i In the scene s Time period t The discomfort costs associated with internal air conditioning; For producers and consumers i In the scene s Time period t The discomfort costs arising from internally regulated flexible loads; For producers and consumers i In the scene s Time period t Internal fuel cell cost; For producers and consumers i During the period t Internal transaction costs with the power grid company; The objective function of the multi-prosumer distributed transaction model based on conditional value at risk is: ； In the formula: The total cost of multi-prosumer distributed transactions based on conditional value of risk; risk preference coefficient. This indicates the consumer's attitude towards risk, and its value ranges from [value range missing]. , The higher the value, the more risk-averse consumers are. This refers to network loss costs.
2. The multi-prosumer distributed transaction method based on conditional value of risk according to claim 1, characterized in that, Step (1.1) The prosumer group includes photovoltaics, energy storage, central air conditioning, flexible loads, and fuel cells. Photovoltaic output data is used. The specific constraints of the distributed transaction of the prosumer group are listed as follows: Energy storage constraints: ； In the formula: Indicates photovoltaic power output scenarios; Indicates the trading session; Various scenarios s Time period t As the output of photovoltaic power varies, variables such as energy storage, central air conditioning, flexible loads, and fuel cells within the producer-consumer system are adjusted accordingly. and Producers and consumers respectively i In the scene s Time period t The amount of energy stored internally for charging and discharging; and Producers and consumers respectively i Maximum charging and discharging power of internal energy storage; For producers and consumers i In the scene s Time period t State of charge of internal energy storage; and Producers and consumers respectively i Minimum and maximum storage capacity of internal energy storage; and For producers and consumers i The charging and discharging efficiency of internal energy storage; For producers and consumers i In the scene s Time period t Internal energy storage cost, which is the cost of battery wear and tear during charging and discharging; and Producers and consumers respectively i The charge and discharge dissipation coefficient of internal energy storage; Central air conditioning constraints: ； ； ； In the formula: For producers and consumers i During the period t The parameters describe the building's cold storage characteristics and outdoor temperature. For producers and consumers i Parameters related to the cold storage of building exterior walls and windows, and the fresh air volume of air conditioning; For producers and consumers i Parameters related to cold storage in building interior walls; For producers and consumers i In the scene s Time period t Indoor temperature for users; For producers and consumers i Energy efficiency ratio of internal central air conditioning refrigeration units; For producers and consumers i In the scene s Time period t Cooling capacity of the internal central air conditioning system; and These are the minimum and maximum permissible indoor temperatures, respectively. For producers and consumers i In the scene s Time period t The discomfort costs associated with internal air conditioning; To adjust the discomfort coefficient for central air conditioning users; For producers and consumers i The most comfortable temperature for the user's body; Flexible load constraints: ； In the formula: For producers and consumers i In the scene s Time period t Flexible load values for internal users; For producers and consumers i During the period t Internal user load baseline value; and Producers and consumers respectively i During the period t The lower and upper limits of the internal flexible load are adjustable; For producers and consumers i In the scene s Time period t The discomfort costs arising from internally regulated flexible loads; To adjust the user discomfort coefficient under flexible loads; Fuel cell constraints: ； In the formula: For producers and consumers i In the scene s Time period t The power generation capacity of the internal fuel cell; and Producers and consumers respectively i Minimum and maximum output power of the internal fuel cell; and For producers and consumers i Uphill and downhill ramp rates of the internal fuel cell; For producers and consumers i In the scene s Time period t Internal fuel cell cost; For producers and consumers i The unit power generation cost of internal fuel cells; Distributed transaction constraints: ； In the formula: For producers and consumers i During the period t The energy used for transactions with other producers and consumers; Transaction constraints with power grid companies: ； In the formula: For producers and consumers i In the scene s Time period t Internally predicted photovoltaic output; and Producers and consumers respectively i During the period t Electricity purchased and sold from the power grid company; For producers and consumers i During the period t Internal transaction costs with the power grid company; and From the power grid company during the time period t Prices for electricity purchased internally and electricity sold internally.
3. The multi-prosumer distributed transaction method based on conditional value at risk according to claim 2, characterized in that, Step (3) The specific process of solving the multi-prosumer distributed transaction model based on conditional value at risk using the ADMM algorithm is as follows: (3.1) List the Lagrange equation for the objective function of the multi-prosumer distributed transaction model based on conditional value of risk: ； In the formula: , , For producers and consumers i During the period t Internal dual variables; As a penalty factor; , , For producers and consumers i During the period t Internal auxiliary variables, representing the electricity purchased and sold from the grid by prosumers under the supervision of the distribution network operator, respectively, satisfy the following constraints: ； (3.2) The ADMM algorithm is used in GAMS software to iteratively update the producer-consumer, distribution network operator, and dual variables as follows: (a) Prosumer variable k+1 Next update: ； (b) Distribution network operator variable k+1 Next update: ； (c) The dual variable k+1 Next update: ； (3.3) Calculate the original residual and dual residual after each iteration: ； In the formula: , , Producers and consumers respectively i During the period t Inner The original residuals of the next iteration of distributed transactions between prosumers and consumers, and the electricity purchased and sold by the grid; , , Producers and consumers respectively i During the period t Inner The next iteration of distributed transaction volume of prosumers and consumers, and the dual residuals of grid-purchased and sold electricity; (3.4) Determine whether the ADMM algorithm has converged based on the magnitude of the original residual and the dual residual. If it has converged, output the transaction decisions of each producer and consumer. If it has not converged, return to step (3.2) and continue iterating until it converges. The iteration stopping condition is as follows: ； In the formula: and These are the tolerance upper limits for the original residual and the dual residual, respectively; if the above conditions are met, the algorithm converges.
4. The multi-prosumer distributed transaction method based on conditional value at risk according to claim 3, characterized in that, Step (4) Establishing a market clearing model and allocating distributed transaction costs among prosumers based on the generalized Nash equilibrium method is as follows: (4.1) Allocate network transmission costs based on the transaction volumes of each prosumer obtained from the conditional value-at-risk (VAT) distributed transaction model: When each producer-consumer only transacts with the power grid, the transaction costs for each producer-consumer are: The corresponding network transmission cost is When each producer-consumer transacts with the power grid and other producer-consumers, the transaction cost for each producer-consumer is... The corresponding network transmission cost is The increased network transmission costs associated with distributed transactions will be distributed among the prosumers according to the proportion of transaction volume between them. ； In the formula: For producers and consumers i The allocated network transmission costs; Cost allocation constraints: ； In the formula: For producers and consumers i The cost of redistributed distributed transactions; (4.2) A market clearing model is established based on the generalized Nash equilibrium method to satisfy the distributed transaction cost constraints between producers and consumers: Market liquidation model objective function: ； In the formula: For producers and consumers i Market power, the value of which is determined by market operators, satisfies... ; The constraints of the market liquidation model are: ； Taking the logarithm and negative of the objective function of the market clearing model, we obtain the following minimization problem: ； In the formula: For producers and consumers i Cost savings from participating in distributed transactions; (4.3) The market clearing model is solved using the Lagrange multiplier method to calculate the distributed transaction costs for each producer and consumer: 1) The objective function of the market liquidation model, the Lagrange equation, is: ； 2) Taking the first-order partial derivative of the above equation, we get: ； 3) and Substituting into the above equation, we get: ； 4) Calculate the distributed transaction cost for prosumers based on the cost savings and market power after participating in distributed transactions: 。