Two-dimensional electric power spot market quotation system and method based on trend and competition

Abstract

The invention discloses an electric power spot market quotation system based on trend and competition two dimensions, and the system comprises a data input module which is used for receiving, analyzing, verifying and formatting data needed by quotation optimization from a heterogeneous data source; the intelligent decision-making unit is used for receiving data from the data input module or a simulation result of the market clearing simulation unit and generating an adjusted quotation curve; the market clearing simulation unit is used for simulating the clearing process of the real electricity market; and the quotation output module is used for converting the final optimized quotation curve calculated by the intelligent decision-making unit from the system internal data structure into a standard format conforming to the technical specification of the power market operator, and outputting the final optimized quotation curve to the market transaction platform. According to the invention, defects in the prior art can be improved, and quotation accuracy is improved.

Description

Technical Field

[0001] This invention relates to the field of electricity market competitive bidding technology, and in particular to an electricity spot market quotation system and method based on both power flow and competition. Background Technology

[0002] Against the backdrop of power market reform, the day-ahead spot market has become the core platform for resource allocation. Power generation companies, as independent market players, are shifting their profit model from traditional planned power generation to obtaining revenue through market bidding. Under this model, power generation companies must submit a segmented bid curve consisting of multiple capacity-price pairs for each time period within the next 24 hours. Market operators then use a complex market clearing model to uniformly optimize the clearing process, aiming to minimize the social cost of electricity purchase, ultimately determining the winning output of each unit and the LMP (Lower Maximum Power Pair) for each node. Therefore, a bidding strategy that can accurately respond to market signals, fully utilize physical constraints, and effectively cope with competition is the core competitiveness for power generation companies to survive and thrive in the market.

[0003] Existing technologies for electricity spot market pricing strategies suffer from several technical shortcomings. First, existing pricing methods fail to effectively couple the physical power flow constraints of the power grid with market economic competition models, resulting in an inability to accurately respond to local market price signals generated by physical conditions such as line congestion. This not only leads to insufficient computational accuracy but also causes them to miss significant potential economic benefits. Second, when facing market environments where competitor pricing information is uncertain, existing technologies lack a systematic framework capable of adaptively switching decision logic based on information transparency. This results in poor robustness of the generated pricing strategies, making it difficult to adapt to the dynamic and volatile nature of the real market. Furthermore, most existing methods employ static or single-step optimization open-loop decision-making models, lacking a closed-loop iterative mechanism that can provide feedback, learn, and correct based on market clearing results. This leads to insufficient optimization depth, preventing them from continuously approaching the optimal solution. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a power spot market quotation system and method based on both current flow and competition, which can overcome the shortcomings of the existing technology and improve the accuracy of quotations.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows.

[0006] A power spot market pricing system based on both current flow and competition, comprising: The data input module is used to receive, parse, validate, and format the data required for quote optimization from heterogeneous data sources; The intelligent decision-making unit is used to receive data from the data input module or simulation results from the market clearing simulation unit and generate an adjusted price curve. The market clearing simulation unit is used to simulate the clearing process in the real electricity market;

[0007] The quotation output module is used to convert the final optimized quotation curve calculated by the intelligent decision-making unit from the internal data structure of the system into a standard format that conforms to the technical specifications of the power market operator, and then output it to the market trading platform.

[0008] Preferably, the data input module processes the following data types: Static data of the power grid includes the network topology of node connections, the reactance and thermal stability limits of the lines, as well as the minimum technical output of generator sets, the maximum technical output of generator sets, the ramp rate, the landslide rate, and the cost curve. Market dynamics data, including system load forecast curves for the next 24 hours, renewable energy output forecast curves, and demand for ancillary services;

[0009] Competitor data includes real-time publicly available competitor quotes, historical quote databases, and competitor intelligence obtained through third-party subscriptions.

[0010] Preferably, the objective function of the market clearing simulation unit is: Where t is the set of time periods and i is the set of conventional generator sets. Is it a conventional unit i in terms of output? The quoted cost at that time, where VOLL is the value of power loss. It represents the load reduction at node i during time period t. The gas turbine unit j is at an output of The cost of electricity generation at that time; the constraints are as follows. Node power balance constraint: Ensure that the power injection of each node at any given time equals the power outflow. , in, It is the set of generators connected to node k. and These are the charging and discharging power of the energy storage unit. It is the wind power output at node k. It is the load demand of node k. This refers to the line power flow from node k to node j, and the dual variable of this constraint. This refers to the marginal electricity price of node k in time period t;

[0011] Power flow constraints on transmission lines: Based on power flow models, power transmission on the lines is limited to prevent it from exceeding its physical limits. , , in, It is the susceptance of line ij. It is the voltage phase angle at node i. This is the upper limit of the transmission capacity of line ij. When the line power flow reaches the upper limit, this constraint becomes a tight constraint, causing congestion and directly affecting the LMP of the relevant nodes. Unit operating constraints: including upper and lower limits of generator unit output constraints The constraints include ramp rate constraints, landslide rate constraints, state of charge update constraints for energy storage units, charge and discharge power limits, and mutual exclusion constraints.

[0012] A pricing method for a dual-dimensional electricity spot market pricing system based on power flow and competition, as described above, includes the following steps: The data input module collects static power grid data, dynamic market data, and competitor data. After the data collection is completed, it generates a conservative initial segmented pricing strategy for the target generator set. The intelligent decision-making unit generates an adjusted price curve based on the data input module; The market clearing simulation unit executes a market clearing simulation operation based on the price curve. The intelligent decision-making unit analyzes the simulation results, including the expected profit of the target unit in this simulation and the congestion index of the critical transmission lines connected to the node where the target unit is located. The congestion index is defined as follows: Where L is the set of critical paths, This is the actual power flow of line l. This is its transmission limit; The intelligent decision-making unit generates a new price curve based on the analysis results of the simulation results. When the expected profit of the current price curve is greater than or equal to the expected profit of the previous iteration, or when the number of iterations reaches the preset upper limit, the latest price curve is sent to the price output module.

[0013] The quotation output module will ultimately optimize the quotation curve by converting it from the system's internal data structure into a standard format that conforms to the technical specifications of the power market operator, and then output it to the market trading platform.

[0014] Preferably, the intelligent decision-making unit generates the price quote curve by including the following steps. Neural network encoders process historical market data input time series. For each time step t, a hidden state is generated. The set of hidden states It represents the encoded information from the entire input sequence; Use a feedforward neural network for each hidden state Calculate an alignment score The alignment score quantifies the relevance of information at time step t to the current prediction task. ,in , and These are learnable weight parameters; The scores are normalized to attention weights using the softmax function. , , It is a value between 0 and 1, and ; The context vector c is the weighted average of the encoder's hidden states. ; Input the context vector c into the neural network decoder to generate the price quote curve; Using the initial pricing curve, initial temperature T, cooling rate, and maximum number of iterations. Begin iteratively calculating the new price curve, and in the k-th iteration, calculate according to the step size. Update the price curve. , The temperature at the k-th iteration. This is the preset adjustment amount; Calculate the new profit based on the new price curve. If the new profit is greater than the profit in the previous iteration, accept the new price curve; otherwise, accept the new price curve with probability p. , The profit difference between two adjacent iterations; After each iteration, the temperature is reduced according to the cooling plan. , When the temperature The number of iterations decreases to a preset threshold or the maximum number of iterations is reached. At that time, the current price curve is output as the final optimized price curve.

[0015] As a preferred option, preset adjustment amount The calculation method for the dynamically adjusted amount is as follows: , in, This is the preset adjustment amount for the k-th iteration. This is the preset adjustment amount for the (k+1)th iteration. Step size, As weight, Let U be the gradient of the direction of the fastest rate of change between the current price and the initial price after the k-th iteration, and U be a preset adjustment reference function.

[0016] The beneficial effects of adopting the above technical solution are as follows: This invention embeds a market clearing model containing complete power grid physical constraints into the bidding decision-making process, enabling the bidding strategy to respond in real time to physical signals such as line congestion. By establishing an iterative feedback optimization loop of bidding-clearing-analysis-correction, the bidding strategy can continuously approach the optimal solution through self-learning. In the iterative optimization loop, this invention transforms a single sequential information processor into a content-based historical information retrieval system. Compared to the inefficient traditional method of relying on a single hidden state to gradually transmit and compress all historical information, this invention can establish a parallel direct connection between the hidden state at each historical moment and the prediction task, thereby achieving selective weighting and dynamic retrieval of complete historical information. This mechanism allows the model to maintain the modeling advantages of local temporal context while achieving global information retrieval similar to database queries, significantly improving the ability to capture complex seasonal patterns and sudden events, especially suitable for application scenarios with multi-level time characteristics and abnormal disturbances, such as power market bidding prediction. This invention synergistically couples the adaptive step size mechanism with the temperature parameters of annealing scheduling, creating a single, unified control parameter. When the algorithm is in the high-temperature exploration phase, it automatically uses large price adjustments; when it cools down and enters the exploitation phase, it naturally tightens the search range with smaller adjustments. This creates a more coherent and efficient search process than using two independent control mechanisms, simplifies the algorithm tuning process, and provides a more robust optimization tool tailored to the continuous variable nature of price bidding. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a specific embodiment of the present invention. Detailed Implementation

[0018] Reference Figure 1 This invention provides a power spot market pricing system based on both current flow and competition, comprising: The data input module is used to receive, parse, validate, and format the data required for quote optimization from heterogeneous data sources.

[0019] The data input module processes the following data types: The static data of the power grid includes the network topology of node connections, the reactance and thermal stability limits of the lines, as well as the minimum technical output of the generator sets, the maximum technical output of the generator sets, the ramp rate, the landslide rate, and the cost curve.

[0020] Market dynamics data, including system load forecast curves for the next 24 hours, renewable energy output forecast curves, and demand for ancillary services.

[0021] Competitor data includes real-time publicly available competitor quotes, historical quote databases, and competitor intelligence obtained through third-party subscriptions.

[0022] The intelligent decision-making unit is used to receive data from the data input module or simulation results from the market clearing simulation unit and generate an adjusted price curve.

[0023] The market clearing simulation unit is used to simulate the clearing process in a real electricity market.

[0024] The objective function of the market clearing simulation unit is Where t is the set of time periods and i is the set of conventional generator sets. Is it a conventional unit i in terms of output? The quoted cost at that time, where VOLL is the value of power loss. It represents the load reduction at node i during time period t. The gas turbine unit j is at an output of The cost of generating electricity at that time. The constraints are as follows: Node power balance constraint: Ensure that the power injection of each node at any given time equals the power outflow. , in, It is the set of generators connected to node k. and These are the charging and discharging power of the energy storage unit. It is the wind power output at node k. It is the load demand of node k. This refers to the line power flow from node k to node j, and the dual variable of this constraint. This refers to the marginal electricity price of node k in time period t.

[0025] Power flow constraints on transmission lines: Based on power flow models, power transmission on the lines is limited to prevent it from exceeding its physical limits. , , in, It is the susceptance of line ij. It is the voltage phase angle at node i. It is the upper limit of the transmission capacity of line ij. When the line power flow reaches the upper limit, this constraint becomes a tight constraint, causing congestion and directly affecting the LMP of the relevant nodes.

[0026] Unit operating constraints: including upper and lower limits of generator unit output constraints The constraints include ramp rate constraints, landslide rate constraints, state of charge update constraints for energy storage units, charge and discharge power limits, and mutual exclusion constraints.

[0027] The quotation output module is used to convert the final optimized quotation curve calculated by the intelligent decision-making unit from the internal data structure of the system into a standard format that conforms to the technical specifications of the power market operator, and then output it to the market trading platform.

[0028] A pricing method for a dual-dimensional electricity spot market pricing system based on power flow and competition, as described above, includes the following steps: The data input module collects static power grid data, dynamic market data, and competitor data. After the data collection is completed, it generates a conservative initial segmented pricing strategy for the target generator set.

[0029] The intelligent decision-making unit generates an adjusted price curve based on the data input module.

[0030] The market clearing simulation unit performs a market clearing simulation operation based on the price curve.

[0031] The intelligent decision-making unit analyzes the simulation results, including the expected profit of the target unit in this simulation and the congestion index of the critical transmission lines connected to the node where the target unit is located. The congestion index is defined as follows: Where L is the set of critical paths, This is the actual power flow of line l. This is its transmission limit.

[0032] The intelligent decision-making unit generates a new pricing curve based on the analysis results of the simulation results. When the expected profit of the current pricing curve is greater than or equal to the expected profit of the previous iteration, or when the number of iterations reaches the preset upper limit, the latest pricing curve is sent to the pricing output module.

[0033] The quotation output module will ultimately optimize the quotation curve by converting it from the system's internal data structure into a standard format that conforms to the technical specifications of the power market operator, and then output it to the market trading platform.

[0034] The key to this process lies in how the intelligent decision-making unit generates the price quotation curve. This invention presents the following method for generating the price quotation curve.

[0035] Neural network encoders process historical market data input time series. For each time step t, a hidden state is generated. The set of hidden states It represents the encoded information from the entire input sequence; Use a feedforward neural network for each hidden state Calculate an alignment score The alignment score quantifies the relevance of information at time step t to the current prediction task. ,in , and These are learnable weight parameters; The scores are normalized to attention weights using the softmax function. , , It is a value between 0 and 1, and ; The context vector c is the weighted average of the encoder's hidden states. ; Input the context vector c into the neural network decoder to generate the price quote curve; Using the initial pricing curve, initial temperature T, cooling rate, and maximum number of iterations. Begin iteratively calculating the new price curve, and in the k-th iteration, calculate according to the step size. Update the price curve. , The temperature at the k-th iteration. This is the preset adjustment amount; Calculate the new profit based on the new price curve. If the new profit is greater than the profit in the previous iteration, accept the new price curve; otherwise, accept the new price curve with probability p. , The profit difference between two adjacent iterations; After each iteration, the temperature is reduced according to the cooling plan. , When the temperature The number of iterations decreases to a preset threshold or the maximum number of iterations is reached. At that time, the current price curve is output as the final optimized price curve.

[0036] This invention can also The design incorporates dynamically adjustable values ​​to achieve a bivariate dynamic step size, thereby improving search efficiency.

[0037] , in, This is the preset adjustment amount for the k-th iteration. This is the preset adjustment amount for the (k+1)th iteration. Step size, As weight, Let U be the gradient of the direction of the fastest rate of change between the current price and the initial price after the k-th iteration, and U be a preset adjustment reference function.

[0038] The bivariate dynamic step size fundamentally changes the dynamic characteristics of the system, transforming it from a simple local improvement search into a stable trajectory search.

[0039] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A spot electricity market pricing system based on both current flow and competition, characterized in that: include, The data input module is used to receive, parse, validate, and format the data required for quote optimization from heterogeneous data sources; The intelligent decision-making unit is used to receive data from the data input module or simulation results from the market clearing simulation unit and generate an adjusted price curve. The market clearing simulation unit is used to simulate the clearing process in the real electricity market; The quotation output module is used to convert the final optimized quotation curve calculated by the intelligent decision-making unit from the internal data structure of the system into a standard format that conforms to the technical specifications of the power market operator, and then output it to the market trading platform.

2. The electricity spot market quotation system based on a dual-dimensional approach of power flow and competition as described in claim 1, characterized in that: The data input module processes the following data types: Static data of the power grid includes the network topology of node connections, the reactance and thermal stability limits of the lines, as well as the minimum technical output of generator sets, the maximum technical output of generator sets, the ramp rate, the landslide rate, and the cost curve. Market dynamics data, including system load forecast curves for the next 24 hours, renewable energy output forecast curves, and demand for ancillary services; Competitor data includes real-time publicly available competitor quotes, historical quote databases, and competitor intelligence obtained through third-party subscriptions.

3. The electricity spot market quotation system based on a dual-dimensional approach of power flow and competition as described in claim 1, characterized in that: The objective function of the market clearing simulation unit is: Where t is the set of time periods and i is the set of conventional generator sets. Is it a conventional unit i in terms of output? The quoted cost at that time, where VOLL is the value of power loss. It represents the load reduction at node i during time period t. The gas turbine unit j is at an output of The cost of electricity generation at that time; the constraints are as follows. Node power balance constraint: Ensure that the power injection of each node at any given time equals the power outflow. ， in, It is the set of generators connected to node k. and These are the charging and discharging power of the energy storage unit. It is the wind power output at node k. It is the load demand of node k. This refers to the line power flow from node k to node j, and the dual variable of this constraint. This refers to the marginal electricity price of node k in time period t; Power flow constraints on transmission lines: Based on power flow models, power transmission on the lines is limited to prevent it from exceeding its physical limits. ， ， in, It is the susceptance of line ij. It is the voltage phase angle at node i. This is the upper limit of the transmission capacity of line ij. When the line power flow reaches the upper limit, this constraint becomes a tight constraint, causing congestion and directly affecting the LMP of the relevant nodes. Unit operating constraints: including upper and lower limits of generator unit output constraints The constraints include ramp rate constraints, landslide rate constraints, state of charge update constraints for energy storage units, charge and discharge power limits, and mutual exclusion constraints.

4. A pricing method for a dual-dimensional electricity spot market pricing system based on power flow and competition, as described in any one of claims 1-3, characterized in that... Includes the following steps: The data input module collects static power grid data, dynamic market data, and competitor data. After the data collection is completed, it generates a conservative initial segmented pricing strategy for the target generator set. The intelligent decision-making unit generates an adjusted price curve based on the data input module; The market clearing simulation unit executes a market clearing simulation operation based on the price curve. The intelligent decision-making unit analyzes the simulation results, including the expected profit of the target unit in this simulation and the congestion index of the critical transmission lines connected to the node where the target unit is located. The congestion index is defined as follows: Where L is the set of critical paths, This is the actual power flow of line l. This is its transmission limit; The intelligent decision-making unit generates a new price curve based on the analysis results of the simulation results. When the expected profit of the current price curve is greater than or equal to the expected profit of the previous iteration, or when the number of iterations reaches the preset upper limit, the latest price curve is sent to the price output module. The quotation output module will ultimately optimize the quotation curve by converting it from the system's internal data structure into a standard format that conforms to the technical specifications of the power market operator, and then output it to the market trading platform.

5. The pricing method for a dual-dimensional electricity spot market pricing system based on power flow and competition as described in claim 4, characterized in that: The intelligent decision-making unit generates the price quote curve through the following steps: Neural network encoders process historical market data input time series. For each time step t, a hidden state is generated. The set of hidden states It represents the encoded information from the entire input sequence; Use a feedforward neural network for each hidden state Calculate an alignment score The alignment score quantifies the relevance of information at time step t to the current prediction task. ,in , and These are learnable weight parameters; The scores are normalized to attention weights using the softmax function. , ， It is a value between 0 and 1, and ; The context vector c is the weighted average of the encoder's hidden states. ； Input the context vector c into the neural network decoder to generate the price quote curve; Using the initial pricing curve, initial temperature T, cooling rate, and maximum number of iterations. Begin iteratively calculating the new price curve, and in the k-th iteration, calculate according to the step size. Update the price curve. , The temperature at the k-th iteration. This is the preset adjustment amount; Calculate the new profit based on the new price curve. If the new profit is greater than the profit in the previous iteration, accept the new price curve; otherwise, accept the new price curve with probability p. , The profit difference between two adjacent iterations; After each iteration, the temperature is reduced according to the cooling plan. , When the temperature The number of iterations decreases to a preset threshold or the maximum number of iterations is reached. At that time, the current price curve is output as the final optimized price curve.

6. The pricing method for a dual-dimensional electricity spot market pricing system based on power flow and competition as described in claim 5, characterized in that: Preset adjustment amount The calculation method for the dynamically adjusted amount is as follows: ， in, This is the preset adjustment amount for the k-th iteration. This is the preset adjustment amount for the (k+1)th iteration. Step size, As weight, Let U be the gradient of the direction of the fastest rate of change between the current price and the initial price after the k-th iteration, and U be a preset adjustment reference function.

Images