Search

CN-121484926-B - Dynamic pricing method, device, equipment and storage medium for distributed photovoltaic power station

CN121484926BCN 121484926 BCN121484926 BCN 121484926BCN-121484926-B

Abstract

The embodiment of the invention relates to the technical field of artificial intelligence and new energy, and discloses a dynamic pricing method, a dynamic pricing device, dynamic pricing equipment and a dynamic pricing storage medium for a distributed photovoltaic power station, wherein the method comprises a comprehensive health index determining step for determining the comprehensive health index of the photovoltaic power station based on multi-source heterogeneous data; the method comprises the steps of predicting the residual service life, obtaining a historical health state sequence of photovoltaic power station equipment, predicting the residual service life of the equipment according to the health state sequence and a preset sequence prediction model, constructing a pricing model, constructing a dynamic pricing model according to comprehensive health indexes and the residual service life, predicting the future failure probability of the photovoltaic power station, determining the future maintenance cost of the photovoltaic power station according to the future failure probability, and determining the pricing of the photovoltaic power station according to the future maintenance cost and the market correction value in the pricing model. By the aid of the method, accuracy and efficiency of pricing of the photovoltaic power station can be improved.

Inventors

  • ZHAO ZHANGFENG

Assignees

  • 深圳市润世华智数技术有限公司

Dates

Publication Date
20260508
Application Date
20260106

Claims (9)

  1. 1. A method for dynamic pricing of a distributed photovoltaic power plant, the method comprising: a comprehensive health index determining step of collecting multi-source heterogeneous data of a distributed photovoltaic power station, and determining the comprehensive health index of the photovoltaic power station based on the multi-source heterogeneous data, wherein the multi-source heterogeneous data at least comprises equipment data, operation monitoring data, environment data, historical maintenance data, performance evaluation data and market data of the photovoltaic power station; A residual service life prediction step of obtaining a historical health state sequence of the photovoltaic power station equipment and predicting the residual service life of the equipment according to the health state sequence and a preset sequence prediction model; a pricing model constructing step, namely constructing a dynamic pricing model according to the comprehensive health index and the residual service life, wherein the pricing model constructing step comprises the following steps: Determining a basic value of the photovoltaic power station according to the residual service life; correcting the basic value according to the comprehensive health index to obtain a health correction value; Obtaining a market dynamic factor of the photovoltaic power station, and carrying out market correction according to the market dynamic factor and the health correction value to obtain a market correction value; A pricing step of predicting the future failure probability of the photovoltaic power station, determining the future maintenance cost of the photovoltaic power station according to the future failure probability, and determining the pricing of the photovoltaic power station according to the future maintenance cost and the market correction value in the pricing model; The determining the base value of the photovoltaic power plant from the remaining useful life comprises: Obtaining a life cost function based on the residual service life, the design life of the photovoltaic power station and a deep learning model, and determining the basic value based on the life cost function and the market price of the newly built photovoltaic power station with the same scale; Wherein, the basic value v=v_new_f_life (R, l_design), v_new is the market price of the newly built photovoltaic power plant with the same scale, f_life is the life cost function, R is the remaining service life, and l_design is the design life of the photovoltaic power plant; f_life (R, l_design) =a (R/l_design) + (1-a) (R/l_design)/(2, a being the fitting parameter; The base value is corrected according to the comprehensive health index to obtain a health correction value, which comprises the steps of determining a health correction function based on the comprehensive health index and the variance of the comprehensive health index, determining the health correction value based on the health correction function and the base value, wherein the health correction value V_health=V x g_health (H, sigma), Wherein g_health is a health correction function, H is the comprehensive health index, and sigma is the variance of the comprehensive health index; The market dynamic factors comprise market supply and demand balance factors and technical iteration depreciation factors, the market supply and demand balance factors are obtained based on a preset time sequence network model prediction, the market dynamic factors of the photovoltaic power station are obtained, market correction is carried out according to the market dynamic factors and the health correction value, and the market correction value is obtained, wherein the market correction is carried out according to the market supply and demand balance factors, the technical iteration depreciation factors and the health correction value, and the market correction value V_mark is obtained: V_market=V_health*S_market*D_tech, wherein S_mark is the market supply and demand balance factor, and D_tech is the technology iteration depreciation factor.
  2. 2. The method of claim 1, wherein the integrated health index determining step comprises: Collecting multi-source heterogeneous data of a distributed photovoltaic power station, determining a component health index, an inverter health index, a system health index and a reliability index of the photovoltaic power station based on the multi-source heterogeneous data and a preset deep learning model, wherein the training of the deep learning model adopts a multi-task learning strategy for training; Obtaining a weight combination based on training automatic learning of the deep learning model, and determining the comprehensive health index H according to the weight combination, the component health index, the inverter health index, the system health index and the reliability index: , H1 is the component health index, H2 is the inverter health index, H3 is the system health index, H4 is the reliability index, and the weights in the weight combination include w1, w2, w3 and w4, w1+w2+w3+w4=1.
  3. 3. The method of claim 1, wherein the sequence prediction model is a time series convolution network model comprising a plurality of blocks in series, the blocks comprising a residual connection and a plurality of causal expansion convolution layers having different expansion rates, the plurality of causal expansion convolution layers capturing time series characteristics of the health state sequence on a plurality of time scales from short term to long term simultaneously, fusing outputs of the plurality of causal expansion convolution layers as output characteristics of the blocks.
  4. 4. The method of claim 3, wherein the time series convolution network model is trained by combining a loss function of a physical law, the loss function L being l=l_pred+γ x l_physics, wherein l_pred is a prediction error, l_physics is a physical constraint term, γ is a weight parameter, the physical constraint term is constructed based on a physical degradation model of the device, and performance degradation of the device is subject to a weibull distribution.
  5. 5. The method of claim 1, wherein the pricing step comprises calculating a pricing P for the photovoltaic power plant by the formula: , Wherein V_market is the market revision value, PV (C_future) is the future maintenance cost, C_future is the future failure probability, and μ is the profit margin of the seller.
  6. 6. The method according to claim 1, wherein the method further comprises: a deep reinforcement learning algorithm is adopted to construct a selling occasion decision model of the photovoltaic power station, and the optimal selling time of the photovoltaic power station is determined based on pricing of the photovoltaic power station and the selling occasion decision model, wherein the selling occasion decision model determines the optimal selling time t by maximizing a cumulative discount prize: Wherein gamma is a discount factor, and R (t) is a rewarding function, which is obtained based on pricing, accumulated holding cost, opportunity cost and punishment items of the photovoltaic power station.
  7. 7. A distributed photovoltaic power plant dynamic pricing device, the device comprising: the comprehensive health index determining module is used for collecting multi-source heterogeneous data of the distributed photovoltaic power station, determining the comprehensive health index of the photovoltaic power station based on the multi-source heterogeneous data, wherein the multi-source heterogeneous data at least comprises equipment data, operation data, environment data, historical maintenance data, performance evaluation data and market data of the photovoltaic power station; The residual service life prediction module is used for acquiring a health state sequence of the photovoltaic power station equipment history and predicting the residual service life of the equipment according to the health state sequence and a preset sequence prediction model; the building module is used for building a dynamic pricing model according to the comprehensive health index and the residual service life, and the building module is specifically used for: determining the basic value of the photovoltaic power station according to the residual service life: Obtaining a life cost function based on the residual service life, the design life of the photovoltaic power station and a deep learning model, and determining the basic value based on the life cost function and the market price of the newly built photovoltaic power station with the same scale; Wherein, the basic value v=v_new_f_life (R, l_design), v_new is the market price of the newly built photovoltaic power plant with the same scale, f_life is the life cost function, R is the remaining service life, and l_design is the design life of the photovoltaic power plant; f_life (R, l_design) =a (R/l_design) + (1-a) (R/l_design)/(2, a being the fitting parameter; Correcting the basic value according to the comprehensive health index to obtain a health correction value, wherein a health correction function is determined based on the comprehensive health index and the variance of the comprehensive health index, the health correction value is determined based on the health correction function and the basic value, the health correction value V_health=V.g_health (H, sigma), Wherein g_health is a health correction function, H is the comprehensive health index, and sigma is the variance of the comprehensive health index; Obtaining a market supply and demand balance factor and a technology iteration depreciation factor of the photovoltaic power station, and carrying out market correction according to the market supply and demand balance factor, the technology iteration depreciation factor and the health correction value to obtain a market correction value V_mark, wherein the market supply and demand balance factor is predicted based on a preset time sequence network model to obtain: V_market=V_health*S_market*D_tech, Wherein S_mark is the market supply and demand balance factor, and D_tech is the technology iteration depreciation factor; And the pricing module is used for predicting the future failure probability of the photovoltaic power station, determining the future maintenance cost of the photovoltaic power station according to the future failure probability, and determining the pricing of the photovoltaic power station according to the future maintenance cost and the market correction value in the pricing model.
  8. 8. The computer equipment is characterized by comprising a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete communication with each other through the communication bus; The memory is configured to hold at least one executable instruction that causes the processor to perform the method of any one of claims 1-6.
  9. 9. A computer readable storage medium having stored therein at least one executable instruction which, when executed on a computer device, causes the computer device to perform the method of any of claims 1-6.

Description

Dynamic pricing method, device, equipment and storage medium for distributed photovoltaic power station Technical Field The embodiment of the invention relates to the technical field of artificial intelligence and new energy, in particular to a dynamic pricing method, a dynamic pricing device, dynamic pricing equipment and a dynamic pricing storage medium for a distributed photovoltaic power station. Background With the acceleration of energy conversion, distributed photovoltaic power stations are rapidly developing as an important component of clean energy. The current distributed photovoltaic installed capacity has exceeded 200GW, creating a vast stock asset market. However, in pricing of photovoltaic power plants, the following technical problems exist: The existing pricing method is static, mainly adopts manual experience, a simple annual fold method, a capacity unit price method or a cash value method and the like, and cannot accurately quantify the actual health state of equipment due to the fact that single index or single model is adopted for pricing, and neglects multidimensional factors such as nonlinear attenuation characteristics of photovoltaic modules, complex aging mechanisms of inverters, accumulated influences of environmental factors on service life of equipment and the like, so that the pricing model is single, and the pricing result accuracy and efficiency are low. Disclosure of Invention In view of the above problems, embodiments of the present invention provide a method, an apparatus, a device, and a storage medium for dynamic pricing of a distributed photovoltaic power station, which are used to solve the problems existing in the prior art. According to an aspect of an embodiment of the present invention, there is provided a distributed photovoltaic power plant dynamic pricing method, the method comprising: a comprehensive health index determining step of collecting multi-source heterogeneous data of a distributed photovoltaic power station, and determining the comprehensive health index of the photovoltaic power station based on the multi-source heterogeneous data, wherein the multi-source heterogeneous data at least comprises equipment data, operation monitoring data, environment data, historical maintenance data, performance evaluation data and market data of the photovoltaic power station; A residual service life prediction step of obtaining a historical health state sequence of the photovoltaic power station equipment and predicting the residual service life of the equipment according to the health state sequence and a preset sequence prediction model; a pricing model constructing step, namely constructing a dynamic pricing model according to the comprehensive health index and the residual service life, wherein the pricing model constructing step comprises the following steps: Determining a basic value of the photovoltaic power station according to the residual service life; correcting the basic value according to the comprehensive health index to obtain a health correction value; Obtaining a market dynamic factor of the photovoltaic power station, and carrying out market correction according to the market dynamic factor and the health correction value to obtain a market correction value; And a pricing step of predicting the future failure probability of the photovoltaic power station, determining the future maintenance cost of the photovoltaic power station according to the future failure probability, and determining the pricing of the photovoltaic power station according to the future maintenance cost and the market correction value in the pricing model. In an alternative manner, the comprehensive health index determining step includes: Collecting multi-source heterogeneous data of a distributed photovoltaic power station, determining a component health index, an inverter health index, a system health index and a reliability index of the photovoltaic power station based on the multi-source heterogeneous data and a preset deep learning model, wherein the training of the deep learning model adopts a multi-task learning strategy for training; Obtaining a weight combination based on training automatic learning of the deep learning model, and determining the comprehensive health index H according to the weight combination, the component health index, the inverter health index, the system health index and the reliability index: , H1 is the component health index, H2 is the inverter health index, H3 is the system health index, H4 is the reliability index, and the weights in the weight combination include w1, w2, w3 and w4, w1+w2+w3+w4=1. In an alternative manner, the sequence prediction model is a time sequence convolution network model, and the time sequence convolution network model comprises a plurality of blocks connected in series, wherein the blocks comprise residual connection and a plurality of causal expansion convolution layers with different expansion rates, the causal expa