Intelligent evaluation system of photovoltaic power generation system
Through the intelligent evaluation system of photovoltaic power generation system, the PST-Transformer model and multi-feature fusion strategy are used to solve the problem of lack of time-variability considerations in the existing photovoltaic power generation system evaluation methods, and high-precision photovoltaic power generation system prediction and reliability evaluation are achieved.
Patent Information
- Application Number
- CN202510040697.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-10
- Publication Date
- 2025-05-13
AI Technical Summary
The reliability evaluation methods of existing photovoltaic power generation systems lack consideration of time-variability of photovoltaic power generation, resulting in inaccurate evaluation results, and traditional methods are difficult to capture instantaneous fluctuations and complex nonlinear relationships.
A smart evaluation system for photovoltaic power generation system is proposed. Through the photovoltaic intelligent prediction and diagnosis subsystem and the photovoltaic power generation multi-state evaluation subsystem, combined with the PST-Transformer model and multi-feature fusion strategy, a prediction and diagnosis probability model of the system output power is established, and the reliability evaluation of the photovoltaic power generation system is carried out.
It improves the accuracy of power prediction of photovoltaic power generation systems, can predict future failure trends based on the historical operating status of the system, provides scientific energy allocation and planning decision support, and improves the reliability evaluation capabilities of the system.
Smart Images
Figure CN119989042A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of photovoltaic power generation reliability assessment and relates to an intelligent assessment system for a photovoltaic power generation system. Background Art
[0002] With the large-scale development of photovoltaic power generation, the reliability level of photovoltaic power stations has received more and more attention. However, the rapid development of photovoltaic power generation has also brought corresponding problems, among which the power generation performance and system reliability of photovoltaic power generation systems are the most prominent. At present, the reliability assessment methods for photovoltaic power generation systems mainly include analytical method and simulation method. The analytical method equates photovoltaic power stations to conventional power generation systems with multiple states and establishes a system outage capacity probability model. However, this method lacks consideration of the time-varying nature of photovoltaic power generation, so its assessment results cannot fully reflect photovoltaic power generation, a power generation system with output that varies over time.
[0003] Photovoltaic power prediction is crucial for smart grids and energy dispatching, which can optimize energy utilization and reduce costs. However, there are unstable factors such as weather changes and shadow effects in the prediction process, which lead to inaccurate predictions. In addition, traditional methods have difficulty in modeling complex nonlinear relationships and capturing instantaneous fluctuations. Therefore, it is necessary to combine advanced data-driven algorithms for high-precision photovoltaic power generation system modeling and diagnosis to improve the accuracy of system power generation prediction and predict future fault development trends based on the system's historical operating status. The power prediction results are integrated with the diagnosis results to judge the reliability of the photovoltaic system, providing important decision-making support for relevant departments to deploy scientific energy allocation and planning in advance. Summary of the invention
[0004] In view of the above problems, the purpose of the present invention is to propose an intelligent evaluation system for photovoltaic power generation system, which analyzes the operation mechanism of photovoltaic power generation system, component failure mode and its impact on output power through photovoltaic intelligent predictive diagnosis, establishes a predictive diagnosis probability model of system output power, thereby obtaining the reliability evaluation index of photovoltaic power generation system, and constructs a six-state space model of photovoltaic power generation system to guide the scientific operation and maintenance of actual photovoltaic power generation system.
[0005] The technical solution of the present invention is: a photovoltaic power generation system intelligent evaluation system described in the present invention includes a photovoltaic intelligent prediction and diagnosis subsystem and a photovoltaic power generation multi-state evaluation subsystem;
[0006] The photovoltaic intelligent prediction and diagnosis subsystem includes a data acquisition unit, a data processing unit, an intelligent prediction unit and a fault diagnosis unit that are interconnected;
[0007] The data acquisition unit is used to collect longitude and latitude, illumination angle, solar radiation, temperature, air pressure, atmospheric humidity, system voltage, and current characteristic information to provide a basis for input into the model for modeling and prediction;
[0008] The data processing unit is used for interpolation of missing data caused by, but not limited to, sensor failure, network freeze and electromagnetic interference, and smoothing of abnormal data;
[0009] The intelligent forecasting unit models and predicts the collected characteristic information, and determines the availability status of the photovoltaic power generation system through the predicted photovoltaic power;
[0010] The fault diagnosis unit monitors the system components in real time, determines the system operation status through the current and voltage changes, and predicts the time of future faults in order to arrange planned maintenance of the system.
[0011] Furthermore, the photovoltaic intelligent prediction and diagnosis subsystem takes into account the time-varying and spatial heterogeneity of illumination, monitors the characteristic information of illumination angle, solar radiation, temperature, air pressure, and atmospheric humidity at different longitudes and latitudes, and inputs the information into the PST-Transformer model established in the intelligent forecasting unit after multi-feature fusion to predict photovoltaic power. At the same time, the voltage and current characteristic information of the system is collected through sensors, and input into the fault diagnosis unit after multi-feature fusion. By establishing the PST-Transformer model, the operating status of the equipment is accurately identified, the fault trend is predicted, and planned fault inspection and maintenance are carried out.
[0012] Furthermore, the photovoltaic power generation multi-state assessment subsystem divides the photovoltaic power generation system state into multiple types, determines the availability state of the photovoltaic power generation system through the prediction data output by the photovoltaic intelligent prediction and diagnosis subsystem, and performs reliability assessment on the photovoltaic power generation system.
[0013] Furthermore, in the photovoltaic power generation multi-state assessment subsystem, the photovoltaic power generation system status is divided into two categories: operation and shutdown; the operation status is divided into full-rated status and reduced-rated status according to the size of the photovoltaic power; the reduced-rated status can be divided into two types according to its cause: low power and partial failure of photovoltaic components; the shutdown status can be divided into two types: shutdown under power restriction and shutdown caused by failure.
[0014] Furthermore, the multi-feature fusion is achieved through three strategies. The first strategy is the multi-scale spatial feature reorganization strategy MFS, which is used to reorganize the spatial scale information of the input feature variable; the second strategy is the attention mechanism strategy AS, which is used to extract the local information of the input feature variable; the last strategy is the input feature information cross fusion strategy CS. The three strategies are characterized as follows:
[0015]
[0016] In the formula, is the input feature, i represents the time scale, j represents the spatial position, I is the dimension of the feature variable, Y ASis the attention weight matrix that records the target position information, Output features for the multi-scale information extraction module.
[0017] Furthermore, the calculation formula of the PST-Transformer model is as follows:
[0018]
[0019] Where W1,b1 and W2,b2 are the weights and bias terms of the two fully connected layers, Q, K and V are the query matrix, key matrix and value matrix respectively, and W Q , W K and W V are the weight matrices of query, key, and value, respectively. T represents transposition, and d k It is the dimension of attention.
[0020] Furthermore, the intelligent forecasting unit uses the PST-Transformer model to forecast the photovoltaic power generation system and uses the PO algorithm to find the optimal parameter set of key parameters of the PST-Transformer model;
[0021] The PO algorithm uses exploration, development, update and reverse search operations to calculate and sort the initial key parameter group of the PST-Transformer model, and randomly exhibits one of the four behaviors during each iteration. After meeting the maximum number of iterations, the optimal initial key parameter group of the PST-Transformer model is output. The formula is as follows:
[0022] Key parameter groups in the PST-Transformer model Initially set to:
[0023]
[0024] The calculation formula for searching the optimal parameter group based on the initial parameters is:
[0025]
[0026] Development formula:
[0027]
[0028] According to the size of the random number P, the key parameters of the PST-Transformer model are updated, and the calculation formula is:
[0029]
[0030] The formula for reverse search of the key parameter group of the PST-Transformer model is:
[0031]
[0032] Where lb is the lower limit of the search space for key parameters of the PST-Transformer model, ub is the upper limit of the search space for key parameters of the PST-Transformer model, rand(0,1) and P are random numbers in the range of [0,1]. is the initial stage PST-Transformer model parameter group, is the current model parameter set, is the model parameter group to be updated later, ones(1,dim) is a vector of all 1s with dimension dim, is the average value of the key parameter group, Levy(dim) is the Levy distribution, which represents the change process of the key parameter group of the PST-Transformer model, N best is the optimal position of the key parameter group of the PST-Transformer model, Max iter is the maximum number of iterations, and t is the current number of iterations.
[0033] Furthermore, the key parameter group of the PST-Transformer model includes but is not limited to the weights and biases of the self-attention layer, the weights and biases of the MLP layer, the scaling parameters and the translation parameters.
[0034] Furthermore, the PST-Transformer model uses temporal position coding and spatial position coding; temporal position coding assigns the same time coding to data collected at the same time, helping the self-attention mechanism to better identify data acquired at the same time; and spatial position coding assigns the same position coding to data acquired at the same location, so that the model understands the relationship between data from the same monitoring site;
[0035] These two position codes are connected to the input data and then input into the PST-Transformer model for processing, so that the model can more comprehensively consider the spatiotemporal information in the data. The formula is as follows:
[0036]
[0037] Where, PE (pos) is the original position encoding of Transformer, pos is the position of feature information data in the input sequence, a is the position of the input sequence, c is the dimension index of the model, d is the dimension of the position encoding, T_PE (pos) is the time position code, T_V is the time when each data in the data set was collected, T_E is a list with time codes, S_PE (pos)is the spatial position code, L_V is the location where each data in the dataset is collected, and L_E is a list with position codes.
[0038] Furthermore, the reliability evaluation adopts the following reliability index: failure rate λ T , that is, the number of failures that occurred in the photovoltaic power generation system during the evaluation period; the average repair time r T , which is the average repair time after a photovoltaic power generation system fails; the actual availability rate A a , that is, the probability that the photovoltaic power generation system is in normal operation during the assessment period; the full operation rate R FR , that is, the probability that the photovoltaic power generation system can produce corresponding output under the corresponding power conditions; the resource limitation derating rate R RCP , that is, the probability of derating due to power restriction during normal operation of the photovoltaic power generation system; the derating rate of equipment failure R FCP , that is, the probability that the photovoltaic power generation system will operate at a reduced rating due to a fault.
[0039] The beneficial effects of the present invention are: 1. The present invention predicts photovoltaic power by using the PST-Transformer model, taking into account the time-varying and spatial heterogeneity of illumination. After multi-feature fusion, the photovoltaic prediction is more accurate, and photovoltaic prediction can also be performed on photovoltaic power generation sites in different regions; 2. The present invention introduces innovative spatiotemporal position coding. Compared with the traditional position coding method, it not only considers time information, but also takes into account spatial information, so that the model can more accurately capture the spatiotemporal dependency in the data; 3. Based on the structure, failure mode and light intensity of the photovoltaic power generation system, the present invention establishes the output probability, failure probability and six-state space model of the photovoltaic power generation system and proposes a reliability evaluation index for the photovoltaic power generation system, which provides a reference for analyzing the impact of the photovoltaic power generation system on the grid reliability after the grid connection. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 It is a structural block diagram of the photovoltaic power generation system intelligent evaluation system of the present invention;
[0041] Figure 2 It is a flow chart of the PO algorithm of the present invention. DETAILED DESCRIPTION
[0042] The specific technical scheme of the present invention is further described in detail below with reference to specific examples.
[0043] As shown in the figure, a photovoltaic power generation system intelligent evaluation system according to the present invention has a structural schematic diagram as shown in Figure 1 As shown; the system includes:
[0044] The photovoltaic intelligent prediction and diagnosis subsystem takes into account the time-varying and spatial heterogeneity of light. It monitors the characteristic information of light angle, solar radiation, temperature, air pressure, and atmospheric humidity at different longitudes and latitudes, and inputs the information into the PST-Transformer model established in the intelligent forecasting unit after multi-feature fusion to predict photovoltaic power. At the same time, the system voltage and current characteristic information is collected through sensors, and input into the fault diagnosis unit after multi-feature fusion. The PST-Transformer model is established to accurately identify the equipment operation status and predict the fault trend, so as to carry out planned fault inspection and maintenance.
[0045] The photovoltaic power generation multi-state assessment subsystem divides the photovoltaic power generation system status into multiple types, judges the availability status of the photovoltaic power generation system through the prediction data output by the photovoltaic intelligent prediction and diagnosis subsystem, and conducts reliability assessment on the photovoltaic power generation system.
[0046] Furthermore, the multi-feature fusion is achieved through three strategies. The first strategy is the multi-scale spatial feature reorganization strategy MFS, which is mainly used to reorganize the spatial scale information of the input feature variable; the second strategy is the attention mechanism strategy AS, which is used to extract the local information of the input feature variable; the last strategy is the input feature information cross fusion strategy CS. The three strategies are characterized as follows:
[0047]
[0048] In the formula, is the input feature, i represents the time scale, j represents the spatial position, I is the dimension of the feature variable, Y AS is the attention weight matrix that records the target position information, Output features for the multi-scale information extraction module.
[0049] Furthermore, the calculation formula of the PST-Transformer model is as follows:
[0050]
[0051] Where W1,b1 and W2,b2 are the weights and bias terms of the two fully connected layers, Q, K and V are the query matrix, key matrix and value matrix respectively, and W Q , W K and W V are the weight matrices of query, key, and value, respectively. T represents transposition, and d k It is the dimension of attention.
[0052] Furthermore, the intelligent forecasting unit uses the PST-Transformer model to predict the photovoltaic power generation system. In order to avoid the subjectivity of manual parameter adjustment and improve the intelligence of the system, the PO algorithm is used to find the optimal parameter set of the key parameters of the PST-Transformer model;
[0053] The key parameter groups of the PST-Transformer model include but are not limited to the weights and biases of the self-attention layer, the weights and biases of the MLP layer, the scaling parameters, and the translation parameters.
[0054] The PO algorithm of the present invention has a flow chart as shown in FIG. Figure 2 As shown; the PO algorithm uses exploration, development, update and reverse search operations to calculate and sort the initial key parameter group of the PST-Transformer model, and randomly exhibits one of the four behaviors during each iteration. After meeting the maximum number of iterations, the optimal initial key parameter group of the PST-Transformer model is output. The formula is as follows:
[0055] Key parameter groups in the PST-Transformer model Initially set to:
[0056]
[0057] The calculation formula for searching the optimal parameter group based on the initial parameters is:
[0058]
[0059] Development formula:
[0060]
[0061] According to the size of the random number P, the key parameters of the PST-Transformer model are updated, and the calculation formula is:
[0062]
[0063] The formula for reverse search of the key parameter group of the PST-Transformer model is:
[0064]
[0065] Where lb is the lower limit of the search space for key parameters of the PST-Transformer model, ub is the upper limit of the search space for key parameters of the PST-Transformer model, rand(0,1) and P are random numbers in the range of [0,1]. is the initial stage PST-Transformer model parameter group, is the current model parameter set, is the model parameter group to be updated later, ones(1,dim) is a vector of all 1s with dimension dim, is the average value of the key parameter group, Levy(dim) is the Levy distribution, which represents the change process of the key parameter group of the PST-Transformer model, N best is the optimal position of the key parameter group of the PST-Transformer model, Max iter is the maximum number of iterations, and t is the current number of iterations.
[0066] Furthermore, the PST-Transformer model uses temporal position coding and spatial position coding; temporal position coding assigns the same time coding to data collected at the same time, helping the self-attention mechanism to better identify data acquired at the same time; and spatial position coding assigns the same position coding to data acquired at the same location, so that the model can understand the relationship between data from the same monitoring site; these two position codings are connected with the input data and then input into the PST-Transformer model for processing, so that the model can more comprehensively consider the temporal and spatial information in the data, and the formula is as follows:
[0067]
[0068] Where, PE (pos) is the original position encoding of Transformer, pos is the position of feature information data in the input sequence, a is the position of the input sequence, c is the dimension index of the model, d is the dimension of the position encoding, T_PE (pos) is the time position code, T_V is the time when each data in the data set was collected, T_E is a list with time codes, S_PE (pos) is the spatial position code, L_V is the location where each data in the dataset is collected, and L_E is a list with position codes.
[0069] Furthermore, the photovoltaic intelligent prediction and diagnosis subsystem includes a data acquisition unit, a data processing unit, an intelligent prediction unit and a fault diagnosis unit;
[0070] The data acquisition unit is used to collect longitude and latitude, illumination angle, solar radiation, temperature, air pressure, atmospheric humidity, system voltage, and current characteristic information to provide a basis for input into the model for modeling and prediction;
[0071] The data processing unit is used for interpolation of missing data caused by, but not limited to, sensor failure, network freeze and electromagnetic interference, and smoothing of abnormal data;
[0072] The intelligent forecasting unit models and predicts the collected characteristic information, and determines the availability status of the photovoltaic power generation system through the predicted photovoltaic power;
[0073] The fault diagnosis unit monitors the system components in real time, determines the system operation status through the current and voltage changes, and predicts the time of future faults in order to arrange planned maintenance of the system.
[0074] Furthermore, in the photovoltaic power generation multi-state assessment system, the photovoltaic power generation system state can be divided into two categories: operation and shutdown; the operation state can be divided into full-rated state and reduced-rated state according to the size of the photovoltaic power; the reduced-rated state can be divided into two types according to its cause: low power and partial failure of photovoltaic components; the shutdown state can be divided into two types: shutdown under power restriction and shutdown caused by failure.
[0075] Furthermore, the reliability evaluation adopts the following reliability index: failure rate λ T , that is, the number of failures that occurred in the photovoltaic power generation system during the evaluation period; the average repair time r T , which is the average repair time after a photovoltaic power generation system fails; the actual availability rate A a , that is, the probability that the photovoltaic power generation system is in normal operation during the assessment period; the full operation rate R FR , that is, the probability that the photovoltaic power generation system can produce corresponding output under the corresponding power conditions; the resource limitation derating rate R RCP , that is, the probability of derating due to power restriction during normal operation of the photovoltaic power generation system; the derating rate of equipment failure R FCP , that is, the probability that the photovoltaic power generation system will operate at a reduced rating due to a fault.
Claims
1. A photovoltaic power generation system intelligent evaluation system, characterized in that: Including photovoltaic intelligent prediction and diagnosis subsystem and photovoltaic power generation multi-state evaluation subsystem; The photovoltaic intelligent prediction and diagnosis subsystem includes a data acquisition unit, a data processing unit, an intelligent prediction unit and a fault diagnosis unit that are interconnected; The data acquisition unit collects information on longitude and latitude, illumination angle, solar radiation, temperature, air pressure, atmospheric humidity, system voltage and current characteristics to provide a basis for input into the model for modeling and prediction; The data processing unit interpolates missing data and smoothes abnormal data including but not limited to those caused by sensor failure, network freeze and electromagnetic interference; The intelligent forecasting unit models and predicts the collected characteristic information, and determines the availability status of the photovoltaic power generation system through the predicted photovoltaic power; The fault diagnosis unit monitors the system components in real time, determines the system operation status through the current and voltage changes, and predicts the time of future faults in order to arrange planned maintenance of the system.
2. A photovoltaic power generation system intelligent evaluation system according to claim 1, characterized in that: The photovoltaic intelligent prediction and diagnosis subsystem considers the time-varying and spatial heterogeneity of illumination, monitors the characteristic information of illumination angle, solar radiation, temperature, air pressure, and atmospheric humidity at different longitudes and latitudes, and inputs the characteristic information into the PST-Transformer model established in the intelligent forecasting unit after multi-feature fusion to predict photovoltaic power. At the same time, the characteristic information of system voltage and current is collected through sensors, and input into the fault diagnosis unit after multi-feature fusion. By establishing the PST-Transformer model, the operating status of the equipment is accurately identified and the fault trend is predicted, so as to carry out planned fault inspection and maintenance.
3. A photovoltaic power generation system intelligent evaluation system according to claim 1, characterized in that: The photovoltaic power generation multi-state assessment subsystem divides the photovoltaic power generation system state into multiple types, judges the availability state of the photovoltaic power generation system through the prediction data output by the photovoltaic intelligent prediction and diagnosis subsystem, and performs reliability assessment on the photovoltaic power generation system.
4. A photovoltaic power generation system intelligent evaluation system according to claim 3, characterized in that: In the photovoltaic power generation multi-state assessment subsystem, the photovoltaic power generation system status is divided into two categories: operation and shutdown; the operation status is divided into full-rated status and reduced-rated status according to the size of the photovoltaic power; the reduced-rated status is divided into low power and partial failure of photovoltaic components according to its cause; the shutdown status is divided into shutdown under power restriction and shutdown caused by failure.
5. The photovoltaic power generation system intelligent evaluation system according to claim 1, characterized in that: The multi-feature fusion is achieved through three strategies. The first strategy is the multi-scale spatial feature reorganization strategy MFS, which is used to reorganize the spatial scale information of the input feature variable; the second strategy is the attention mechanism strategy AS, which is used to extract the local information of the input feature variable; the last strategy is the input feature information cross fusion strategy CS. The three strategies are characterized as follows: In the formula, represents the input feature, i represents the time scale, j represents the spatial position, I represents the dimension of the feature variable, and Y AS Represents the attention weight matrix that records the target position information, Represents the output features of the multi-scale information extraction module.
6. A photovoltaic power generation system intelligent evaluation system according to claim 2, characterized in that: The calculation formula of the PST-Transformer model is as follows: Where W1, b1 and W2, b2 represent the weights and bias terms of the two fully connected layers, Q, K and V represent the query matrix, key matrix and value matrix respectively, and W Q , W K and W V They represent the weight matrices of query, key, and value, respectively. T represents transposition. k Representation is the dimension of attention.
7. A photovoltaic power generation system intelligent evaluation system according to claim 2, characterized in that: The intelligent forecasting unit uses the PST-Transformer model to predict the photovoltaic power generation system and uses the PO algorithm to find the optimal parameter set of key parameters of the PST-Transformer model; The PO algorithm uses exploration, development, update and reverse search operations to calculate and sort the initial key parameter group of the PST-Transformer model, and randomly exhibits one of the four behaviors during each iteration. After meeting the maximum number of iterations, the optimal initial key parameter group of the PST-Transformer model is output, and its formula is as follows: Key parameter groups in the PST-Transformer model Initially set to: The calculation formula for searching the optimal parameter group based on the initial parameters is: Development formula: According to the size of the random number P, the key parameters of the PST-Transformer model are updated, and the calculation formula is: The formula for reverse search of the key parameter group of the PST-Transformer model is: In the formula, lb represents the lower limit of the search space of key parameters of the PST-Transformer model, ub represents the upper limit of the search space of key parameters of the PST-Transformer model, rand(0,1) and P are random numbers in the range of [0,1]. represents the initial stage PST-Transformer model parameter group, Represents the current model parameter group, Represents the model parameter group to be updated subsequently, ones(1,dim) represents a vector of all 1s with dimension dim, represents the average value of the key parameter group, Levy(dim) Levy distribution, represents the change process of the key parameter group of the PST-Transformer model, N best Indicates the optimal position of the key parameter group of the PST-Transformer model, Max iter represents the maximum number of iterations, and t represents the current number of iterations.
8. A photovoltaic power generation system intelligent evaluation system according to claim 7, characterized in that: The key parameter groups of the PST-Transformer model include but are not limited to the weights and biases of the self-attention layer, the weights and biases of the MLP layer, the scaling parameters, and the translation parameters.
9. A photovoltaic power generation system intelligent evaluation system according to claim 7, characterized in that: The PST-Transformer model uses temporal position encoding and spatial position encoding; temporal position encoding assigns the same time encoding to data collected at the same time, helping the self-attention mechanism to better identify data acquired at the same time; while spatial position encoding assigns the same position encoding to data acquired at the same location, understanding the relationship between data from the same monitoring site; The two position codes are connected to the input data and then input into the PST-Transformer model for processing, so that the model takes into account the spatiotemporal information in the data. The formula is as follows: Where, PE (pos) represents the original position encoding of Transformer, pos represents the position of feature information data in the input sequence, a represents the position of the input sequence, c represents the dimension index of the model, d represents the dimension of the position encoding, T_PE (pos) represents the time position code, T_V represents the time when each data in the data set was collected, T_E represents a list with time codes, S_PE (pos) Represents spatial position coding, L_V represents the location where each data in the data set is collected, and L_E represents a list with position coding.
10. A photovoltaic power generation system intelligent evaluation system according to claim 3, characterized in that: The reliability evaluation adopts the following reliability index: failure rate λ T , that is, the number of failures that occurred in the photovoltaic power generation system during the evaluation period; the average repair time r T , which is the average repair time after a photovoltaic power generation system fails; the actual availability rate A a , that is, the probability that the photovoltaic power generation system is in normal operation during the assessment period; the full operation rate R FR , that is, the probability of the photovoltaic power generation system to produce corresponding output under corresponding power conditions; the resource limitation derating rate R RCP , that is, the probability of derating due to power restriction during normal operation of the photovoltaic power generation system; the derating rate of equipment failure R FCP , that is, the probability that the photovoltaic power generation system will operate at a reduced rating due to a fault.