A photovoltaic power generation power prediction method, system, device and medium

Abstract

The application discloses a photovoltaic power generation power prediction method, system and device and a medium, relates to the photovoltaic power generation prediction technical field, and the photovoltaic power generation power prediction method comprises the following steps: obtaining the actual value of the environmental parameter of each time in the current period; the environmental parameter includes: the light intensity, temperature and photovoltaic power generation power; the actual value of the environmental parameter of each time in the current period is input into a photovoltaic power generation power prediction model to obtain the predicted value of the photovoltaic power generation power of each time in the future period; the photovoltaic power generation power prediction model is obtained by training a time convolution network using a joint loss function; the joint loss function includes a data fitting loss and a physical constraint loss calculated based on the photovoltaic power generation physical mechanism.The application improves the photovoltaic power generation power prediction accuracy.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation prediction technology, and in particular to a photovoltaic power generation prediction method, system, device and medium. Background Technology

[0002] With the global energy structure transitioning towards cleaner energy sources, the installed capacity of photovoltaic (PV) power generation has continued to grow rapidly. However, PV power output exhibits significant intermittency and volatility, and its power generation is affected by the nonlinear coupling of multiple meteorological factors such as solar irradiance, ambient temperature, and cloud cover. This uncertainty poses a severe challenge to grid dispatching and electricity market transactions, necessitating the development of high-precision PV power generation forecasting technologies to ensure the safe and economical operation of the power system.

[0003] Current mainstream photovoltaic power prediction methods can be divided into two categories: physical mechanism models and data-driven models. Physical mechanism models construct a theoretical prediction framework based on physical laws such as the photoelectric conversion equation and module temperature characteristics, and have clear physical interpretability. However, in practical applications, this type of model faces problems such as difficulty in collecting meteorological parameters, complexity in system efficiency calibration, and lag in dynamic correction. Especially under non-ideal weather conditions such as cloudy skies and sudden changes, the prediction accuracy decreases significantly.

[0004] Data-driven models make predictions by mining the mapping relationship between historical operational data and meteorological characteristics, avoiding complex physical modeling processes. However, traditional data-driven methods typically use data fitting error as the optimization objective during model training, belonging to a pure data-driven paradigm. These methods ignore the physical laws governing photovoltaic power generation, such as energy conservation and the variation of photoelectric conversion efficiency with temperature. In scenarios with insufficient training data coverage (such as extreme weather, component aging, or pollution), pure data-driven models may output predictions that violate physical laws, resulting in low prediction accuracy. Summary of the Invention

[0005] The purpose of this application is to provide a method, system, device, and medium for predicting photovoltaic power generation, in order to solve the problem of low accuracy in existing photovoltaic power generation prediction methods.

[0006] To achieve the above objectives, this application provides the following solution.

[0007] In a first aspect, this application provides a method for predicting photovoltaic power generation, including: Obtain the actual values ​​of environmental parameters at each moment in the current time period; environmental parameters include: light intensity, temperature, and photovoltaic power generation. The actual values ​​of environmental parameters at each moment in the current time period are input into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each moment in the future time period. The photovoltaic power generation prediction model is obtained by training a temporal convolutional network using a joint loss function. The joint loss function includes data fitting loss and physical constraint loss calculated based on the physical mechanism of photovoltaic power generation.

[0008] In one embodiment, the process of determining the photovoltaic power generation prediction model includes: Obtain a training set; the training set includes multiple sets of training data; the training data includes the actual values ​​of environmental parameters at each moment in the first historical period and the actual values ​​of photovoltaic power generation at each moment in the second historical period; Construct the temporal convolutional network; The temporal convolutional network is trained using a joint loss function, with the actual values ​​of environmental parameters at each moment of the previous historical period in the training data in the training set as input and the actual values ​​of photovoltaic power generation at each moment of the corresponding subsequent historical period as output, to obtain the photovoltaic power generation prediction model.

[0009] In one embodiment, the temporal convolutional network includes: four residual blocks connected in sequence and a fully connected layer; Each residual block includes: an adder, a 1×1 convolutional layer, and a dilated causal convolutional layer, a weight normalization layer, a linear rectifier unit, and a random deactivation layer connected in sequence; The output of the random deactivation layer of the a-th residual block and the output of the 1×1 convolutional layer of the a-th residual block are both connected to the input of the adder of the a-th residual block; a=1,2,3,4; The output of the adder in the b-th residual block is connected to the input of the dilated causal convolutional layer of the (b+1)-th residual block and the input of the 1×1 convolutional layer of the (b+1)-th residual block, respectively; b=1,2,3; The output of the adder in the fourth residual block is connected to the input of the fully connected layer.

[0010] In one embodiment, the actual values ​​of environmental parameters at each time point in the current time period are input into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each time point in the future time period, including: The actual values ​​of environmental parameters at each moment in the current time period are input into the first residual block of the photovoltaic power generation prediction model to obtain the first time series feature sequence; The first time-series feature sequence is input into the second residual block of the photovoltaic power generation prediction model to obtain the second time-series feature sequence; The second time-series feature sequence is input into the third residual block of the photovoltaic power generation prediction model to obtain the third time-series feature sequence; The third time-series feature sequence is input into the fourth residual block of the photovoltaic power generation prediction model to obtain the fourth time-series feature sequence; The fourth time-series feature sequence is input into the fully connected layer of the photovoltaic power generation prediction model to obtain the predicted value of photovoltaic power generation at each time point in the future period.

[0011] In one embodiment, the training data further includes the actual values ​​of environmental parameters at each time point in the subsequent historical period; the environmental parameters include: irradiance, temperature coefficient, and temperature; The joint loss function includes: ; ; ; ; in, For joint losses; For data fitting loss; This is the balance coefficient; Loss due to physical constraints; The number of training data sets in the training set; It is a vector composed of the predicted photovoltaic power generation values ​​at each time point in the next historical period of the i-th training data set; It is a vector composed of the actual values ​​of photovoltaic power generation at each time point in the next historical period of the i-th training data set; The physical constraint loss is calculated based on the physical mechanism of photovoltaic power generation. It is a vector consisting of the theoretical values ​​of photovoltaic power generation at each moment in the next historical period of the i-th training data set; This is the preset reference efficiency; It is a vector consisting of the actual values ​​of irradiance at each time point in the next historical period of the i-th training data set; This is a vector consisting of the actual values ​​of the temperature coefficient at each moment in the next historical period of the i-th training data set; It is a vector consisting of the actual temperature values ​​at each time point in the next historical period of the i-th training data set; This is the preset reference temperature.

[0012] Secondly, this application provides a photovoltaic power generation prediction system for implementing a photovoltaic power generation prediction method, the photovoltaic power generation prediction system comprising: The actual value acquisition module is used to acquire the actual values ​​of environmental parameters at each moment in the current time period; environmental parameters include: light intensity, temperature, and photovoltaic power generation. The prediction module is used to input the actual values ​​of environmental parameters at each time point in the current period into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each time point in the future period. The photovoltaic power generation prediction model is obtained by training a temporal convolutional network using a joint loss function. The joint loss function includes data fitting loss and physical constraint loss calculated based on the physical mechanism of photovoltaic power generation.

[0013] In one embodiment, the process of determining the photovoltaic power generation prediction model includes: Obtain a training set; the training set includes multiple sets of training data; the training data includes the actual values ​​of environmental parameters at each moment in the first historical period and the actual values ​​of photovoltaic power generation at each moment in the second historical period; Construct the temporal convolutional network; The temporal convolutional network is trained using a joint loss function, with the actual values ​​of environmental parameters at each moment of the previous historical period in the training data in the training set as input and the actual values ​​of photovoltaic power generation at each moment of the corresponding subsequent historical period as output, to obtain the photovoltaic power generation prediction model.

[0014] In one embodiment, the temporal convolutional network includes: four residual blocks connected in sequence and a fully connected layer; Each residual block includes: an adder, a 1×1 convolutional layer, and a dilated causal convolutional layer, a weight normalization layer, a linear rectifier unit, and a random deactivation layer connected in sequence; The output of the random deactivation layer of the a-th residual block and the output of the 1×1 convolutional layer of the a-th residual block are both connected to the input of the adder of the a-th residual block; a=1,2,3,4; The output of the adder in the b-th residual block is connected to the input of the dilated causal convolutional layer of the (b+1)-th residual block and the input of the 1×1 convolutional layer of the (b+1)-th residual block, respectively; b=1,2,3; The output of the adder in the fourth residual block is connected to the input of the fully connected layer.

[0015] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the photovoltaic power generation prediction method described above.

[0016] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the photovoltaic power generation prediction method described above.

[0017] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application discloses a method, system, device, and medium for predicting photovoltaic (PV) power generation. It obtains the irradiance, temperature, and PV power generation at various times during the current period as input parameters and feeds them into a PV power generation prediction model. This model uses a temporal convolutional network as its basic architecture, enabling efficient extraction of long-range temporal dependencies between irradiance, temperature, and PV power generation. More importantly, the joint loss function introduced during model training includes both data fitting loss and physical constraint loss calculated based on the physical mechanisms of PV power generation. This allows the model to learn and adhere to physical laws such as energy conservation and the change in photoelectric conversion efficiency with temperature while fitting historical data. Compared to purely data-driven models that only optimize data fitting and may output results that violate physical laws under conditions not covered by training data, this application constrains the model output through physical constraint loss, ensuring that the prediction results always lie within the physically feasible region. This maintains accurate prediction capabilities even under extrapolated scenarios such as extreme weather and component aging, significantly improving the accuracy of PV power generation prediction. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of a photovoltaic power generation prediction method provided in an embodiment of this application.

[0020] Figure 2 This is a schematic diagram of a temporal convolutional network structure.

[0021] Figure 3 This is a schematic diagram of the residual block structure.

[0022] Figure 4 This is a comparison chart of photovoltaic power generation prediction results.

[0023] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] The purpose of this application is to provide a method, system, device and medium for predicting photovoltaic power generation, with the aim of improving the accuracy of photovoltaic power generation prediction.

[0026] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] In one exemplary embodiment, such as Figure 1 As shown, a method for predicting photovoltaic power generation is provided, including: Step 1: Obtain the actual values ​​of environmental parameters at each time point in the current time period; environmental parameters include: light intensity, temperature, and photovoltaic power generation. Step 2: Input the actual values ​​of environmental parameters at each time point in the current period into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each time point in the future period; The photovoltaic power generation prediction model is obtained by training a temporal convolutional network using a joint loss function. The joint loss function includes data fitting loss and physical constraint loss calculated based on the physical mechanism of photovoltaic power generation.

[0028] As an optional implementation method, the process of determining the photovoltaic power generation prediction model includes: Obtain the training set; the training set includes multiple sets of training data; the training data includes the actual values ​​of environmental parameters at each moment in the previous historical period and the actual values ​​of photovoltaic power generation at each moment in the next historical period; Construct a temporal convolutional network; By using a joint loss function, taking the actual values ​​of environmental parameters at each moment of the previous historical period in the training data of each training data in the training set as input, and taking the actual values ​​of photovoltaic power generation at each moment of the corresponding subsequent historical period as output, a temporal convolutional network is trained to obtain a photovoltaic power generation prediction model.

[0029] As an optional implementation method, such as Figure 2 As shown, the temporal convolutional network consists of four residual blocks connected in sequence and a fully connected layer; like Figure 3As shown, each residual block includes: an adder, a 1×1 convolutional layer, and a dilated causal convolutional layer, a weight normalization layer, a linear rectifier unit, and a random deactivation layer connected in sequence; The output of the random deactivation layer of the a-th residual block and the output of the 1×1 convolutional layer of the a-th residual block are both connected to the input of the adder of the a-th residual block; a=1,2,3,4; The output of the adder in the b-th residual block is connected to the input of the dilated causal convolutional layer of the (b+1)-th residual block and the input of the 1×1 convolutional layer of the (b+1)-th residual block, respectively; b=1,2,3; The output of the adder in the fourth residual block is connected to the input of the fully connected layer.

[0030] Specifically, the expansion coefficients of the four residual blocks are 1, 2, 4, and 8, respectively.

[0031] The principle of the dilated causal convolutional layer in each residual block is as follows: ; in, For the first in the output sequence One output value; This represents the total number of weight parameters in the convolution kernel; The first in the convolution kernel One weight parameter; For the first in the input sequence One input value; is the coefficient of thermal expansion.

[0032] As an optional implementation, step 2 includes: Step 21: Input the actual values ​​of environmental parameters at each time point in the current time period into the first residual block of the photovoltaic power generation prediction model to obtain the first time series feature sequence; Step 22: Input the first time-series feature sequence into the second residual block of the photovoltaic power generation prediction model to obtain the second time-series feature sequence; Step 23: Input the second time series feature sequence into the third residual block of the photovoltaic power generation prediction model to obtain the third time series feature sequence; Step 24: Input the third time series feature sequence into the fourth residual block of the photovoltaic power generation prediction model to obtain the fourth time series feature sequence; Step 25: Input the fourth time series feature sequence into the fully connected layer of the photovoltaic power generation prediction model to obtain the predicted value of photovoltaic power generation at each time in the future period.

[0033] Specifically, after extracting features, each residual block passes its output to the next residual block, allowing the entire temporal convolutional network to learn information at different time scales. This multi-layered structure helps to better capture long-term dependencies between data in long sequences, thereby improving the information extraction capability of the photovoltaic power generation prediction model. Inside the temporal convolutional network, by stacking multiple residual blocks and using an exponentially growing inflation coefficient, the network can effectively fuse local details with global context. Finally, the feature sequence output from the fourth temporal feature sequence is input into the fully connected layer to obtain the predicted photovoltaic power generation at each time point in the future.

[0034] As an optional implementation, the training data also includes the actual values ​​of environmental parameters at each time point in the next historical period; the environmental parameters include: irradiance, temperature coefficient, and temperature; Joint loss functions include: ; ; ; ; in, For joint losses; For data fitting loss; This is the balance coefficient; Loss due to physical constraints; The number of training data sets in the training set; It is a vector composed of the predicted photovoltaic power generation values ​​at each time point in the next historical period of the i-th training data set; It is a vector composed of the actual values ​​of photovoltaic power generation at each time point in the next historical period of the i-th training data set; The physical constraint loss is calculated based on the physical mechanism of photovoltaic power generation. It is a vector consisting of the theoretical values ​​of photovoltaic power generation at each moment in the next historical period of the i-th training data set; This is the preset reference efficiency; It is a vector consisting of the actual values ​​of irradiance at each time point in the next historical period of the i-th training data set; This is a vector consisting of the actual values ​​of the temperature coefficient at each moment in the next historical period of the i-th training data set; It is a vector consisting of the actual temperature values ​​at each time point in the next historical period of the i-th training data set; This is the preset reference temperature.

[0035] Specifically, by minimizing the joint loss, the output of the photovoltaic power generation prediction model can both fit historical data and satisfy the physical laws of photovoltaic power generation, thereby improving the prediction reliability of the photovoltaic power generation prediction model in extrapolation scenarios.

[0036] Specifically, after obtaining the photovoltaic power generation prediction model, two statistical indicators were used to quantitatively evaluate and compare the performance of the photovoltaic power generation prediction model: the coefficient of determination (COP). ) and mean absolute error ( These indicators collectively constitute a comprehensive evaluation system for the prediction accuracy of photovoltaic power generation prediction models, helping to understand the predictive capabilities of these models from different perspectives. Generally speaking, The closer the value is to 1, The smaller the value, the better the predictive performance of the photovoltaic power generation prediction model. The calculation formulas for the indicators are as follows: ; ; in, The actual value of the environmental parameter corresponding to the j-th test sample in the test set; Let be the predicted value of the photovoltaic power generation corresponding to the j-th test sample in the test set; To take the absolute value; This represents the number of test samples in the test set.

[0037] Furthermore, this application compares and analyzes the photovoltaic power generation prediction results on the test set. The prediction results are as follows: Figure 4 As shown, the solid blue line represents the actual value, and the dashed red line represents the predicted value. The calculated coefficient of determination on the test set is 0.9945, and the mean absolute error is 0.0098, indicating that the photovoltaic power generation prediction model has good prediction accuracy and good fit.

[0038] In one exemplary embodiment, a photovoltaic power generation prediction system is provided for implementing a photovoltaic power generation prediction method. The photovoltaic power generation prediction system includes: The actual value acquisition module is used to acquire the actual values ​​of environmental parameters at each moment in the current time period; environmental parameters include: light intensity, temperature, and photovoltaic power generation. The prediction module is used to input the actual values ​​of environmental parameters at each moment in the current time period into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each moment in the future time period. The photovoltaic power generation prediction model is obtained by training a temporal convolutional network using a joint loss function. The joint loss function includes data fitting loss and physical constraint loss calculated based on the physical mechanism of photovoltaic power generation.

[0039] As an optional implementation method, the process of determining the photovoltaic power generation prediction model includes: Obtain the training set; the training set includes multiple sets of training data; the training data includes the actual values ​​of environmental parameters at each moment in the previous historical period and the actual values ​​of photovoltaic power generation at each moment in the next historical period; Construct a temporal convolutional network; By using a joint loss function, taking the actual values ​​of environmental parameters at each moment of the previous historical period in the training data of each training data in the training set as input, and taking the actual values ​​of photovoltaic power generation at each moment of the corresponding subsequent historical period as output, a temporal convolutional network is trained to obtain a photovoltaic power generation prediction model.

[0040] As an optional implementation, the temporal convolutional network includes: four residual blocks connected in sequence and a fully connected layer; Each residual block includes: an adder, a 1×1 convolutional layer, and a dilated causal convolutional layer, a weight normalization layer, a linear rectifier unit, and a random deactivation layer connected in sequence; The output of the random deactivation layer of the a-th residual block and the output of the 1×1 convolutional layer of the a-th residual block are both connected to the input of the adder of the a-th residual block; a=1,2,3,4; The output of the adder in the b-th residual block is connected to the input of the dilated causal convolutional layer of the (b+1)-th residual block and the input of the 1×1 convolutional layer of the (b+1)-th residual block, respectively; b=1,2,3; The output of the adder in the fourth residual block is connected to the input of the fully connected layer.

[0041] In one exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a photovoltaic power generation prediction method.

[0042] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements a photovoltaic power generation prediction method.

[0043] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements a photovoltaic power generation prediction method.

[0044] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 5As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media to run. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a photovoltaic power generation prediction method.

[0045] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0046] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0047] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0048] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0049] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0050] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for predicting photovoltaic power generation, characterized in that, The photovoltaic power generation prediction method includes: Obtain the actual values ​​of environmental parameters at each moment in the current time period; environmental parameters include: light intensity, temperature, and photovoltaic power generation. The actual values ​​of environmental parameters at each moment in the current time period are input into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each moment in the future time period. The photovoltaic power generation prediction model is obtained by training a temporal convolutional network using a joint loss function. The joint loss function includes data fitting loss and physical constraint loss calculated based on the physical mechanism of photovoltaic power generation.

2. The photovoltaic power generation prediction method according to claim 1, characterized in that, The process of determining the photovoltaic power generation prediction model includes: Obtain a training set; the training set includes multiple sets of training data; the training data includes the actual values ​​of environmental parameters at each moment in the first historical period and the actual values ​​of photovoltaic power generation at each moment in the second historical period; Construct the temporal convolutional network; The temporal convolutional network is trained using a joint loss function, with the actual values ​​of environmental parameters at each moment of the previous historical period in the training data in the training set as input and the actual values ​​of photovoltaic power generation at each moment of the corresponding subsequent historical period as output, to obtain the photovoltaic power generation prediction model.

3. The photovoltaic power generation prediction method according to claim 1, characterized in that, The temporal convolutional network comprises: four residual blocks connected in sequence and a fully connected layer; Each residual block includes: an adder, a 1×1 convolutional layer, and a dilated causal convolutional layer, a weight normalization layer, a linear rectifier unit, and a random deactivation layer connected in sequence; The output of the random deactivation layer of the a-th residual block and the output of the 1×1 convolutional layer of the a-th residual block are both connected to the input of the adder of the a-th residual block; a=1,2,3,4; The output of the adder in the b-th residual block is connected to the input of the dilated causal convolutional layer of the (b+1)-th residual block and the input of the 1×1 convolutional layer of the (b+1)-th residual block, respectively; b=1,2,3; The output of the adder in the fourth residual block is connected to the input of the fully connected layer.

4. The photovoltaic power generation prediction method according to claim 3, characterized in that, The actual values ​​of environmental parameters at each moment in the current time period are input into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each moment in the future time period, including: The actual values ​​of environmental parameters at each moment in the current time period are input into the first residual block of the photovoltaic power generation prediction model to obtain the first time series feature sequence; The first time-series feature sequence is input into the second residual block of the photovoltaic power generation prediction model to obtain the second time-series feature sequence; The second time-series feature sequence is input into the third residual block of the photovoltaic power generation prediction model to obtain the third time-series feature sequence; The third time-series feature sequence is input into the fourth residual block of the photovoltaic power generation prediction model to obtain the fourth time-series feature sequence; The fourth time-series feature sequence is input into the fully connected layer of the photovoltaic power generation prediction model to obtain the predicted value of photovoltaic power generation at each time point in the future period.

5. The photovoltaic power generation prediction method according to claim 1, characterized in that, The training data also includes the actual values ​​of environmental parameters at each time point in the next historical period; the environmental parameters include: irradiance, temperature coefficient, and temperature. The joint loss function includes: ； ； ； ； in, For joint losses; For data fitting loss; This is the balance coefficient; Loss due to physical constraints; The number of training data sets in the training set; It is a vector composed of the predicted photovoltaic power generation values ​​at each time point in the next historical period of the i-th training data set; It is a vector composed of the actual values ​​of photovoltaic power generation at each time point in the next historical period of the i-th training data set; The physical constraint loss is calculated based on the physical mechanism of photovoltaic power generation. It is a vector consisting of the theoretical values ​​of photovoltaic power generation at each moment in the next historical period of the i-th training data set; This is the preset reference efficiency; It is a vector consisting of the actual values ​​of irradiance at each time point in the next historical period of the i-th training data set; This is a vector consisting of the actual values ​​of the temperature coefficient at each moment in the next historical period of the i-th training data set; It is a vector consisting of the actual temperature values ​​at each time point in the next historical period of the i-th training data set; This is the preset reference temperature.

6. A photovoltaic power generation prediction system, used to implement the photovoltaic power generation prediction method as described in any one of claims 1-5, characterized in that, The photovoltaic power generation prediction system includes: The actual value acquisition module is used to acquire the actual values ​​of environmental parameters at each moment in the current time period; environmental parameters include: light intensity, temperature, and photovoltaic power generation. The prediction module is used to input the actual values ​​of environmental parameters at each time point in the current period into the photovoltaic power generation prediction model to obtain the predicted values ​​of photovoltaic power generation at each time point in the future period. The photovoltaic power generation prediction model is obtained by training a temporal convolutional network using a joint loss function. The joint loss function includes data fitting loss and physical constraint loss calculated based on the physical mechanism of photovoltaic power generation.

7. The photovoltaic power generation prediction system according to claim 6, characterized in that, The process of determining the photovoltaic power generation prediction model includes: Obtain a training set; the training set includes multiple sets of training data; the training data includes the actual values ​​of environmental parameters at each moment in the first historical period and the actual values ​​of photovoltaic power generation at each moment in the second historical period; Construct the temporal convolutional network; The temporal convolutional network is trained using a joint loss function, with the actual values ​​of environmental parameters at each moment of the previous historical period in the training data in the training set as input and the actual values ​​of photovoltaic power generation at each moment of the corresponding subsequent historical period as output, to obtain the photovoltaic power generation prediction model.

8. The photovoltaic power generation prediction system according to claim 6, characterized in that, The temporal convolutional network comprises: four residual blocks connected in sequence and a fully connected layer; Each residual block includes: an adder, a 1×1 convolutional layer, and a dilated causal convolutional layer, a weight normalization layer, a linear rectifier unit, and a random deactivation layer connected in sequence; The output of the random deactivation layer of the a-th residual block and the output of the 1×1 convolutional layer of the a-th residual block are both connected to the input of the adder of the a-th residual block; a=1,2,3,4; The output of the adder in the b-th residual block is connected to the input of the dilated causal convolutional layer of the (b+1)-th residual block and the input of the 1×1 convolutional layer of the (b+1)-th residual block, respectively; b=1,2,3; The output of the adder in the fourth residual block is connected to the input of the fully connected layer.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the photovoltaic power generation prediction method according to any one of claims 1-5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the photovoltaic power generation prediction method according to any one of claims 1-5.

Images