Green power demand determination method and device, nonvolatile storage medium and electronic equipment

Abstract

The invention discloses a green power demand determination method and device, a nonvolatile storage medium and electronic equipment. The method comprises the steps that total power data of a target park are acquired, and the power data comprise green power data using clean energy; based on the time sequence characteristics of the power data, determining the total power consumption demand of the target park; based on the business volume change information of the target park, correcting a predetermined green electricity proportion, and determining a correction proportion; and determining the green electricity demand of the target park according to the correction proportion and the total electricity demand. According to the invention, the technical problem of non-ideal accuracy and qualitative property of green power demand prediction is solved.

Description

Technical Field

[0001] The present application relates to the field of power demand forecasting, and in particular, to a method and device for determining green power demand, a non-volatile storage medium, and an electronic device. Background Art

[0002] Green electricity mainly comes from renewable energy sources such as solar energy, wind energy, hydropower, and biomass energy. Green electricity helps reduce dependence on fossil fuels, reduce pollution, and mitigate climate change. At present, the green electricity demand forecasting method mainly relies on the power load forecasting method, including the unit consumption method, the per capita electricity consumption method, and the time series forecast. However, these methods have exposed obvious technical limitations when faced with the complexity and uncertainty of the green electricity market. The forecasting method dominated by expert experience reduces the objectivity and accuracy of the forecast results. The existing methods fail to fully consider the dynamic factors affecting the demand for green electricity, such as fluctuations in order volume, seasonal changes, and analysis of historical electricity price data, resulting in insufficient forecasting accuracy. There are technical problems in the relevant technologies that the accuracy of green electricity demand forecasting is not ideal.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide a method, device, non-volatile storage medium and electronic device for determining green electricity demand, so as to at least solve the technical problem that the accurate characterization of green electricity demand prediction is not ideal.

[0005] According to one aspect of an embodiment of the present application, a method for determining green electricity demand is provided, comprising: obtaining total power data of a target park, wherein the power data includes green electricity data using clean energy; determining the total electricity demand of the target park based on time series characteristics of the power data; correcting a predetermined green electricity ratio based on business volume change information of the target park, and determining a correction ratio; and determining the green electricity demand of the target park according to the correction ratio and the total electricity demand.

[0006] Optionally, obtaining the total power data of the target park includes: using multiple sensors to collect data on the target park respectively to obtain collected data corresponding to the multiple sensors; and integrating the collected data corresponding to the multiple sensors to obtain the total power data.

[0007] Optionally, the collected data corresponding to the multiple sensors respectively include at least: load data, weather data, time information data, power equipment information, park business volume, and business transformation requirements.

[0008] Optionally, determining the total electricity demand of the target park based on the timing characteristics of the power data includes: determining hyperparameters of an initial long short-term memory network, wherein the hyperparameters are parameters determined before training the initial long short-term memory network; training the initial long short-term memory network with the hyperparameters based on historical data of the target park in a historical period to obtain a target long short-term memory network; and using the target long short-term memory network to determine the total electricity demand based on the timing characteristics.

[0009] Optionally, there are multiple hyperparameters, including the number of neurons, a learning rate, and the number of training iterations. The determining of the hyperparameters of the initial long short-term memory network includes: determining a fitness function based on the mean absolute percentage error between the actual value and the predicted value of the electric quantity; using a particle swarm algorithm to determine the fitness function values ​​corresponding to a plurality of particles, wherein the plurality of particles represent a combination of the number of neurons, the learning rate, and the number of training iterations; determining a target particle based on the fitness function value that minimizes the mean absolute percentage error between the actual value and the predicted value of the electric quantity; and determining a hyperparameter based on a combination of the number of neurons, the learning rate, and the number of training iterations indicated by the target particle.

[0010] Optionally, the method of correcting the predetermined green electricity ratio based on the business volume change information of the target park and determining the correction ratio includes: setting a hidden layer of an initial neural network based on the business volume change information of the target park to obtain a set initial neural network; using a historical green electricity ratio for the same period that matches the cycle characteristics of the current cycle to train the set initial neural network to obtain a target neural network; and based on the predetermined green electricity ratio, using the target neural network for processing to obtain the correction ratio.

[0011] Optionally, the historical green electricity ratio in the same period that matches the cycle characteristics of the current period is used to train the set initial neural network to obtain the target neural network, including: performing data normalization processing based on historical change information before the current cycle to obtain normalized data, wherein the historical change information corresponds to the historical green electricity ratio in the same period; inputting the normalized data into the initial neural network for training to obtain the target neural network.

[0012] According to another aspect of an embodiment of the present application, a green electricity demand determination device is provided, including: a power data acquisition module, used to acquire total power data of a target park, wherein the power data includes green electricity data using clean energy; a power demand determination module, used to determine the total power demand of the target park based on the timing characteristics of the power data; a correction ratio determination module, used to correct a predetermined green electricity ratio based on business volume change information of the target park and determine a correction ratio; a green electricity demand determination module, used to determine the green electricity demand of the target park according to the correction ratio and the total power demand.

[0013] According to another aspect of an embodiment of the present application, a non-volatile storage medium is provided, wherein the non-volatile storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing any one of the green electricity demand determination methods.

[0014] According to another aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the green electricity demand determination methods.

[0015] In the embodiment of the present application, by obtaining the total power data of the target park, wherein the power data includes green power data using clean energy; based on the time series characteristics of the power data, the total power demand of the target park is determined; based on the business volume change information of the target park, the predetermined green power ratio is corrected to determine the correction ratio; according to the correction ratio and the total power demand, the green power demand of the target park is determined. The purpose of improving energy utilization is achieved, and the technical effect of improving the accuracy of green power demand prediction is achieved, thereby solving the technical problem of unsatisfactory accuracy of green power demand prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 is a flow chart of an optional method for determining green electricity demand provided according to an embodiment of the present application;

[0018] Figure 2 is a first schematic diagram of an optional method for determining green electricity demand provided according to an embodiment of the present application;

[0019] Figure 3is a second schematic diagram of an optional method for determining green electricity demand provided according to an embodiment of the present application;

[0020] Figure 4 It is a schematic diagram of an optional green electricity demand determination device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:

[0024] Long Short-Term Memory (LSTM) is used to solve the gradient vanishing and gradient exploding problems in long sequence data processing, and can effectively capture long-term dependencies in data. LSTM can remember contextual information with a longer time span when processing sequence data.

[0025] Particle Swarm Optimization (PSO) is an optimization algorithm based on swarm intelligence, inspired by the foraging behavior of bird flocks, and used to solve complex optimization problems. Each solution (i.e., "particle") searches for the best position in the solution space, and through information sharing and interaction between particles, the entire particle swarm gradually moves towards the optimal solution.

[0026] Mean Absolute Percentage Error (MAPE) is used to measure the difference between the predicted value and the true value, and is particularly suitable for the prediction evaluation of ratio or percentage data. The smaller the value of MAPE, the higher the prediction accuracy.

[0027] BP neural network (Back Propagation Neural Network) consists of an input layer, one or more hidden layers and an output layer. Each layer is connected by weights, and the neurons in each layer are fully connected to the neurons in the next layer. Its core idea is to calculate the output result through forward propagation, and then adjust the network weights and biases through back propagation to minimize the error between the network output and the target output.

[0028] According to an embodiment of the present application, a method embodiment for determining green electricity demand is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] Figure 1 is a flow chart of an optional method for determining green electricity demand provided according to an embodiment of the present application, such as Figure 1 As shown, the method comprises the following steps:

[0030] Step S102, obtaining total power data of the target park, wherein the power data includes green power data using clean energy;

[0031] It can be understood that collecting the total electricity data of the target park, including green electricity data using clean energy, can fully grasp the power consumption of the park, provide basic data support for subsequent predictive analysis, ensure the integrity and real-time nature of the input data of the predictive model, and improve the accuracy and reliability of the prediction.

[0032] In an optional embodiment, the total power data of the target park is obtained, including: using multiple sensors to collect data from the target park respectively, and obtaining collected data corresponding to the multiple sensors; integrating the collected data corresponding to the multiple sensors to obtain the total power data.

[0033] It is understandable that multiple sensors are used to collect data from the target park, and then the data collected by the sensors are integrated to obtain the total power data. The use of multiple sensors can capture subtle changes in power consumption in the park, including the use of green electricity, and other key parameters related to power consumption, such as temperature, humidity, light intensity and other environmental factors. Data integration ensures that all collected information can be effectively used to build a more comprehensive power consumption portrait, providing a solid foundation for subsequent predictive analysis.

[0034] In an optional embodiment, the collected data corresponding to the multiple sensors include at least: load data, weather data, time information data, power equipment information, campus business volume, and business transformation requirements.

[0035] It can be understood that the collected data at least covers load data, weather data, time information data, power equipment information, park business volume and business transformation needs. By collecting multi-dimensional data, the changing patterns and influencing factors of green electricity demand can be accurately reflected. Load data provides basic information on park electricity consumption, weather data and time information data consider the impact of natural conditions and time cycles on green electricity supply and demand, and power equipment information helps to understand the specific sources and characteristics of electricity consumption. The comprehensive use of data enables the prediction model to capture the dynamic changes in green electricity demand.

[0036] Optionally, the collected data needs to be preprocessed, that is, data with obvious deviations and problems are eliminated, and the data is normalized and scaled to the same interval. The preprocessing process mainly includes data cleaning and data standardization. Data preprocessing directly affects the accuracy and stability of the prediction model.

[0037] Optionally, the above data cleaning methods can be multiple, for example, through statistical analysis, visual inspection and the use of anomaly detection algorithms, to remove obvious outliers and noise in the data set to ensure the data quality of the input model. Outliers may be caused by sensor failure, data transmission errors or human misoperation, which will distort the true distribution of the data and lead to deviations in the prediction results.

[0038] Optionally, the above data normalization method can be multiple, for example, scaling the data to the same interval, usually [0, 1] or [-1, 1]. Data normalization can eliminate the dimensional differences between different input features, so that the model will not be biased towards features with larger values ​​when processing data. Data normalization is also called data normalization. The mathematical expression for normalized data processing is:

[0039]

[0040] Where X max and X minRespectively represent the maximum and minimum values ​​of the indicator x; x N is the normalized load data, with a value between 0 and 1. The purpose of data normalization is to eliminate the influence of dimension.

[0041] Optionally, the data preprocessing method at least includes: feature engineering, that is, generating more meaningful features by transforming and combining the original data to enhance the prediction ability of the model. Seasonal indicators and weekday / weekend indicators can be extracted from the time information data, or the cyclical fluctuation characteristics of the business can be calculated from the order change. In addition, the preprocessing step can also include missing value filling, data smoothing (such as moving average) and extraction of time series features.

[0042] Step S104, determining the total power demand of the target park based on the time series characteristics of the power data;

[0043] It can be understood that the total electricity demand is determined based on the time series characteristics of the target park's power data. By identifying and analyzing the time series attributes of power data, capturing the seasonal changes in power demand, and the impact of sudden events on power consumption, the forecasting model can more accurately predict future power demand and provide accurate decision-making basis for the park's energy management.

[0044] In an optional embodiment, the total electricity demand of the target park is determined based on the time series characteristics of the power data, including: determining the hyperparameters of the initial long short-term memory network, wherein the hyperparameters are parameters determined before the initial long short-term memory network is trained; based on the historical data of the target park in the historical period, the initial long short-term memory network of the hyperparameters is trained to obtain the target long short-term memory network; using the target long short-term memory network, the total electricity demand is determined based on the time series characteristics.

[0045] It can be understood that when determining the total electricity demand of the target park based on the time series characteristics of the power data, the optimized long short-term memory network model is used. Before model training, the hyperparameters of the initial long short-term memory network, such as the number of neurons, learning rate, number of training iterations, etc., are determined. Model training is performed based on the historical data of the target park, and the model parameters are optimized so that the model can better understand and predict the time series pattern of power demand. Through the above training process, the optimized target long short-term memory network model is obtained. The target long short-term memory network is used to determine the total electricity demand based on the time series characteristics. Optimizing the hyperparameters of the model can enable the model to learn the complex time series relationships in the power data more efficiently, reduce overfitting, and improve the generalization ability of the model.

[0046] In an optional embodiment, there are multiple hyperparameters, including the number of neurons, the learning rate, and the number of training iterations. The hyperparameters of the initial long short-term memory network are determined, including: determining a fitness function based on the mean absolute percentage error between the actual value and the predicted value of the power; using a particle swarm algorithm to determine the fitness function values ​​corresponding to a plurality of particles, wherein the plurality of particles represent a combination of the number of neurons, the learning rate, and the number of training iterations; determining a target particle based on the fitness function value that minimizes the mean absolute percentage error between the actual value and the predicted value of the power; and determining a hyperparameter based on a combination of the number of neurons, the learning rate, and the number of training iterations indicated by the target particle.

[0047] It can be understood that the determination of hyperparameters adopts an intelligent optimization method based on the particle swarm optimization algorithm. Particles represent different combinations of the number of neurons, learning rate and number of training iterations. The fitness function is set to the mean absolute percentage error, which directly reflects the degree of deviation between the green electricity demand predicted by the model and the actual demand. The target particle is determined by the minimum fitness function value, and the hyperparameter is determined based on the combination of the number of neurons indicated by the target particle, the learning rate and the number of training iterations. Using the particle swarm algorithm to optimize the hyperparameters of the long short-term memory network can improve the accuracy of green electricity demand prediction and the robustness of the model.

[0048] Optionally, the fitness function at least includes: taking three hyperparameters of the long short-term memory network as parameters to be optimized for PSO, and setting the fitness function to:

[0049] F(x)=min(MAPE)

[0050] The mean absolute percentage error (MAPE) refers to the average value of the percentage error between the actual value and the predicted value of electricity consumption.

[0051]

[0052] Where Y a Indicates the actual value of the load, Y p represents the predicted value of the load, n represents the number of predicted load points, and t represents the tth predicted load point.

[0053] Step S106, based on the business volume change information of the target park, the predetermined green electricity ratio is corrected to determine the correction ratio;

[0054] It can be understood that according to the business volume change information of the target park, the pre-set green electricity ratio is dynamically revised to determine a more reasonable green electricity demand correction ratio. By collecting and analyzing business volume change information, the prediction of the green electricity ratio is corrected to make the prediction result closer to the actual current needs of the park.

[0055] In an optional embodiment, based on the business volume change information of the target park, the predetermined green electricity ratio is corrected to determine the correction ratio, including: based on the business volume change information of the target park, setting the hidden layer of the initial neural network to obtain the set initial neural network; using the historical green electricity ratio in the same period that matches the cycle characteristics of the current cycle to train the set initial neural network to obtain the target neural network; based on the predetermined green electricity ratio, using the target neural network for processing to obtain the correction ratio.

[0056] It can be understood that first, based on the business volume change information of the park, such as business order volume, production activity level, etc., the hidden layer settings of the neural network model are adjusted to enhance the model's sensitivity and adaptability to these changes. Subsequently, the adjusted neural network is trained using the green electricity share data of the same period in history, in order to obtain a target neural network that can accurately reflect the impact of business volume changes on green electricity demand. Finally, the park's scheduled green electricity share is input into this target neural network, and after processing, a correction ratio is output to adjust the green electricity demand forecast to make it closer to actual business needs. By incorporating business volume change information into the structural design and training process of the neural network, the model can learn the intrinsic relationship between business volume and green electricity demand, so that it can automatically and intelligently evaluate the impact of current business volume changes on green electricity demand during prediction, and output more reasonable and closer to actual demand prediction results.

[0057] In an optional embodiment, the historical green electricity ratio in the same period that matches the cycle characteristics of the current period is used to train the set initial neural network to obtain a target neural network, including: performing data normalization processing based on historical change information before the current period to obtain normalized data, wherein the historical change information corresponds to the historical green electricity ratio in the same period; inputting the normalized data into the initial neural network for training to obtain a target neural network.

[0058] It can be understood that the historical change information before the current cycle is normalized, and the data is converted into a unified range to eliminate the impact caused by the difference in dimensions and values, thereby ensuring the consistency and comparability of the training data. These historical change information at least includes the business volume, order volume, production activity level, etc. of the past cycles, which are directly or indirectly related to the proportion of green electricity in the same period of history, and can reflect the dynamic characteristics of green electricity demand with the fluctuation of business volume. The normalized data, that is, the normalized data, is then input into the initial neural network with a hidden layer. Through training, the mapping relationship between the normalized data and the green electricity demand is learned, and its weight and structure are gradually optimized, and finally the target neural network that can efficiently predict the green electricity demand is obtained. Data normalization ensures the quality of the training data, avoids the interference of numerical differences on model training, and enables the neural network to focus more on learning the intrinsic connection between data, rather than being confused by the surface numerical size.

[0059] Optionally, the neural network at least includes: BP neural network. The green electricity percentage input by the user is used as the input layer, and the BP neural network is used to correct the previous prediction, and the green electricity demand for the next period is predicted and calculated; the implementation steps of the BP neural network model are as follows: network initialization, processing the input data as data between (-1,1), setting relevant values; input sample data and corresponding expected output, learning each sample, and calculating the actual output and the output of the hidden layer node according to the input sample; calculating the error, by calculating the difference between the actual output and the expected output, calculating the error between the output layer and the hidden layer; making corrections based on the obtained error and the input connection weights of each neuron; calculating the global error E, and judging whether E reaches the preset accuracy or number of learning times; ending the algorithm if it reaches it, otherwise returning to the second step and continuing learning.

[0060] Use the mathematical expression to determine the number of hidden layer neurons:

[0061]

[0062] Among them, n2 is the number of hidden layer neurons, n1 is the number of input layer nodes, m is the number of output layer nodes, and a is an integer between 1 and 10.

[0063] The tanh activation function of the BP neural network is:

[0064]

[0065] The tanh activation function, full name Hyperbolic Tangent function, is used to convert the input signal of the neuron into the output signal. The output range of the tanh function is (-1, 1), and the function can map any real number input to this interval. xand e -x It is the positive and negative exponential form of the natural exponential function.

[0066] Step S108, determining the green electricity demand of the target park according to the correction ratio and the total electricity demand.

[0067] It can be understood that the green electricity demand of the park is determined based on the correction ratio and combined with the total electricity demand of the target park. The correction ratio reflects the adjustment of the green electricity demand of the park relative to the initial forecast value after taking into account the impact of business volume changes, seasonal factors, etc. on green electricity demand. By applying this ratio to the total electricity demand, a forecast value that is closer to the actual green electricity demand of the park can be obtained. The accuracy and practicality of the green electricity demand forecast are improved, providing a more refined decision-making basis for the energy management of the park.

[0068] Through the above-mentioned step S102, the total power data of the target park is obtained, wherein the power data includes green electricity data using clean energy; step S104, based on the time series characteristics of the power data, the total power demand of the target park is determined; step S106, based on the business volume change information of the target park, the predetermined green electricity ratio is corrected to determine the correction ratio; step S108, according to the correction ratio and the total power demand, the green electricity demand of the target park is determined, which can achieve the purpose of improving energy utilization and realize the technical effect of improving the accuracy of green electricity demand prediction, thereby solving the technical problem of unsatisfactory accuracy of green electricity demand prediction.

[0069] Based on the above embodiments and optional embodiments, the present application proposes an optional implementation method to construct a prediction method based on green electricity demand, provide a green electricity demand prediction and analysis method for park users, guide users to better participate in the green electricity market, and improve the prediction accuracy of green electricity demand load forecasting in the target geographical area. First, in the data collection module, the power data set collected from users in the park is bundled and analyzed, and then the overall electricity demand in the park is predicted based on the improved time series analysis prediction (PSO-LSTM) method. Finally, based on the green electricity proportion reported by the user, the green electricity proportion is corrected based on the neural network model, taking into account order changes, seasonal changes, historical data, etc., to obtain the user's corrected green electricity demand. Figure 2 is a first schematic diagram of an optional method for determining green electricity demand provided according to an embodiment of the present application, such as Figure 2 As shown, the specific steps of the prediction method are demonstrated. Figure 3 is a second schematic diagram of an optional method for determining green electricity demand provided according to an embodiment of the present application, such as Figure 3 As shown in Figure 2, the process of calculating the proportion of green electricity is demonstrated.

[0070] Step S1, the system performs data bundling analysis on the user's data based on the power data collected in the park;

[0071] Collect data, using sensor equipment to collect load data, weather data, time information data, power equipment information, and other information including order volume, business transformation needs, etc., and bundle them according to the load. The collected data needs to be preprocessed, that is, to remove data with obvious deviations and problems, normalize the data, and scale it to the same range. The preprocessing process mainly includes data cleaning and data standardization. Data preprocessing directly affects the accuracy and stability of the prediction model.

[0072] Through statistical analysis, visual inspection and the use of anomaly detection algorithms, we can remove obvious outliers and noise in the data set to ensure the data quality of the input model. Outliers may be caused by sensor failure, data transmission errors or human error. Outliers will distort the true distribution of the data and lead to biased prediction results.

[0073] Scaling the data to the same interval, usually [0, 1] or [-1, 1], data standardization can eliminate the dimensional differences between different input features, so that the model will not be biased towards features with larger values ​​when processing data. Data standardization is also called data normalization. The mathematical expression for normalized data processing is:

[0074]

[0075] Where X max and X min Respectively represent the maximum and minimum values ​​of the indicator x; x N is the normalized load data, with a value between 0 and 1. The purpose of data normalization is to eliminate the influence of dimension.

[0076] Data preprocessing methods include at least feature engineering, which is to generate more meaningful features by transforming and combining the original data to enhance the model's predictive ability. Seasonal indicators and weekday / weekend indicators can be extracted from time information data, or the cyclical fluctuation characteristics of the business can be calculated from the order change volume. In addition, the preprocessing steps can also include missing value filling, data smoothing (such as moving average) and extraction of time series features.

[0077] Since the demand change trend of green power load is not only affected by its inherent characteristics, but also by a variety of interference factors, especially meteorological factors and social activities, temperature and humidity are the most direct factors affecting the load among meteorological factors. In different seasons and even different time periods, the growth of clean energy will directly affect the demand for green electricity in the park. In addition, social activities cannot be quantitatively analyzed, but social activities are strongly correlated with the number of business orders, and there will be great differences in the load levels and change trends of different date types such as different months, working days, weekends and holidays. For the purpose of short-term prediction of green electricity demand at a certain day and time, the sample data is reconstructed to determine the input and output of the prediction model. Specifically, Table 1 is a schematic table of an optional green electricity demand determination method provided according to an embodiment of the present application. As shown in Table 1, X1 to X17 respectively represent various load-related information that need to be considered, such as the green electricity load information of the day before the time T to be tested on the prediction day, and other inputs are similar and will not be repeated. The final output Y is the green electricity demand at the time T of the prediction day.

[0078] Table 1

[0079]

[0080] Step S2: predicting the power demand data in the park based on the above data results using an improved time series prediction analysis algorithm;

[0081] Build an LSTM model. Determine the parameters that need to be optimized, the number of neurons L1, the learning rate ε, and the number of training iterations k, and determine their respective optimization ranges.

[0082] Initialize PSO parameters and determine the fitness function of particles, including the initial velocity and position of particles, learning weights, number of training times and scale, etc.; use the MAPE value of the prediction model as the fitness function of particles to find the optimal model parameters.

[0083] Compare the fitness values ​​of particles, find the individual optimal position and the global optimal position, and update the optimal fitness value.

[0084] Determine whether the maximum number of iterations has been reached. If so, pass the optimal parameters to the LSTM model for training and prediction. If not, return to step 1 to compare the fitness values ​​of the particles.

[0085] The three hyperparameters of LSTM are used as the parameters to be optimized for PSO, and the mathematical expression of the fitness function is as follows:

[0086] F(x)=min(MAPE)

[0087] The mean absolute percentage error (MAPE) refers to the average value of the percentage error between the actual value and the predicted value of electricity consumption:

[0088]

[0089] Where Y a Indicates the actual value of the load, Y p represents the predicted value of the load, n represents the number of predicted load points, and t represents the tth predicted load point.

[0090] Step S3, predicting the green electricity proportion of the user in the next period based on the system prediction result and the green electricity proportion prediction provided by the user.

[0091] Based on the data collected from previous surveys, such as order volume, seasonal characteristics, time information, business transformation demand willingness, and green electricity ratio data, data training was conducted to obtain appropriate weights;

[0092] The green electricity percentage input by the user is used as the input layer, and the BP neural network is used to correct the previous prediction, and the green electricity demand for the next period is predicted and calculated. The implementation steps of the BP neural network model are as follows: network initialization, processing the input data as data between (-1,1), and setting relevant values; input sample data and the corresponding expected output, learning each sample, and calculating the actual output and the output of the hidden layer node based on the input sample; calculating the error, by calculating the difference between the actual output and the expected output, the error between the output layer and the hidden layer is calculated; correction is made based on the obtained error and the input connection weights of each neuron; calculating the global error E, and judging whether E reaches the preset accuracy or learning times; if it reaches it, the algorithm ends, otherwise it returns to the second step and continues learning.

[0093] Use the mathematical expression to determine the number of hidden layer neurons:

[0094]

[0095] Among them, n2 is the number of hidden layer neurons, n1 is the number of input layer nodes, m is the number of output layer nodes, and a is an integer between 1 and 10.

[0096] The tanh activation function of the BP neural network is:

[0097]

[0098] The tanh activation function, full name Hyperbolic Tangent function, is used to convert the input signal of the neuron into the output signal. The output range of the tanh function is (-1, 1), and the function can map any real number input to this interval. x and e-x It is the positive and negative exponential form of the natural exponential function.

[0099] The green demand forecasting system includes a data collection module, a data processing module and a data feedback module.

[0100] The data collection module is used to obtain the initial green electricity information data. It collects and pre-processes load data, weather data, time information data, power equipment information, and other information including the company's foreign trade order volume, company transformation needs, etc.

[0101] The data processing module is used to calculate data related to green electricity demand. Based on the preprocessed data, the user's load data is summed to obtain the total load curve (user bundling); based on the green electricity data set, the particle swarm algorithm (PSO) is used to optimize the three hyperparameters of the LSTM model, namely the number of neurons, learning rate, and number of training times. The optimized hyperparameter data is substituted into the LSTM model to predict electricity demand based on load bundling data, time information, weather information, etc.

[0102] The data feedback module is used to help users predict the next green electricity demand. According to the green electricity percentage input by the user, the artificial network neural model (BP neural network) is used to make prediction corrections; the data collected from previous surveys, such as order volume, seasonal characteristics, time information, business transformation demand willingness, etc., are used as hidden layers, and data training is performed based on the green electricity percentage data of the same period. After obtaining the appropriate weights, the green electricity percentage input by the user is used as the input layer, and the green electricity percentage forecast for the next period is finally calculated.

[0103] The above optional implementation methods can achieve at least the following effects: by predicting future green electricity demand and rationally planning energy supply, we can better understand the trends and changes in the energy market, thereby formulating a reasonable energy supply plan and ensuring a stable supply of energy; adjust electricity consumption in production and life, improve energy efficiency, reduce waste, and reduce energy costs; understand the market demand and potential for green energy, promote the development and application of green energy, and promote sustainable economic development; effectively evaluate various energy projects and optimize energy investment decisions.

[0104] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0105] In this embodiment, a green power demand determination device is also provided, which is used to implement the above embodiments and preferred implementation modes, and will not be repeated hereafter. As used below, the terms "module" and "device" may be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0106] According to an embodiment of the present application, a device embodiment for implementing a method for determining green electricity demand is also provided. Figure 4 is a schematic diagram of an optional green electricity demand determination device provided according to an embodiment of the present application, such as Figure 4 As shown, the green electricity demand determination device includes a power data acquisition module 402, a power demand determination module 404, a correction ratio determination module 406, and a green electricity demand determination module 408. The device is described below.

[0107] The power data acquisition module 402 is used to acquire the total power data of the target park, wherein the power data includes green power data using clean energy;

[0108] The power demand determination module 404 is connected to the power data acquisition module 402 and is used to determine the total power demand of the target park based on the time series characteristics of the power data;

[0109] The correction ratio determination module 406 is connected to the power demand determination module 404 and is used to correct the predetermined green power ratio based on the business volume change information of the target park and determine the correction ratio;

[0110] The green electricity demand determination module 408 is connected to the correction ratio determination module 406 and is used to determine the green electricity demand of the target park according to the correction ratio and the total electricity demand.

[0111] In a green electricity demand determination device provided by an embodiment of the present application, a power data acquisition module 402 is set to acquire the total power data of the target park, wherein the power data includes green electricity data using clean energy; a power demand determination module 404 is connected to the power data acquisition module 402, and is used to determine the total power demand of the target park based on the time series characteristics of the power data; a correction ratio determination module 406 is connected to the power demand determination module 404, and is used to correct the predetermined green electricity ratio based on the business volume change information of the target park and determine the correction ratio; a green electricity demand determination module 408 is connected to the correction ratio determination module 406, and is used to determine the green electricity demand of the target park according to the correction ratio and the total power demand. The purpose of improving energy utilization is achieved, and the technical effect of improving the accuracy of green electricity demand prediction is achieved, thereby solving the technical problem of unsatisfactory accuracy of green electricity demand prediction.

[0112] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0113] It should be noted that the power data acquisition module 402, power demand determination module 404, correction ratio determination module 406, and green power demand determination module 408 correspond to steps S102 to S108 in the embodiment, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments. It should be noted that the modules as part of the device can be run in a computer terminal.

[0114] It should be noted that the optional or preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0115] The green electricity demand determination device may further include a processor and a memory. The power data acquisition module 402, the power demand determination module 404, the correction ratio determination module 406, the green electricity demand determination module 408, etc. are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions.

[0116] The processor includes a kernel, which retrieves the corresponding program unit from the memory. There can be one or more kernels. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0117] An embodiment of the present application provides a non-volatile storage medium having a program stored thereon, and the program, when executed by a processor, implements a method for determining green electricity demand.

[0118] The embodiment of the present application provides an electronic device, which includes a processor, a memory, and a program stored in the memory and can be run on the processor. When the processor executes the program, the following steps are implemented: obtaining the total power data of the target park, wherein the power data includes green power data using clean energy; determining the total power demand of the target park based on the time series characteristics of the power data; correcting the predetermined green power ratio based on the business volume change information of the target park and determining the correction ratio; determining the green power demand of the target park according to the correction ratio and the total power demand. The device in this article can be a server, a PC, etc.

[0119] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: obtaining total power data of a target park, wherein the power data includes green power data using clean energy; determining the total power demand of the target park based on the time series characteristics of the power data; correcting a predetermined green power ratio based on business volume change information of the target park, and determining a correction ratio; and determining the green power demand of the target park according to the correction ratio and the total power demand.

[0120] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0121] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0122] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0123] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0124] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0125] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0126] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0127] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0128] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0129] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for determining green electricity demand, characterized in that: include: Obtaining total power data of the target park, wherein the power data includes green power data using clean energy; Determining the total power demand of the target park based on the time series characteristics of the power data; Based on the business volume change information of the target park, the predetermined green electricity ratio is corrected to determine the correction ratio; The green electricity demand of the target park is determined according to the correction ratio and the total electricity demand.

2. The method according to claim 1, characterized in that The obtaining of total power data of the target park includes: Using a plurality of sensors to collect data from the target park respectively, and obtaining the collected data corresponding to the plurality of sensors respectively; The total power data is obtained by integrating the collected data corresponding to the various sensors.

3. The method according to claim 2, characterized in that The collected data corresponding to the various sensors respectively include at least: load data, weather data, time information data, power equipment information, park business volume, and business transformation requirements.

4. The method according to claim 1, characterized in that The determining the total power demand of the target park based on the time series characteristics of the power data includes: Determine a hyperparameter of an initial long short-term memory network, wherein the hyperparameter is a parameter determined before training the initial long short-term memory network; Based on the historical data of the target park in the historical period, the initial long short-term memory network of the hyperparameter is trained to obtain a target long short-term memory network; The target long short-term memory network is used to determine the total power demand based on the time series characteristics.

5. The method according to claim 4, characterized in that The hyperparameters are multiple, including the number of neurons, the learning rate, and the number of training iterations. The hyperparameters for determining the initial long short-term memory network include: Determine the fitness function based on the mean absolute percentage error between the actual value and the predicted value of the power; Using a particle swarm algorithm, determining fitness function values ​​corresponding to a plurality of particles, wherein the plurality of particles respectively represent combinations of the number of neurons, the learning rate, and the number of training iterations; Determine the target particle based on the fitness function value that minimizes the mean absolute percentage error between the actual value and the predicted value of the electric quantity; A hyperparameter is determined based on a combination of the number of neurons indicated by the target particle, the learning rate, and the number of training iterations.

6. The method according to claim 1, characterized in that The step of revising the predetermined green electricity ratio based on the business volume change information of the target park and determining the correction ratio includes: Based on the business volume change information of the target park, a hidden layer of the initial neural network is set to obtain the set initial neural network; The green electricity proportion in the same period of history matching the cycle characteristics of the current cycle is used to train the initial neural network after the setting to obtain a target neural network; Based on the predetermined green electricity proportion, the target neural network is used for processing to obtain the correction ratio.

7. The method according to claim 6, characterized in that The green electricity proportion in the same period of history matching the cycle characteristics of the current cycle is used to train the set initial neural network to obtain the target neural network, including: Performing data normalization processing based on the historical change information before the current period to obtain normalized data, wherein the historical change information corresponds to the green electricity proportion in the same period of history; The normalized data is input into the initial neural network for training to obtain the target neural network.

8. A device for determining green electricity demand, characterized in that: include: A power data acquisition module is used to acquire total power data of a target park, wherein the power data includes green power data using clean energy; A power demand determination module, configured to determine the total power demand of the target park based on the time series characteristics of the power data; A correction ratio determination module is used to correct the predetermined green electricity ratio based on the business volume change information of the target park and determine the correction ratio; The green electricity demand determination module is used to determine the green electricity demand of the target park according to the correction ratio and the total electricity demand.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the green electricity demand determination method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: include: One or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the green electricity demand determination method according to any one of claims 1 to 7.

Images