Load power prediction method and device for industrial park, terminal and storage medium

Abstract

The invention provides a load power prediction method and device for an industrial park, a terminal and a storage medium. The method comprises the following steps: acquiring historical industrial load power data and power related data of a target park; wherein the power related data comprises meteorological information, social economic data and calendar information; performing K-Means clustering on the power related data according to the meteorological information to obtain a plurality of clusters; taking the electric power related data of the sunny cluster as a reference, respectively calculating the cosine similarity between the electric power related data of each cluster and the electric power related data of the sunny cluster, and obtaining a meteorological saturation factor of each cluster; the industrial load power value of the HBI-LSTM model at the previous moment and the correction coefficient of the backward process of the HBI-LSTM model are obtained, and the industrial load power is calculated and updated; and when the correction coefficient of the backward process of the HBI-LSTM model is smaller than a correction coefficient threshold value, stopping calculation, and obtaining a load power prediction value of the target park. According to the invention, the load power prediction rate can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method, device, terminal and storage medium for predicting load power in an industrial park. Background Art

[0002] With the increasing application of big data technology, power load forecasting is one of the important scenarios of the energy data platform and the basis for the economic operation of the power system. Power load forecasting is of great significance to the development and stability of the power industry. It can provide important decision-making support for power production, dispatching and operation, help formulate more reasonable energy planning and power production plans, and provide effective data support for development. At the same time, power load forecasting can also improve the ability of power emergency response, promote the stable operation of power supply, and reduce the risk of power supply.

[0003] In the prior art, when predicting the electric load power of an industrial park, most of the existing historical load power data is used and various influencing factors are considered to find out the changing pattern of load power, and different methods are used to establish models to predict the load power situation at future times. Since multiple influencing factors are considered, it is necessary to obtain multiple types of data, resulting in low computational efficiency of the load power prediction model. Summary of the invention

[0004] The embodiments of the present invention provide a method, device, terminal and storage medium for load power prediction of an industrial park to solve the problem of low computational efficiency of a load power prediction model.

[0005] In a first aspect, an embodiment of the present invention provides a method for predicting load power of an industrial park, comprising:

[0006] Obtain historical industrial load power data and power-related data of the target park; power-related data includes meteorological information, socio-economic data, and calendar information;

[0007] K-Means clustering is performed on the power-related data according to the meteorological information to obtain multiple clusters; the clusters include the clear day cluster, the cloudy day cluster, the strong wind cluster, the rainy day cluster and the snowy day cluster. The clear day cluster is the cluster with the highest industrial load power at the centroid among all clusters.

[0008] Taking the power-related data of the clear cluster as the benchmark, the cosine similarity between the power-related data of each cluster and the power-related data of the clear cluster is calculated to obtain the meteorological saturation factor of each cluster.

[0009] Obtain the industrial load power value of the previous moment of the HBI-LSTM model and the correction coefficient of the backward process of the HBI-LSTM model, input the socio-economic data, calendar information, meteorological saturation factor and the industrial load power value of the previous moment into the HBI-LSTM model, and calculate and update the industrial load power;

[0010] When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value of the target park.

[0011] In a possible implementation, obtaining a correction coefficient of a backward process of the HBI-LSTM model includes:

[0012] Obtain the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model;

[0013] According to the confidence of the forward process of the HBI-LSTM model, the prior knowledge of the forward process of the HBI-LSTM model is obtained;

[0014] Combined with the preset formula, and based on the prior knowledge of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model, the correction coefficient of the backward process of the HBI-LSTM model is calculated. The preset formula is:

[0015]

[0016] Where P(y) is the prior knowledge of the forward process of the HBI-LSTM model, P(x) is the prior knowledge of the backward process of the HBI-LSTM model, and P(x|y) is the probability of x occurring when y occurs.

[0017] In a possible implementation, obtaining the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model includes:

[0018] Obtain the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model;

[0019] Calculate the first cosine similarity of the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model, and obtain the confidence of the forward process of the HBI-LSTM model according to the first cosine similarity;

[0020] Obtain the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process of the HBI-LSTM model;

[0021] The second cosine similarity between the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process is calculated, and the prior knowledge of the backward process of the HBI-LSTM model is obtained according to the second cosine similarity.

[0022] In a possible implementation, the correction coefficient threshold is determined according to the running rate, hardware resource consumption and data missing degree of the HBi-LSTM model.

[0023] In a possible implementation, before obtaining power-related data, the following steps are also included:

[0024] Acquiring initial data of power-related data; wherein the initial data includes initial data of meteorological information;

[0025] Performing smoothing and median filtering on the initial data of the meteorological information to obtain processed meteorological information, and obtaining processed power-related data based on the processed meteorological information;

[0026] Standardizing the processed power-related data to obtain standard power-related data;

[0027] The missing values ​​of standard power-related data are filled to obtain power-related data.

[0028] In a possible implementation, missing value filling includes: forward filling and linear interpolation.

[0029] In one possible implementation, the meteorological information includes temperature, humidity, light intensity, and wind speed;

[0030] Socioeconomic data include electricity prices and socioeconomic curves;

[0031] Calendar information includes holidays and seasons.

[0032] In a second aspect, an embodiment of the present invention provides a load power prediction device for an industrial park, comprising:

[0033] A data acquisition module is used to acquire historical industrial load power data and power-related data of the target park; wherein the power-related data includes meteorological information, socio-economic data and calendar information;

[0034] The clustering module is used to perform K-Means clustering on the power-related data according to the meteorological information to obtain multiple clusters; wherein the clusters include a clear day cluster, a cloudy day cluster, a strong wind cluster, a rainy day cluster, and a snowy day cluster, and the clear day cluster is the cluster with the highest industrial load power at the centroid among all clusters;

[0035] A meteorological saturation factor calculation module is used to calculate the cosine similarity between the power-related data of each cluster and the power-related data of the clear cluster based on the power-related data of the clear cluster, so as to obtain the meteorological saturation factor of each cluster;

[0036] The load power prediction module is used to obtain the industrial load power value of the previous moment of the HBI-LSTM model and the correction coefficient of the backward process of the HBI-LSTM model, input the socio-economic data, calendar information, meteorological saturation factor and the industrial load power value of the previous moment into the HBI-LSTM model, and calculate and update the industrial load power;

[0037] When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value of the target park.

[0038] In a third aspect, an embodiment of the present invention provides a terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the first aspect or any possible implementation method of the first aspect are implemented.

[0039] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the first aspect or any possible implementation method of the first aspect.

[0040] The embodiment of the present invention provides a load power prediction method, device, terminal and storage medium for an industrial park, clusters power-related data according to meteorological information, thereby dividing the power-related data into five clusters, and taking the clear cluster as a benchmark, respectively calculates the cosine similarity of the power-related data of each cluster with the power-related data of the clear cluster, and obtains the meteorological saturation factor of each cluster. In this way, the meteorological information is converted into the meteorological saturation factor, which simplifies the complexity of the data and improves the operation efficiency. The socio-economic data, calendar information, meteorological saturation factor and the industrial load power value at the previous moment are input into the HBI-LSTM model, and the power load power is calculated and updated, wherein a correction coefficient is set in the HBI-LSTM model, and when the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value, thereby reducing the prediction time of the HBI-LSTM model. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0042] Figure 1 It is a flow chart of the implementation of the load power prediction method for an industrial park provided by an embodiment of the present invention;

[0043] Figure 2 It is a structural schematic diagram of a load power prediction device for an industrial park provided by an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0046] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0047] See also Figure 1 , which shows a flow chart of the implementation of the load power prediction method for an industrial park provided by an embodiment of the present invention, which is described in detail as follows:

[0048] Step 101, obtaining historical industrial load power data and power-related data of a target park; wherein the power-related data includes meteorological information, socio-economic data, and calendar information.

[0049] In this embodiment, the historical industrial load power data is the power usage in the target park over the past period of time. These data can help understand the power demand pattern of the park, such as which periods are peak and which are low. A lot of power consumption in the industrial park (such as air conditioning and heating) is closely related to meteorological information. Socioeconomic data helps to analyze the relationship between power consumption and socioeconomic activities and predict future power demand. Calendar information helps to identify seasonal and cyclical changes in power consumption, such as holidays that may affect industrial production.

[0050] In a possible implementation, the meteorological information includes temperature, humidity, light intensity and wind speed; the socio-economic data includes electricity prices and socio-economic curves; and the calendar information includes holidays and season information.

[0051] In this embodiment, these data can help park managers or energy planners make more accurate predictions, develop more effective energy management strategies, optimize energy efficiency, and reduce energy costs.

[0052] Step 102, performing K-Means clustering on the power-related data according to the meteorological information to obtain multiple clusters; wherein the clusters include a clear day cluster, a cloudy day cluster, a strong wind cluster, a rainy day cluster and a snowy day cluster, and the clear day cluster is the cluster with the highest industrial load power at the centroid among all clusters.

[0053] In this embodiment, K-Means (K-means clustering) is an unsupervised machine learning algorithm. The present invention uses the K-Means clustering algorithm to classify power-related data according to meteorological information, and divides the power-related data into several different categories, each of which represents a specific weather condition. Such classification helps to identify power consumption patterns under different weather conditions. In clear weather, the industrial load power of the park reaches its peak. This analysis is very useful for energy management and predicting power demand, and can help the park plan and dispatch power resources more effectively.

[0054] Specifically, in K-Means clustering, each cluster has a centroid, which is the center position of all points in the cluster and represents the average level of industrial load power in the cluster. Among them, the "clear cluster" is the cluster with the highest industrial load power in the centroid among all clusters, which means that under clear weather conditions, the industrial load power of the target park reaches the highest. The specific clustering method is as follows:

[0055] (1) Format the power-related data to obtain a power-related data set D.

[0056]

[0057] Where d is the total number of days of electricity-related data, seq is the total number of points of electricity-related data per day, and each data point D d,seq Including maximum temperature, minimum temperature, humidity, light intensity, wind speed, electricity price, socio-economic curves, holiday information and seasonal information.

[0058] (2) Selection of the initial cluster center. When selecting, choose the most dispersed center points. First, randomly select a point as the center point of the first initial cluster. Then select the point farthest from this point as the center point of the second initial cluster. Then select the point with the largest distance from the first two points as the center point of the third initial cluster. The distance between the two cluster centers is measured by Euclidean distance. Euclidean distance is a measure of the straight-line distance between two points. It is applicable to two-dimensional or higher-dimensional spaces. The formula is as follows:

[0059]

[0060] Where L is the Euclidean distance, x m ,y m is the center point of the cluster, and m is the data dimension.

[0061] This process is repeated until K initial cluster centers are selected. The K initial center points are recorded as

[0062] (3) For each point in the data set, assign it to the cluster where the nearest centroid is located. For each cluster, recalculate its centroid (i.e., the mean of all points in the cluster). The loss function is defined as the sum of squared errors of each sample from the center point of the cluster to which it belongs. The formula is as follows:

[0063]

[0064] Where, d i is the i-th data, c i Yes i The cluster to which it belongs, is the center point of the cluster, and M is the total number of samples. Repeat step (3) until the function converges and the position of the centroid no longer changes significantly.

[0065] (4) Determination of K value. When determining K value, the weighted average value R of the radius of all clusters and the weighted average value L of the average centroid distance of all clusters are selected as the basis. When K value is 2 to 9, R and L no longer change significantly after K value is greater than 5, so K value is selected as 5.

[0066] (5) After multiple iterations of the model, five clusters and their centroids were obtained. According to the centroid of each cluster, its meteorological type was manually labeled as sunny, cloudy, windy, rainy, and snowy. The cluster with the highest industrial load power consumption was marked as sunny and recorded as 1.

[0067] Step 103, taking the power-related data of the clear cluster as a reference, respectively calculate the cosine similarity between the power-related data of each cluster and the power-related data of the clear cluster, and obtain the meteorological saturation factor of each cluster.

[0068] In this embodiment, by using cosine similarity, the similarity between power-related data under different meteorological conditions and power-related data under clear weather can be quantified, thereby obtaining an indicator reflecting the impact of meteorological conditions on power-related data - meteorological saturation factor. This indicator helps to understand the impact of different meteorological conditions on power consumption patterns, and thus provides a basis for energy management and optimization.

[0069] Specifically, taking the electricity-related data of the "clear cluster" as a benchmark, the meteorological saturation factor of the "clear cluster" is set to 1. The electricity-related data under all other meteorological conditions will be compared with the data under clear weather. For each meteorological cluster (cloudy cluster, strong wind cluster, rainy cluster, snowy cluster), the cosine similarity of its electricity-related data and the electricity-related data of the clear cluster is calculated respectively. The cosine similarity value obtained by calculation can be regarded as the "meteorological saturation factor" (between 0-1) of each cluster, and the meteorological saturation factor represents the correlation between such weather and industrial load power. The meteorological information is converted into a meteorological saturation factor to simplify the complexity of the data and improve the computing efficiency. The present invention makes good use of the meteorology, deeply mines the meteorological information, and quantifies the weather data by defining the meteorological saturation factor.

[0070] Step 104, obtain the industrial load power value of the previous moment of the HBI-LSTM model and the correction coefficient of the backward process of the HBI-LSTM model, input the socio-economic data, calendar information, meteorological saturation factor and the industrial load power value of the previous moment into the HBI-LSTM model, and calculate and update the industrial load power; when the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, stop the calculation and obtain the load power forecast value of the target park.

[0071] In this embodiment, the load power of the target park is predicted by combining socio-economic data, calendar information, meteorological saturation factor and historical load power data through the HBI-LSTM model, and the correction coefficient and correction coefficient threshold are set. When the correction coefficient exceeds the correction coefficient threshold, the calculation is stopped to obtain the final prediction result, which can improve the accuracy and reliability of load power prediction, which is crucial for the planning and operation of the power system.

[0072] Specifically, the industrial load power value at the previous moment is the historical data used by the model to predict the load power at the current moment. The correction coefficient of the backward process is the coefficient used by the model to adjust the prediction result during the prediction process to improve the accuracy of the prediction. Socioeconomic data, calendar information, meteorological saturation factors, and industrial load power values ​​at the previous moment are input into the HBI-LSTM model as input features. The HBI-LSTM model uses these input data to calculate and update the predicted value of industrial load power. This model is usually able to capture long-term dependencies in time series data and improve the accuracy of prediction through bidirectional data flow (i.e., past and future data). In the backward process of the model, a correction coefficient is calculated. If this correction coefficient is less than the preset correction coefficient threshold, the model will consider the prediction result to be accurate enough and no further correction is required. The predicted value obtained at this time is considered to be the final load power prediction value. Finally, the model outputs the load power prediction value of the target park, which can be used for energy management and scheduling decisions. The specific formula of the HBI-LSTM model is as follows:

[0073] f t =σ g (W f x t +U f h t-1 +b f )

[0074] i t =σ g (W i x t +U i h t-1 +b i )

[0075] o t =σ g (W o x t +U o h t-1 +b o )

[0076]

[0077] h t =o t ⊙σ h (c t )

[0078] In the formula, f t 、i t , o t They are forget gate, input gate and output gate, x tis the input vector of the LSTM logic gate, including the power load power and characteristic data at time t. The characteristic data includes meteorological saturation factor, power price, socio-economic curve, holiday information and seasonal information. t-1 is the output value of LSTM at the previous moment, and b is the bias term of each logic gate.

[0079] In a possible implementation, the correction coefficient of the backward process of the HBI-LSTM model is obtained, and the following processing is performed: the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model are obtained; according to the confidence of the forward process of the HBI-LSTM model, the prior knowledge of the forward process of the HBI-LSTM model is obtained, and the specific processing is as follows: combining a preset formula, and according to the prior knowledge of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model, the correction coefficient of the backward process of the HBI-LSTM model is calculated, and the preset formula is:

[0080]

[0081] Where P(y) is the prior knowledge of the forward process of the HBI-LSTM model, P(x) is the prior knowledge of the backward process of the HBI-LSTM model, and P(x|y) is the probability of x occurring when y occurs.

[0082] In this embodiment, a preset formula is used to calculate the correction factor of the backward process by combining the confidence and prior knowledge of the forward and backward processes of the model. This correction factor can be used to adjust the prediction of the model to make it more accurate. P(x|y) is the probability of the forward process occurring when the backward process occurs.

[0083] In a possible implementation, the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model are obtained, and the specific processing is as follows: the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model are obtained; the first cosine similarity of the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model is calculated, and the confidence of the forward process of the HBI-LSTM model is obtained according to the first cosine similarity; the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process of the HBI-LSTM model are obtained; the second cosine similarity of the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process is calculated, and the prior knowledge of the backward process of the HBI-LSTM model is obtained according to the second cosine similarity.

[0084] In this embodiment, by calculating the cosine similarity between the industrial load power values ​​at different times in the forward process and the backward process, the confidence and prior knowledge of the model can be obtained. This information provides a measure of the accuracy and stability of the model prediction, which is crucial for adjusting and optimizing the prediction performance of the HBI-LSTM model. The calculation formula for the confidence of the forward process of the HBI-LSTM model is as follows:

[0085]

[0086] In the formula, h t Industrial load power value at the current moment, h t-1 It is the industrial load power value at the previous moment.

[0087] The state of the improved HBI-LSTM model at each time point is expressed as:

[0088]

[0089] t m is the time point when the correction coefficient crosses the correction coefficient threshold, is the forward industrial load power value, It is the backward industrial load power value.

[0090] In a possible implementation, the correction coefficient threshold is determined according to the running rate, hardware resource consumption and data missing degree of the HBi-LSTM model.

[0091] In this embodiment, when determining the correction coefficient threshold of the HBi-LSTM model, it is necessary to comprehensively consider the three factors of the model's operating rate, hardware resource consumption, and data missing degree, so as to ensure that the model can run efficiently and adapt to the incompleteness of the data while ensuring the accuracy of the prediction. The setting of the correction coefficient threshold helps to optimize the performance and resource utilization of the model. Based on the prior Bayes, the present invention improves the efficiency of BI-LSTM and reduces the hardware operation cost.

[0092] Specifically, the running rate of the model is affected by factors such as model complexity, data volume, and computing resources. When determining the correction factor threshold, the running rate of the model needs to be considered to ensure that the model can respond quickly while maintaining the accuracy of the prediction. It involves the computing resources required for the model to run, such as CPU, memory, and storage. Hardware resource consumption directly affects the performance and scalability of the model. When determining the correction factor threshold, it is necessary to balance the prediction accuracy and resource consumption of the model to avoid excessive consumption of hardware resources. The proportion of missing values ​​in the input data during model training and prediction. Missing data may affect the prediction performance of the model, especially in time series prediction. When determining the correction factor threshold, it is necessary to consider the degree of data missing so that the model can adapt to the incompleteness of the data and minimize the prediction error. The correction factor threshold is a preset value used to determine when to stop the backward process correction of the model. When the correction factor is lower than this threshold, the model believes that the prediction result is accurate enough and no further correction is required.

[0093] The embodiment of the present invention clusters the power-related data according to the meteorological information, thereby dividing the power-related data into five clusters, and taking the clear cluster as the benchmark, respectively calculates the cosine similarity of the power-related data of each cluster with the power-related data of the clear cluster, and obtains the meteorological saturation factor of each cluster. In this way, the meteorological information is converted into the meteorological saturation factor, which simplifies the complexity of the data and improves the operation efficiency. The socio-economic data, calendar information, meteorological saturation factor and the industrial load power value at the previous moment are input into the HBI-LSTM model to calculate and update the power load power, wherein a correction coefficient is set in the HBI-LSTM model. When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value, thereby reducing the prediction time of the HBI-LSTM model.

[0094] In a possible implementation, before obtaining power-related data, the following processing may be performed: obtaining initial data of power-related data; wherein the initial data includes initial data of meteorological information; performing smoothing and median filtering on the initial data of meteorological information to obtain processed meteorological information, and obtaining processed power-related data based on the processed meteorological information; performing standardization on the processed power-related data to obtain standard power-related data; and performing missing value filling on missing values ​​of standard power-related data to obtain power-related data.

[0095] In this embodiment, before actually using the power-related data, a series of data preprocessing steps are required to improve the quality of the data, including smoothing, median filtering, standardization, and missing value filling, to ensure the accuracy and availability of the data.

[0096] Specifically, moving average is used to reduce random fluctuations and noise in meteorological data and make the data smoother. Median filtering replaces each data point with the median of the point and its neighboring points to reduce the impact of outliers. After smoothing and median filtering, cleaner and more accurate meteorological information is obtained. Standardization scales the data to a specific range (such as 0 to 1) or with a specific mean and standard deviation, which helps with subsequent data analysis and model training. When filling missing values, if missing values ​​are found, appropriate methods (such as forward filling and linear interpolation, etc.) need to be used to fill these missing values ​​to ensure data integrity.

[0097] In a possible implementation, missing values ​​are filled by the following process: forward filling and linear interpolation.

[0098] In this embodiment, in practical applications, these two methods can be used alone or in combination to meet different data processing requirements. For example, linear interpolation can be used to estimate missing values ​​first, and then forward filling can be used to process the remaining missing values.

[0099] In other possible implementations, HBI-LSTM may adopt other methods. This application mainly uses HBI-LSTM as an example for illustration.

[0100] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0101] The following is an embodiment of the device of the present invention. For details not described in detail therein, reference may be made to the corresponding method embodiment described above.

[0102] Figure 2 The following is a schematic diagram showing the structure of a load power prediction device for an industrial park provided by an embodiment of the present invention. For ease of explanation, only the parts related to the embodiment of the present invention are shown, which are described in detail as follows:

[0103] like Figure 2 As shown, the load power prediction device 2 of the industrial park includes:

[0104] The data acquisition module 21 is used to acquire the historical industrial load power data and power-related data of the target park; wherein the power-related data includes meteorological information, socio-economic data and calendar information;

[0105] The clustering module 22 is used to perform K-Means clustering on the power-related data according to the meteorological information to obtain multiple clusters; wherein the clusters include a clear sky cluster, a cloudy sky cluster, a strong wind cluster, a rainy sky cluster and a snowy sky cluster, and the clear sky cluster is the cluster with the highest industrial load power at the centroid among all clusters;

[0106] The meteorological saturation factor calculation module 23 is used to calculate the cosine similarity between the power-related data of each cluster and the power-related data of the clear cluster based on the power-related data of the clear cluster, so as to obtain the meteorological saturation factor of each cluster;

[0107] The load power prediction module 24 is used to obtain the industrial load power value of the previous moment of the HBI-LSTM model and the correction coefficient of the backward process of the HBI-LSTM model, input the socio-economic data, calendar information, meteorological saturation factor and the industrial load power value of the previous moment into the HBI-LSTM model, and calculate and update the industrial load power;

[0108] When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value of the target park.

[0109] In a possible implementation, the data acquisition module 21 may be used to:

[0110] Acquiring initial data of power-related data; wherein the initial data includes initial data of meteorological information;

[0111] Performing smoothing and median filtering on the initial data of the meteorological information to obtain processed meteorological information, and obtaining processed power-related data based on the processed meteorological information;

[0112] Standardizing the processed power-related data to obtain standard power-related data;

[0113] The missing values ​​of standard power-related data are filled to obtain power-related data.

[0114] In a possible implementation, the load power prediction module 24 may be used to:

[0115] Get the correction coefficients for the backward process of the HBI-LSTM model, including:

[0116] Obtain the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model;

[0117] According to the confidence of the forward process of the HBI-LSTM model, the prior knowledge of the forward process of the HBI-LSTM model is obtained;

[0118] Combined with the preset formula, and based on the prior knowledge of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model, the correction coefficient of the backward process of the HBI-LSTM model is calculated. The preset formula is:

[0119]

[0120] Where P(y) is the prior knowledge of the forward process of the HBI-LSTM model, P(x) is the prior knowledge of the backward process of the HBI-LSTM model, and P(x|y) is the probability of x occurring when y occurs.

[0121] In a possible implementation, the load power prediction module 24 may be used to:

[0122] Obtain the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model, including:

[0123] Obtain the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model;

[0124] Calculate the first cosine similarity of the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model, and obtain the confidence of the forward process of the HBI-LSTM model according to the first cosine similarity;

[0125] Obtain the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process of the HBI-LSTM model;

[0126] The second cosine similarity between the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process is calculated, and the prior knowledge of the backward process of the HBI-LSTM model is obtained according to the second cosine similarity.

[0127] The embodiment of the present invention clusters the power-related data according to the meteorological information, thereby dividing the power-related data into five clusters, and taking the clear cluster as the benchmark, respectively calculates the cosine similarity of the power-related data of each cluster with the power-related data of the clear cluster, and obtains the meteorological saturation factor of each cluster. In this way, the meteorological information is converted into the meteorological saturation factor, which simplifies the complexity of the data and improves the operation efficiency. The socio-economic data, calendar information, meteorological saturation factor and the industrial load power value at the previous moment are input into the HBI-LSTM model to calculate and update the power load power, wherein a correction coefficient is set in the HBI-LSTM model. When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value, thereby reducing the prediction time of the HBI-LSTM model.

[0128] Figure 3 is a schematic diagram of a terminal provided by an embodiment of the present invention. Figure 3As shown, the terminal 3 of this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, the steps in the load power prediction method embodiments of each industrial park described above are implemented, for example Figure 1 Alternatively, when the processor 30 executes the computer program 32, the functions of the modules in the above-mentioned device embodiments are realized, for example Figure 2 The functions of each module are shown.

[0129] Exemplarily, the computer program 32 may be divided into one or more modules, which are stored in the memory 31 and executed by the processor 30 to implement the present invention. The one or more modules may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program 32 in the terminal 3. For example, the computer program 32 may be divided into Figure 2 The modules shown.

[0130] The terminal 3 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal 3 may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art will appreciate that Figure 3 It is only an example of terminal 3 and does not constitute a limitation on terminal 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal may also include input and output devices, network access devices, buses, etc.

[0131] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0132] The memory 31 may be an internal storage unit of the terminal 3, such as a hard disk or memory of the terminal 3. The memory 31 may also be an external storage device of the terminal 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal 3. Further, the memory 31 may also include both an internal storage unit of the terminal 3 and an external storage device. The memory 31 is used to store the computer program and other programs and data required by the terminal. The memory 31 may also be used to temporarily store data that has been output or is to be output.

[0133] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0134] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0135] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0136] In the embodiments provided by the present invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0137] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0138] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0139] If the integrated module is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the load power prediction method embodiment of each industrial park can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0140] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for predicting load power in an industrial park, characterized in that: include: Acquire historical industrial load power data and power-related data of the target park; wherein the power-related data includes meteorological information, socio-economic data, and calendar information; K-Means clustering is performed on the power-related data according to the meteorological information to obtain multiple clusters; wherein the clusters include a clear day cluster, a cloudy day cluster, a strong wind cluster, a rainy day cluster, and a snowy day cluster, and the clear day cluster is the cluster with the highest industrial load power at the centroid among all clusters; Taking the power-related data of the clear cluster as a reference, respectively calculating the cosine similarity between the power-related data of each cluster and the power-related data of the clear cluster, and obtaining the meteorological saturation factor of each cluster; Obtaining the industrial load power value of the HBI-LSTM model at the previous moment and the correction coefficient of the backward process of the HBI-LSTM model, inputting the socio-economic data, the calendar information, the meteorological saturation factor and the industrial load power value at the previous moment into the HBI-LSTM model, and calculating and updating the industrial load power; When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value of the target park.

2. The load power prediction method for an industrial park according to claim 1, characterized in that: Get the correction coefficients for the backward process of the HBI-LSTM model, including: Acquire the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model; Obtaining prior knowledge of the forward process of the HBI-LSTM model according to the confidence of the forward process of the HBI-LSTM model; Combined with the preset formula, and based on the prior knowledge of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model, the correction coefficient of the backward process of the HBI-LSTM model is calculated. The preset formula is: Wherein, P(y) is the prior knowledge of the forward process of the HBI-LSTM model, P(x) is the prior knowledge of the backward process of the HBI-LSTM model, and P(x|y) is the probability of x occurring when y occurs.

3. The load power prediction method for an industrial park according to claim 2, characterized in that: The obtaining of the confidence of the forward process of the HBI-LSTM model and the prior knowledge of the backward process of the HBI-LSTM model includes: Obtaining the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model; Calculating the first cosine similarity of the industrial load power value at the previous moment and the industrial load power value at the current moment of the forward process of the HBI-LSTM model, and obtaining the confidence of the forward process of the HBI-LSTM model according to the first cosine similarity; Obtaining the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process of the HBI-LSTM model; The second cosine similarity between the industrial load power value at the current moment and the industrial load power value at the previous moment of the backward process is calculated, and the prior knowledge of the backward process of the HBI-LSTM model is obtained according to the second cosine similarity.

4. The load power prediction method for an industrial park according to claim 1, characterized in that: The correction coefficient threshold is determined according to the operating rate, hardware resource consumption and data missing degree of the HBi-LSTM model.

5. The load power prediction method for an industrial park according to claim 1, characterized in that: Before obtaining the power-related data, the method further includes: Acquire initial data of power-related data; wherein the initial data includes initial data of meteorological information; Performing smoothing and median filtering on the initial data of the meteorological information to obtain processed meteorological information, and obtaining processed power-related data according to the processed meteorological information; Performing standardization processing on the processed power-related data to obtain standard power-related data; The missing values ​​of the standard power-related data are filled to obtain the power-related data.

6. The load power prediction method for an industrial park according to claim 5, characterized in that: The missing value filling includes: forward filling and linear interpolation.

7. The load power prediction method for an industrial park according to claim 1, characterized in that: The meteorological information includes temperature, humidity, light intensity and wind speed; The socio-economic data include electricity prices and socio-economic curves; The calendar information includes holidays and seasons information.

8. A load power prediction device for an industrial park, characterized in that: include: A data acquisition module, used to acquire historical industrial load power data and power-related data of the target park; wherein the power-related data includes meteorological information, socio-economic data and calendar information; A clustering module, used for performing K-Means clustering on the power-related data according to the meteorological information to obtain a plurality of clusters; wherein the clusters include a clear sky cluster, a cloudy sky cluster, a strong wind cluster, a rainy sky cluster and a snowy sky cluster, and the clear sky cluster is the cluster with the highest industrial load power at the centroid among the clusters; A meteorological saturation factor calculation module, used to calculate the cosine similarity between the power-related data of each cluster and the power-related data of the clear cluster based on the power-related data of the clear cluster, so as to obtain the meteorological saturation factor of each cluster; A load power prediction module, used to obtain the industrial load power value of the HBI-LSTM model at the previous moment and the correction coefficient of the backward process of the HBI-LSTM model, input the socio-economic data, the calendar information, the meteorological saturation factor and the industrial load power value at the previous moment into the HBI-LSTM model, and calculate and update the industrial load power; When the correction coefficient of the backward process of the HBI-LSTM model is less than the correction coefficient threshold, the calculation is stopped to obtain the load power prediction value of the target park.

9. A terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method as claimed in any one of claims 1 to 7 are implemented.

Images