Energy internet-oriented distributed power consumption measurement evaluation system

Abstract

The invention relates to the technical field of power supply consumption measurement evaluation, and particularly discloses an energy internet-oriented distributed power supply consumption measurement evaluation system, which specifically comprises a related influence screening module, an energy internet-oriented distributed power supply consumption measurement evaluation module, an energy internet-oriented distributed power supply consumption measurement evaluation module, an energy internet-oriented distributed power supply consumption measurement evaluation module and an energy internet-oriented distributed power supply consumption measurement evaluation module, and the collected power supply output power is analyzed to obtain a power supply output value, and related influence data which greatly influence the power supply output value is screened out through analysis. By introducing the power supply output prediction model and combining power grid load analysis, the uncertainty of distributed power supply output can be more comprehensively considered, and by collecting related influence factor data in real time and inputting the data into the prediction model, the power supply output value in the future time period can be calculated. According to the evaluation method based on prediction and real-time data, the accuracy of consumption measurement evaluation of the distributed power supply can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power consumption metering and evaluation, and in particular to a distributed power consumption metering and evaluation system for energy internet. Background Art

[0002] Distributed power sources usually refer to independent power sources with a voltage level of 35kV or below that are arranged near users to meet the power needs of specific power users. These power sources can be renewable energy power generation facilities such as solar photovoltaic, wind power, and small hydropower. After the distributed power sources are connected to the power system, it is necessary to measure the electric energy they generate and evaluate the power system's absorption capacity, that is, the absorption capacity;

[0003] The Chinese invention patent with the publication number CN115330125A, the distributed power consumption metering evaluation method and system for energy internet, discloses that by analyzing the consumption characteristics of new rural power grids, considering the influence of the uncertainty of distributed power output, organically combining the influencing factors such as power quality, power supply reliability and operation safety, a multi-dimensional index evaluation system for measuring the consumption metering capacity of distributed power sources is established; secondly, the hierarchical analysis method and the improved grey correlation analysis are used to combine and weight the indicators; finally, the Monte Carlo model is used to determine the consumption capacity of rural energy internet. This method can comprehensively and accurately evaluate the consumption capacity of large-scale distributed power sources in new rural power grids, and also provide data support for solving the problem of new energy consumption in power grids;

[0004] However, due to the uncertainty of the output of distributed power sources in the energy Internet, such as wind power generation and photovoltaic power generation are greatly affected by weather conditions, it is necessary to specially consider the impact of this uncertainty on the absorption capacity during the evaluation. Ordinary distributed power absorption metering evaluation is often based on the determined power output and rarely involves uncertainty analysis. This evaluation method may not accurately reflect the actual impact of the uncertainty of distributed power output on the grid absorption capacity, resulting in inaccurate evaluation results. Summary of the invention

[0005] The purpose of the present invention is to provide a distributed power consumption metering and evaluation system for energy Internet to solve the technical problems in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] The present invention provides a distributed power consumption metering and evaluation system for energy internet, which specifically includes:

[0008] Related influence screening module: The sensor installed on the power supply equipment collects the power output power and related influencing factor data in real time, analyzes the collected power output power, obtains the power output value, and screens out the related influencing data that has a greater impact on the power output value through analysis;

[0009] Prediction model training module: Based on the selected data of relevant factors that have a greater impact on the power output coefficient, the random forest algorithm is selected to build a power output prediction model;

[0010] Power supply and demand degree identification module: Based on the power output prediction model, the power output prediction coefficient is output, and combined with the grid load value for analysis, to identify the large signal of the imbalance degree of output power supply and demand;

[0011] Absorption capacity assessment module: Based on the large signal of the imbalance degree of power supply and demand, the absorption capacity is assessed, the absorption capacity difference signal is identified and output, and the absorption capacity difference signal is optimized.

[0012] As a further solution of the present invention: the specific process of analyzing and selecting the relevant influencing data that has a greater impact on the power output value is:

[0013] Obtain the number of retained absolute slope deviation values, the mean deviation degree, and the mean absolute slope deviation;

[0014] Substitute into the formula The synchronous correlation value GL is calculated, where SL is the number of retained absolute slope deviation values, PJ is the mean deviation degree, XJ is the mean absolute slope deviation, a1, a2, and a3 are weight coefficients, a1 is 0.363, a2 is 0.231, and a is 0.406;

[0015] A synchronization correlation threshold is set, and the synchronization correlation values ​​between any two variables are arranged in order from large to small, and the synchronization correlation values ​​corresponding to the synchronization correlation values ​​greater than the synchronization correlation threshold are retained. The relevant influencing factors corresponding to the retained synchronization correlation values ​​are the relevant influencing factor data that have a greater impact on the power output value.

[0016] As a further solution of the present invention: the number of retained absolute slope deviation values, the mean value of the deviation degree and the mean value of the absolute slope deviation are obtained as follows:

[0017] Obtaining the absolute slope value of the power output sub-curve and the absolute slope value of the relevant influencing factor sub-curve respectively;

[0018] The absolute slope value of the power output value sub-curve corresponding to the same time period is calculated by difference with the absolute slope value of the relevant influencing factor sub-curve, and the absolute value of the difference is taken to obtain the absolute slope deviation value;

[0019] All the absolute slope deviation values ​​are summed and averaged to obtain the absolute slope deviation mean, all the absolute slope deviation values ​​are arranged in order from small to large, and the absolute slope deviation values ​​that are smaller than the absolute slope deviation mean are retained;

[0020] The number of retained absolute slope deviation values ​​is counted, and the difference between the retained absolute slope deviation values ​​and the mean absolute slope deviation is calculated respectively. The absolute value of the difference is taken to obtain the deviation degree value. All the deviation degree values ​​are summed and averaged to obtain the deviation degree mean.

[0021] As a further solution of the present invention: the process of obtaining the absolute slope value of the power output value sub-curve and the absolute slope value of the relevant influencing factor sub-curve is:

[0022] Install high-precision sensors on distributed power supply equipment to collect power output data and related influencing factor data;

[0023] Calculate the ratio of the power supply output power to the rated power to obtain the power supply output value;

[0024] Extract the power output value and the corresponding relevant influencing factor data from the historical data;

[0025] Taking the power output as one variable and any related influencing factor as another variable, the synchronous correlation value between any two variables is calculated. The specific process is as follows:

[0026] With time as X value, power output value and related influencing factors as Y values, draw the power output value curve and the corresponding curve of related influencing factors;

[0027] The power output value curve and the curve corresponding to the relevant influencing factors are respectively divided into a plurality of sub-curves to obtain a power output value sub-curve and a relevant influencing factor sub-curve;

[0028] The slope values ​​of each power output sub-curve and the relevant influencing factor sub-curve are calculated respectively, and the absolute values ​​of the slope values ​​are taken to obtain the absolute slope values.

[0029] As a further solution of the present invention: the process of constructing the power output prediction model is:

[0030] Obtain the selected relevant influencing factor data and power output values, and divide the data set into a training set and a test set in a ratio of 8:2;

[0031] Specify the number of decision trees in the random forest;

[0032] The random forest model training set data is used to build a random forest model. Each decision tree is built based on different sub-samples and features. At each decision tree node, a subset of features is randomly selected for splitting. The random forest model is then trained using the random forest model training set data. Each tree fits the data and learns the relationship between the relevant influencing factor data and the power output value.

[0033] Use the test set to evaluate the trained random forest model and evaluate the model;

[0034] Based on the evaluation results, adjustments are made until the most optimized model settings are found;

[0035] After the model training is completed, it becomes a power output prediction model, which is used to predict power output in real time.

[0036] As a further solution of the present invention: the process of generating a large signal of imbalance degree is:

[0037] Obtaining a power supply-demand ratio difference, and comparing the power supply-demand ratio difference with a power supply-demand ratio difference threshold;

[0038] If the power supply-demand ratio difference is greater than the power supply-demand ratio difference threshold, a large signal of the degree of power supply-demand imbalance is generated.

[0039] As a further solution of the present invention: the process of obtaining the power supply-demand ratio difference is:

[0040] Obtain power output and power load values ​​for future time periods;

[0041] Divide the future time period into several analysis time points, calculate the ratio of the power output and power load value corresponding to each analysis time point, and obtain the power supply and demand ratio;

[0042] The difference between the electricity supply-demand ratio and 1 is calculated, and the absolute value of the difference is taken to obtain the electricity supply-demand ratio difference.

[0043] As a further solution of the present invention: the process of obtaining the power output value and the power load value in the future time period is:

[0044] The screened relevant influencing factor data is collected in real time and transmitted to the power output prediction model. The power output prediction model will calculate the power output value in the future time period based on the input data;

[0045] Extract the power load data from the historical data, and divide the historical power load data into a training set and a test set in a ratio of 8:2;

[0046] Select the long short-term memory network for training;

[0047] Use a deep learning framework to build the model and use the training set data to train the model until the model achieves satisfactory prediction performance;

[0048] Use the test set data to verify the trained model and evaluate the model's predictive performance;

[0049] The model is optimized according to the verification results to obtain the power load time series model;

[0050] The time in the future time period to be predicted is input into the trained power load time series model to obtain the power load value of the future time period.

[0051] As a further solution of the present invention: the process of generating the output absorption capacity difference signal is:

[0052] Obtaining a consumption evaluation value, and comparing the consumption evaluation value with a consumption evaluation threshold;

[0053] If the absorption evaluation value is greater than the absorption evaluation threshold, a absorption capacity difference signal is generated;

[0054] Based on the generated absorption capacity difference signal, the balance of electricity supply and demand is optimized.

[0055] As a further solution of the present invention: the process of obtaining the consumption evaluation value is as follows:

[0056] Obtain the analysis time point corresponding to the signal with a large degree of imbalance, mark it as the analysis time point with a large degree, count the number of analysis time points with a large degree, perform ratio processing on the number of analysis time points with a large degree and the total number of analysis time points, and obtain the proportion of the number with a large degree;

[0057] The difference between the power supply and demand ratio difference at the time point with the largest degree of analysis is calculated respectively with the power supply and demand ratio difference threshold, the difference is processed with the power supply and demand ratio difference threshold to obtain the power supply and demand ratio deviation ratio, all the power supply and demand ratio deviation ratios are summed and averaged to obtain the deviation degree value;

[0058] The absorption evaluation value is obtained by multiplying the proportion of the large number of cases by the deviation degree value.

[0059] Beneficial effects of the present invention:

[0060] The present invention introduces a power output prediction model and combines it with grid load analysis to more comprehensively consider the uncertainty of distributed power output. By collecting relevant influencing factor data in real time and inputting it into the prediction model, the power output value in the future time period can be calculated, thereby more accurately reflecting the actual situation of distributed power output. This evaluation method based on prediction and real-time data can significantly improve the accuracy of distributed power consumption metering evaluation;

[0061] The present invention introduces quantitative indicators such as the power supply and demand ratio difference and the absorption evaluation value in the aspects of power supply and demand degree identification and absorption capacity assessment, which can reflect the imbalance degree of power supply and demand and the strength of absorption capacity, and provide timely and accurate early warning information for staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention will be further described below in conjunction with the accompanying drawings.

[0063] Figure 1 This is an architecture diagram of a distributed power consumption metering and evaluation system for energy internet in the present invention;

[0064] Figure 2 It is a flow chart of the process of obtaining the consumption evaluation value in a distributed power consumption metering and evaluation system for energy internet according to the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] Embodiment 1:

[0067] See also Figure 1 As shown, a distributed power consumption metering and evaluation system for energy internet according to an embodiment of the present invention specifically includes the following steps:

[0068] Related influence screening module: The sensor installed on the power supply equipment collects the power output power and related influencing factor data in real time, analyzes the collected power output power, obtains the power output value, and screens out the related influencing data that has a greater impact on the power output value through analysis;

[0069] Install high-precision sensors on distributed power supply equipment to collect power output data and related influencing factor data;

[0070] Among them, distributed power supply equipment includes but is not limited to: solar panels, wind turbines; related influencing factors include but are not limited to: temperature, light intensity, wind speed;

[0071] Calculate the ratio of the power supply output power to the rated power to obtain the power supply output value, where the rated power is the rated output power provided by the power supply manufacturer;

[0072] Extract the power output value and the corresponding relevant influencing factor data from the historical data;

[0073] Taking the power output as one variable and any related influencing factor as another variable, the synchronous correlation value between any two variables is calculated. The specific process is as follows:

[0074] With time as X value, power output value and related influencing factors as Y values, draw the power output value curve and the corresponding curve of related influencing factors;

[0075] It should be noted that the timestamps corresponding to the power output value and the relevant influencing factors must be the same to ensure data synchronization;

[0076] The power output value curve and the curve corresponding to the relevant influencing factors are respectively divided into a plurality of sub-curves to obtain a power output value sub-curve and a relevant influencing factor sub-curve;

[0077] Calculate the slope value of each sub-curve separately, and take the absolute value of the slope value to obtain the absolute slope value;

[0078] The absolute slope value of the power output value sub-curve corresponding to the same time period is calculated by difference with the absolute slope value of the relevant influencing factor sub-curve, and the absolute value of the difference is taken to obtain the absolute slope deviation value;

[0079] All the absolute slope deviation values ​​are summed and averaged to obtain the absolute slope deviation mean, all the absolute slope deviation values ​​are arranged in order from small to large, and the absolute slope deviation values ​​that are smaller than the absolute slope deviation mean are retained;

[0080] Count the number of retained absolute slope deviation values, and perform difference calculations on the retained absolute slope deviation values ​​and the mean absolute slope deviation values, take the absolute value of the difference to obtain the deviation degree value, sum and average all the deviation degree values ​​to obtain the deviation degree mean;

[0081] It should be explained that the mean value of the degree of deviation is calculated by the difference between the absolute slope deviation value and the mean value of the absolute slope deviation. The larger the absolute value of the difference is, the smaller the absolute slope deviation is. The closer the absolute value of the slope between the corresponding power output value sub-curve and the relevant influencing factor sub-curve is, the greater the synchronization is.

[0082] Substitute into the formula The synchronous correlation value GL is calculated, where SL is the number of retained absolute slope deviation values, PJ is the mean deviation degree, XJ is the mean absolute slope deviation, a1, a2, and a3 are weight coefficients, a1 is 0.363, a2 is 0.231, and a is 0.406;

[0083] Set the synchronization correlation threshold, arrange the synchronization correlation values ​​between any two variables in descending order, retain the synchronization correlation values ​​corresponding to the synchronization correlation threshold, and the relevant influencing factors corresponding to the retained synchronization correlation values ​​are the relevant influencing factor data with greater influence on the power output value;

[0084] Prediction model training module: Based on the selected data of relevant factors that have a greater impact on the power output coefficient, the random forest algorithm is selected to build a power output prediction model;

[0085] Obtain the selected relevant influencing factor data and power output values, and divide the data set into a training set and a test set in a ratio of 8:2. Usually, cross-validation is used or a portion of the data is reserved for testing.

[0086] Specify the number of decision trees in the random forest. The choice of the number of decision trees is usually adjusted according to the actual problem and computing resources. In general, the more trees in the random forest, the better the performance of the model is likely to be.

[0087] The random forest model training set data is used to build a random forest model. Each decision tree is built based on different sub-samples and features. At each decision tree node, a subset of features is randomly selected for splitting. The random forest model is then trained using the random forest model training set data. Each tree fits the data and learns the relationship between the relevant influencing factor data and the power output value.

[0088] The random forest model is trained using the test set and evaluated. Common evaluation indicators include but are not limited to: mean square error (MSE), mean absolute error (MAE), coefficient of determination (R 2 );

[0089] If the model performs poorly, you may need to go back and adjust the model parameters or reconsider feature selection;

[0090] Based on the evaluation results, continue to adjust the model parameters or try different feature combinations until the optimal model settings are found;

[0091] Once the model training is completed and reaches a satisfactory performance level, it becomes a power output prediction model, which is used to predict power output in real time;

[0092] The technical solution of the embodiment of the present invention is mainly as follows: high-precision sensors installed on distributed power supply equipment (such as solar panels and wind turbines) are used to collect power output power and related influencing factor data in real time, such as temperature, light intensity, wind speed, etc., and historical data are extracted to screen out related influencing factor data that have a greater impact on power output value using a synchronous correlation value calculation method. Based on the screened data, a random forest algorithm is selected to construct a power output prediction model, and a power output prediction model with satisfactory performance is obtained, which is used to predict power output in real time;

[0093] Through correlation analysis, the relevant influencing factors that have a greater impact on the power output value can be accurately identified, so that the relevant influencing factors that have a greater impact on the power output value can be identified, so as to more accurately evaluate the absorption capacity of distributed power sources and provide a reliable data basis for model construction. At the same time, the random forest algorithm is used to construct the prediction model, which makes full use of the advantages of random forest in processing complex nonlinear relationships and high-dimensional data, improves the accuracy and robustness of the prediction model, and finally obtains the power output prediction model.

[0094] Embodiment 2:

[0095] Based on Example 1, please refer to Figure 1 and Figure 2 As shown, a distributed power consumption metering and evaluation system for energy internet according to an embodiment of the present invention further includes the following steps:

[0096] Power supply and demand degree identification module: Based on the power output prediction model, the power output prediction coefficient is output, and combined with the grid load analysis, the output power supply and demand imbalance signal is identified;

[0097] Among them, the power supply and demand imbalance degree signal includes a power supply and demand imbalance degree small signal and a power supply and demand imbalance degree large signal;

[0098] The screened relevant influencing factor data is collected in real time and transmitted to the power output prediction model. The power output prediction model will calculate the power output value in the future time period based on the input data;

[0099] Among them, the future time period is determined by the needs of power system operation;

[0100] Based on the time series characteristics of historical power load data, a power load time series model is constructed to predict the power load value in the future time period. The specific process is as follows:

[0101] Extracting power load data from historical data, wherein the power load data is obtained from a power database, and dividing the historical power load data into a training set and a test set in a ratio of 8:2;

[0102] Select the long short-term memory network (LSTM) for training;

[0103] Long short-term memory network (LSTM) is a deep learning model based on recurrent neural network (RNN), which is suitable for processing long-term dependency problems in time series data. The LSTM model can capture complex patterns and trends in time series data by introducing memory units and recursive structures. In power load forecasting, the LSTM model can show high prediction accuracy and stability.

[0104] Use a deep learning framework (such as TensorFlow, PyTorch, etc.) to build the model, and use the training set data to train the model until the model achieves satisfactory prediction performance. During the training process, you can use cross-validation, early stopping and other techniques to prevent overfitting;

[0105] Use the test set data to verify the trained model and evaluate the prediction performance of the model. Common evaluation indicators include but are not limited to: mean square error (MSE), root mean square error (RMSE), and mean absolute error (MAE);

[0106] The model is optimized according to the verification results to obtain a power load time series model, wherein the process of optimizing the model includes but is not limited to: adjusting model parameters, adding feature variables, and using data enhancement technology;

[0107] Input the time in the future time period to be predicted into the trained power load time series model to obtain the power load value in the future time period;

[0108] Divide the future time period into several analysis time points, calculate the ratio of the power output value and the power load value corresponding to each analysis time point, and obtain the power supply and demand ratio;

[0109] Calculate the difference between the power supply-demand ratio and 1, take the absolute value of the difference, and obtain the power supply-demand ratio difference;

[0110] It needs to be explained that the power supply-demand ratio reflects the relative relationship between power output and power demand (i.e., power load), and the power supply-demand ratio difference reflects the degree of imbalance between power output and power demand;

[0111] Comparing the power supply-demand ratio difference with the power supply-demand ratio difference threshold, wherein the power supply-demand ratio difference threshold is set by a person skilled in the art based on a summary of multiple historical experimental data;

[0112] If the power supply-demand ratio difference is less than or equal to the power supply-demand ratio difference threshold, a small signal of the power supply-demand imbalance degree is generated; if the power supply-demand ratio difference is greater than the power supply-demand ratio difference threshold, a large signal of the power supply-demand imbalance degree is generated;

[0113] Absorption capacity evaluation module: Based on the signal of imbalance degree of power supply and demand, it evaluates the absorption capacity, identifies and outputs the signal of poor absorption capacity or the signal of normal absorption capacity, and optimizes the signal of poor absorption capacity;

[0114] Obtain the analysis time point corresponding to the signal with a large degree of imbalance, mark it as the analysis time point with a large degree, count the number of analysis time points with a large degree, perform ratio processing on the number of analysis time points with a large degree and the total number of analysis time points, and obtain the proportion of the number with a large degree;

[0115] The difference between the power supply and demand ratio difference at the time point with the largest degree of analysis is calculated respectively with the power supply and demand ratio difference threshold, the difference is processed with the power supply and demand ratio difference threshold to obtain the power supply and demand ratio deviation ratio, all the power supply and demand ratio deviation ratios are summed and averaged to obtain the deviation degree value;

[0116] The proportion of the number of large degrees is multiplied by the deviation degree value to obtain the absorption evaluation value;

[0117] Compare the consumption evaluation value with the consumption evaluation threshold, wherein the consumption evaluation threshold is set by a person skilled in the art based on a summary of multiple historical experimental data;

[0118] If the absorption evaluation value is less than the absorption evaluation threshold, a normal absorption capacity signal is generated; if the absorption evaluation value is greater than the absorption evaluation threshold, a poor absorption capacity signal is generated;

[0119] Based on the generation of the absorption capacity difference signal, it indicates that the power grid has a weak absorption capacity for power sources in the future period, and there is a large imbalance between power output and power demand. Adjustment measures need to be taken to optimize the balance between power supply and demand. The adjustment measures may include but are not limited to: adjusting the power output plan, optimizing the grid dispatching strategy, and adding energy storage facilities;

[0120] In this embodiment, the operation status of the power grid after the adjustment measures are taken can also be further detected to evaluate the effectiveness of the adjustment measures. The specific process is as follows:

[0121] After formulating and implementing adjustment measures to optimize the balance of power supply and demand based on the generated absorption capacity difference signal;

[0122] Monitor the operation of the power grid in real time and collect operation data, including power output and power load values;

[0123] Using the collected data, re-identify the signal of the imbalance between power supply and demand and re-evaluate the power consumption capacity, and calculate the new power consumption evaluation value;

[0124] Compare the consumption evaluation values ​​before and after the implementation of the adjustment measures to evaluate the effectiveness of the strategy. If the new consumption evaluation value is lower and lower than the set consumption evaluation threshold, the strategy is considered effective; otherwise, further adjustment and optimization of the strategy is required;

[0125] The technical solution of the embodiment of the present invention is mainly as follows: by predicting power output and power load, identifying the signal of the degree of imbalance between power supply and demand; evaluating the power grid's absorption capacity based on the signal of the degree of imbalance between power supply and demand, and outputting the corresponding signal, and proposing optimization suggestions and implementing adjustment measures for the absorption capacity difference signal, and conducting real-time monitoring and evaluation of the power grid operation after taking the adjustment measures to verify the effectiveness of the strategy, so as to evaluate the power grid's absorption capacity based on the predicted power output value and power load value, and propose advance optimization suggestions based on the evaluation results, which will help power grid operators to more reasonably configure power sources and energy storage facilities and optimize the balance between power supply and demand.

[0126] Embodiment 3:

[0127] On the basis of Example 1 and Example 2, a distributed power consumption metering evaluation method for energy internet described in an embodiment of the present invention specifically includes the following steps:

[0128] Step 1: The power output power and related influencing factor data are collected in real time by sensors installed on the power supply equipment, and the power output value is obtained from the collected power output power, and the related influencing data with a greater impact on the power output value is screened out through analysis;

[0129] Step 2: Based on the screened data of relevant factors that have a greater impact on the power output coefficient, select the random forest algorithm to build a power output prediction model;

[0130] Step 3: Based on the power output prediction model, output the power output coefficient, and combine it with the grid load analysis to identify the signal of the imbalance degree of output power supply and demand;

[0131] Among them, the power supply and demand imbalance degree signal includes a power supply and demand imbalance degree small signal and a power supply and demand imbalance degree large signal;

[0132] Step 4: Based on the signal of the imbalance degree of electricity supply and demand, evaluate the absorption capacity, identify and output the signal of poor absorption capacity or the signal of normal absorption capacity, and optimize the signal of poor absorption capacity.

[0133] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A distributed power consumption metering and evaluation system for energy internet, characterized in that: Specifically include: Related influence screening module: The sensor installed on the power supply equipment collects the power output power and related influencing factor data in real time, analyzes the collected power output power, obtains the power output value, and screens out the related influencing data that has a greater impact on the power output value through analysis; Prediction model training module: Based on the selected data of relevant factors that have a greater impact on the power output coefficient, the random forest algorithm is selected to build a power output prediction model; Power supply and demand degree identification module: Based on the power output prediction model, the power output prediction coefficient is output, and combined with the grid load value for analysis, to identify the large signal of the imbalance degree of output power supply and demand; Absorption capacity assessment module: Based on the large signal of the imbalance degree of power supply and demand, the absorption capacity is assessed, the absorption capacity difference signal is identified and output, and the absorption capacity difference signal is optimized.

2. According to the energy internet-oriented distributed power consumption metering and evaluation system according to claim 1, it is characterized in that: The specific process of analyzing and screening out the relevant influencing data that has a greater impact on the power output value is as follows: Obtain the number of retained absolute slope deviation values, the mean deviation degree, and the mean absolute slope deviation; Substitute into the formula The synchronous correlation value GL is calculated, where SL is the number of retained absolute slope deviation values, PJ is the mean deviation degree, XJ is the mean absolute slope deviation, a1, a2, and a3 are weight coefficients, a1 is 0.363, a2 is 0.231, and a is 0.406; A synchronization correlation threshold is set, and the synchronization correlation values ​​between any two variables are arranged in order from large to small, and the synchronization correlation values ​​corresponding to the synchronization correlation values ​​greater than the synchronization correlation threshold are retained. The relevant influencing factors corresponding to the retained synchronization correlation values ​​are the relevant influencing factor data that have a greater impact on the power output value.

3. According to the energy internet-oriented distributed power consumption metering and evaluation system of claim 2, it is characterized in that: The number of retained absolute slope deviation values, the mean deviation degree and the mean absolute slope deviation are obtained as follows: Obtaining the absolute slope value of the power output sub-curve and the absolute slope value of the relevant influencing factor sub-curve respectively; The absolute slope value of the power output value sub-curve corresponding to the same time period is calculated by difference with the absolute slope value of the relevant influencing factor sub-curve, and the absolute value of the difference is taken to obtain the absolute slope deviation value; All the absolute slope deviation values ​​are summed and averaged to obtain the absolute slope deviation mean, all the absolute slope deviation values ​​are arranged in order from small to large, and the absolute slope deviation values ​​that are smaller than the absolute slope deviation mean are retained; The number of retained absolute slope deviation values ​​is counted, and the difference between the retained absolute slope deviation values ​​and the mean absolute slope deviation is calculated respectively. The absolute value of the difference is taken to obtain the deviation degree value. All the deviation degree values ​​are summed and averaged to obtain the deviation degree mean.

4. A distributed power consumption metering and evaluation system for energy internet according to claim 3, characterized in that: The process of obtaining the absolute slope value of the power output sub-curve and the absolute slope value of the relevant influencing factor sub-curve is as follows: Install high-precision sensors on distributed power supply equipment to collect power output data and related influencing factor data; Calculate the ratio of the power supply output power to the rated power to obtain the power supply output value; Extract the power output value and the corresponding relevant influencing factor data from the historical data; Taking the power output value as one variable and any related influencing factor as another variable, the synchronous correlation value between any two variables is calculated. The specific process is as follows: With time as X value, power output value and related influencing factors as Y values, draw the power output value curve and the corresponding curve of related influencing factors; The power output value curve and the curve corresponding to the relevant influencing factors are respectively divided into a plurality of sub-curves to obtain a power output value sub-curve and a relevant influencing factor sub-curve; The slope values ​​of each power output sub-curve and the relevant influencing factor sub-curve are calculated respectively, and the absolute values ​​of the slope values ​​are taken to obtain the absolute slope values.

5. According to the energy internet-oriented distributed power consumption metering and evaluation system of claim 1, it is characterized in that: The process of constructing the power output prediction model is as follows: Obtain the selected relevant influencing factor data and power output values, and divide the data set into a training set and a test set in a ratio of 8:2; Specify the number of decision trees in the random forest; The random forest model training set data is used to build a random forest model. Each decision tree is built based on different sub-samples and features. At each decision tree node, a subset of features is randomly selected for splitting. The random forest model is then trained using the random forest model training set data. Each tree fits the data and learns the relationship between the relevant influencing factor data and the power output value. Use the test set to evaluate the trained random forest model and evaluate the model; Based on the evaluation results, adjustments are made until the most optimized model settings are found; After the model training is completed, it becomes a power output prediction model, which is used to predict power output in real time.

6. The distributed power consumption metering and evaluation system for energy internet according to claim 1 is characterized in that: The process of generating the large signal of the degree of imbalance is: Obtaining a power supply-demand ratio difference, and comparing the power supply-demand ratio difference with a power supply-demand ratio difference threshold; If the power supply-demand ratio difference is greater than the power supply-demand ratio difference threshold, a large signal of the degree of power supply-demand imbalance is generated.

7. A distributed power consumption metering and evaluation system for energy internet according to claim 6, characterized in that: The process of obtaining the power supply-demand ratio difference is as follows: Obtain power output and power load values ​​for future time periods; Divide the future time period into several analysis time points, calculate the ratio of the power output and power load value corresponding to each analysis time point, and obtain the power supply and demand ratio; The difference between the electricity supply-demand ratio and 1 is calculated, and the absolute value of the difference is taken to obtain the electricity supply-demand ratio difference.

8. A distributed power consumption metering and evaluation system for energy internet according to claim 7, characterized in that: The process of obtaining the power output value and the power load value in the future time period is as follows: The screened relevant influencing factor data is collected in real time and transmitted to the power output prediction model. The power output prediction model will calculate the power output value in the future time period based on the input data; Extract the power load data from the historical data, and divide the historical power load data into a training set and a test set in a ratio of 8:2; Select the long short-term memory network for training; Use a deep learning framework to build the model and use the training set data to train the model until the model achieves satisfactory prediction performance; Use the test set data to verify the trained model and evaluate the model's predictive performance; The model is optimized according to the verification results to obtain the power load time series model; The time in the future time period to be predicted is input into the trained power load time series model to obtain the power load value of the future time period.

9. The distributed power consumption metering and evaluation system for energy internet according to claim 1 is characterized in that: The process of generating the output absorption capacity difference signal is as follows: Obtaining a consumption evaluation value, and comparing the consumption evaluation value with a consumption evaluation threshold; If the absorption evaluation value is greater than the absorption evaluation threshold, a absorption capacity difference signal is generated; Based on the generated absorption capacity difference signal, the balance of electricity supply and demand is optimized.

10. A distributed power consumption metering and evaluation system for energy internet according to claim 9, characterized in that: The process of obtaining the consumption evaluation value is as follows: Obtain the analysis time point corresponding to the signal with a large degree of imbalance, mark it as the analysis time point with a large degree, count the number of analysis time points with a large degree, perform ratio processing on the number of analysis time points with a large degree and the total number of analysis time points, and obtain the proportion of the number with a large degree; The difference between the power supply and demand ratio difference at the time point with the largest degree of analysis is calculated respectively with the power supply and demand ratio difference threshold, the difference is processed with the power supply and demand ratio difference threshold to obtain the power supply and demand ratio deviation ratio, all the power supply and demand ratio deviation ratios are summed and averaged to obtain the deviation degree value; The absorption evaluation value is obtained by multiplying the proportion of the large number of cases by the deviation degree value.

Images