Photovoltaic energy storage management system based on enterprise electrical load
By introducing power load modules, photovoltaic power generation modules and adaptation management modules into the photovoltaic energy storage management system, combined with a variety of prediction models and algorithms, the problems of low prediction accuracy and poor flexibility in adjusting energy storage strategies are solved, and efficient and flexible photovoltaic energy storage management is achieved.
Patent Information
- Application Number
- CN202411826697.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-12
- Publication Date
- 2025-05-13
AI Technical Summary
The traditional photovoltaic energy storage management system based on enterprise electricity load has low prediction accuracy and is difficult to cope with random changes in external factors, resulting in poor flexibility in adjusting energy storage strategies.
A photovoltaic energy storage management system based on power load module, photovoltaic power generation module and adaptation management module was designed. By analyzing the correlation coefficient between enterprise power consumption and various influencing factors, combining ARIMA and HMM models to predict electricity consumption, and using Gaussian-CNN-GRU model to predict electricity generation, and finally optimizing the allocation of energy storage resources through adaptation index.
It realizes accurate prediction of enterprise electricity load and photovoltaic power generation, improves energy storage resource utilization, enhances system flexibility and adaptability, and reduces electricity consumption costs and carbon emissions.
Smart Images
Figure CN119990573A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of photovoltaic energy storage management, and in particular to a photovoltaic energy storage management system based on enterprise electricity load. Background Art
[0002] Photovoltaic energy storage management improves the efficiency and stability of photovoltaic power generation systems through real-time monitoring, remote control and intelligent scheduling. The comprehensive use of photovoltaic and energy storage systems is of great significance to manufacturing companies. First, photovoltaic power generation can reduce the power load of the power grid, reduce electricity expenses, and reduce carbon emissions. Secondly, the characteristics of "low storage and high release" and time-of-use electricity prices of energy storage systems can effectively solve the problem of over-capacity electricity consumption in some enterprises, which not only ensures the safe use of electricity by enterprises, but also reduces the electricity costs of enterprises. It has the economic value of effectively reducing electricity costs and the social value of low-carbon emission reduction. Energy storage technology can solve the intermittent and uncontrollable factors of renewable energy and improve the flexibility and reliability of energy systems. With the advancement of science and technology and the growth of demand, energy storage systems are currently developing in the direction of greater energy density, higher efficiency and stronger stability, lower cost and longer life.
[0003] In addition, there is another important significance, which is that it can effectively solve the problem of overcapacity electricity consumption in enterprises. Overcapacity electricity consumption refers to electricity consumption by electricity users that exceeds the capacity agreed in the contract. It will not only cause electrical equipment such as transformers to overload and burn, causing fires, but also cause overload of public power supply lines, endangering the safety of the public power grid. The survey found that some companies will only experience short-term overloads during equipment startup, peak production, etc. In order to meet the safe electricity demand during peak loads, they have to expand the capacity of the transformer. The expansion process is not only long but also costly. According to conservative estimates, an expansion of 100kW will cost about 200,000 yuan, and the basic electricity fee will also increase. The photovoltaic system can reduce the power load of the power grid during the day. Combined with the energy storage system's "peak shaving and valley filling", it can effectively reduce the power load of the power grid during other periods, thereby avoiding overcapacity electricity consumption. At present, traditional photovoltaic energy storage management systems based on corporate electricity loads rely on historical electricity consumption data and power generation data for electricity forecasting, and often ignore the impact of external factors such as economic conditions, weather conditions, and seasonal changes on corporate electricity consumption and photovoltaic power generation, resulting in low prediction accuracy. When dealing with random changes in influencing factors, there is a lack of efficient optimization algorithms, making it difficult to adjust energy storage strategies in a timely manner, and poor flexibility and applicability. Summary of the invention
[0004] 1. Technical issues to be resolved In view of the shortcomings of the prior art, the present invention provides a photovoltaic energy storage management system based on enterprise electricity load, which has the advantages of high prediction accuracy of comprehensive analysis of influencing factors and high utilization rate of intelligent management of energy storage resources. It solves the problem that the traditional photovoltaic energy storage management system based on enterprise electricity load has low prediction accuracy and difficulty in timely adjustment of energy storage strategies.
[0005] (II) Technical solution To achieve the above-mentioned purpose, the present invention provides the following technical solutions: A photovoltaic energy storage management system based on the enterprise power load includes a power load module, a photovoltaic power generation module and an adaptation management module; The power load module is composed of an enterprise data unit, a related data unit and a power consumption prediction unit. The enterprise data unit collects a load data set through a network connection to the distribution network. The load data set includes the power consumption of all enterprises. The related data unit collects an impact data set through a network connection to the big data platform. The impact data set includes a variety of related factors that affect the power consumption of enterprises. The power consumption prediction unit is set with a fixed-duration monitoring cycle. , combined with the load data set and the impact data set, analyze the correlation coefficient between the enterprise's electricity consumption and each related factor , and calculate the next monitoring cycle of the enterprise Predicted power consumption within ; The photovoltaic power generation module is composed of a photovoltaic data unit, an environmental data unit and a power generation prediction unit. The photovoltaic data unit is connected to the photovoltaic power station management system through a network to collect photovoltaic data sets, and the photovoltaic data sets include the power generation of the photovoltaic power station at all time points. The environmental data unit is connected to the sensor device through a network to collect environmental data sets, and the environmental data sets include a variety of related factors that affect the photovoltaic power generation. The power generation prediction unit analyzes the monitoring period according to the photovoltaic data set and the environmental data set. The system efficiency of the photovoltaic power station , and the next monitoring cycle The predicted power generation value within ; The adaptation management module is based on the power consumption prediction value , analyze the adaptation index of each enterprise , and according to the adaptation index The adaptation list is arranged from high to low, and the adaptation management module is based on the power generation forecast value. , giving priority to enterprises with the highest ranking in the adaptation list.
[0006] Preferably, the expression of the load data set is: , to The first to the The electricity consumption of an enterprise, Indicates the specific time for obtaining the electricity consumption of a single enterprise.
[0007] Preferably, the expression affecting the data set is , represents the gross regional product, represents the gross industrial output value, Indicates the regional temperature, represents the power elasticity coefficient, Indicates the specific time for obtaining factors affecting the enterprise's electricity consumption.
[0008] Preferably, the correlation coefficient The calculation process is as follows: S11. Extract monitoring cycle based on load data set within, no. The electricity consumption of each enterprise is marked as , to Respectively From the first time point to the The electricity consumption at a point in time; S12. Extract monitoring cycles based on impact data sets Internal, affecting The regional GDP of the electricity consumption of enterprises is marked as , to From the first time point to the The gross regional product at a point in time; S13, calculate the The correlation coefficient between the electricity consumption of enterprises and the regional GDP ; In the formula, Indicates The company in The power consumption at a point in time, Indicates monitoring cycle within, no. The average electricity consumption of an enterprise, Indicates The GDP of a region at a point in time, Indicates monitoring cycle The average value of regional GDP is According to the covariance and standard deviation calculation formula, we get Pearson correlation coefficient between electricity consumption of enterprises and regional GDP.
[0009] Preferably, the power consumption forecast value The calculation process is as follows: S21. Based on the load data set, the ARIMA(1,1) model is used to perform time series analysis; In the formula, represents the random variable at the current moment in the ARIMA(1,1) model, i.e. Enterprises at the moment of electricity consumption, Represents the first-order autoregressive coefficient, which represents the influence coefficient of the time series value at the previous moment on the value at the current moment. represents the error term at the current moment, Represents the first-order moving average coefficient, which represents the influence coefficient of the error term at the previous moment on the value at the current moment. Indicates that the ARIMA(1,1) model is The predicted value one step ahead, i.e. Enterprises at the moment The power consumption prediction value is iterated successively. Indicates that the ARIMA(1,1) model is forward The predicted value at the step time, i.e. Enterprises at the moment The predicted value of electricity consumption; S22. Based on the correlation coefficient between the enterprise's electricity consumption and each related factor , the numerical error of the ARIMA(1,1) model is predicted by the HMM model. The HMM model uses the Viterbi algorithm of dynamic programming to predict the state sequence of the numerical error. The modeling process is as follows: Initialization: The initialization expression is: In the formula, Indicates status When it is observed The probability of state When the first The correlation coefficient between the electricity consumption of enterprises and the regional GDP The probability of being the maximum value, Indicates status The initial probability of Recursion: For , the maximum probability expression of the observation path obtained by recursion is: In the formula, Indicated in When, status It was observed that , and in When, status It was observed that The maximum probability of the path is Represents the previous state of the maximum probability path The probability of Termination: The optimal path expression for the termination of the optimization is: In the formula, represents the probability of the optimal path, Indicates the state of the optimal path; Optimal path backtracking: For , the state sequence of the optimal path is obtained in reverse: ; S23, the predicted value obtained by the ARIMA(1,1) model is corrected by the state sequence predicted by the HMM model. In the formula, Represents the evaluation weight for the ARIMA(1,1) model prediction value. Represents the evaluation weight for the HMM model state sequence, , Indicates according to and Weight, comprehensive prediction value and state sequence, get the The next monitoring cycle of enterprises The predicted power consumption within .
[0010] Preferably, the expression of the photovoltaic data set is: , to They are the first time point to the second time point of the photovoltaic power station. The amount of electricity generated at a point in time, Indicates the specific time for obtaining the power generation of the PV power station.
[0011] Preferably, the expression of the environmental data set is , represents the irradiance, represents the surface temperature, Indicates the temperature of the photovoltaic module, Indicates the ambient humidity. Indicates the specific time for obtaining the factors affecting photovoltaic power generation.
[0012] Preferably, the system efficiency The calculation process is as follows: S31. Statistical monitoring cycle based on photovoltaic data set The total power generation within is marked as ; S32. Statistical monitoring cycle based on environmental data set The total radiation exposure within is marked as ; S33. Calculate monitoring cycle The system efficiency of the photovoltaic power station ; In the formula, Indicates the irradiance under standard test conditions, Indicates monitoring cycle The total amount of electricity actually output by the photovoltaic power station is represents the installed capacity of the photovoltaic power station, Indicates monitoring cycle The total amount of electrical energy that the photovoltaic power station is estimated to output.
[0013] Preferably, the power generation forecast value The calculation process is as follows: According to the photovoltaic data set, the Gaussian-CNN-GRU prediction model is used to predict the power generation. The Gaussian-CNN-GRU prediction model consists of an input layer, three hidden layers, and a Gaussian output layer. The first hidden layer is a one-dimensional convolutional neural network, the second hidden layer is a gated recurrent unit, and the third hidden layer is a fully connected layer. The Gaussian output layer contains two neurons, which output the power generation prediction value and the variance of the power generation prediction uncertainty respectively. Input layer: input photovoltaic data set and environmental data set; The first hidden layer: The CNN layer is used to explore the potential relationship between photovoltaic power generation and each related factor and extract high-dimensional features. The convolution layer is used to extract high-dimensional features. convolution kernels, mapping the power generation at all time points in the photovoltaic dataset to time series, and shorten the time series length to , output Matrix to GRU layer; The second hidden layer: The GRU layer is used to extract the time series characteristics, calculate the power generation at all time points in the photovoltaic data set time by time step, and the last step neuron outputs The power generation vector in the format is sent to the third hidden layer; The expressions of the third hidden layer and Gaussian output layer are: In the formula, The variance of the uncertainty in power generation forecast is the uncertainty in power generation forecast due to the randomness of the environmental data set. represents the power generation vector output by the GRU layer, represents the weight matrix of the third hidden layer, represents the bias vector of the third hidden layer, represents the weight matrix of the Gaussian output layer, Represents the bias vector of the Gaussian output layer.
[0014] Preferably, the adaptation index The calculation process is as follows: S41. Based on the power consumption forecast value and power generation forecast , Statistics The growth rate of electricity consumption of each enterprise in the next year is marked as , will The ratio of the maximum power consumption of an enterprise in the next year to 80% of the rated capacity of the transformer is marked as , will The average standard deviation of daily electricity load of each enterprise in the next year is marked as ; S42: Standardize the parameters in S41 to obtain hyperparameters , and ; In the formula, and Respectively The minimum and maximum growth rates of electricity consumption of each enterprise in the next year. and Respectively The minimum and maximum ratios of the maximum power consumption of an enterprise to 80% of the rated capacity of the transformer in the next year. and Respectively The minimum and maximum values of the average standard deviation of daily electricity load for each enterprise in the next year; S43. Calculate the first The fit index of an enterprise ; In the formula, Indicates according to , and Fixed weight of 2, comprehensive hyperparameters , and , get the The adaptation index of an enterprise.
[0015] Compared with the prior art, the present invention provides a photovoltaic energy storage management system based on enterprise power load, which has the following beneficial effects: 1. The present invention collects the power consumption of all enterprises and various related factors affecting the power consumption of enterprises through the power load module, and then sets a monitoring cycle with a fixed duration. , analyze the correlation coefficient between the enterprise's electricity consumption and each related factor , and calculate the next monitoring cycle of the enterprise Predicted power consumption within , and realizes the accurate prediction of the enterprise's future electricity load. The photovoltaic power generation module collects the power generation of the photovoltaic power station at all time points, as well as various related factors affecting the photovoltaic power generation, and then analyzes the monitoring cycle The system efficiency of the photovoltaic power station , and the next monitoring cycle The predicted power generation value within , fully considering the impact of various random factors, comprehensively analyzing the influencing factors and achieving high prediction accuracy.
[0016] 2. The present invention uses the adaptation management module to predict the power consumption value , combined with the future growth rate of electricity consumption, the future overcapacity ratio of electricity consumption and the fluctuation of electricity load, analyze the adaptation index of each enterprise , and according to the adaptation index The adaptation list is arranged from high to low, and the adaptation management module is based on the power generation forecast value. , giving priority to the companies with the highest ranking in the adaptation list, targeted screening of target companies suitable for photovoltaic energy storage new energy systems, creating greater investment value, proposing the best discharge strategy, and intelligently managing energy storage resources with high utilization rate. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 It is a schematic diagram of the system flow of the present invention; Figure 2 This is an analysis diagram of the power load and influencing factors of the enterprise system of the present invention; Figure 3 It is a schematic diagram of the Gaussian-CNN-GRU prediction model structure of the system of the present invention. DETAILED DESCRIPTION
[0018] 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.
[0019] Since the traditional photovoltaic energy storage management system based on enterprise power load relies on historical power consumption data and power generation data for power forecasting, it often ignores the impact of external factors such as economic conditions, weather conditions, and seasonal changes on enterprise power consumption and photovoltaic power generation, resulting in low prediction accuracy. When dealing with random changes in influencing factors, there is a lack of efficient optimization algorithms, making it difficult to adjust energy storage strategies in a timely manner, and the flexibility and applicability are poor. Therefore, a photovoltaic energy storage management system based on enterprise power load is provided. Please refer to Figure 1 , a photovoltaic energy storage management system based on the enterprise power load, including a power load module, a photovoltaic power generation module and an adaptation management module; The power load module consists of enterprise data unit, related data unit and power consumption forecast unit. The enterprise data unit collects load data set through network connection to distribution network. The load data set includes the power consumption of all enterprises. The expression of load data set is: , to The first to the The electricity consumption of an enterprise, Indicates the specific time of obtaining the electricity consumption of a single enterprise. Collecting the enterprise's long-term electricity consumption data in time series will help to analyze the cycle trend in a targeted manner later; The relevant data unit collects the impact data set through the network connection big data platform. The impact data set includes a variety of related factors that affect the electricity consumption of enterprises. The expression of the impact data set is , represents the gross regional product, represents the gross industrial output value, Indicates the regional temperature, represents the power elasticity coefficient, It indicates obtaining the specific time of factors affecting the electricity consumption of enterprises and comprehensively collecting various influencing factors, which will help to compare and analyze the influence of different factors on the electricity consumption of enterprises in the future; The power consumption prediction unit is set with a fixed-length monitoring cycle , combined with the load data set and the impact data set, analyze the correlation coefficient between the enterprise's electricity consumption and each related factor , and use multiple forecasting models to calculate the next monitoring cycle of the enterprise Predicted power consumption within , which takes into account both the compliance of the power load in the time series and the interference of random fluctuations; Example 1 See also Figure 2 , correlation coefficient The calculation process is as follows: S11. Extract monitoring cycle based on load data set within, no. The electricity consumption of each enterprise is marked as , to Respectively From the first time point to the The electricity consumption at a point in time; S12. Extract monitoring cycles based on impact data sets Internal, affecting The regional GDP of the electricity consumption of enterprises is marked as , to From the first time point to the The regional GDP at each time point is unified on the time axis, and the electricity consumption at the same time point and the related factors affecting electricity consumption are screened, which can quickly analyze the linear relationship between the two variables; S13, calculate the The correlation coefficient between the electricity consumption of enterprises and the regional GDP ; In the formula, Indicates The company in The power consumption at a point in time, Indicates monitoring cycle within, no. The average electricity consumption of an enterprise, Indicates The GDP of a region at a point in time, Indicates monitoring cycle The average value of regional GDP is According to the covariance and standard deviation calculation formula, we get Pearson correlation coefficient between electricity consumption of enterprises and regional GDP; In this embodiment, the correlation coefficient As a dimensionless indicator, it is not affected by the unit and magnitude of the variable. , indicating that there is a completely positive correlation between the two variables, that is, when one variable increases, the other variable also increases at a fixed ratio. If the correlation coefficient , indicating that there is a completely negative correlation between the two variables, that is, when one variable increases, the other variable decreases at a fixed ratio. If the correlation coefficient , indicating that there is no linear correlation between the two variables, through the correlation coefficient , which clarifies that the urban economy is a significant factor affecting corporate electricity consumption.
[0020] Example 2 This embodiment is based on the explanation of embodiment 1. Specifically, the power consumption prediction value The calculation process is as follows: S21. Based on the load data set, the ARIMA(1,1) model is used to perform time series analysis; In the formula, represents the random variable at the current moment in the ARIMA(1,1) model, i.e. Enterprises at the moment The power consumption, Represents the first-order autoregressive coefficient, which represents the influence coefficient of the time series value at the previous moment on the value at the current moment. Represents the error term at the current moment, which is usually assumed to be a normally distributed random variable with a mean of 0. Represents the first-order moving average coefficient, which represents the influence coefficient of the error term at the previous moment on the value at the current moment. Indicates that the ARIMA(1,1) model is The predicted value one step ahead, i.e. Enterprises at the moment The power consumption prediction value is iterated successively. Indicates that the ARIMA(1,1) model is forward The predicted value at the step time, i.e. Enterprises at the moment The predicted value of electricity consumption; S22. Based on the correlation coefficient between the enterprise's electricity consumption and each related factor , the numerical error of the ARIMA(1,1) model is predicted by the HMM model. The HMM model uses the Viterbi algorithm of dynamic programming to predict the state sequence of the numerical error. The modeling process is as follows: Initialization: The initialization expression is: In the formula, Indicates status When it is observed The probability of state When the first The correlation coefficient between the electricity consumption of enterprises and the regional GDP The probability of being the maximum value, Indicates status The initial probability of Recursion: For , the maximum probability expression of the observation path obtained by recursion is: In the formula, Indicated in When, status It was observed that , and in When, status It was observed that The maximum probability of the path is Represents the previous state of the maximum probability path The probability of Termination: The optimal path expression for the termination of the optimization is: In the formula, represents the probability of the optimal path, Indicates the state of the optimal path; Optimal path backtracking: For , the state sequence of the optimal path is obtained in reverse: ; S23, the predicted value obtained by the ARIMA(1,1) model is corrected by the state sequence predicted by the HMM model. In the formula, Represents the evaluation weight for the ARIMA(1,1) model prediction value. Represents the evaluation weight for the HMM model state sequence, , Indicates according to and Weight, comprehensive prediction value and state sequence, get the The next monitoring cycle of enterprises The predicted value of electricity consumption within In this embodiment, a time-axis-based electricity consumption forecasting model is established through the ARIMA (1,1) model. Then, by comprehensively considering factors such as economy, meteorology, and electricity policy, the HMM model is used to perform vertical recursive difference corrections on the electricity load forecast results for the same month, thereby achieving accurate prediction of the company's future electricity load.
[0021] Example 3 This embodiment is an explanation based on Embodiment 2. Specifically, the photovoltaic power generation module is composed of a photovoltaic data unit, an environmental data unit and a power generation prediction unit. The photovoltaic data unit is connected to the photovoltaic power station management system through a network to collect photovoltaic data sets. The photovoltaic data set includes the power generation of the photovoltaic power station at all time points. The expression of the photovoltaic data set is: , to They are the first time point to the second time point of the photovoltaic power station. The amount of electricity generated at a point in time, Indicates the specific time of obtaining the power generation of the photovoltaic power station; The environmental data unit collects environmental data sets through network-connected sensor devices. The environmental data sets include various factors that affect photovoltaic power generation. The expression of the environmental data set is: , represents the irradiance, represents the surface temperature, Indicates the temperature of the photovoltaic module, Indicates the ambient humidity. Indicates the specific time for obtaining the factors affecting photovoltaic power generation; The power generation prediction unit analyzes the monitoring period based on the photovoltaic data set and environmental data set The system efficiency of the photovoltaic power station , and the next monitoring cycle The predicted power generation value within ; System efficiency The calculation process is as follows: S31. Statistical monitoring cycle based on photovoltaic data set The total power generation within is marked as ; S32. Statistical monitoring cycle based on environmental data set The total radiation exposure within is marked as ,in, ; S33. Calculate monitoring cycle The system efficiency of the photovoltaic power station ; In the formula, Indicates the irradiance under standard test conditions, Indicates monitoring cycle The total amount of electricity actually output by the photovoltaic power station is represents the installed capacity of the photovoltaic power station, Indicates monitoring cycle The total amount of electricity that the photovoltaic power station estimates to output, and the ratio of the actual total amount of electricity to the estimated total amount of electricity, directly reflect the production efficiency of the photovoltaic power station; See also Figure 3 , power generation forecast value The calculation process is as follows: According to the photovoltaic data set, the Gaussian-CNN-GRU prediction model is used to predict the power generation. The Gaussian-CNN-GRU prediction model consists of an input layer, three hidden layers, and a Gaussian output layer. The first hidden layer is a one-dimensional convolutional neural network, the second hidden layer is a gated recurrent unit, and the third hidden layer is a fully connected layer. The Gaussian output layer contains two neurons, which output the power generation prediction value and the variance of the power generation prediction uncertainty respectively. Input layer: input photovoltaic data set and environmental data set; The first hidden layer: The CNN layer is used to explore the potential relationship between photovoltaic power generation and each related factor and extract high-dimensional features. The convolution layer is used to extract high-dimensional features. convolution kernels, mapping the power generation at all time points in the photovoltaic dataset to time series, and shorten the time series length to , output The matrix to the GRU layer, high-dimensional features include but are not limited to: when the light intensity is higher, the radiation is greater, the photovoltaic power generation is greater, the surface temperature is higher, the more sunlight is sufficient, the photovoltaic value is higher, and as the temperature of the component changes, the open circuit voltage, short circuit current and peak power of the photovoltaic component will change, resulting in system efficiency. The influence of relative humidity on photovoltaics is mainly that the moisture in the air will reflect and scatter part of the solar radiation; The second hidden layer: The GRU layer is used to extract the time series characteristics, calculate the power generation at all time points in the photovoltaic data set time by time step, and the last step neuron outputs The power generation vector in the format is sent to the third hidden layer; The expressions of the third hidden layer and Gaussian output layer are: In the formula, The variance of the uncertainty in power generation forecast is the uncertainty in power generation forecast due to the randomness of the environmental data set. represents the power generation vector output by the GRU layer, represents the weight matrix of the third hidden layer, represents the bias vector of the third hidden layer, represents the weight matrix of the Gaussian output layer, Represents the bias vector of the Gaussian output layer. In actual application, the trained Gaussian-CNN-GRU prediction model can also fully consider the impact of various random factors on corporate electricity consumption, and can also predict corporate electricity load, comprehensively analyzing the influencing factors with high prediction accuracy; The adaptation management module is based on the power consumption prediction value , analyze the adaptation index of each enterprise , the calculation process is as follows: S41. Based on the power consumption forecast value and power generation forecast , Statistics The growth rate of electricity consumption of each enterprise in the next year is marked as , will The ratio of the maximum power consumption of an enterprise in the next year to 80% of the rated capacity of the transformer is marked as , will The average standard deviation of daily electricity load of each enterprise in the next year is marked as ; S42: Standardize the parameters in S41 to obtain hyperparameters , and ; In the formula, and Respectively The minimum and maximum growth rates of electricity consumption of each enterprise in the next year. and Respectively The minimum and maximum ratios of the maximum power consumption of an enterprise to 80% of the rated capacity of the transformer in the next year. and Respectively The minimum and maximum values of the average standard deviation of daily electricity load for each enterprise in the next year; S43. Calculate the first The fit index of an enterprise ; In the formula, Indicates according to , and Fixed weight of 2, comprehensive hyperparameters , and , get the The adaptation index of each enterprise; Adaptation management module according to the adaptation index The adaptation list is arranged from high to low, and the adaptation management module is based on the power generation forecast value. , giving priority to the enterprises with the highest ranking in the adaptation list, and intelligently managing energy storage resources with high utilization rate.
[0022] In this embodiment, the photovoltaic power generation module is used to effectively predict the future photovoltaic power generation power while considering multiple influencing factors. The enterprise's new energy system adaptation index is constructed by combining the three factors of the enterprise's future electricity consumption growth rate, future electricity overcapacity ratio and electricity load fluctuation. Target enterprises suitable for photovoltaic energy storage new energy systems are selected in a targeted manner, creating greater investment value and proposing the best discharge strategy.
[0023] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. Photovoltaic energy storage management system based on enterprise power load, characterized by: It includes power load module, photovoltaic power generation module and adaptation management module; The power load module is composed of an enterprise data unit, a related data unit and a power consumption prediction unit. The enterprise data unit collects a load data set through a network connection to the distribution network. The load data set includes the power consumption of all enterprises. The related data unit collects an impact data set through a network connection to the big data platform. The impact data set includes a variety of related factors that affect the power consumption of enterprises. The power consumption prediction unit is set with a fixed-duration monitoring cycle. , combined with the load data set and the impact data set, analyze the correlation coefficient between the enterprise's electricity consumption and each related factor , and calculate the next monitoring cycle of the enterprise Predicted power consumption within ; The photovoltaic power generation module is composed of a photovoltaic data unit, an environmental data unit and a power generation prediction unit. The photovoltaic data unit is connected to the photovoltaic power station management system through a network to collect photovoltaic data sets, and the photovoltaic data sets include the power generation of the photovoltaic power station at all time points. The environmental data unit is connected to the sensor device through a network to collect environmental data sets, and the environmental data sets include a variety of related factors that affect the photovoltaic power generation. The power generation prediction unit analyzes the monitoring period according to the photovoltaic data set and the environmental data set. The system efficiency of photovoltaic power plants , and the next monitoring cycle The predicted power generation value within ; The adaptation management module is based on the power consumption prediction value , analyze the adaptation index of each enterprise , and according to the adaptation index The adaptation list is arranged from high to low, and the adaptation management module is based on the power generation forecast value. , giving priority to enterprises with the highest ranking in the adaptation list.
2. The photovoltaic energy storage management system based on enterprise power load according to claim 1 is characterized in that: The expression of the load data set is , to The first to the The electricity consumption of an enterprise, Indicates the specific time for obtaining the electricity consumption of a single enterprise.
3. The photovoltaic energy storage management system based on enterprise power load according to claim 2 is characterized in that: The expression affecting the data set is: , represents the gross regional product, represents the gross industrial output value, Indicates the regional temperature, represents the power elasticity coefficient, Indicates the specific time for obtaining factors affecting the enterprise's electricity consumption.
4. The photovoltaic energy storage management system based on enterprise power load according to claim 3 is characterized in that: The correlation coefficient The calculation process is as follows: S11. Extract monitoring cycle based on load data set within, no. The electricity consumption of each enterprise is marked as , to Respectively From the first time point to the The electricity consumption at a point in time; S12. Extract monitoring cycles based on impact data sets Internal, affecting The regional GDP of the electricity consumption of enterprises is marked as , to From the first time point to the Gross regional product at a point in time; S13, calculate the The correlation coefficient between the electricity consumption of enterprises and the regional GDP ; In the formula, Indicates The company in The power consumption at a point in time, Indicates monitoring cycle within, no. The average electricity consumption of an enterprise, Indicates The GDP of a region at a point in time, Indicates monitoring cycle The average value of regional GDP is According to the covariance and standard deviation calculation formula, we get Pearson correlation coefficient between electricity consumption of enterprises and regional GDP.
5. The photovoltaic energy storage management system based on enterprise power load according to claim 4 is characterized in that: The power consumption forecast value The calculation process is as follows: S21. Based on the load data set, the ARIMA(1,1) model is used to perform time series analysis; In the formula, represents the random variable at the current moment in the ARIMA(1,1) model, i.e. Enterprises at the moment The power consumption, Represents the first-order autoregressive coefficient, which represents the influence coefficient of the time series value at the previous moment on the value at the current moment. represents the error term at the current moment, Represents the first-order moving average coefficient, which represents the influence coefficient of the error term at the previous moment on the value at the current moment. Indicates that the ARIMA(1,1) model is The predicted value one step ahead, i.e. Enterprises at the moment The power consumption prediction value is iterated successively. Indicates that the ARIMA(1,1) model is forward The predicted value at the step time, i.e. Enterprises at the moment The predicted value of electricity consumption; S22. Based on the correlation coefficient between the enterprise's electricity consumption and each related factor , the numerical error of the ARIMA(1,1) model is predicted by the HMM model. The HMM model uses the Viterbi algorithm of dynamic programming to predict the state sequence of the numerical error. The modeling process is as follows: Initialization: The initialization expression is: In the formula, Indicates status When it is observed The probability of state When the first The correlation coefficient between the electricity consumption of enterprises and the regional GDP The probability of being the maximum value, Indicates status The initial probability of Recursion: For , the maximum probability expression of the observation path obtained by recursion is: In the formula, Indicated in When, status It was observed that , and in When, status It was observed that The maximum probability of the path is Represents the previous state of the maximum probability path probability; Termination: The optimal path expression for the termination of the optimization is: In the formula, represents the probability of the optimal path, Indicates the state of the optimal path; Optimal path backtracking: For , the state sequence of the optimal path is obtained in reverse: ; S23, the predicted value obtained by the ARIMA(1,1) model is corrected by the state sequence predicted by the HMM model. In the formula, Represents the evaluation weight for the ARIMA(1,1) model prediction value. Represents the evaluation weight for the HMM model state sequence, , Indicates according to and Weight, comprehensive prediction value and state sequence, get the The next monitoring cycle of enterprises The predicted power consumption within .
6. The photovoltaic energy storage management system based on enterprise power load according to claim 5 is characterized in that: The expression of the photovoltaic data set is , to They are the first time point to the second time point of the photovoltaic power station. The amount of electricity generated at a point in time, Indicates the specific time for obtaining the power generation of the PV power station.
7. The photovoltaic energy storage management system based on enterprise power load according to claim 6 is characterized in that: The expression of the environmental data set is , represents irradiance, represents the surface temperature, Indicates the temperature of the photovoltaic module, Indicates the ambient humidity. Indicates the specific time for obtaining the factors affecting photovoltaic power generation.
8. The photovoltaic energy storage management system based on enterprise power load according to claim 7 is characterized in that: The system efficiency The calculation process is as follows: S31. Statistical monitoring cycle based on photovoltaic data set The total power generation within is marked as ; S32. Statistical monitoring cycle based on environmental data set The total radiation exposure within is marked as ; S33. Calculate monitoring cycle The system efficiency of photovoltaic power plants ; In the formula, Indicates the irradiance under standard test conditions, Indicates monitoring cycle The total amount of electricity actually output by the photovoltaic power station is represents the installed capacity of the photovoltaic power station, Indicates monitoring cycle The total amount of electrical energy that the photovoltaic power station is estimated to output.
9. The photovoltaic energy storage management system based on enterprise power load according to claim 8 is characterized in that: The power generation forecast value The calculation process is as follows: According to the photovoltaic data set, the Gaussian-CNN-GRU prediction model is used to predict the power generation. The Gaussian-CNN-GRU prediction model consists of an input layer, three hidden layers, and a Gaussian output layer. The first hidden layer is a one-dimensional convolutional neural network, the second hidden layer is a gated recurrent unit, and the third hidden layer is a fully connected layer. The Gaussian output layer contains two neurons, which output the power generation prediction value and the variance of the power generation prediction uncertainty respectively. Input layer: input photovoltaic data set and environmental data set; The first hidden layer: The CNN layer is used to explore the potential relationship between photovoltaic power generation and each related factor and extract high-dimensional features. The convolution layer is used to extract high-dimensional features. convolution kernels, mapping the power generation at all time points in the photovoltaic dataset to time series, and shorten the time series length to , output Matrix to GRU layer; The second hidden layer: The GRU layer is used to extract the time series characteristics, calculate the power generation at all time points in the photovoltaic data set time by time step, and the last step neuron outputs The power generation vector in the format is sent to the third hidden layer; The expressions of the third hidden layer and Gaussian output layer are: In the formula, The variance of the uncertainty in power generation forecast is the uncertainty in power generation forecast due to the randomness of the environmental data set. represents the power generation vector output by the GRU layer, represents the weight matrix of the third hidden layer, represents the bias vector of the third hidden layer, represents the weight matrix of the Gaussian output layer, Represents the bias vector of the Gaussian output layer.
10. The photovoltaic energy storage management system based on enterprise power load according to claim 9 is characterized in that: The fitness index The calculation process is as follows: S41. Based on the predicted power consumption and power generation forecast , Statistics The growth rate of electricity consumption of each enterprise in the next year is marked as , will The ratio of the maximum power consumption of an enterprise in the next year to 80% of the rated capacity of the transformer is marked as , will The average standard deviation of daily electricity load of each enterprise in the next year is marked as ; S42: Standardize the parameters in S41 to obtain hyperparameters , and ; In the formula, and Respectively The minimum and maximum growth rates of electricity consumption of each enterprise in the next year. and Respectively The minimum and maximum ratios of the maximum power consumption of an enterprise to 80% of the rated capacity of the transformer in the next year. and Respectively The minimum and maximum values of the average standard deviation of daily electricity load for each enterprise in the next year; S43. Calculate the first The fit index of an enterprise ; In the formula, Indicates according to , and Fixed weight of 2, comprehensive hyperparameters , and , get the The adaptation index of an enterprise.