Energy efficiency evaluation method and system for source network load storage multi-element load

Abstract

The invention discloses an energy efficiency evaluation method and system for a source network load storage multi-element load, relates to the technical field of energy efficiency metering, and aims to solve the energy efficiency metering problem of the source network load storage multi-element load, and the method comprises the following steps: S1, calculating an optimal configuration scheme of an intelligent sensing device, and forming an industrial park energy efficiency metering general scheme; s2, performing adaptive synchronous sampling on each energy consumption measurement node to obtain equipment energy consumption data; s3, training a load energy consumption prediction model and a photovoltaic output prediction model based on the equipment energy consumption data to form an overall energy efficiency prediction model; and S4, according to the energy efficiency metering general scheme and the overall energy efficiency prediction model, forming an industrial park energy efficiency metering and intelligent analysis system, and realizing energy consumption prediction of different links of source network load storage of the industrial park and overall energy efficiency metering and evaluation of the industrial park.

Description

Technical Field

[0001] The present invention relates to the technical field of energy efficiency measurement, and particularly relates to an energy efficiency evaluation method and system for multi-source, network, load, and storage multi-loads. Background Art

[0002] Industrial energy consumption in China accounts for about 65% of the total social energy consumption. Industrial parks are the gathering places of industrial users and are even more the core objects for electricity consumption monitoring. The "integration of power sources, grids, loads, and energy storage" industrial parks are the development templates advocated by the state. Accurate and reliable energy consumption monitoring and energy efficiency measurement are the basis for realizing the low-carbon and efficient operation of industrial parks and the dual control of carbon emissions, and are also the key to supporting "power supply + energy efficiency services". However, at present, industrial parks lack digital models that can accurately reflect the energy consumption of each link in the park and the overall energy efficiency characteristics, and the influence mechanism of the operation mode and dynamic load of the park on the energy consumption measurement of key equipment such as power sources, grids, loads, and energy storage is not clear. Moreover, the multi-source, network, load, and storage multi-loads involve a large scope and various types, showing mixed characteristics such as non-linearity, randomness, and time-variation, and it is difficult to directly obtain a stable and convergent response output through a simple single prediction model.

[0003] For example, there is a Chinese patent with the publication number CN105182126B, which relates to an improved energy efficiency measurement and detection method for distribution transformers. First, an equivalent two-port network model of the distribution transformer is constructed, then an energy efficiency calculation model of the distribution transformer is determined based on the equivalent two-port network model, and finally, an energy efficiency detection device for the distribution transformer under actual working conditions is constructed based on the energy efficiency calculation model to measure the energy efficiency value of the distribution transformer under actual working conditions, which can accurately measure the energy efficiency value of the distribution transformer under actual working conditions and be used as an important reference for energy-saving data, and be applied to the precise analysis, calculation, and energy efficiency level judgment of the distribution transformer for energy efficiency measurement; however, the energy efficiency measurement method of the Chinese patent with the publication number CN105182126B cannot be applied to the energy efficiency measurement of non-linear complex multi-loads. Summary of the Invention

[0004] In order to solve the problem of energy efficiency measurement of multi-source, network, load, and storage multi-loads, the present invention proposes an energy efficiency evaluation method and system for multi-source, network, load, and storage multi-loads, which can realize the energy consumption prediction of different links of power sources, networks, loads, and energy storage in industrial parks, as well as the energy efficiency measurement and evaluation of the whole industrial park.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: An energy efficiency evaluation method for multi-source, network, load, and storage multi-loads, comprising the following steps: S1: Calculate the optimal configuration scheme of intelligent sensing devices to form a general energy efficiency measurement scheme for industrial parks; S2: Perform adaptive synchronous sampling on each energy consumption measurement node to obtain equipment energy consumption data; S3: Train the load energy consumption prediction model and the photovoltaic output prediction model based on the device energy consumption data to form an overall energy efficiency prediction model; S4: According to the general energy efficiency measurement plan and the overall energy efficiency prediction model, form the industrial park energy efficiency measurement and intelligent analysis system.

[0006] In this technical solution, by establishing a load energy consumption prediction model and a photovoltaic output prediction model, an overall energy efficiency prediction model that accurately reflects the energy consumption of each link in the park and the overall energy efficiency characteristics is formed, realizing the energy efficiency measurement and low-carbon and efficient operation of the "source-network-load-storage integration" industrial park.

[0007] Preferably, the step S1 includes: S11, construct an intelligent perception device parameter optimization model, apply the whale optimization algorithm to optimize the parameters of the intelligent perception device, and obtain the optimal configuration plan of the intelligent perception device; S12, according to the optimal configuration plan of the intelligent perception device, form the general energy efficiency measurement plan for the industrial park.

[0008] Preferably, the step S2 includes: S21, obtain a high-precision time synchronization signal, that is, the PPS signal, through a synchronization signal receiver to provide a time reference for sampling; S22, adaptively adjust the sampling control parameters; S23, form a sampling module based on a polyphase interpolation filter to obtain device energy consumption data.

[0009] Preferably, the step S22 includes: real-time monitoring of the local crystal oscillator frequency, and adaptively adjusting the sampling control parameters according to the local crystal oscillator frequency deviation and the time reference provided by the PPS signal.

[0010] Preferably, the step S3 includes: S31, collect and decouple the device operation data, obtain a number of operation parameters, and form historical data; S32, according to the historical data, establish and train the load energy consumption prediction model and the photovoltaic output prediction model; S33, real-time collect various parameters of the device's current operation to obtain actual data, and analyze the historical operation data and the actual operation data to obtain the prediction error; S34, according to the feedback of the prediction error, correct and compensate the load energy consumption prediction model to form an overall energy efficiency prediction model.

[0011] Preferably, in step S31, the historical data includes historical operation data and historical weather data; conduct correlation analysis on the historical weather data and the device energy consumption data in the corresponding period, quantify the influence weight of weather factors on the device load, and obtain the weather influence factor.

[0012] Preferably, the step S32 includes: S321, training a generalized regression neural network according to historical operation data, establishing a load energy consumption prediction model, performing load energy consumption prediction based on the generalized regression neural network, and obtaining an intraday ultra-short-term load energy consumption prediction result; S322, establishing a photovoltaic power output prediction model according to the intraday ultra-short-term load energy consumption prediction result and weather impact factors.

[0013] Preferably, the step S34 includes: dynamically adjusting the prediction model according to the prediction error feedback by using an adaptive algorithm, and the adaptive algorithm includes Kalman filtering and recursive least squares.

[0014] The present invention also adopts the following technical solution: an energy efficiency evaluation and analysis system for a source-network-load-storage multi-source load, adopting the above-mentioned energy efficiency evaluation method for a source-network-load-storage multi-source load, and is characterized by including: A data acquisition module, which uses intelligent sensing devices to collect consumption data of various energies, and classifies, organizes, and stores the data; An energy efficiency analysis module, which performs fusion processing on the collected multi-source data and applies an energy efficiency prediction model to perform energy consumption prediction; An optimization module, which proposes optimization suggestions for the links and equipment with unreasonable energy consumption according to the analysis results of the energy efficiency analysis module; An early warning module, which is built-in with an energy consumption threshold and issues an early warning when the energy consumption exceeds the threshold.

[0015] Preferably, the intelligent sensing devices include intelligent electric meters, sensors, and energy efficiency monitoring terminals.

[0016] The beneficial effect of the present invention is that through the overall energy efficiency prediction model, the "source-network-load-storage" multi-source load with a large coverage, various types, and presenting mixed characteristics such as non-linearity, randomness, and time-variation can obtain a stable and convergent response output, realizing accurate energy efficiency measurement and evaluation of the multi-source load. Description of the Drawings

[0017] Figure 1 is a flowchart of an energy efficiency evaluation method for a source-network-load-storage multi-source load of the present invention. Detailed Embodiments

[0018] To make the purpose, technical solution, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described here are only the best embodiments of the present invention, only used to explain the present invention, and do not limit the protection scope of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0019] Embodiment 1 This embodiment provides an energy efficiency evaluation method for a source-network-load-storage multi-source load, and the flow chart is as Figure 1 shown, which can realize the energy consumption prediction of different links of the source-network-load-storage in the industrial park, as well as the energy efficiency measurement and evaluation of the whole industrial park, mainly including the following several steps.

[0020] Step S1, calculate the optimal configuration scheme of the intelligent sensing device to form a general energy efficiency measurement scheme for the industrial park; for the specific process of S1, it includes the following several sub-steps.

[0021] Step S11, construct a parameter optimization model for the intelligent sensing device, and apply the whale optimization algorithm to optimize the parameters of the intelligent sensing device to obtain the optimal configuration scheme of the intelligent sensing device. Among them, applying the whale optimization algorithm to optimize the parameters of the intelligent sensing device includes: discriminability analysis, non-dominated sorting and crowding degree calculation, constraint condition construction, and improvement to improve the global optimal solution ability of the algorithm.

[0022] The intelligent sensing device plays a key role in the energy efficiency measurement and management of the industrial park. The goal of its parameter optimization is to improve the energy efficiency, coverage, data transmission efficiency and overall performance of the device, while reducing energy consumption and cost.

[0023] Since the parameter optimization of the intelligent sensing device needs to optimize multiple parameters including but not limited to the deployment location of sensors, sampling frequency and communication method at the same time, and the environment of the industrial park is complex and changeable, an algorithm that can dynamically adjust parameters is required. Therefore, this embodiment adopts the whale optimization algorithm.

[0024] The whale optimization algorithm is an efficient swarm intelligence optimization algorithm that finds the optimal solution by simulating the foraging behavior of whales.

[0025] In this embodiment, the elite opposition-based learning strategy is adopted in the initialization stage to improve the quality of the initial population. At the same time, an adaptive weight mechanism is introduced to dynamically adjust the optimization weights of each parameter, and a chaotic mapping strategy is combined to enhance the global search ability of the algorithm.

[0026] An energy efficiency evaluation method for a source-network-load-storage multi-source load in this embodiment optimizes the deployment and parameter configuration of sensor nodes through the whale optimization algorithm, significantly improving the network coverage and energy efficiency.

[0027] Step S12, form a general energy efficiency measurement scheme for the industrial park according to the optimal configuration scheme of the intelligent sensing device.

[0028] Step S2, perform adaptive synchronous sampling on each energy consumption measurement node to obtain equipment energy consumption data; for the specific process of S2, it includes the following several sub-steps.

[0029] Step S21: Obtain a high-precision time synchronization signal, i.e., the PPS signal, through a synchronization signal receiver to provide a time reference for sampling.

[0030] The PPS signal is a very precise time synchronization signal, which is output in the form of one electrical pulse per second. The PPS timing device relies on receiving the timing signal sent by the GPS satellite to convert the pulse signal into a synchronization signal that can be recognized and utilized by electronic devices. In this embodiment, the synchronization signal receiver captures the GPS signal, extracts accurate clock information on the premise of ensuring signal quality, and outputs it as the PPS signal. In the energy consumption measurement nodes in the industrial park, after the synchronization signal receiver obtains the PPS signal, it uses it as the time reference for the entire sampling system. Each measurement node starts sampling energy consumption data based on the received PPS signal to ensure that all nodes are synchronized in time.

[0031] Step S22: Monitor the local oscillator frequency in real time, and adaptively adjust the sampling control parameters according to the local oscillator frequency deviation and the time reference provided by the PPS signal.

[0032] The local oscillator is a key component that generates the clock signal inside the energy consumption measurement node. The stability of its frequency directly affects the accuracy and synchronization of sampling. However, the frequency of the oscillator will drift due to environmental factors such as temperature and voltage, so it needs to be monitored in real time.

[0033] In this embodiment, a frequency counter is used to measure the output signal of the local oscillator in real time to obtain the actual frequency value of the oscillator. Compare the measured frequency value with the nominal frequency of the oscillator to calculate the frequency deviation. Then, combine the accurate time reference provided by the PPS signal and the real-time monitoring result of the local oscillator frequency to adaptively adjust the sampling control parameters.

[0034] Step S23: Form a sampling module based on a polyphase interpolation filter to obtain the device energy consumption data.

[0035] The polyphase interpolation filter can improve the sampling resolution and accuracy without increasing the sampling frequency. Its basic principle is to decompose the input signal according to different phases, then perform interpolation processing on the signals of each phase through a group of filters, and finally recombine the interpolated signals to obtain a sampling signal with higher resolution.

[0036] In the technical solution of step S2 in this embodiment, the collected energy consumption simulation signal is input into the sampling module based on a polyphase interpolation filter after analog-to-digital conversion. The sampling module based on the polyphase interpolation filter performs polyphase interpolation processing on the digital signal according to the PPS signal and the clock information provided by the local crystal oscillator. By adjusting the parameters of the filter and the interpolation algorithm, the detailed information of the energy consumption data can be obtained with higher accuracy at different sampling times, improving the accuracy and reliability of the energy consumption measurement, and further optimizing the adaptive synchronous sampling effect of each node.

[0037] Step S3: Train a load energy consumption prediction model and a photovoltaic power output prediction model based on the device energy consumption data to form an overall energy efficiency prediction model; for this step, it mainly includes the following multiple sub-steps.

[0038] Step S31: Collect and decouple the device operation data to obtain several operation parameters and form historical data.

[0039] The collection of device operation data is the basis of the entire process. A variety of parameters during device operation are monitored and collected through various sensors including temperature sensors, pressure sensors, and current transformers. The operation parameters include but are not limited to the operation temperature, working pressure, and energy consumption current of the device. These analog or digital signals are converted into a data form for subsequent analysis and processing to obtain historical data.

[0040] In step S31, the historical data includes historical operation data and historical weather data.

[0041] Perform correlation analysis on the historical weather data and the device energy consumption data in the corresponding period, quantify the influence weight of weather factors on the device load, and obtain the weather influence factor.

[0042] During the operation of the device, there are often complex interaction effects among various parameters. Therefore, it is necessary to sort out these interrelated and entangled data relationships, that is, to perform decoupling of the device operation data.

[0043] In this embodiment, multiple parameters are separated through data decoupling, and the independent influence of each parameter on the device performance and energy consumption is accurately found, providing a basis for accurate modeling and optimization.

[0044] Step S32: Establish and train a load energy consumption prediction model and a photovoltaic power output prediction model according to the historical data.

[0045] For the process of step S32, it specifically includes the following steps.

[0046] The first step: Train a generalized regression neural network according to the historical operation data, establish a load energy consumption prediction model, perform load energy consumption prediction based on the generalized regression neural network, and obtain the intra-day ultra-short-term load energy consumption prediction result.

[0047] The generalized regression neural network has a powerful non - linear mapping ability and is suitable for complex and variable prediction scenarios such as load energy consumption. Using the collected and decoupled historical operation data, the equipment operation conditions, time series, etc. are used as input neurons, and the load energy consumption is used as the output neuron to train the network.

[0048] By learning the internal laws of the data, the network can quickly predict the load energy consumption in the future period under given new working conditions. Compared with the traditional linear regression model, the load energy consumption prediction model based on the generalized regression neural network has a better fitting effect on complex non - linear energy consumption trends.

[0049] Apply the load energy consumption prediction model based on the generalized regression neural network to conduct intra - day ultra - short - term load energy consumption prediction. Intra - day ultra - short - term load energy consumption prediction focuses on predicting the load energy consumption within a very short time interval within a day. This ultra - short - term prediction can accurately capture the load fluctuations of the equipment caused by factors such as production rhythm and environmental changes at different times of the day, providing support for immediately adjusting the production plan and optimizing the start - stop of the equipment, making the energy distribution and use more time - efficient and reducing unnecessary energy waste.

[0050] In this embodiment, the time interval for load energy consumption prediction is 15 minutes. In some other embodiments, a half - hour or other time interval lengths can also be adopted according to actual needs.

[0051] Weather has a significant impact on the equipment load energy consumption. For example, the temperature directly affects the energy consumption of refrigeration and heating equipment, the light intensity affects the output of the photovoltaic system and then changes the power grid load distribution, and the wind speed and direction also have an impact on the operation conditions and energy consumption of ventilation and wind power generation equipment.

[0052] Correlate and analyze the historical weather data with the equipment energy consumption data in the corresponding period, quantify the influence weight of weather factors on the equipment load, obtain the weather influence factor, and incorporate it into the prediction model, which can improve the prediction accuracy.

[0053] It should be noted that in this embodiment, the historical weather data includes but is not limited to temperature, humidity, light, and wind speed.

[0054] In the second step, according to the intra - day ultra - short - term load energy consumption prediction results and the weather influence factor, establish a photovoltaic output prediction model.

[0055] Considering that in the actual scenario, the equipment load is restricted by multiple factors. In addition to the equipment's own working conditions and weather, it also involves factors such as raw material supply and market order demand.

[0056] Therefore, in this embodiment, multi - source information is integrated for multi - load prediction, data is input into the model from different dimensions, and a comprehensive load prediction result is output to fully reflect the equipment energy consumption demand.

[0057] Meanwhile, for photovoltaic power generation equipment, a photovoltaic power output prediction model was constructed based on local meteorological historical data, photovoltaic panel characteristic parameters, combined with machine learning and physical modeling methods to estimate the power generation power under different weather and times, assisting the power grid dispatching and energy management.

[0058] It should be noted that in this embodiment, the photovoltaic panel characteristic parameters include but are not limited to the photoelectric conversion efficiency, tilt angle, and orientation.

[0059] Step S33: Collect various parameters of the device running at present in real time to obtain actual data, and analyze the historical operation data and actual operation data to obtain the prediction error. Real-time data collection: Continuously capture various parameters of the device at the moment of running, obtain actual data that can timely reflect the change of the device's operation state, compare the actual data with the historical data, and once the device has abnormal energy consumption increase and working condition fluctuation, the real-time data can trigger an alarm immediately, providing support for the operation and maintenance personnel to quickly intervene in the fault and ensure the efficient operation of the device.

[0060] Since the model cannot perfectly simulate complex and changeable real working conditions, there must be a deviation between the predicted value and the actual value, and the prediction error is obtained by comparing the historical data and the actual data.

[0061] Analyzing the error sources, such as data noise, model structure defects, and sudden working condition changes not covered by the model, is the key to optimizing the model.

[0062] By comparing the magnitude and distribution law of the prediction error, specifically improving the selection of model input variables and adjusting model parameters, the prediction can be made closer to the actual energy consumption situation.

[0063] Step S34: Modify and compensate the load energy consumption prediction model according to the prediction error feedback to form an overall energy efficiency prediction model.

[0064] According to the prediction error feedback, an adaptive algorithm is used to dynamically adjust the prediction model, and the adaptive algorithm includes Kalman filtering and recursive least squares.

[0065] Modifying and compensating the load energy consumption prediction model can further improve the prediction accuracy of the model.

[0066] In this embodiment, according to the prediction error feedback, an adaptive algorithm is used to dynamically adjust the prediction model.

[0067] When it is found that the prediction error is large in a certain period, the model automatically retraces the data of the relevant period to re-learn, modifies the internal weight parameters, or introduces additional compensation factors to improve the model, continuously improving the prediction accuracy.

[0068] It should be noted that the additional compensation factor here can be a special scenario identification variable such as a holiday or sudden equipment maintenance. The selectable adaptive algorithms include, but are not limited to, Kalman filtering and recursive least squares. In this embodiment, the Kalman filtering algorithm is adopted as the adaptive algorithm.

[0069] After correcting and compensating the load energy consumption prediction model, from the macroscopic perspective of the park, comprehensively considering various types of equipment such as production equipment, lighting, air conditioning, and power transmission in the park, different energy supply systems including electricity, gas, and heat, and the production coordination relationship between enterprises, the operation data of each subsystem and equipment is collected and summarized. Combining external factors such as the park's production plan, energy price trend, and weather changes, an integrated energy efficiency prediction model is constructed to provide a strategic decision-making basis for the park's energy planning, energy-saving transformation, and energy cost optimization, and promote the improvement of the overall energy utilization efficiency of the park.

[0070] Step S4, according to the general energy efficiency measurement plan and the overall energy efficiency prediction model, an energy efficiency measurement and analysis system for industrial parks is formed.

[0071] In this embodiment, an energy efficiency measurement and analysis system for the multi-source-network-load-storage multi-load in industrial parks is carried out according to the general energy efficiency measurement plan and the overall energy efficiency prediction model.

[0072] This embodiment provides an energy efficiency evaluation method for the multi-source-network-load-storage multi-load. By establishing an overall energy efficiency prediction model for industrial parks, the energy consumption prediction of different links of the multi-source-network-load-storage in industrial parks and the energy efficiency measurement and evaluation of the overall industrial park are realized.

[0073] Embodiment 2 This embodiment provides an energy efficiency evaluation and analysis system for the multi-source-network-load-storage multi-load, which adopts the above-mentioned energy efficiency evaluation method for the multi-source-network-load-storage multi-load, including a data acquisition module, an energy efficiency analysis module, an optimization module, and an early warning module. Applied in industrial parks, it can monitor and analyze the energy consumption of various enterprises and public facilities in the park, realize the unified management and optimal dispatching of park energy, reduce the overall energy consumption and emissions of the park, and not only reduce costs but also promote the green development of the park.

[0074] The data acquisition module uses intelligent sensing devices to collect the consumption data of various energies, classify, organize, and store the data. The intelligent sensing devices include, but are not limited to, intelligent electric meters, sensors, and energy efficiency monitoring terminals, and are used to collect power data, environmental data, equipment operation status data, etc. at each link of the energy production end, power network, load end, and energy storage system.

[0075] Among them, the power data includes, but is not limited to, voltage, current, power, and electricity consumption.

[0076] The data acquisition module includes the sampling module based on a polyphase interpolation filter described in step S23 of the above-mentioned energy efficiency evaluation method for a source-network-load-storage multi-source load. For the detailed content of this module, refer to the introduction of the above system regarding step S23.

[0077] In this embodiment, the system accurately measures different types of loads and loads of different energy forms through the data acquisition module, realizing the separate monitoring and statistics of various loads.

[0078] Among them, different types of loads include industrial loads, commercial loads, and residential loads, and loads of different energy forms include electric loads, thermal loads, and gas loads.

[0079] The energy efficiency analysis module performs fusion processing on the collected multi-source data and applies an energy efficiency prediction model for energy consumption prediction.

[0080] In this embodiment, the energy efficiency analysis module cleans, stores, analyzes, and mines the transmitted data, calculates various energy efficiency indicators such as energy utilization rate, load rate, energy consumption distribution, and power factor, and identifies load characteristics and energy consumption patterns.

[0081] The energy efficiency analysis module can comprehensively analyze the energy efficiency of each link of the source-network-load-storage, evaluate the energy utilization efficiency of the overall system and each part, find out the links and reasons for low energy efficiency, provide a basis for energy-saving transformation and optimized operation, and provide data support for subsequent energy problem optimization.

[0082] The optimization module can put forward optimization suggestions for the links and equipment with unreasonable energy consumption according to the analysis results of the energy efficiency analysis module.

[0083] Based on the real-time data monitoring and energy efficiency analysis of each link of the source-network-load-storage, put forward optimization suggestions for the links and equipment with unreasonable energy consumption, realize the optimized allocation and efficient utilization of energy, and reduce energy costs.

[0084] The warning module is built-in with an energy consumption threshold and issues a warning when the energy consumption exceeds the threshold. The warning module can monitor the operating status of the energy system in real time and give warnings about abnormal situations and potential risks. For example, when there is a load overload, voltage abnormality, or energy storage system failure, it will issue a warning in time to notify the user to take measures to ensure the safe and stable operation of the energy system.

[0085] For the specific content of the warning module, refer to step S33.

[0086] The energy efficiency evaluation and analysis system for a source-network-load-storage multi-source load in this embodiment can achieve the deep integration and efficient utilization of energy, reduce energy consumption, and improve energy utilization efficiency through an energy efficiency prediction model and optimized control.

Claims

1. An energy efficiency evaluation method for a multi-source-network-load-storage multi-load, characterized in that, Including: S1: Calculate the optimal configuration plan of the intelligent perception device to form a general plan for energy efficiency measurement in the industrial park; S2: Perform adaptive synchronous sampling on each energy consumption measurement node to obtain device energy consumption data; S3: Train a load energy consumption prediction model and a photovoltaic output prediction model based on the device energy consumption data to form an overall energy efficiency prediction model; S4: According to the general plan for energy efficiency measurement and the overall energy efficiency prediction model, form an energy efficiency measurement and intelligent analysis system for the industrial park.

2. The energy efficiency evaluation method for a multi-source-network-load-storage multi-load according to claim 1, wherein The step S1 includes: S11. Construct an intelligent perception device parameter optimization model, apply the whale optimization algorithm to optimize the parameters of the intelligent perception device, and obtain the optimal configuration plan of the intelligent perception device; S12. According to the optimal configuration plan of the intelligent perception device, form a general plan for energy efficiency measurement in the industrial park.

3. The energy efficiency evaluation method for a multi-source-network-load-storage multi-load according to claim 1, characterized in that, The step S2 includes: S21. Obtain a high-precision time synchronization signal, i.e., the PPS signal, through a synchronization signal receiver to provide a time reference for sampling; S22. Adaptively adjust the sampling control parameters; S23. Form a sampling module based on a polyphase interpolation filter to obtain device energy consumption data.

4. The energy efficiency evaluation method for a multi-source-network-load-storage multi-load according to claim 3, characterized in that The step S22 includes: Real-time monitor the local crystal oscillator frequency, and adaptively adjust the sampling control parameters according to the local crystal oscillator frequency deviation and the time reference provided by the PPS signal.

5. The energy efficiency evaluation method for a multi-source-network-load-storage multi-load according to claim 1, characterized in that, The step S3 includes: S31. Collect and decouple the device operation data to obtain several operation parameters and form historical data; S32. According to the historical data, establish and train a load energy consumption prediction model and a photovoltaic output prediction model; S33. Real-time collect various parameters of the device's current operation to obtain actual data, and analyze the historical operation data and the actual operation data to obtain the prediction error; S34. According to the feedback of the prediction error, correct and compensate the load energy consumption prediction model to form an overall energy efficiency prediction model.

6. The energy efficiency evaluation method for a source-network-load-storage multi-load according to claim 5, characterized in that In step S31, the historical data includes historical operation data and historical weather data; Correlate and analyze the historical weather data and the device energy consumption data in the corresponding period, quantify the influence weight of weather factors on the device load, and obtain the weather influence factor.

7. The energy efficiency evaluation method for a multi-source-network-load-storage multi-load according to claim 6, characterized in that, The step S32 includes: S321. Train a generalized regression neural network according to the historical operation data, establish a load energy consumption prediction model, and perform load energy consumption prediction based on the generalized regression neural network to obtain the intra-day ultra-short-term load energy consumption prediction result; S322. According to the intra-day ultra-short-term load energy consumption prediction result and the weather influence factor, establish a photovoltaic output prediction model.

8. An energy efficiency evaluation method for a multi-source-network-load-storage multi-load, according to claim 5 or 6 or 7, characterized in that The step S34 includes: According to the feedback of the prediction error, dynamically adjust the prediction model using an adaptive algorithm, and the adaptive algorithm includes Kalman filtering and recursive least squares.

9. An energy efficiency evaluation and analysis system for a multi-source-network-load-storage multi-load, adopting the energy efficiency evaluation method for a multi-source-network-load-storage multi-load described in any one of claims 1-8, characterized in that, Including: Data acquisition module, which uses an intelligent perception device to collect the consumption data of various energies, and classifies, organizes, and stores the data; Energy efficiency analysis module, which performs fusion processing on the collected multi-source data and applies the energy efficiency prediction model to predict energy consumption; Optimization module, which proposes optimization suggestions for the links and devices with unreasonable energy consumption according to the analysis results of the energy efficiency analysis module; Early warning module, which has an energy consumption threshold built in, and issues an early warning when the energy consumption exceeds the threshold.

10. An energy efficiency evaluation and analysis system for a source-network-load-storage multi-load, according to claim 9, characterized in that, The intelligent sensing device includes a smart meter, a sensor, and an energy efficiency monitoring terminal.

Images