Air conditioning use algorithm and energy consumption efficiency optimization method based on micro-grid strategy

Abstract

The invention discloses an air conditioning use algorithm and an energy consumption efficiency optimization method based on a micro-grid strategy, and relates to the technical field of resource allocation, and the method comprises the following steps: obtaining historical comprehensive data of an energy system, carrying out the preprocessing of the obtained historical comprehensive data, and extracting the load characteristics of the comprehensive energy system based on a correlation analysis method, obtaining load characteristic data of the integrated energy system; taking the load characteristic data of the integrated energy system as the input of the multi-layer feed-forward network model, and training the multi-layer feed-forward network model to obtain an optimized air conditioner resource allocation scheme; and on the basis of the trained multi-layer feedforward network model, an optimized air conditioner resource allocation strategy is obtained, allocation operation of the unit is adjusted according to the optimized air conditioner resource allocation strategy, and air conditioner resource allocation is dynamically adjusted by monitoring the unit in real time. According to the method, through algorithm optimization, the energy consumption is reduced, and the influence on the environment is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of resource allocation, and more specifically, to an air conditioning usage algorithm based on a microgrid strategy and an energy consumption efficiency optimization method. Background Art

[0002] Energy resources refer to natural substances that provide energy for humans. Any substance or the movement of a substance that can provide a certain form of energy can be called an energy source. Nature endows humans with a variety of energy sources. One is the energy from the sun, including solar radiation energy and coal, biomass energy, etc. that indirectly come from solar energy; the second is the energy from the earth itself, such as geothermal energy and atomic energy; the third is the energy generated by the interaction between the earth and other celestial bodies, such as tidal energy. In a microgrid, distributed power sources (such as solar energy, wind energy, etc.) and loads are uniformly managed through a microgrid control system.

[0003] In the context of efficient energy utilization and sustainable development, the air conditioning system plays a crucial role in the application of the heating, ventilation, and air conditioning (HVAC) field. By adopting advanced energy-saving technologies and optimization strategies, the air conditioning system can achieve more efficient and intelligent energy management, contributing to the sustainable development of society.

[0004] In today's society, the efficient utilization of energy and sustainable development have become important topics. However, in terms of energy efficiency, certain challenges still exist. As an important part of building energy consumption, the energy consumption of the HVAC system accounts for a relatively large proportion of the total social energy consumption. Traditional HVAC systems often adopt relatively simple control methods and cannot perform precise energy allocation and control according to actual situations, resulting in serious energy waste.

[0005] Regarding the problems in the related technologies, no effective solutions have been proposed yet. Summary of the Invention

[0006] Regarding the problems in the related technologies, the present invention proposes an air conditioning usage algorithm based on a microgrid strategy and an energy consumption efficiency optimization method to overcome the above-mentioned technical problems existing in the existing related technologies.

[0007] To this end, the specific technical solutions adopted by the present invention are as follows:

[0008] According to one aspect of the present invention, there is provided an air conditioning usage algorithm based on a microgrid strategy and an energy consumption efficiency optimization method, and the optimization method includes the following steps:

[0009] S1. Obtain the historical comprehensive data of the energy system, preprocess the obtained historical comprehensive data, and extract the load characteristics of the integrated energy system based on the correlation analysis method to obtain the load characteristic data of the integrated energy system;

[0010] S2. Use the load characteristic data of the integrated energy system as the input of the multi-layer feedforward network model, train the multi-layer feedforward network model, and obtain an optimized air-conditioning resource allocation plan;

[0011] S3. Based on the trained multi-layer feedforward network model, obtain an optimized air-conditioning resource allocation strategy, adjust the distribution operation of the units according to the optimized air-conditioning resource allocation strategy, and dynamically adjust the air-conditioning resource allocation through real-time monitoring of the units.

[0012] Furthermore, to obtain the historical comprehensive data of the energy system, preprocess the obtained historical comprehensive data, and extract the load characteristics of the integrated energy system based on the correlation analysis method, the steps to obtain the load characteristic data of the integrated energy system are as follows:

[0013] S11. Obtain the historical comprehensive data of the energy system, perform data screening on the obtained historical comprehensive data, remove abnormal and missing historical comprehensive data, and perform standardization processing on the screened historical comprehensive data to obtain the standardized historical comprehensive data;

[0014] S12. Based on the correlation analysis method, combine the standardized historical comprehensive data and calculate the Pearson correlation coefficient of the historical comprehensive data;

[0015] S13. Select the historical comprehensive data with a higher Pearson correlation coefficient and input it into the convolutional neural network-long short-term memory network prediction model for feature extraction to obtain the load characteristics of the integrated energy system.

[0016] Furthermore, the historical comprehensive data of the energy system includes: historical energy load data, weather data, and date data.

[0017] Furthermore, the steps to select the historical comprehensive data with a higher Pearson correlation coefficient and input it into the convolutional neural network-long short-term memory network prediction model for feature extraction to obtain the load characteristics of the integrated energy system are as follows:

[0018] S131. Select the historical comprehensive data with a higher Pearson correlation coefficient and extract the local load characteristics of the integrated energy system through the convolutional neural network model;

[0019] S132. Based on the long short-term memory network model, combine the extraction of local load characteristics of the integrated energy system and extract the time series characteristics;

[0020] S133. Integrate the extracted local load characteristics of the integrated energy system and the time series characteristics into the load characteristics of the integrated energy system.

[0021] Furthermore, the steps to use the load characteristic data of the integrated energy system as the input of the multi-layer feedforward network model, train the multi-layer feedforward network model, and obtain an optimized air-conditioning resource allocation plan are as follows:

[0022] S21. Divide the load characteristic data of the integrated energy system into a training set, a validation set, and a test set;

[0023] S22. Based on the data in the training set, use the typical load day fitting method of normal distribution to obtain the annual daily average load prediction curve that meets the typical characteristic constraints. Take values according to time periods to obtain the corresponding load prediction situation;

[0024] S23. According to the data in the training set and combined with the load prediction situation, obtain the corresponding cooling water temperature range, and based on the performance curve of the air conditioning equipment, obtain the maximum power and maximum cooling capacity of the equipment;

[0025] S24. Based on the performance of the equipment and the cooling water temperature range, use the logical deduction method to loop in reverse time to obtain the optimized state situation corresponding to the corresponding time period, and combined with the electricity price factor, obtain the optimized air conditioning resource allocation plan.

[0026] Further, the performance curve of the air conditioning equipment includes: the machine operation performance curve with the equipment performance parameters existing at the factory for each equipment and the performance curve obtained through the data acquired by intelligent metering monitoring of the equipment operation parameters.

[0027] Further, based on the data in the training set, using the typical load day fitting method of normal distribution to obtain the annual daily average load prediction curve that meets the typical characteristic constraints and taking values according to time periods to obtain the corresponding load prediction situation includes the following steps:

[0028] S221. Based on the data in the training set, use the typical load day fitting method of normal distribution to fit the training set data at the same moment of different load days into a normal distribution curve, and take the expectation of the normal distribution curve as the load value at this moment to fit the typical daily average load curve;

[0029] S222. Eliminate the date influence on the typical daily average load curve of historical years to obtain the annual daily average load shape factor prediction curve;

[0030] S223. Based on the pre-assumed load typical characteristic prediction value, combine the annual daily average load shape factor prediction curve with the typical characteristic prediction value to obtain the annual daily average load prediction curve that meets the typical characteristic constraints, and take values according to time periods to obtain the corresponding load prediction situation.

[0031] Further, based on the data in the training set, using the typical load day fitting method of normal distribution to fit the training set data at the same moment of different load days into a normal distribution curve and taking the expectation of the normal distribution curve as the load value at this moment to fit the typical daily average load curve includes the following steps:

[0032] S2211. Statistically analyze the training set data at the same time every day, calculate the frequency of each load value, and obtain the probability distribution of the load at this time;

[0033] S2212. Use statistical methods to test whether the load values at the same time on different days follow a normal distribution;

[0034] S2213. For the load data that follows a normal distribution, use the maximum likelihood estimation method to fit it into a normal distribution curve;

[0035] S2214. According to the fitted normal distribution curve, set evaluation indicators. According to the set evaluation indicators, select typical load days that meet the conditions from the historical load data, and fit to obtain the annual daily average load prediction curve that meets the typical feature constraints.

[0036] Furthermore, obtaining the annual daily average load shape factor prediction curve by eliminating the date influence on the typical daily average load curve of historical years includes the following steps:

[0037] S2221. Use the periodic smoothing processing method to perform periodic smoothing processing on the daily average load curve of historical years;

[0038] S2222. Use the calendar features of holidays in historical years and the target year to be predicted to generate a dynamic time anchor matrix;

[0039] S2223. Based on the daily average load curve of historical years after periodic smoothing processing, combined with the dynamic time anchor matrix, obtain the annual daily average load shape factor prediction curve.

[0040] Further, after preprocessing the obtained historical comprehensive data, it further includes obtaining screened historical comprehensive data; the process of obtaining the screened historical comprehensive data is as follows: obtaining preprocessed historical comprehensive data by preprocessing the historical comprehensive data; obtaining a heat change evaluation index by combining the temperature change amount, humidity change amount, preset maximum temperature change value and preset maximum humidity change value obtained from the database; obtaining an electricity consumption evaluation index by combining the average power of the device, device usage duration, heat change evaluation index within the reference compliance change range and preset maximum electricity consumption value obtained from the database; obtaining a load evaluation index by combining the air flow, device heat, heat change evaluation index within the reference compliance change range and specific heat capacity of air, preset heat load value and preset cold load value obtained from the database; screening the preprocessed historical comprehensive data based on the electricity consumption evaluation index and load evaluation index to obtain the screened historical comprehensive data; the screened historical comprehensive data represents the preprocessed historical comprehensive data corresponding to the electricity consumption evaluation index within the reference electricity consumption range and the load evaluation index within the reference load range; the heat change evaluation index is used to evaluate the degree of heat change; the electricity consumption evaluation index is used to evaluate the electricity consumption of the preset device; the load evaluation index is used to reflect the load condition of the preset device; the preprocessed historical comprehensive data includes temperature change amount, humidity change amount, average power of the device, device usage duration, air flow and device heat.

[0041] Further, the specific process of obtaining the heat change evaluation index is as follows: obtaining a temperature change compliance value by performing a ratio operation on the sum of the temperature change amount and the preset maximum temperature change value and twice the preset maximum temperature change value; obtaining a humidity change compliance value by performing a ratio operation on the sum of the humidity change amount and the preset maximum humidity change value and twice the preset maximum humidity change value; obtaining a heat change evaluation index by combining the temperature change compliance value, humidity change compliance value, heat first influence weight and heat second influence weight obtained from the database.

[0042] Further, the specific process of obtaining the electricity consumption evaluation index is as follows: obtaining an electricity consumption compliance value by performing a ratio operation on the sum of the initial electricity consumption value and the preset maximum electricity consumption value and twice the preset maximum electricity consumption value; the initial electricity consumption value is represented by the result of multiplying the average power of the device and the device usage duration; obtaining an electricity consumption evaluation index by performing an operation on the electricity consumption compliance value and the heat change evaluation index within the reference compliance change range.

[0043] Further, the specific process for obtaining the load assessment index is as follows: Multiply the air flow rate, the specific heat capacity of air, and the temperature change amount to obtain the initial heat load value; Divide the sum of the initial heat load value and the preset heat load value by twice the preset heat load value to obtain the heat load compliance value; Add the initial heat load value and the equipment heat to obtain the initial cooling load value; Divide the sum of the initial cooling load value and the preset cooling load value by twice the preset cooling load value to obtain the cooling load compliance value; Combine the heat load compliance value, the cooling load compliance value, and the heat change assessment index within the reference compliance change range to obtain the load assessment index.

[0044] The beneficial effects of the present invention are as follows:

[0045] 1. Through algorithm optimization, the present invention comprehensively considers factors such as historical data, temperature, electricity price, and performance, and can achieve more precise energy distribution and control of the HVAC system; This not only helps to reduce energy consumption, reduce the impact on the environment, but also saves energy costs for users, improves economic benefits, and improves the implementation of energy consumption efficiency. At the same time, it also avoids the overuse of energy and pollution emissions, achieving the goal of saving resources and reducing pollution.

[0046] 2. By analyzing the operation data of the past HVAC system, the present invention can understand the energy demand patterns under different seasons, different time periods, and different climate conditions; Based on this historical data, the algorithm can establish a prediction model to predict future energy demand in advance, so as to reasonably arrange energy supply; At the same time, the historical data can also help to discover problems and potential energy-saving spaces in the operation of the HVAC system.

[0047] 3. By real-time monitoring of the indoor and outdoor temperature changes and incorporating the temperature data into the algorithm, the present invention can more precisely control the operation of the HVAC system. In addition, the temperature data can be combined with other factors, such as historical data and electricity price data, to further optimize the energy distribution strategy; According to the real-time electricity price information, combined with the energy demand and performance characteristics of the HVAC system, automatically adjust the operation strategy.

[0048] 4. Through the analysis of performance data, the algorithm can select the optimal equipment combination and operation parameters to achieve efficient utilization of energy, reasonably distribute energy, ensure that each area can obtain a comfortable indoor environment, and at the same time minimize energy consumption. In addition, the performance data can also be used to monitor and diagnose the operation status of HVAC equipment, timely discover faults and problems, and ensure the stable operation of the system.

[0049] 5. The present invention combines the temperature change amount, the humidity change amount, the preset maximum temperature change value and the preset maximum humidity change value obtained from the database to obtain a heat change evaluation index, realizing the precise quantification of the degree of heat change evaluation, and further improving the accuracy of the degree of heat change evaluation.

[0050] 6. The present invention combines the average power of the device, the device usage duration, the heat change evaluation index within the reference compliance change range and the preset maximum power consumption value obtained from the database to obtain a power consumption evaluation index, realizing the precise quantification of the power consumption situation of the preset device, and further improving the reliability of the power consumption situation evaluation of the preset device.

[0051] 7. The present invention combines the air flow rate, the device heat, the heat change evaluation index within the reference compliance change range, the specific heat capacity of air, the preset heat load value and the preset cooling load value obtained from the database to obtain a load evaluation index, realizing the precise quantification of the load situation of the preset device, and further improving the accuracy of the load situation evaluation of the preset device. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0053] Figure 1 is a flowchart of an air conditioning usage algorithm and energy consumption efficiency optimization method based on a microgrid strategy according to an embodiment of the present invention;

[0054] Figure 2 is one of the partial analysis diagrams of the algorithm process in the air conditioning usage algorithm and energy consumption efficiency optimization method based on a microgrid strategy according to an embodiment of the present invention;

[0055] Figure 3 is another partial analysis diagram of the algorithm process in the air conditioning usage algorithm and energy consumption efficiency optimization method based on a microgrid strategy according to an embodiment of the present invention;

[0056] Figure 4 is the third partial analysis diagram of the algorithm process in the air conditioning usage algorithm and energy consumption efficiency optimization method based on a microgrid strategy according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] To further illustrate the embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention.

[0058] According to an embodiment of the present invention, there is provided an air-conditioning usage algorithm and an energy consumption efficiency optimization method based on a microgrid strategy.

[0059] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation manners. As Figure 1 shown, according to an embodiment of the present invention, there is provided an air-conditioning usage algorithm and an energy consumption efficiency optimization method based on a microgrid strategy. The optimization method includes the following steps:

[0060] S1. Obtain the historical comprehensive data of the energy system, preprocess the obtained historical comprehensive data, and extract the load characteristics of the integrated energy system based on the correlation analysis method to obtain the load characteristic data of the integrated energy system.

[0061] Specifically, obtaining the historical comprehensive data of the energy system, preprocessing the obtained historical comprehensive data, and extracting the load characteristics of the integrated energy system based on the correlation analysis method to obtain the load characteristic data of the integrated energy system includes the following steps:

[0062] S11. Obtain the historical comprehensive data of the energy system, perform data screening on the obtained historical comprehensive data, remove abnormal and missing historical comprehensive data, and perform standardization processing on the screened historical comprehensive data to obtain the standardized historical comprehensive data;

[0063] S12. Based on the correlation analysis method, combine the standardized historical comprehensive data to calculate the Pearson correlation coefficient of the historical comprehensive data;

[0064] S13. Select the historical comprehensive data with a higher Pearson correlation coefficient and input it into a convolutional neural network-long short-term memory network prediction model for feature extraction to obtain the load characteristics of the integrated energy system.

[0065] Specifically, the historical comprehensive data of the energy system includes: energy historical load data, weather data, and date data.

[0066] It should be noted that the historical energy load data is based on the data collection of the energy station energy management system. It can help us collect, store, process, and analyze a large amount of energy data. Through algorithms, we can find out the historical energy operation conditions. Through the learning and simulation of algorithms and continuous curve fitting of numbers, we can complete the guidance for favorable energy consumption prediction basis and distribution control, serving as a favorable basis for distribution. In the absence of a large amount of data, the algorithm can use the performance curve of the equipment as a basis and simulate and learn based on the actual functional area. As the performance curve of the equipment decreases with the usage time, it is corrected synchronously to achieve double simulation training of history and equipment performance curve.

[0067] It should be noted that the weather data is the integration of real-time weather data, which is crucial for energy management. The algorithm corrects the data based on the real-time sunny or cloudy situation, and is corroborated by the same conditions over the years to improve the accuracy of the algorithm for resource allocation. By real-time monitoring of indoor and outdoor temperatures and combining historical data and weather forecasts, the algorithm can predict future temperature changes, thereby adjusting the energy supply in advance to maintain the best indoor comfort and energy efficiency.

[0068] Specifically, the steps of inputting the historical comprehensive data with a relatively high Pearson correlation coefficient into the convolutional neural network-long short-term memory network prediction model for feature extraction to obtain the comprehensive energy system load characteristics are as follows:

[0069] S131. Select the historical comprehensive data with a relatively high Pearson correlation coefficient, and extract the local comprehensive energy system load characteristics through the convolutional neural network model;

[0070] S132. Based on the long short-term memory network model, combined with the extraction of local comprehensive energy system load characteristics, extract the time series characteristics;

[0071] S133. Integrate the extracted local comprehensive energy system load characteristics and time series characteristics into the comprehensive energy system load characteristics.

[0072] S2. Use the comprehensive energy system load characteristic data as the input of the multi-layer feedforward network model, train the multi-layer feedforward network model, and obtain an optimized air-conditioning resource allocation plan.

[0073] It should be noted that for evaluation and optimization, the performance of the model is evaluated using the validation set. Common evaluation metrics include Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The multi-layer feedforward neural network model is trained using the training set data. Then, the model is evaluated using the validation set, and common evaluation metrics are calculated, including MAE, MSE, RMSE, and MAPE, which can reflect the accuracy and error magnitude of the model prediction. Next, the hyperparameters of the model, such as the learning rate and the number of hidden layers, are adjusted according to the evaluation results to optimize the model performance. Among them, Genetic Algorithm Particle Swarm Optimization (GAPSO) is used to optimize the parameters in the prediction model to find the optimal initial learning rate and the number of neurons in the hidden layer, and the model is retrained using the new hyperparameters. Finally, the robustness of the model is further verified through K-Fold Cross-Validation and a final evaluation is conducted on an independent test set to ensure that the model has good generalization ability. This process helps to improve the accuracy and reliability of the model in predicting energy demand. The proposed model is tested using actual data to verify the effectiveness and practicality of the model. Through case analysis, the prediction accuracy and generalization ability of the model are obtained.

[0074] After determining the model parameters, data visualization is carried out based on the energy management system to intuitively view the changes in algorithm parameters and prediction results, so as to improve the accuracy and practicality of the prediction. The comparison between the prediction results and the actual data is shown through charts to help understand the accuracy of the model prediction.

[0075] Specifically, taking the load characteristic data of the integrated energy system as the input of the multi-layer feedforward network model, training the multi-layer feedforward network model, and obtaining an optimized air-conditioning resource allocation scheme includes the following steps:

[0076] S21. Divide the load characteristic data of the integrated energy system into a training set, a validation set, and a test set;

[0077] S22. Based on the data in the training set, use the typical load day fitting method of the normal distribution to obtain the annual daily average load prediction curve that meets the typical characteristic constraints, and obtain the corresponding load prediction situation according to the time period;

[0078] S23. According to the training set data and combined with the load prediction situation, obtain the corresponding cooling water temperature range, and based on the performance curve of the air-conditioning equipment, obtain the maximum power and maximum cooling capacity of the equipment;

[0079] S24. Based on the performance of the device and the range of cooling water temperature, use the logical deduction method to loop in reverse time to obtain the optimized state under the corresponding time period, and combine with the electricity price factor to obtain an optimized air-conditioning resource allocation plan.

[0080] It should be noted that for the historical load data of the energy system, data screening is mainly carried out to remove outliers such as abnormal data, unreasonable data, and dirty data. The cumulative average of all loads within a time period (per hour) is calculated to obtain the relative load for the corresponding time period. The two-dimensional date and time period are flattened into one-dimensional time period data using a mathematical matrix. Here, a matrix is a basic concept in mathematics, especially important in linear algebra. A matrix is a rectangular array arranged by numbers, symbols, or expressions, usually used to represent linear equations, transformations, linear mappings, etc. At the same time, weights are assigned according to the proximity of the date. The closer the date is to the prediction date, the closer the corresponding weight is to 1.

[0081] Fit the collected historical load data, including timestamp, load value, and date information, through a fitting function. Clean the historical load data, handle missing values and outliers, and convert the timestamp into a format suitable for analysis. Extract time features such as hour, day, week, etc. These features are crucial for predicting the load. Consider converting the date into a numerical type so that the model can better understand and learn the time pattern.

[0082] Train the model using historical data. During the training process, the sliding window method can be used. Each time a new observation value is added to the training set, this observation value is deleted from the beginning of the set to keep the training set window fixed. This method is called walk-forward validation, which is the gold standard for time series model evaluation. The method of fitting typical load days based on the normal distribution fits the load data at the same moment on different load days into a normal distribution curve, and then uses the expectation of this distribution curve as the load value at this moment to fit the typical load day curve. Obtain the load curve, take values according to the time period, and get the corresponding predicted load situation.

[0083] According to the dynamic programming method, the algorithm continuously deduces the results and trains to obtain a result dataset. According to the device operation parameters, such as the data of intelligent metering monitoring implementation, the power of the unit's basic settings, and the supply and return water temperatures, determine the result set. Loop in reverse time to obtain the external temperature for each time period, predict the load. First, collect historical load data, meteorological data such as external temperature, and date information. These data will serve as the basis for model training. Extract time features such as hour, day, week, etc., and meteorological features such as external temperature. These features are crucial for predicting the load.

[0084] Specifically, the performance curves of the air-conditioning equipment include: the machine operation performance curves with equipment performance parameters existing at the factory for each piece of equipment, and the performance curves obtained through the intelligent metering and monitoring of equipment operation parameters to implement data.

[0085] It should be noted that under the corresponding time conditions, the optimal allocation of equipment is based on historical experience data to obtain the corresponding cooling water temperature range, and at the same time, the performance curves of the air-conditioning equipment are processed; there are two ways to obtain the equipment performance curves: one is that the performance curves of each piece of equipment existing at the factory usually include the machine operation performance curves under different parameters such as temperature, power, and voltage; the other is that through the intelligent metering and monitoring of equipment operation parameters to implement data, and the power, supply and return water temperatures of the basic settings of the unit, and the historical operation data can obtain the equipment performance curves through algorithms.

[0086] Specifically, based on the data of the training set, using the typical load day fitting method of normal distribution, the annual daily average load prediction curve that meets the typical feature constraints is obtained, and the corresponding load prediction situation is obtained by taking values according to time periods, including the following steps:

[0087] S221. Based on the training set data, using the typical load day fitting method of normal distribution, fit the training set data at the same moment on different load days into a normal distribution curve, and take the expectation of the normal distribution curve as the load value at this moment to fit the typical daily average load curve;

[0088] S222. Eliminate the date influence on the typical daily average load curve of the historical year to obtain the annual daily average load shape factor prediction curve;

[0089] S223. Based on the pre-assumed load typical feature prediction value, combine the annual daily average load shape factor prediction curve with the typical feature prediction value to obtain the annual daily average load prediction curve that meets the typical feature constraints, and take values according to time periods to obtain the corresponding load prediction situation.

[0090] Specifically, based on the training set data, using the typical load day fitting method of normal distribution, fitting the training set data at the same moment on different load days into a normal distribution curve, and taking the expectation of the normal distribution curve as the load value at this moment to fit the typical daily average load curve includes the following steps:

[0091] S2211. Statistically analyze the training set data at the same moment every day, calculate the frequency of each load value appearing, and obtain the probability distribution of the load at this moment;

[0092] S2212. Use the statistical method to test whether the load values at the same moment on different days follow the normal distribution;

[0093] S2213. For the load data that follows the normal distribution, use the maximum likelihood estimation method to fit it into a normal distribution curve;

[0094] S2214. Based on the normal distribution curve obtained by fitting, set evaluation indicators. According to the set evaluation indicators, select typical load days that meet the conditions from historical load data, and fit to obtain an annual daily average load prediction curve that meets the constraints of typical characteristics.

[0095] Specifically, eliminating the date influence on the typical daily average load curve of historical years to obtain the annual daily average load shape factor prediction curve includes the following steps:

[0096] S2221. Use the periodic smoothing processing method to perform periodic smoothing processing on the daily average load curve of historical years;

[0097] S2222. Use the calendar characteristics of holidays in historical years and the target year to be predicted to generate a dynamic time anchor matrix;

[0098] S2223. Based on the daily average load curve of historical years after periodic smoothing processing, combined with the dynamic time anchor matrix, obtain the annual daily average load shape factor prediction curve.

[0099] It should be noted that through the performance curve of the equipment, the highest power and the highest cooling capacity of the equipment are obtained by continuous combined training of the algorithm under the ideal external temperature; the circulating chilled water temperature obtains the best state of the equipment under the current outdoor temperature condition. Taking the coefficient of performance (COP) as the standard, COP is an index to measure the energy efficiency of an air-conditioning system, which is defined as the ratio of the output power to the input power during refrigeration or heating; the higher the COP, the better the energy efficiency of the air-conditioning system, that is, less electrical energy consumption can generate more cooling or heating; the higher the COP, the less the equipment consumes, and an equipment allocation method that meets the current load is obtained. The best air-conditioning resource allocation plan is obtained at each chilled water temperature, mainly calculating the ratio of the cooling capacity (British thermal units per hour BTU / h or kilowatts kw) to the input power; by installing intelligent meters on the equipment, it is returned to the platform in real time, once every minute. The platform calculates and obtains the corresponding values in real time through the reported data; for the allocation of air-conditioning equipment under the obtained chilled water temperature conditions, logical deduction is used, and continuous calculation is carried out through an example algorithm; thus, the best situation is obtained. First, clarify the problem definition and objectives, and select a suitable algorithm framework; second, collect and preprocess relevant data, including feature extraction and data cleaning; then, use the algorithm to model the data and evaluate the model performance through methods such as cross-validation; next, adjust the model parameters according to the performance evaluation results and optimize the model structure; finally, iterate this process until the optimal model configuration is found to achieve the best prediction or decision-making results; this process requires a deep understanding of the algorithm and the ability to flexibly select and adjust the algorithm according to the problem characteristics and data characteristics to achieve the best solution.

[0100] S3. Based on the trained multi-layer feedforward network model, obtain an optimized air-conditioning resource allocation strategy, adjust the allocated operation of the units according to the optimized air-conditioning resource allocation strategy, and dynamically adjust the air-conditioning resource allocation through real-time monitoring of the units.

[0101] It should be noted that the dynamic adjustment of air-conditioning resource allocation is based on the results of algorithm optimization, obtaining information such as the operating load of the allocated machines, the cooling capacity required within a specified time, and the number of machines to be turned on for several days. The relevant information is used to control the controlled devices through network communication. The operating parameters required by the devices obtained by the algorithm are sent to the relevant machines; and through real-time monitoring of the machine operation and the on-site environment, dynamic adjustments are made.

[0102] First, determine the operating load, required cooling capacity, and number of days of machine startup; then, send these parameters to the control device through network communication so that the machine can operate automatically according to the preset parameters; at the same time, through real-time monitoring of the machine operation status and on-site environment changes, dynamically adjust the operating parameters to optimize performance and efficiency; this process involves multiple steps such as data collection, algorithm calculation, network communication control, and real-time monitoring and adjustment to ensure that the machine operates in the best state; for example, when it is predicted that the energy demand will increase, the system can increase the energy supply in advance to avoid energy shortages; on the contrary, when it is predicted that the energy demand will decrease, the system can reduce the supply to avoid waste.

[0103] The present invention aims to solve the problem of low operating efficiency caused by uneven resource allocation among units; the resource allocation algorithm can be implemented through, for example, Java programming, call third-party libraries to improve the calculation efficiency, and write the main body of the algorithm, adopting various algorithmic thinking (such as greedy algorithm, dynamic programming, simulated annealing, etc.) to optimize the resource allocation scheme; the algorithm dynamically adjusts the resource allocation strategy according to the real-time operation data of the units to ensure the efficient collaborative work of each unit.

[0104] The technical features include the design of the main body of the algorithm, the call of third-party libraries, the integration of multiple algorithms, and real-time dynamic adjustment; after implementation, it significantly improves the operating efficiency of the units, reduces the energy consumption cost, and realizes the optimal allocation of resources.

[0105] Specific steps: Clearly define the resource allocation problem to be solved, including the types of units, the types of resources, and the goals of efficiency and cost; Design the overall architecture of the system, including modules such as data collection, processing, algorithm execution, and result output; Develop or integrate a data collection system to collect the operation data of the units in real time; Clean, format, and standardize the collected data for easy algorithm processing; Design an algorithm framework according to the characteristics of the problem, and determine the input, output, and basic logic of the algorithm; For the short-term resource allocation problem, implement a greedy algorithm to quickly obtain an approximate optimal solution. The greedy algorithm mainly trains on the training set model, assuming that only one host runs and undertakes the main operation tasks; For problems with phased characteristics, apply the dynamic programming algorithm to find the global optimal solution; According to external factors such as changes in external temperature, weather changes, and electricity price changes, dynamically allocate resources; K-Fold Cross-Validation divides the dataset into K equal (or almost equal) subsets. Each time, one subset is left as the test set, and the other K-1 subsets are used as the training set. This process is repeated K times; BP neural network: The BP neural network is a powerful machine learning model that can learn and simulate complex non-linear relationships.

[0106] Third-party library calls: Call a mathematical calculation library, such as Apache Commons Math, to support complex mathematical operations; Use a parallel processing library, such as the Java parallel streams (Streams API), to improve the computing efficiency; Integrate a machine learning library, such as Weka or Deeplearning4j, to achieve data-driven intelligent resource allocation.

[0107] Real-time dynamic adjustment: Real-time monitoring, monitor the operation status of the units and the resource usage in real time; Dynamic adjustment strategy, dynamically adjust the resource allocation strategy according to the real-time data to adapt to environmental changes and improve efficiency.

[0108] Testing and optimization: Unit testing, conduct unit testing on each module of the algorithm to ensure the correctness of each module; Performance testing, conduct performance testing to evaluate the efficiency and response time of the algorithm; Optimization and adjustment, optimize and adjust the algorithm according to the test results to improve the overall performance.

[0109] This invention mainly relies on the energy management system. Through the microservices architecture, it can support multiple platforms, breaking through the limitations of different system platforms and being compatible with all platforms; Integrate a set of algorithms, load forecasting, and algorithm libraries into a heating and ventilation intelligent control algorithm, which is suitable for migration and use in various heating and ventilation intelligent control scenarios; Based on the energy management platform, monitor the heating and ventilation to verify the accuracy of the algorithm; Install intelligent meters on the on-site machines to monitor the operation data in real time. The platform can view the operation status in real time through data visualization to verify the correctness of the algorithm.

[0110] In summary, historical energy load data is an important basis for algorithm optimization. By analyzing the past operation data of the HVAC system, the energy demand patterns under different seasons, time periods, and climate conditions can be understood. For example, by analyzing the heating data in winter over the past few years, it can be found that the lower the temperature, the greater the heating demand and the corresponding increase in energy consumption. Based on these historical data, the algorithm can establish a prediction model to predict future energy demand in advance, thereby reasonably arranging energy supply. At the same time, it can also help identify problems and potential energy-saving spaces in the operation of the HVAC system. For example, by analyzing the relationship between energy consumption and parameters such as indoor temperature and humidity, the links of energy waste can be found, and corresponding improvement measures can be taken.

[0111] Temperature is one of the key factors affecting the energy distribution and control of the HVAC system. Under different indoor and outdoor temperature conditions, the energy demand of the HVAC system varies greatly. In winter, when the outdoor temperature is low, more heat needs to be provided to maintain the indoor temperature. In summer, when the outdoor temperature is high, more energy needs to be consumed for cooling. By real-time monitoring the changes in indoor and outdoor temperature and incorporating the temperature data into the algorithm, the operation of the HVAC system can be controlled more precisely. For example, when the outdoor temperature rises, the algorithm can automatically adjust the operation parameters of the cooling equipment to improve the cooling efficiency and reduce energy consumption at the same time. In addition, the temperature data can be combined with other factors, such as historical data and electricity price data, to further optimize the energy distribution strategy.

[0112] Electricity price is also an important factor affecting the energy distribution and control of the HVAC system. The electricity price varies greatly at different time periods. Reasonably utilizing the electricity price fluctuations can reduce the operation cost of the HVAC system. The algorithm can automatically adjust the operation strategy according to the real-time electricity price information, combined with the energy demand and performance characteristics of the HVAC system. For example, during the low electricity price period, the algorithm can start the heat storage equipment to store heat or cold for use during the high electricity price period, thereby reducing the energy cost. At the same time, the electricity price data can also be used to evaluate the economy of different energy supply schemes and provide a basis for selecting the optimal energy distribution scheme. Performance data is also crucial for achieving precise energy distribution and control of the HVAC system. Different HVAC equipment and systems have different performance characteristics, such as cooling efficiency, heating efficiency, response time, etc. By analyzing these performance data, the algorithm can select the optimal equipment combination and operation parameters to achieve efficient utilization of energy.

[0113] Data training is carried out through price optimization and energy optimization, and finally the optimal control scheme is obtained; specifically, the implementation is based on the calculation of the working costs of machines with different powers at different stages of electricity prices. For example, when the electricity price is 0.5 yuan and the machine power is 800 kw, how to calculate the working cost of the total price, and when the electricity price is 0.5 yuan and the machine power is 850 kw, how to calculate the working cost of the total price. By continuously trying combinations to judge the total price, the obtained results are used to train the set and fit the curve of electricity price and power, so as to determine the relatively optimal control scheme.

[0114] For example, for the HVAC system of a large commercial building, the algorithm can reasonably allocate energy according to the energy demands and equipment performances of different areas to ensure that each area can obtain a comfortable indoor environment while minimizing energy consumption; in addition, the performance data can also be used to monitor and diagnose the operating status of HVAC equipment, timely detect faults and problems, and ensure the stable operation of the system.

[0115] In order to achieve more precise HVAC energy allocation and control, the algorithm needs to be continuously optimized; first of all, a more accurate prediction model should be established, fully considering the influence of various factors to improve the accuracy of energy demand prediction; secondly, advanced optimization algorithms such as genetic algorithms and particle swarm optimization algorithms should be adopted to find the optimal energy allocation scheme and operating parameters; in addition, the real-time performance and adaptability of the algorithm need to be strengthened to be able to quickly respond to various changes and emergencies.

[0116] The algorithm is split into independent services by using the microservices architecture and interacts through lightweight communication mechanisms (such as RESTful API or message queues); frameworks such as Spring Boot can be used to quickly build microservices; for example, create a Spring Boot project, implement the algorithm as a service, and expose the RESTful API; other services can call this algorithm service through HTTP requests; the advantages are high flexibility, easy to expand and maintain; each microservice can be independently deployed and upgraded without affecting other services; the disadvantage is that it increases the complexity of the system and the communication and coordination problems between services need to be handled.

[0117] Furthermore, preprocess the obtained historical comprehensive data, and then it also includes obtaining screened historical comprehensive data. The process of obtaining the screened historical comprehensive data is as follows: Obtain preprocessed historical comprehensive data by preprocessing the historical comprehensive data. The preprocessing includes removing abnormal missing historical comprehensive data and standardization processing. The preprocessed historical comprehensive data includes temperature change amount, humidity change amount, average device power, device usage duration, air flow rate, and device heat. The temperature change amount represents the absolute value of the difference between the indoor temperature and the outdoor temperature. The humidity change amount represents the absolute value of the difference between the indoor humidity and the outdoor humidity. The device heat represents the heat generated by the device itself. Combine the temperature change amount, humidity change amount, and the preset maximum temperature change value and preset maximum humidity change value obtained from the database to obtain a heat change evaluation index. Combine the average device power, device usage duration, heat change evaluation index within the reference compliance change range, and the preset maximum power consumption value obtained from the database to obtain a power consumption evaluation index. Combine the air flow rate, device heat, heat change evaluation index within the reference compliance change range, and the specific heat capacity of air, preset heat load value, and preset cooling load value obtained from the database to obtain a load evaluation index. Screen the preprocessed historical comprehensive data based on the power consumption evaluation index and the load evaluation index to obtain the screened historical comprehensive data. The screened historical comprehensive data represents the preprocessed historical comprehensive data corresponding to the power consumption evaluation index within the reference power consumption range and the load evaluation index within the reference load range. The heat change evaluation index is used to evaluate the degree of heat change. The power consumption evaluation index is used to evaluate the power consumption of the preset device. The load evaluation index is used to reflect the load condition of the preset device.

[0118] In this embodiment, determine whether the heat change evaluation index is within the reference compliance change range obtained from the database, and obtain the heat change evaluation index within the reference compliance change range. Determine whether the power consumption evaluation index is within the reference power consumption range obtained from the database. Determine whether the load evaluation index is within the reference load range obtained from the database. Obtain the preprocessed historical comprehensive data with the power consumption evaluation index within the reference power consumption range and the load evaluation index within the reference load range. When the power consumption evaluation index is not within the reference power consumption range or the load evaluation index is not within the reference load range, send a prompt to the preset personnel for power transmission through superconducting materials (such as yttrium barium copper oxide, bismuth strontium calcium copper oxide, etc.). Obtain the screened historical comprehensive data that meets the requirements, which can more accurately reflect the load demand and change trend of the system, and realizes the improvement of the accuracy of extracting the load characteristics of the integrated energy system.

[0119] It should be noted that the aforementioned database is a database established before the design of the air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy provided by the embodiments of the present application for storing various setting data. The database includes but is not limited to power consumption, temperature change amount, humidity change amount, etc. Various values therein are directly set by technicians. For example, the preset maximum temperature change is represented by the maximum value of the temperature change amount in the historical time period in the database, the preset maximum humidity change is represented by the maximum value of the humidity change amount in the historical time period in the database, the preset maximum power consumption is represented by the maximum value of the power consumption amount in the historical time period in the database, the preset heat load value is represented by the average value of the heat load in the historical time period in the database, the preset cooling load value is represented by the average value of the cooling load in the historical time period in the database, the upper limit of the reference compliance change range is represented by the maximum value of the heat change evaluation index in the historical time period in the database, the lower limit of the reference compliance change range is represented by the minimum value of the heat change evaluation index in the historical time period in the database, the upper limit of the reference power consumption range is represented by the maximum value of the power consumption evaluation index in the historical time period in the database, the lower limit of the reference power consumption range is represented by the minimum value of the power consumption evaluation index in the historical time period in the database, the upper limit of the reference load range is represented by the maximum value of the load evaluation index in the historical time period in the database, and the lower limit of the reference load range is represented by the minimum value of the load evaluation index in the historical time period in the database.

[0120] Further, the specific process for obtaining the heat change evaluation index is as follows: The temperature change compliance value (i.e., WDF in the heat change evaluation index limit expression) is obtained by performing a ratio operation on the sum of the temperature change amount and the preset maximum temperature change and twice the preset maximum temperature change. w ) The temperature change compliance value is used to reflect the compliance situation of the temperature change; The humidity change compliance value (i.e., SDF in the heat change evaluation index limit expression) is obtained by performing a ratio operation on the sum of the humidity change amount and the preset maximum humidity change and twice the preset maximum humidity change. w ) The humidity change compliance value is used to reflect the compliance situation of the humidity change; The heat change evaluation index is obtained by combining the temperature change compliance value, the humidity change compliance value, and the first heat influence weight and the second heat influence weight obtained from the database; The first heat influence weight is used to evaluate the influence of the temperature change on the heat change evaluation index; The second heat influence weight is used to evaluate the influence of the humidity change on the heat change evaluation index.

[0121] Among them, the heat change evaluation index is obtained by the following method:

[0122]

[0123]

[0124]

[0125] Wherein, represents the heat change evaluation index of the preset device in the w-th preset historical time period, w = 1, 2,..., r, w represents the number of the preset historical time period, r represents the total number of the preset historical time periods, WDF w represents the temperature change compliance value of the preset device in the w-th preset historical time period, SDF w represents the humidity change compliance value of the preset device in the w-th preset historical time period, |ΔT w | represents the temperature change amount of the preset device in the w-th preset historical time period, |ΔSD w | represents the humidity change amount of the preset device in the w-th preset historical time period, ΔT0 represents the maximum preset temperature change, ΔSD0 represents the maximum preset humidity change, μ1 represents the first influence weight of heat, and μ2 represents the second influence weight of heat.

[0126] In this embodiment, the algorithm of this embodiment combines and comprehensively analyzes the preprocessed historical comprehensive data to obtain the heat change evaluation index. The preprocessed historical comprehensive data in the algorithm of this embodiment does not exist independently and has mutual relevance. The change in temperature will affect the saturation vapor pressure of water vapor in the air, thereby affecting the relative humidity. When the temperature rises, the saturation vapor pressure of water vapor in the air increases, which means that more water vapor can be contained in the air without reaching the saturation state, resulting in a decrease in relative humidity, and further leading to a decrease in the heat change evaluation index. The change in humidity will also affect the heat capacity and thermal conductivity of the air, thereby affecting the change speed and amplitude of the temperature. The increase in humidity will lead to an increase in the water vapor content in the air, thereby slowing down the rising speed of the temperature. The parameters of the algorithm of this embodiment need to jointly consider the influence on the result.

[0127] It should be added that in this embodiment, μ1 and μ2 respectively represent the first influence weight of heat and the second influence weight of heat. μ1 and μ2 are respectively the weights corresponding to the temperature change difference and the humidity change difference in the database. μ1 and μ2 respectively describe the numerical values corresponding to the influence degree of the temperature change difference and the humidity change difference on the heat change evaluation index, and can be directly obtained from the database during use. Their corresponding relationship is preset. For example, the temperature change difference and the humidity change difference form a mapping set with the weights corresponding to the preset temperature change difference and humidity change difference in the database. The real-time temperature change difference and humidity change difference are input into the mapping set to obtain the weights corresponding to the temperature change difference and the humidity change difference. The mapping relationship therein can be a one-to-one or many-to-one relationship, and the value range in this embodiment is 0-1.

[0128] Specifically, assume that the temperature change compliance value WDF wranges from 0.6 to 1, and the humidity change compliance value is SDF w ranges from 0.6 to 1, the first heat influence weight μ1 is fixed at 0.5, and the second heat influence weight μ2 is fixed at 0.5. As shown in Table 1, it is the change statistical table of the heat change evaluation index provided by the embodiment of the present application:

[0129] Table 1 Change statistical table of heat change evaluation index

[0130]

[0131] As can be seen from the above table, as the temperature change compliance value and the humidity change compliance value gradually increase, the heat change evaluation index gradually increases, which means that the degree of heat change gradually increases, realizing the accurate quantification of the degree of heat change, and further realizing the improvement of the accuracy of extracting the load characteristics of the integrated energy system.

[0132] Furthermore, the specific acquisition process of the power consumption evaluation index is as follows: The power consumption compliance value (i.e., XH in the power consumption evaluation index limit expression) is obtained by performing a ratio operation on the sum of the initial power consumption value and the preset maximum power consumption value and twice the preset maximum power consumption value w ); The power consumption compliance value is used to reflect the compliance situation of power consumption within the preset historical time period; The initial power consumption value is represented by the result of multiplying the average power of the device and the device usage duration; The power consumption evaluation index is obtained by performing an operation combining the power consumption compliance value and the heat change evaluation index within the reference compliance change range.

[0133] Among them, the power consumption evaluation index is obtained by the following method:

[0134]

[0135]

[0136] In the formula, represents the power consumption evaluation index of the preset device in the w-th preset historical time period, w = 1, 2,..., r, w represents the number of the preset historical time period, r represents the total number of the preset historical time periods, XH w represents the power consumption compliance value of the preset device in the w-th preset historical time period, represents the heat change evaluation index of the preset device in the w-th preset historical time period within the reference compliance change range, P w represents the average power of the preset device in the w-th preset historical time period, t w represents the device usage duration of the preset device in the w-th preset historical time period, E0 represents the preset maximum power consumption, and e represents the natural constant.

[0137] In this embodiment, the algorithm of this embodiment comprehensively analyzes the preprocessed historical comprehensive data and the heat change evaluation index to obtain the power consumption evaluation index. In the algorithm of this embodiment, the preprocessed historical comprehensive data and the heat change evaluation index do not exist independently and are interrelated. The average power and usage duration of the device jointly determine the total amount of electrical energy consumed by the device. The higher the power and the longer the usage duration, the greater the total power consumption. The longer the device is used, the more heat is generated. The longer the time runs, it may cause the device temperature to rise, which in turn affects the power consumption. The higher the device power, the more heat is generated during operation, and the higher the requirement for the heat dissipation system. If the heat dissipation is poor and the device temperature rises, it may lead to performance degradation and increased energy consumption, and then cause the power consumption evaluation index to decrease. The parameters of the algorithm in this embodiment need to jointly consider the impact on the results; it realizes the accurate quantification of the power consumption of the preset device, and further realizes the improvement of the accuracy of extracting the load characteristics of the integrated energy system.

[0138] In a microgrid, by accurately quantifying the power consumption situation, energy resources can be more reasonably allocated. For example, in the case of energy shortage, the power consumption needs of key devices can be prioritized to ensure the stability and reliability of the power system.

[0139] Further, the specific process of obtaining the load evaluation index is as follows: Multiply the air flow rate, the specific heat capacity of air, and the temperature change amount to obtain the initial heat load value; Divide the sum of the initial heat load value and the preset heat load value by twice the preset heat load value to obtain the heat load compliance value (i.e., RFH in the load evaluation index limit expression W ); The heat load compliance value is used to reflect the heat load situation of the preset device within the preset historical time period; Add the initial heat load value and the device heat to obtain the initial cold load value; Divide the sum of the initial cold load value and the preset cold load value by twice the preset cold load value to obtain the cold load compliance value (i.e., LFH in the load evaluation index limit expression W ); The cold load compliance value is used to reflect the cold load situation of the preset device within the preset historical time period; Combine the heat load compliance value, the cold load compliance value, and the heat change evaluation index within the reference compliance change range to obtain the load evaluation index.

[0140] Among them, the load evaluation index is obtained through the following method:

[0141]

[0142]

[0143]

[0144] In the formula, Denote the load assessment index of the preset device in the w-th preset historical time period, where w represents the number of the preset historical time period, r represents the total number of the preset historical time periods, RFH W Denote the heat load compliance value of the preset device in the w-th preset historical time period, LFH W Denote the cold load compliance value of the preset device in the w-th preset historical time period, Denote the heat change assessment index within the reference compliance change range of the preset device in the w-th preset historical time period, M w Denote the air flow rate of the preset device in the w-th preset historical time period, |ΔT w | Denote the temperature change amount of the preset device in the w-th preset historical time period, Q w Denote the device heat of the preset device in the w-th preset historical time period, c0 represents the specific heat capacity of air, RFH0 represents the preset heat load value, LFH0 represents the preset cold load value, and e represents the natural constant.

[0145] In this embodiment, the algorithm of this embodiment combines the preprocessed historical comprehensive data and the heat change assessment index for comprehensive analysis to obtain the load assessment index. The preprocessed historical comprehensive data and the heat change assessment index in the algorithm of this embodiment do not exist independently and are interrelated. The magnitude of the air flow rate directly affects the heat dissipation effect of the device. When the air flow rate increases, the device can more effectively dissipate the heat generated inside to the external environment. When the air flow rate decreases, the heat dissipation effect of the device will weaken. The higher the device heat, the faster the change rate of the device heat, which in turn usually leads to a larger heat change assessment index. When the air flow rate increases, the convective heat transfer coefficient increases, the heat transfer rate accelerates, and the heat change assessment index will also increase accordingly, which in turn leads to an increase in the load assessment index. The parameters of the algorithm of this embodiment need to jointly consider the influence on the result; the accurate quantification of the load condition of the preset device is realized, and then the accuracy of extracting the load characteristics of the integrated energy system is improved.

[0146] It should be noted that, Figure 2 the line where point a in Figure 3 is connected to the line where point a in Figure 3 the line where point b in Figure 4 is connected to the line where point b in

[0147] In summary, by means of the above technical solutions of the present invention, through algorithm optimization and comprehensive consideration of factors such as historical data, temperature, electricity price, and performance, the present invention can achieve more precise energy distribution and control of the HVAC system; this not only helps to reduce energy consumption, reduce the impact on the environment, but also saves energy costs for users, improves economic benefits, and improves the implementation of energy consumption efficiency. At the same time, it also avoids the overuse of energy and pollution emissions, achieving the goal of saving resources and reducing pollution; by analyzing the past operation data of the HVAC system, the energy demand patterns in different seasons, different time periods, and different climate conditions can be understood; based on these historical data, the algorithm can establish a prediction model to predict future energy demand in advance, so as to reasonably arrange energy supply; at the same time, historical data can also help to discover problems and potential energy-saving spaces in the operation of the HVAC system; by real-time monitoring of the indoor and outdoor temperature changes and incorporating the temperature data into the algorithm, the operation of the HVAC system can be controlled more precisely. In addition, the temperature data can be combined with other factors, such as historical data and electricity price data, to further optimize the energy distribution strategy; according to the real-time electricity price information, combined with the energy demand and performance characteristics of the HVAC system, the operation strategy is automatically adjusted; through the analysis of performance data, the algorithm can select the optimal equipment combination and operation parameters to achieve efficient utilization of energy, reasonably distribute energy, ensure that each area can obtain a comfortable indoor environment, and at the same time minimize energy consumption. In addition, performance data can also be used to monitor and diagnose the operation status of HVAC equipment, timely discover faults and problems, and ensure the stable operation of the system.

[0148] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An air conditioning usage algorithm and energy consumption efficiency optimization method based on a microgrid strategy, characterized in that The optimization method includes the following steps: S1. Obtain the historical comprehensive data of the energy system, preprocess the obtained historical comprehensive data, and extract the load characteristics of the integrated energy system based on the correlation analysis method to obtain the load characteristic data of the integrated energy system; S2. Use the load characteristic data of the integrated energy system as the input of the multi-layer feedforward network model, train the multi-layer feedforward network model, and obtain an optimized air-conditioning resource allocation plan; S3. Based on the trained multi-layer feedforward network model, obtain an optimized air-conditioning resource allocation strategy, adjust the distribution operation of the units according to the optimized air-conditioning resource allocation strategy, and dynamically adjust the air-conditioning resource allocation through real-time monitoring of the units.

2. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 1, characterized in that The steps of obtaining the historical comprehensive data of the energy system, preprocessing the obtained historical comprehensive data, and extracting the load characteristics of the integrated energy system based on the correlation analysis method to obtain the load characteristic data of the integrated energy system include the following steps: S11. Obtain the historical comprehensive data of the energy system, screen the obtained historical comprehensive data, remove abnormal and missing historical comprehensive data, and perform standardization processing on the screened historical comprehensive data to obtain the standardized historical comprehensive data; S12. Based on the correlation analysis method, calculate the Pearson correlation coefficient of the historical comprehensive data in combination with the standardized historical comprehensive data; S13. Select the historical comprehensive data with a higher Pearson correlation coefficient and input it into the convolutional neural network-long short-term memory network prediction model for feature extraction to obtain the load characteristics of the integrated energy system.

3. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 2, characterized in that The historical comprehensive data of the energy system includes: historical energy load data, weather data, and date data.

4. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 2, wherein The steps of selecting the historical comprehensive data with a higher Pearson correlation coefficient and inputting it into the convolutional neural network-long short-term memory network prediction model for feature extraction to obtain the load characteristics of the integrated energy system include the following steps: S131. Select the historical comprehensive data with a higher Pearson correlation coefficient, and extract the local load characteristics of the integrated energy system through the convolutional neural network model; S132. Based on the long short-term memory network model, combine the extracted local load characteristics of the integrated energy system to extract time series characteristics; S133. Integrate the extracted local load characteristics of the integrated energy system and the time series characteristics into the load characteristics of the integrated energy system.

5. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 1, characterized in that, The steps of using the load characteristic data of the integrated energy system as the input of the multi-layer feedforward network model, training the multi-layer feedforward network model, and obtaining an optimized air-conditioning resource allocation plan include the following steps: S21. Divide the load characteristic data of the integrated energy system into a training set, a validation set, and a test set; S22. Based on the data of the training set, use the typical load day fitting method of the normal distribution to obtain the annual daily average load prediction curve that meets the typical characteristic constraints, and obtain the corresponding load prediction situation according to the time period; S23. According to the training set data, combined with the load prediction situation, obtain the corresponding cooling water temperature range, and based on the performance curve of the air-conditioning equipment, obtain the maximum power and maximum cooling capacity of the equipment. S24. Based on the performance of the equipment and the range of cooling water temperature, use the logical deduction method to loop backwards in time to obtain the optimized state under the corresponding time period, and combine with the electricity price factor to obtain an optimized air-conditioning resource allocation plan.

6. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 5, characterized in that, The performance curve of the air-conditioning equipment includes: the machine operation performance curve with the equipment performance parameters existing at the factory for each piece of equipment and the performance curve obtained through the intelligent metering and monitoring of equipment operation parameters to implement data.

7. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 5, characterized in that For the data based on the training set, using the typical load day fitting method of normal distribution, to obtain the annual average daily load prediction curve that meets the typical feature constraints, and taking values according to time periods, the corresponding load prediction situation includes the following steps: S221. Based on the training set data, use the typical load day fitting method of normal distribution to fit the training set data at the same moment on different load days into a normal distribution curve, and take the expectation of the normal distribution curve as the load value at this moment to fit the typical daily average load curve; S222. Eliminate the date influence on the typical daily average load curve of historical years to obtain the annual average daily load shape factor prediction curve; S223. Based on the pre-assumed load typical feature prediction value, combine the annual average daily load shape factor prediction curve with the typical feature prediction value to obtain the annual average daily load prediction curve that meets the typical feature constraints, and take values according to time periods to obtain the corresponding load prediction situation.

8. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 5, characterized in that, For the data based on the training set, using the typical load day fitting method of normal distribution to fit the training set data at the same moment on different load days into a normal distribution curve, and take the expectation of the normal distribution curve as the load value at this moment to fit the typical daily average load curve includes the following steps: S2211. Statistically analyze the training set data at the same moment every day, calculate the frequency of each load value appearing, and obtain the probability distribution of the load at this moment; S2212. Use the statistical method to test whether the load values at the same moment on different days follow a normal distribution; S2213. For the load data that follows a normal distribution, use the maximum likelihood estimation method to fit it into a normal distribution curve; S2214. According to the fitted normal distribution curve, set evaluation indicators, and select typical load days that meet the conditions from the historical load data according to the set evaluation indicators to fit the annual average daily load prediction curve that meets the typical feature constraints.

9. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 7, characterized in that Eliminating the date influence on the typical daily average load curve of historical years to obtain the annual average daily load shape factor prediction curve includes the following steps: S2221. Use the periodic smoothing processing method to perform periodic smoothing processing on the daily average load curve of historical years; S2222. Use the calendar features of holidays in historical years and the target year to be predicted to generate a dynamic time anchor point matrix; S2223. Based on the daily average load curve of historical years after periodic smoothing processing, combined with the dynamic time anchor point matrix, obtain the annual average daily load shape factor prediction curve.

10. The air-conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 1, wherein After preprocessing the obtained historical comprehensive data, it also includes obtaining screened historical comprehensive data; The process of obtaining the screened historical comprehensive data is as follows: Obtain the preprocessed historical comprehensive data by preprocessing the historical comprehensive data; Obtain a heat change evaluation index by combining the temperature change amount, the humidity change amount, the preset maximum temperature change value and the preset maximum humidity change value obtained from the database; Obtain an electricity consumption evaluation index by combining the average power of the device, the device usage duration, the heat change evaluation index within the reference compliance change range, and the preset maximum electricity consumption value obtained from the database; Obtain a load evaluation index by combining the air flow rate, the device heat, the heat change evaluation index within the reference compliance change range, and the specific heat capacity of air, the preset heat load value and the preset cooling load value obtained from the database; Screen the preprocessed historical comprehensive data based on the electricity consumption evaluation index and the load evaluation index to obtain the screened historical comprehensive data; The screened historical comprehensive data represents the preprocessed historical comprehensive data corresponding to the electricity consumption evaluation index within the reference electricity consumption range and the load evaluation index within the reference load range; The heat change evaluation index is used to evaluate the degree of heat change; The electricity consumption evaluation index is used to evaluate the electricity consumption of the preset device; The load evaluation index is used to reflect the load condition of the preset device; The preprocessed historical comprehensive data includes the temperature change amount, the humidity change amount, the average power of the device, the device usage duration, the air flow rate and the device heat; 11. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 10, characterized in that, The specific process of obtaining the heat change evaluation index is as follows: Obtain a temperature change compliance value through a ratio operation of the sum of the temperature change amount and the preset maximum temperature change value to twice the preset maximum temperature change value; Obtain a humidity change compliance value through a ratio operation of the sum of the humidity change amount and the preset maximum humidity change value to twice the preset maximum humidity change value; Obtain a heat change evaluation index by combining the temperature change compliance value, the humidity change compliance value, and the first heat influence weight and the second heat influence weight obtained from the database; 12. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 11, characterized in that, The specific process of obtaining the electricity consumption evaluation index is as follows: Obtain an electricity consumption compliance value through a ratio operation of the sum of the initial electricity consumption value and the preset maximum electricity consumption value to twice the preset maximum electricity consumption value; The initial electricity consumption value is represented by the result of multiplying the average power of the device and the device usage duration; Obtain an electricity consumption evaluation index through an operation combining the electricity consumption compliance value and the heat change evaluation index within the reference compliance change range; 13. The air conditioning usage algorithm and energy consumption efficiency optimization method based on the microgrid strategy according to claim 12, characterized in that The specific process of obtaining the load evaluation index is as follows: Obtain an initial heat load value through a product operation of the air flow rate, the specific heat capacity of air and the temperature change amount; Obtain a heat load compliance value through a ratio operation of the sum of the initial heat load value and the preset heat load value to twice the preset heat load value; Obtain an initial cooling load value through an addition operation of the initial heat load value and the device heat; Obtain a cooling load compliance value through a ratio operation of the sum of the initial cooling load value and the preset cooling load value to twice the preset cooling load value; Obtain a load evaluation index by combining the heat load compliance value, the cooling load compliance value and the heat change evaluation index within the reference compliance change range;

Images