Smart city energy optimization management method and system based on deep learning

Abstract

The invention discloses a smart city energy optimization management method and system based on deep learning, and relates to the technical field of data management. The smart city energy optimization management method comprises the following steps: collecting heating energy consumption data of a month or a quarter in a city jurisdiction, and combining population density, building types and heating modes to construct a data set to analyze an energy consumption relationship. And training a prediction model by using the data set, learning influence factors and predicting the heating energy demand in a future time period. Model prediction is beneficial to comparison of energy consumption fluctuations among jurisdictions, identification of high and low consumption peaks and valleys, and formulation of an adjustment strategy for a high-energy-consumption area. According to actual climate and building characteristics, heating is dynamically optimized, and heating requirements of a plurality of areas in the jurisdiction are managed and optimized, so that the maximum benefit of energy use is achieved, energy is reasonably distributed according to the requirement result of the jurisdiction, excessive energy consumption is reduced, and scientificity and performability of energy management decision are improved.

Description

Technical Field

[0001] The present invention relates to the field of data management technology, and in particular to a smart city energy optimization management method and system based on deep learning. Background Art

[0002] With the continuous development of smart cities, the number of urban buildings is increasing. Most cities achieve environmentally friendly heating in winter, but the types of heating energy are different and the effects are different. At present, urban heating energy is not effectively managed, resulting in excessive energy consumption and waste. Therefore, a smart city energy management system is designed. However, traditional energy management systems usually have a one-sided view of the energy consumption management required for urban development. It is difficult to implement personalized urban energy management strategies based on the diversity of urban buildings. As a result, while promoting the construction of smart cities, it does not take into account the maximization of energy utilization, and the energy required for multiple buildings in the city is difficult to be reasonably allocated.

[0003] To sum up, how to effectively manage and reasonably allocate the energy required for heating of urban multi-story buildings and improve energy utilization is an urgent problem to be solved and optimized in the smart city energy optimization management system. Summary of the invention

[0004] The present invention provides a smart city energy optimization management method and system based on deep learning, which solves the technical problem of how to effectively manage and reasonably allocate the energy required for heating of urban multi-story buildings and improve energy utilization efficiency.

[0005] In order to solve the above technical problems, the present invention provides a smart city energy optimization management method and system based on deep learning. The specific technical solutions are as follows: In a first aspect, a smart city energy optimization management method based on deep learning comprises the following steps: S100, collecting heating energy consumption data of the urban area within a preset time range; based on the heating energy consumption data, obtaining a dynamic impact data set within the urban area; the dynamic impact data set includes population density data, building type and quantity data, and heating method data; S200, training the acquired dynamic impact data set to construct an energy prediction model to output the prediction result of energy consumption in the jurisdiction within a preset time range; acquiring the next dynamic impact data set in the jurisdiction, inputting it into the energy prediction model, and the energy prediction model outputs the prediction result of energy consumption in the next jurisdiction within the preset time range; S300, based on the energy consumption prediction result, comparing the change fluctuation difference of the dynamic impact data set between multiple jurisdictions to obtain the energy consumption change trend between the multiple jurisdictions; based on the change trend, obtaining the energy consumption proportion of the high energy consumption jurisdiction to set the heating section adjustment strategy; S400, the heating section adjustment strategy dynamically adjusts the optimal allocation of heating energy in the area in combination with the regional climate temperature changes, so as to improve the indoor heating energy efficiency of the building, so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated.

[0006] As a further optimization scheme of the present invention, the acquired dynamic impact data set is trained to construct an energy prediction model to output the energy consumption prediction results of the jurisdiction within a preset time range, including: The main dynamic factors affecting energy consumption are obtained as an initial dynamic impact data set; the dynamic impact data set is preprocessed to obtain a final dynamic impact data set; Encoding the final dynamic impact data set into sequence data, and inputting the sequence data into the energy prediction model, wherein the energy prediction model comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer; intermediate data of each hidden layer is transmitted to the output layer through the layers, and the final output represents the prediction result of energy consumption of the jurisdiction within a preset time range; By comparing the output of the energy forecast model with the actual urban heating energy consumption data, the deviation is analyzed and the deviation factor is obtained.

[0007] As a further optimization scheme of the present invention, by comparing the output of the energy prediction model with the actual urban area heating energy consumption data, analyzing the deviation and obtaining the deviation factor, including: Comparing the predicted energy consumption data output by the energy prediction model with the actually obtained heating energy consumption data of the urban area to obtain a comparison result, and based on the comparison result, obtaining a deviation between the energy prediction result and the actually obtained energy consumption data; The deviation result is analyzed to obtain the deviation factor; the deviation factor is combined with the final dynamic impact data set obtained originally and input into the energy prediction model again for training to output the prediction result representing the final heating energy.

[0008] As a further optimization scheme of the present invention, an analysis is performed based on the deviation result to obtain a deviation factor, including: pass , to obtain the deviation value between the energy prediction result and the actual energy consumption data; where e represents the deviation value; y true Indicates the actual energy consumption data obtained; y pred represents the energy forecast result; If e=0, there is no bias; if e>0 or e<0, there is bias; based on the biased structure, through ; To analyze the correlation between the dynamic impact data set and the deviation value; where D m represents the correlation coefficient; represents the influence parameter of the dynamic impact data set; Contains data on energy transmission and consumption, types of energy used, and energy usage behavior; Determine the influence of each data item in the dynamic impact data set on the deviation according to the correlation coefficient; m When the correlation coefficient D m The larger the value, the higher the impact on the deviation, so as to obtain the deviation factor in the dynamic impact data set; the deviation factor indicates the influence on the deviation. ; Thus, the deviation factor includes energy delivery consumption data and energy usage behavior data; The extracted deviation factors are added to the dynamic impact dataset to form a new dynamic impact dataset ; In the formula, X' represents the new dynamic impact data set; X represents the dynamic impact data set; Represents the deviation factor.

[0009] As a further optimization scheme of the present invention, based on the energy consumption prediction result, the change fluctuation difference of the dynamic impact data set between multiple jurisdictions is compared to obtain the energy consumption change trend between the multiple jurisdictions, including: pass ; To obtain the energy consumption change rate of each jurisdiction in period t; where, It represents the dynamic change rate of energy consumption in the tth period; represents the energy consumption in period t; represents the energy consumption in the period before t; pass ; To obtain the fluctuation difference between jurisdictions; represents the dynamic change rate of the i-th jurisdiction in the t-th period; represents the dynamic change rate of the j jurisdiction in the tth period; It represents the dynamic fluctuation difference between the i-prefecture and the j-prefecture in the t-th period; when is a positive number; the energy consumption trend between jurisdictions is in a downward state; when If it is a negative number, the energy consumption change trend between the jurisdictions is in an upward state; thus, the energy consumption change trend between multiple jurisdictions is obtained; According to the dynamic fluctuation difference between the jurisdictions , to obtain the energy consumption fluctuation rate between jurisdictions; where, represents volatility; represents the rate of change of total energy consumption in period t; represents the energy change in the period before t; The default value R represents the volatility threshold. ; then it is determined that the jurisdiction i is stable in period t; when ; then it is determined that the jurisdiction i exhibits a violent fluctuation type in period t.

[0010] As a further optimization scheme of the present invention, based on the change trend, the energy consumption proportion of the high energy consumption area is obtained, including: pass Obtain the energy consumption of each data item of the dynamic impact data set of the violent fluctuation type jurisdiction in the t period; mark the highest fluctuation point of the energy consumption of each data item as the first marking point; mark the lowest fluctuation point of the energy consumption as the second marking point; Perform energy demand analysis according to the first marking point and the second marking point to obtain analysis results; based on the analysis results, , to obtain the energy consumption proportion of each data item; where PI represents the energy consumption proportion of the data item; Indicates the energy consumption of each data item that dynamically affects the data set; Represents the total energy consumption in a jurisdiction with severe fluctuations.

[0011] As a further optimization solution of the present invention, the heating section adjustment strategy includes: According to the energy consumption proportion of the high energy consumption area, a heating section adjustment strategy is set; the heating section adjustment strategy obtains the outdoor climate temperature in real time, and heats by gas or electricity in the early stage of heating according to the dynamic impact data set, and the average temperature is controlled within the range of 23.5±1.5℃; Then convert the gas or electric heating mode into hot water or steam heating; , to adjust the outlet temperature of hot water boiler or steam boiler; In the formula, Indicates the outlet temperature of a hot water boiler or steam boiler; represents the outdoor temperature, and ; represents building type data; n represents the number of buildings; the building type data includes building volume ratio, space volume, usage quantity and heat load performance data; Represents population density data; according to , to adjust the return water temperature of the hot water boiler or steam boiler; by dynamically adjusting the hot water temperature in real time, to avoid excessive energy consumption caused by overheating.

[0012] As a further optimization scheme of the present invention, the heating section adjustment strategy combines the regional climate temperature changes to dynamically adjust the optimal allocation of heating energy in the jurisdiction, including: Based on real-time meteorological data and historical heating data, a linear relationship between climate temperature and regional heat load demand is established; and heating sections are divided according to building density, thermal insulation performance and heating network layout in the region; Each zone is adjusted according to the heating zone regulation strategy. ; Dynamically adjust the optimal allocation of heating energy in the jurisdiction, where E i represents the heating energy demand in the i area; c i represents the heating energy cost in the i-th district; P i represents the heat network loss in area i; p s Represents the loss weight coefficient; n represents the number of dynamically changing data items.

[0013] As a further optimization scheme of the present invention, the energy efficiency of indoor heating in buildings is improved so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated, including: The building interior is divided into zones, , to adjust the zone heating temperature; where T set Indicates the target indoor temperature; P density represents the population density weight; E building represents the building heat loss coefficient; A, B and C represent adjustment factors; the adjustment factors include outdoor temperature, indoor space volume and crowd category; Maintain the temperature of the main activity area at 23.5±1.5℃, and the temperature of the secondary area is allowed to be 17.5±1.5℃; control the water pump power at 0.35-0.45W / m 2 , in extremely cold weather T outdoor <-20℃ enables high flow operation, but the power is strictly limited to 0.6W / m 2 Below, to meet the basic heating needs.

[0014] In a second aspect, the system provides an electronic device including a memory, a processor, and a smart city energy optimization management method program based on deep learning stored in the memory and executable on the processor, wherein the smart city energy optimization management method program based on deep learning implements the steps of the smart city energy optimization management method based on deep learning when executed by the processor, and the system includes: Data collection module 201: used to collect heating energy consumption data of the urban area within a preset time range; based on the heating energy consumption data, to obtain a dynamic impact data set within the urban area; the dynamic impact data set includes population density data, building type and quantity data, and heating mode data; Data processing module 202: It is used to train the acquired dynamic impact data set to construct an energy prediction model to output the prediction result of energy consumption in the jurisdiction within a preset time range; obtain the next dynamic impact data set in the jurisdiction and input it into the energy prediction model, and the energy prediction model outputs the prediction result of energy consumption in the next jurisdiction within the preset time range; Data management module 203: used to compare the change fluctuation difference of the dynamic impact data set between multiple jurisdictions based on the energy consumption prediction result, so as to obtain the energy consumption change trend between multiple jurisdictions; based on the change trend, to obtain the energy consumption proportion of the high energy consumption jurisdiction, so as to set the heating section adjustment strategy; Data allocation module 204: It is used to dynamically adjust the optimal allocation of heating energy in the area in combination with the heating section adjustment strategy in the area climate temperature change, so as to improve the indoor heating energy efficiency of the building, so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated.

[0015] The present invention has at least the following beneficial effects: the present invention collects heating energy consumption data within a preset time range (one month, one quarter or half a year) within the city's jurisdiction. This may include the use of different energy sources such as natural gas, electricity, coal, etc.; the collected energy consumption data is combined with the population density data, building types and numbers, and heating methods such as central heating, heat pump heating or household heating within the city's jurisdiction; thus, a data set containing these influencing factors is constructed. This data set will help analyze and predict the relationship between energy consumption and these factors.

[0016] Use the constructed dynamic impact dataset to train an energy forecasting model. This model aims to learn how different factors affect energy consumption and predict future energy demand; the trained model is used to predict energy consumption in the same jurisdiction or different jurisdictions in the future time frame. By inputting the new dynamic impact dataset combined with the obtained deviation factor, the model can output the predicted results of energy consumption.

[0017] Based on the model's prediction results, compare the fluctuations and changing trends of energy consumption between different jurisdictions. This step helps to identify the peaks and troughs of energy consumption, thereby allocating resources more effectively; according to the changing trends of energy consumption, identify high-energy consumption areas and set specific heating section adjustment strategies for these areas.

[0018] Dynamically adjust the distribution and optimization of heating energy according to the actual climate changes in the region to ensure the efficiency and effectiveness of energy use; adjust the indoor heating energy efficiency according to the characteristics of different buildings, manage and optimize the heating needs of multiple areas within the jurisdiction to maximize the benefits of energy use. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of a smart city energy optimization management method based on deep learning provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of a smart city energy optimization management system based on deep learning provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0021] This embodiment provides a smart city energy optimization management method and system based on deep learning, and the specific implementation methods are as follows: like Figure 1 As shown, a smart city energy optimization management method based on deep learning includes the following steps: S100, collecting heating energy consumption data of the urban area within a preset time range; based on the heating energy consumption data, obtaining a dynamic impact data set within the urban area; the dynamic impact data set includes population density data, building type and quantity data, and heating method data; S200, training the acquired dynamic impact data set to construct an energy prediction model to output the prediction result of energy consumption in the jurisdiction within a preset time range; acquiring the next dynamic impact data set in the jurisdiction, inputting it into the energy prediction model, and the energy prediction model outputs the prediction result of energy consumption in the next jurisdiction within the preset time range; S300, based on the energy consumption prediction result, comparing the change fluctuation difference of the dynamic impact data set between multiple jurisdictions to obtain the energy consumption change trend between the multiple jurisdictions; based on the change trend, obtaining the energy consumption proportion of the high energy consumption jurisdiction to set the heating section adjustment strategy; S400, the heating section adjustment strategy dynamically adjusts the optimal allocation of heating energy in the area in combination with the regional climate temperature changes, so as to improve the indoor heating energy efficiency of the building, so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated.

[0022] In this embodiment, within the time range specified in step S100 (one month, one quarter or half a year), heating energy consumption data of the urban area is collected through a sensor network, smart meters or an energy management platform; based on the collected data, combined with the population density in the area (obtained through statistics or real-time population flow data), building types and quantities (obtained through building planning data and actual surveys), and heating method information, a dynamic impact data set is constructed; the energy data is made more comprehensive and multidimensional, providing high-quality input for subsequent model training; the generation of a dynamic impact data set can reveal the key drivers of energy consumption in the area (such as population distribution and building characteristics), and provide data support for decision-making.

[0023] Through step S200, the dynamic impact data set obtained in step S100 is used as training data, and a machine learning algorithm (such as regression analysis, deep learning) is used to build an energy consumption prediction model, so that the model can learn the complex relationship between the heating energy consumption of the jurisdiction and the dynamic influencing factors; the dynamic impact data set of the next jurisdiction is input, and the energy prediction model outputs the energy consumption prediction result within a preset time range according to the training result; it can predict the heating demand of the jurisdiction in advance and reduce the uncertainty of energy allocation; through model training, the accuracy of energy prediction will continue to improve with the accumulation of data, ensuring the rationality and efficiency of energy heating.

[0024] Step S300 analyzes the dynamic trend of energy consumption by comparing the fluctuation differences between the dynamic impact data sets of multiple jurisdictions (such as changes in population density, changes in the proportion of building types, etc.) based on step S200; identifies high-energy consumption jurisdictions based on trend analysis, and calculates their share in total energy consumption; formulates heating section adjustment strategies based on the characteristics of high-energy consumption areas (such as population density and building density); grasps the differences in energy demand between jurisdictions in advance, which helps to reasonably allocate energy resources; reduces energy waste and improves overall heating efficiency; and through dynamic analysis, helps to discover abnormal energy consumption areas and take targeted optimization measures.

[0025] Step S400 dynamically adjusts the heating energy allocation ratio of each jurisdiction based on step S300 in combination with the regional climate temperature changes (obtained through meteorological monitoring) and energy consumption forecast results to avoid energy shortages; optimizes the energy efficiency of the indoor heating system through intelligent adjustment equipment in the building (such as adjustable valves, intelligent thermostats) to reduce unnecessary energy consumption; maximizes the rational use of energy through dynamic adjustment; reduces unnecessary energy waste, thereby reducing the total cost of urban heating; and ensures that the heating quality meets the needs of residents through indoor heating energy efficiency adjustment.

[0026] Through the collaborative work between the above steps, it is driven by dynamic data, using energy consumption prediction models and data fluctuation analysis to achieve intelligent allocation and optimal management of energy; improve energy utilization efficiency and reduce waste; improve the scientific nature and feasibility of energy management decisions; and through dynamic adjustment, enhance the adaptability of the heating system in the jurisdiction to changes in the environment and user needs.

[0027] In a preferred embodiment of the present invention, step S100 may further include the following steps: Step S101, obtaining the main dynamic factors affecting energy consumption, namely the initial dynamic impact data set; preprocessing the dynamic impact data set to obtain the final dynamic impact data set; Step S102, encoding the final dynamic impact data set into sequence data, and inputting the sequence data into the energy prediction model, wherein the energy prediction model comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer; intermediate data of each hidden layer is transmitted to the output layer through the layers, and the final output represents the prediction result of energy consumption of the jurisdiction within a preset time range; Step S103, by comparing the output of the energy prediction model with the actual urban area heating energy consumption data, analyzing the deviation and obtaining the deviation factor.

[0028] In the embodiment of the present invention, step S101 determines the main dynamic factors affecting energy consumption, such as population density, building types and numbers, and heating methods. As climate changes, average temperature and minimum temperature directly determine heating demand. The lower the temperature, the higher the energy consumption. Buildings lose heat faster in high humidity environments, which may increase energy demand. High wind speeds may cause more heat loss and increase heating load. The longer the heating season lasts, the higher the total energy consumption. Divide the impact of time-varying energy supply methods; obtain the initial dynamic impact data set of the jurisdiction through a data acquisition system or an existing database; eliminate missing values ​​or outliers to ensure data integrity; standardize, normalize or classify dynamic influencing factors to facilitate model understanding and calculation; select features that are highly correlated with energy consumption to reduce redundant data interference.

[0029] The cleaned, converted and feature-selected datasets are used as high-quality inputs to ensure the accuracy of subsequent model training and predictions; the quality and consistency of input data are guaranteed to improve the learning effect of the model; data noise and redundancy are reduced to improve computing efficiency; and the main dynamic factors affecting energy consumption are initially identified to provide a basis for in-depth analysis.

[0030] Step S102 converts the final dynamic impact data set determined in step S101 into time series data (such as hourly or daily heating energy consumption and dynamic factor change series); the formatting of the time series ensures that the model can capture the changing patterns in the time dimension.

[0031] Energy forecasting model structure: Input layer: receives time series data and maps it to the feature space of the model; The first hidden layer: preliminarily extracts the basic characteristics between dynamic factors and energy consumption (such as linear relationships); the second hidden layer: deeply extracts the nonlinear relationship between features and establishes complex energy consumption patterns; the third hidden layer: further refines and optimizes the feature representation to generate high-level energy consumption information; the output layer: integrates the feature information transmitted by the hidden layer and outputs the energy consumption forecast results of the jurisdiction within the preset time range.

[0032] Use activation functions (such as ReLU and Sigmoid) to introduce nonlinearity to ensure that the model has strong fitting ability; the output of each layer is used as the input of the next layer, and the transmission and optimization of inter-layer information are adjusted through weights and biases; capture the complex relationship between dynamic factors and energy consumption to improve the accuracy of prediction results; use deep network structures to fully explore multi-dimensional data characteristics and enhance the generalization ability of the model; time series input enables the model to adapt to the cyclical and trend changes in energy consumption.

[0033] Step S103 compares the predicted energy consumption value of the area output by the energy prediction model with the actual heating energy consumption data item by item, and calculates the deviation (such as mean absolute error, mean square error); Bias Factor Analysis: Identify factors that cause bias (e.g. climate change, inefficient building heating equipment, features not covered by the forecast model, etc.).

[0034] Classification bias factors: Bias caused by input data quality issues, such as missing key dynamic factors; insufficient model structure or parameter optimization resulting in prediction results deviating from reality; abnormal situations outside the prediction range (such as sudden extreme weather) leading to inaccurate predictions.

[0035] By analyzing the sources of deviations, the model structure and input data quality are continuously optimized; the credibility of the prediction results and the reference value for decision-making are improved; and a reliable basis is provided for subsequent energy allocation and optimization.

[0036] In a preferred embodiment of the present invention, the above step S103 includes: Step S1031, comparing the energy consumption prediction result data output by the energy prediction model with the actually obtained heating energy consumption data of the urban area to obtain a comparison result, and based on the comparison result, obtaining a deviation between the energy prediction result and the actually obtained energy consumption data; Step S1032, analyzing the deviation result to obtain a deviation factor; the deviation factor is combined with the final dynamic impact data set obtained originally and input into the energy prediction model again for training to output a prediction result representing the final heating energy.

[0037] In an embodiment of the present invention, the result data predicted by the energy prediction model in step S1031 (energy consumption for one month, one quarter or half a year) is compared with the heating energy consumption data obtained from actual operation; the comparison can be performed through a variety of statistical methods, including calculating the error between the predicted value and the actual value (such as the difference, percentage error, etc.).

[0038] The comparison results demonstrate the accuracy of the model's predictive ability. Deviation refers to the difference between the predicted data and the actual data. Deviation calculations usually include mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), etc., which provide key feedback for model optimization. They directly reflect the accuracy and reliability of the prediction model. The size and pattern of the deviation can be used to identify the performance bottleneck of the model under specific conditions or in a specific time period.

[0039] Step S1032 analyzes the possible causes of the deviation based on the deviation obtained in S1031, such as inaccurate input data, insufficient model structure, and changes in external influencing factors; identifies which features or data points are often associated with high deviations, thereby locating potential deviation factors; incorporates deviation factors into the model, which may involve adjusting data weights, adding new features, optimizing model parameters, etc.; uses the corrected data set and model configuration to retrain the energy prediction model to improve its prediction accuracy; during the retraining process, methods such as cross-validation may be used to avoid overfitting and ensure the model's generalization ability on unseen data; by identifying and correcting deviation factors, the model's prediction accuracy will be improved; continuous model training and updating enable the prediction system to adapt to changing environments and conditions, and improve long-term stability and reliability; more accurate prediction results provide more effective support for energy management and deployment of urban heating systems, which helps energy conservation and cost control.

[0040] Step S1031 obtains deviations by comparing the predicted results with the actual data, providing necessary inputs for S1032, namely clear deviation data and comparison results; Step S1032 uses these deviation analysis results to adjust and optimize the model. This feedback-based iterative process helps to fine-tune the performance of the prediction model and reduce future prediction errors.

[0041] In a preferred embodiment of the present invention, the above step S1032 includes: Step S10321, by , to obtain the deviation value between the energy prediction result and the actual energy consumption data; where e represents the deviation value; y true Indicates the actual energy consumption data obtained; y pred represents the energy forecast result; Step S10322, if e=0, then there is no deviation; if e>0 or e<0, then there is deviation; based on the structure with deviation, by ; To analyze the correlation between the dynamic impact data set and the deviation value; where D m represents the correlation coefficient; represents the influence parameter of the dynamic impact data set; Contains data on energy transmission and consumption, types of energy used, and energy usage behavior; Step S10323, judging the influence of each data item in the dynamic impact data set on the deviation according to the correlation coefficient; calculating the correlation coefficients D generated by the multiple dynamic impact data sets. m When the correlation coefficient D m The larger the value, the higher the impact on the deviation, so as to obtain the deviation factor in the dynamic impact data set; the deviation factor indicates the influence on the deviation. ; Thus, the deviation factor includes energy delivery consumption data and energy usage behavior data; Step S10324: adding the extracted deviation factor to the dynamic impact data set to form a new dynamic impact data set ; In the formula, X' represents the new dynamic impact data set; X represents the dynamic impact data set; Represents the deviation factor.

[0042] In this embodiment, the deviation value e is defined as the difference between the energy prediction result and the actual energy consumption data through step S10321, that is, e=actual energy consumption data-energy prediction result; the accuracy of the prediction model is determined by a simple subtraction calculation, where a positive deviation e>0 indicates that the actual consumption exceeds the prediction, and a negative deviation e<0 indicates that the actual consumption is lower than the prediction; the accuracy of the prediction result is directly measured to provide basic data for further analysis; it can quickly discover under what conditions the model performs poorly, facilitating rapid adjustment or optimization.

[0043] Step S10322 determines the accuracy of the prediction result according to the deviation value e defined in step S10321; analyzes the correlation between the deviation and the dynamic impact data set to study the correlation between the deviation value and each parameter in the dynamic impact data set (such as energy transmission consumption, energy type used, etc.), marked as D m; Through correlation analysis, it is possible to identify which factors are significantly correlated with deviations in energy forecasts, which is crucial for adjusting forecast models and improving forecast accuracy; it provides a data-driven foundation for improving models and optimizing energy management strategies.

[0044] Step S10323 compares the correlation coefficients Dm generated by the different dynamic influencing data set items in step S10322 to determine the degree of influence of each data item on the deviation; selecting items with high correlation coefficients as deviation factors means that these factors play a greater role in generating deviations; ensuring that the focus of model optimization is placed on the factors that have the greatest impact on the prediction results, improving resource utilization efficiency and optimizing costs; by continuously adjusting the deviation factor, the model can better adapt to environmental changes and improve the stability and accuracy of the prediction.

[0045] Step S10324 incorporates the deviation factors identified in step S10323 into the existing dynamic impact data set to form a new and more complete dynamic impact data set; the updated data set will be used to retrain and optimize the energy prediction model; by continuously updating the data set and integrating key deviation factors, the model will continue to adapt to new data patterns, thereby improving the accuracy and reliability of the prediction; by adding new data points and deviation factors, the model can better handle changing environments and complex data structures.

[0046] In a preferred embodiment of the present invention, the above step S300 compares the change fluctuation difference of the dynamic impact data set between multiple jurisdictions based on the energy consumption prediction result to obtain the energy consumption change trend between the multiple jurisdictions, including the following steps: Step S301, by ; To obtain the energy consumption change rate of each jurisdiction in period t; where, It represents the dynamic change rate of energy consumption in the tth period; represents the energy consumption in period t; represents the energy consumption in the period before t; Step S302, by ; To obtain the fluctuation difference between jurisdictions; represents the dynamic change rate of the i-th jurisdiction in the t-th period; represents the dynamic change rate of the j jurisdiction in the tth period; It represents the dynamic fluctuation difference between the i-prefecture and the j-prefecture in the t-th period; Step S303, when is a positive number; the energy consumption trend between jurisdictions is in a downward state; when If it is a negative number, the energy consumption change trend between the jurisdictions is in an upward state; thus, the energy consumption change trend between multiple jurisdictions is obtained; Step S304: according to the dynamic fluctuation difference between the jurisdictions, , to obtain the energy consumption fluctuation rate between jurisdictions; where, represents volatility; represents the rate of change of total energy consumption in period t; represents the energy change in the period before t; Step S305: preset R to represent the volatility threshold. ; then it is determined that the jurisdiction i is stable in period t; when ; then it is determined that the jurisdiction i exhibits a violent fluctuation type in period t.

[0047] In this embodiment, step S301 clearly describes the dynamic changes of energy consumption in different time periods, providing a basis for subsequent analysis; provides specific indicators for identifying the rising or falling trend of energy consumption in the jurisdiction; step S302 obtains the fluctuation difference based on step S301, and the fluctuation difference can intuitively reflect whether the energy consumption change trends between different jurisdictions are consistent; supports spatial correlation analysis: provides quantitative indicators for studying the correlation of energy consumption between jurisdictions; step S303 is based on the obtained change fluctuation difference, if > 0, the energy consumption trend of the i district is lower than that of the j district, and the overall trend is in a downward state; if <0, the energy consumption change trend of the i district is higher than that of the j district, and the overall trend is in an upward state, so as to quickly judge the energy consumption change trend among multiple districts, which is helpful to formulate targeted strategies.

[0048] The volatility is obtained through step S304, which reflects the fluctuation of overall energy consumption between jurisdictions and provides a basis for further stability analysis; through the analysis of overall volatility, the synergy between jurisdictions can be studied; through trend analysis, a scientific basis is provided for resource allocation between jurisdictions. Step S305 quickly determines the jurisdictions that show excessive energy consumption based on the volatility threshold; and provides data support for energy management departments to adopt different strategies for stable and violently fluctuating jurisdictions.

[0049] Through the above steps, a complete analysis process is formed, going deeper layer by layer, from a single jurisdiction to overall fluctuations, and then to stability assessment; providing actionable quantitative indicators to support energy management optimization decisions; and adjusting thresholds to identify the needs of energy management goals in different jurisdictions.

[0050] In a preferred embodiment of the present invention, the above step S300, based on the change trend, obtains the energy consumption proportion of the high energy consumption area, including the following steps: Step S306, by Obtain the energy consumption of each data item of the dynamic impact data set of the violent fluctuation type jurisdiction in the t period; mark the highest fluctuation point of the energy consumption of each data item as the first marking point; mark the lowest fluctuation point of the energy consumption as the second marking point; Step S307, performing energy demand analysis according to the first marking point and the second marking point to obtain analysis results; based on the analysis results, , to obtain the energy consumption proportion of each data item; where PI represents the energy consumption proportion of the data item; Indicates the energy consumption of each data item that dynamically affects the data set; Represents the total energy consumption in a jurisdiction with severe fluctuations.

[0051] In this embodiment, the energy consumption of each data item in the dynamic impact data set of the jurisdiction marked as severely fluctuating within the tt period is collected through step S306; in the collected data, the highest and lowest points of energy consumption are found, and marked as the first marked point (highest fluctuation point) and the second marked point (lowest fluctuation point) respectively; by marking the extreme points, the key fluctuation range of energy consumption can be intuitively identified, which is convenient for further detailed analysis; realizing real-time monitoring of energy consumption in severely fluctuating jurisdictions is helpful to timely understand and respond to abnormal consumption situations.

[0052] Step S307 uses the first marking point and the second marking point obtained in step S306 to conduct an in-depth analysis of energy demand, thereby obtaining specific analysis results on changes in energy demand; based on the analysis results, by analyzing the energy consumption proportion of different data items, resource allocation can be optimized to ensure that key areas or equipment have sufficient energy supply; accurately identify high-energy consumption equipment or departments to provide data support for further energy-saving and emission reduction measures.

[0053] Step S306 provides energy consumption extreme point marks for each data item in the violently fluctuating area. These extreme points are the basis for in-depth analysis. Step S307 uses these extreme points to conduct detailed energy demand analysis, and then derives the proportion of each data item relative to total consumption, providing an accurate basis for decision-making.

[0054] In summary, through real-time data monitoring and precise analysis, data-based decision-making is supported to improve the efficiency and effectiveness of energy management; timely identification and response to risks that may arise in areas with severe fluctuations, reducing unnecessary energy waste and possible equipment damage.

[0055] In a preferred embodiment of the present invention, the heating section adjustment strategy in step S400 includes the following steps: Step S401, according to the energy consumption proportion of the high energy consumption area, a heating section adjustment strategy is set; the heating section adjustment strategy obtains the outdoor climate temperature in real time, and heats by gas or electricity in the early stage of heating according to the dynamic impact data set, and the average temperature is controlled within the range of 23.5±1.5℃; Step S402, convert the gas or electric heating mode into hot water or steam heating; , to adjust the outlet temperature of hot water boiler or steam boiler; In the formula, Indicates the outlet temperature of a hot water boiler or steam boiler; represents the outdoor temperature, and ; represents building type data; n represents the number of buildings; the building type data includes building volume ratio, space volume, usage quantity and heat load performance data; Represents population density data; Step S403, according to , to adjust the return water temperature of the hot water boiler or steam boiler; by dynamically adjusting the hot water temperature in real time, to avoid excessive energy consumption caused by overheating.

[0056] In this embodiment, step S401 is used to identify the energy distribution situation in the area according to the proportion of energy consumption in the high energy consumption area, and give priority to heating optimization for the high proportion area; the external temperature is dynamically monitored as a basis for adjusting the heating strategy. In the early stage of heating, gas or electric heating is used to ensure that the indoor temperature remains within a comfortable range of 23.5±1.5℃, and adjustment parameters are provided for gas or electric heating in combination with the energy consumption proportion and environmental data in the area; through a reasonable temperature control range, the comfort of the living environment is guaranteed and energy waste is reduced; the heating strategy is adjusted in real time to adapt to external temperature changes and improve energy utilization efficiency.

[0057] Step S402: After the initial gas or electric heating reaches the target temperature, the heating mode is switched to hot water or steam heating to improve energy efficiency; the outlet temperature T of the hot water boiler or steam boiler is determined by adjusting the boiler outlet temperature. out ; Dynamically adjust the boiler outlet temperature according to building characteristics and outdoor temperature to ensure that heating demand matches energy supply; switch the heating mode after the initial temperature is reached, use hot water or steam to maintain the temperature, and reduce the continuous consumption of gas or electricity.

[0058] Step S403 determines the return water temperature T of the hot water boiler or steam boiler by adjusting the return water temperature. return; Real-time monitoring and adjustment of hot water temperature to avoid excessive energy consumption caused by over-temperature, prevent insufficient heating effect, dynamically adjust return water temperature, reduce unnecessary heat loss during heating, avoid excessive wear of boiler equipment due to overheating, and reduce maintenance and replacement costs.

[0059] In a preferred embodiment of the present invention, the heating section adjustment strategy described in the above step S400 combines the regional climate temperature changes to dynamically adjust the optimal allocation of heating energy in the jurisdiction, including the following steps: Step S404, based on real-time meteorological data and historical heating data, a linear relationship between climate temperature and regional heat load demand is established; and heating sections are divided according to building density, thermal insulation performance and heating network layout in the region; Step S405, each section is adjusted according to the heating section adjustment strategy. ; Dynamically adjust the optimal allocation of heating energy in the jurisdiction, where E i represents the heating energy demand in the i area; c i represents the heating energy cost in the i-th district; P i represents the heat network loss in area i; p s Represents the loss weight coefficient; n represents the number of dynamically changing data items.

[0060] In this embodiment, meteorological data and historical heating data are obtained in real time through step S404: these data are used to establish a linear relationship between outdoor climate temperature and regional heat load demand. This relationship model helps predict heating demand under different temperature conditions; Heating zone division: Building density, thermal insulation performance, and heat network layout: These factors can be taken into account to effectively divide the heating zones. Areas with high building density or poor thermal insulation performance may require more thermal energy, and the layout of the heat network determines the way and efficiency of heating; by establishing the relationship between climate temperature and heat load demand, the heating demand can be more accurately predicted and met; dividing the heating zones according to building density and thermal insulation performance can ensure the rationality and efficiency of thermal energy distribution and reduce energy waste.

[0061] Step S405 dynamically adjusts the energy allocation for each heating zone according to its specific heat load demand, energy cost and heat network loss; cost and loss optimization allows the heating strategy to be adjusted according to real-time data, optimizes energy use and reduces losses; by accurately calculating and predicting the heat energy demand and cost of each area, resources can be used more effectively and operating costs can be reduced; dynamic adjustment strategies ensure optimal energy allocation, reduce energy waste, and improve the overall efficiency of the heating system.

[0062] In summary, the collaborative relationship between step S404 and step S405 is as follows: Step S404 divides the heating sections by analyzing meteorological and historical data as well as regional characteristics, providing the necessary basis and data support for the heating regulation strategy of S405; Step S405 uses the data and divisions provided by step S404 to achieve dynamic heating energy optimization, ensuring that the heating amount of each section matches the actual demand; integrating real-time and historical data to provide a scientific decision-making basis for heating management; taking into account regional characteristics and actual needs, achieving efficient operation and energy distribution of the heating system; optimizing heating energy distribution, reducing energy consumption and economic costs, and improving the quality of life of residents.

[0063] In a preferred embodiment of the present invention, the above step S400 improves the indoor heating energy efficiency of the building so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated, including the following steps: Step S406, divide the building interior into zones, , to adjust the zone heating temperature; where T set Indicates the target indoor temperature; P density represents the population density weight; E building represents the building heat loss coefficient; A, B and C represent adjustment factors; the adjustment factors include outdoor temperature, indoor space volume and crowd category; Step S407: Maintain the temperature of the main activity area at 23.5±1.5℃, and the temperature of the secondary area at 17.5±1.5℃; control the water pump power at 0.35-0.45W / m 2 , in extremely cold weather T outdoor <-20℃ enables high flow operation, but the power is strictly limited to 0.6W / m 2 Below, to meet the basic heating needs.

[0064] In this embodiment, the building interior is zoned through step S406, which is usually based on factors such as frequency of use and density of human traffic, so as to improve the efficiency and comfort of energy use; reduce unnecessary energy consumption by adjusting the temperature according to the specific needs of different areas; ensure that the temperature in frequently used areas and crowded areas is appropriate, and increase the comfort of living or working.

[0065] Step S407 maintains the temperature of the primary area at 23.5±1.5℃, which is suitable for most activities and a comfortable living environment; the temperature of the secondary area is allowed to be 17.5±1.5℃, which is suitable for less used or storage space, etc.; under normal circumstances, it is controlled at 0.35−0.45W / m², and in extreme cold weather (T outdoor<−20℃), high flow operation is enabled, but the power is limited to below 0.6W / m² to prevent excessive power consumption; energy use is optimized and waste is reduced by precisely controlling the temperature and pump power in different areas; system operation is adjusted under extreme climatic conditions to ensure the reliability and effectiveness of the heating system, while taking energy consumption control into account.

[0066] To sum up, the collaborative relationship between step S406 and step S407 is that through the detailed zoning and adjustment strategy of step S406, step S407 can more accurately maintain the temperature in the area to meet actual needs; the temperature adjustment strategy provided by step S406 directly affects the energy allocation and temperature maintenance in step S407, so that the entire system can be flexibly adjusted according to actual needs and external conditions.

[0067] In another preferred embodiment of the present invention, the regional optimization strategy adjusts the heating intensity according to the population density: high population density areas (such as city centers): maintain the upper limit of the hot water temperature curve to ensure the demand for centralized heating; low population density areas (such as suburbs): appropriately reduce the boiler hot water outlet temperature, and supplement building insulation measures.

[0068] Differentiated heating of building structures: High-efficiency buildings: use lower return water temperatures and encourage users to install smart temperature control devices.

[0069] Inefficient buildings: Prioritize renovation plans, including upgrading wall insulation, optimizing door and window tightness, etc., to reduce heating load.

[0070] Optimization of regional pipeline pressure difference: The pressure difference at the end of the pipeline should be maintained at 0.01−0.05MPa, using the following methods: For long-distance pipelines, use distributed circulation pumps to reduce pressure loss; adjust the dynamic flow distribution in the pipeline to give priority to heating in densely populated areas and inefficient buildings.

[0071] Adjust the supply priority of heat stations: give priority to key buildings such as hospitals and schools; Improve energy efficiency: change the heat flow through intelligent distributors and dynamically adjust the boiler group load.

[0072] like Figure 2 As shown, a smart city energy optimization management system based on deep learning includes: Data collection module 201: used to collect heating energy consumption data of the urban area within a preset time range; based on the heating energy consumption data, to obtain a dynamic impact data set within the urban area; the dynamic impact data set includes population density data, building type and quantity data, and heating mode data; Data processing module 202: It is used to train the acquired dynamic impact data set to construct an energy prediction model to output the prediction result of energy consumption in the jurisdiction within a preset time range; obtain the next dynamic impact data set in the jurisdiction and input it into the energy prediction model, and the energy prediction model outputs the prediction result of energy consumption in the next jurisdiction within the preset time range; Data management module 203: used to compare the change fluctuation difference of the dynamic impact data set between multiple jurisdictions based on the energy consumption prediction result, so as to obtain the energy consumption change trend between multiple jurisdictions; based on the change trend, to obtain the energy consumption proportion of the high energy consumption jurisdiction, so as to set the heating section adjustment strategy; Data allocation module 204: It is used to dynamically adjust the optimal allocation of heating energy in the area in combination with the heating section adjustment strategy in the area climate temperature change, so as to improve the indoor heating energy efficiency of the building, so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated.

[0073] When the functions of the above modules are implemented in the form of software functional units and used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks or optical disks.

[0074] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general intelligent device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code that implements the method or system. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

Claims

1. A smart city energy optimization management method based on deep learning, characterized in that: The following steps are involved: S100, collecting heating energy consumption data of the urban area within a preset time range; based on the heating energy consumption data, obtaining a dynamic impact data set within the urban area; the dynamic impact data set includes population density data, building type and quantity data, and heating method data; S200, training the acquired dynamic impact data set to construct an energy prediction model to output the prediction result of energy consumption in the jurisdiction within a preset time range; acquiring the next dynamic impact data set in the jurisdiction, inputting it into the energy prediction model, and the energy prediction model outputs the prediction result of energy consumption in the next jurisdiction within the preset time range; S300, based on the energy consumption prediction result, comparing the change fluctuation difference of the dynamic impact data set between multiple jurisdictions to obtain the energy consumption change trend between the multiple jurisdictions; Based on the changing trend, the energy consumption proportion of high energy consumption areas can be obtained to set the heating section adjustment strategy; S400, the heating section adjustment strategy is combined with the regional climate temperature change to dynamically adjust the optimal allocation of heating energy in the jurisdiction; To improve the energy efficiency of indoor heating in buildings, so that the energy required for heating in multiple jurisdictions can be effectively managed and reasonably allocated.

2. According to the deep learning-based smart city energy optimization management method of claim 1, it is characterized in that: The acquired dynamic impact data set is trained to build an energy forecasting model to output the energy consumption forecast results of the jurisdiction within a preset time range, including: The main dynamic factors affecting energy consumption are obtained as an initial dynamic impact data set; the dynamic impact data set is preprocessed to obtain a final dynamic impact data set; Encoding the final dynamic impact data set into sequence data, and inputting the sequence data into the energy prediction model, wherein the energy prediction model comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer; intermediate data of each hidden layer is transmitted to the output layer through the layers, and the final output represents the prediction result of energy consumption of the jurisdiction within a preset time range; By comparing the output of the energy forecast model with the actual urban heating energy consumption data, the deviation is analyzed and the deviation factor is obtained.

3. According to claim 2, a smart city energy optimization management method based on deep learning is characterized in that: By comparing the output of the energy forecast model with the actual urban heating energy consumption data, the deviation is analyzed and the deviation factors are obtained, including: Comparing the predicted energy consumption data output by the energy prediction model with the actually obtained heating energy consumption data of the urban area to obtain a comparison result, and based on the comparison result, obtaining a deviation between the energy prediction result and the actually obtained energy consumption data; The deviation result is analyzed to obtain the deviation factor; the deviation factor is combined with the final dynamic impact data set obtained originally and input into the energy prediction model again for training to output the prediction result representing the final heating energy.

4. According to the deep learning-based smart city energy optimization management method of claim 3, it is characterized in that: Analyze the deviation results to obtain the deviation factors, including: pass , to obtain the deviation value between the energy prediction result and the actual energy consumption data; where e represents the deviation value; y true Indicates the actual energy consumption data obtained; y pred represents the energy forecast result; If e=0, there is no bias; if e>0 or e<0, there is bias; based on the biased structure, through ; To analyze the correlation between the dynamic impact data set and the deviation value; where D m represents the correlation coefficient; represents the influence parameter of the dynamic impact data set; Contains data on energy transmission and consumption, types of energy used, and energy usage behavior; Determine the influence of each data item in the dynamic impact data set on the deviation according to the correlation coefficient; m When the correlation coefficient D m The larger the value, the higher the impact on the deviation, so as to obtain the deviation factor in the dynamic impact data set; the deviation factor indicates the influence on the deviation. ; Thus, the deviation factor includes energy delivery consumption data and energy usage behavior data; The extracted deviation factors are added to the dynamic impact dataset to form a new dynamic impact dataset ; In the formula, X' represents the new dynamic impact data set; X represents the dynamic impact data set; Represents the deviation factor.

5. According to the deep learning-based smart city energy optimization management method of claim 4, it is characterized in that: Based on the energy consumption prediction results, the fluctuation differences of the dynamic impact data sets between multiple jurisdictions are compared to obtain the energy consumption change trends between the multiple jurisdictions, including: pass ; To obtain the energy consumption change rate of each jurisdiction in period t; where, It represents the dynamic change rate of energy consumption in the tth period; represents the energy consumption in period t; represents the energy consumption in the period before t; pass ; To obtain the fluctuation difference between jurisdictions; It represents the dynamic change rate of the i-th jurisdiction in the t-th period; represents the dynamic change rate of the j jurisdiction in the tth period; It represents the dynamic fluctuation difference between the i-prefecture and the j-prefecture in the t-th period; when is a positive number; the energy consumption trend between jurisdictions is in a downward state; when If it is a negative number, the energy consumption change trend between the jurisdictions is in an upward state; thus, the energy consumption change trend between multiple jurisdictions is obtained; According to the dynamic fluctuation difference between the jurisdictions , to obtain the energy consumption fluctuation rate between jurisdictions; where, represents volatility; represents the rate of change of total energy consumption in period t; represents the energy change in the period before t; The default value R represents the volatility threshold. ; then it is determined that the jurisdiction i is stable in period t; when ; then it is determined that the jurisdiction i exhibits a violent fluctuation type in period t.

6. The smart city energy optimization management method based on deep learning according to claim 5 is characterized in that: Based on the changing trend, the energy consumption proportion of high energy consumption areas is obtained, including: pass Obtain the energy consumption of each data item of the dynamic impact data set of the violent fluctuation type jurisdiction in the t period; mark the highest fluctuation point of the energy consumption of each data item as the first marking point; mark the lowest fluctuation point of the energy consumption as the second marking point; Perform energy demand analysis according to the first marking point and the second marking point to obtain analysis results; based on the analysis results, , to obtain the energy consumption proportion of each data item; where PI represents the energy consumption proportion of the data item; Indicates the energy consumption of each data item that dynamically affects the data set; Represents the total energy consumption in a jurisdiction with severe fluctuations.

7. The smart city energy optimization management method based on deep learning according to claim 6 is characterized in that: The heating section adjustment strategy includes: According to the energy consumption proportion of the high energy consumption area, a heating section adjustment strategy is set; the heating section adjustment strategy obtains the outdoor climate temperature in real time, and heats by gas or electricity in the early stage of heating according to the dynamic impact data set, and the average temperature is controlled within the range of 23.5±1.5℃; Then convert the gas or electric heating mode into hot water or steam heating; , to adjust the outlet temperature of hot water boiler or steam boiler; In the formula, Indicates the outlet temperature of a hot water boiler or steam boiler; represents the outdoor temperature, and ; represents building type data; n represents the number of buildings; the building type data includes building volume ratio, space volume, usage quantity and heat load performance data; Represents population density data; according to , to adjust the return water temperature of the hot water boiler or steam boiler; by dynamically adjusting the hot water temperature in real time, to avoid excessive energy consumption caused by overheating.

8. The smart city energy optimization management method based on deep learning according to claim 7 is characterized in that: The heating section adjustment strategy combines regional climate temperature changes to dynamically adjust the optimal allocation of heating energy in the jurisdiction, including: Based on real-time meteorological data and historical heating data, a linear relationship between climate temperature and regional heat load demand is established; and heating sections are divided according to building density, thermal insulation performance and heating network layout in the region; Each zone is adjusted according to the heating zone regulation strategy. ; Dynamically adjust the optimal allocation of heating energy in the jurisdiction, where E i represents the heating energy demand in the i area; c i represents the heating energy cost in the i-th district; P i represents the heat network loss in area i; p s Represents the loss weight coefficient; n represents the number of dynamically changing data items.

9. The smart city energy optimization management method based on deep learning according to claim 8 is characterized in that: Improve the energy efficiency of indoor heating in buildings so that energy required for heating in multiple jurisdictions can be effectively managed and allocated, including: The building interior is divided into zones, , to adjust the zone heating temperature; where T set Indicates the target indoor temperature; P density represents the population density weight; E building represents the building heat loss coefficient; A, B and C represent adjustment factors; the adjustment factors include outdoor temperature, indoor space volume and crowd category; Maintain the temperature of the main activity area at 23.5±1.5℃, and the temperature of the secondary area is allowed to be 17.5±1.5℃; control the water pump power at 0.35-0.45W / m 2 , in extremely cold weather T outdoor <-20℃ enables high flow operation, but the power is strictly limited to 0.6W / m 2 Below, to meet the basic heating needs.

10. A smart city energy optimization management system based on deep learning, characterized in that: The system sets an electronic device including a memory, a processor, and a smart city energy optimization management method program based on deep learning stored in the memory and executable on the processor. When the smart city energy optimization management method program based on deep learning is executed by the processor, the steps of the smart city energy optimization management method based on deep learning according to any one of claims 1 to 9 are implemented. The system includes: Data collection module: used to collect heating energy consumption data of the urban area within a preset time range; based on the heating energy consumption data, to obtain a dynamic impact data set within the urban area; the dynamic impact data set includes population density data, building type and quantity data, and heating mode data; Data processing module: It is used to train the acquired dynamic impact data set to construct an energy prediction model to output the prediction result of energy consumption in the jurisdiction within a preset time range; obtain the dynamic impact data set in the next jurisdiction and input it into the energy prediction model, and the energy prediction model outputs the prediction result of energy consumption in the next jurisdiction within the preset time range; Data management module: It is used to compare the change fluctuation difference of the dynamic impact data set between multiple jurisdictions based on the energy consumption prediction result to obtain the energy consumption change trend between multiple jurisdictions; based on the change trend, the energy consumption proportion of the high energy consumption jurisdiction is obtained to set the heating section adjustment strategy; Data allocation module: It is used to dynamically adjust the optimal allocation of heating energy in the area in combination with the heating section adjustment strategy in the area climate temperature changes, so as to improve the energy efficiency of indoor heating in buildings, so that the energy required for heating in multiple areas can be effectively managed and reasonably allocated.

Images