Energy loss optimization method and system based on deep learning

Abstract

The invention discloses an energy loss optimization method and system based on deep learning, and relates to the technical field of machine learning, and the method comprises the steps: calculating a cycle period through obtaining secondary pipe network data and circulating water flow data, calibrating a user topological position through combining with a heat supply completion drawing, and carrying out the correlation storage; determining a time reference and spatial topological relation of the heat supply system; a heat supply portrait is constructed based on temperature data collected in a cycle period, historical portraits are integrated to form a data analysis set, a predicted heat supply portrait is obtained by using a time sequence prediction model, and heat supply demand prospective pre-judgment is realized; a mapping relation model of circulating water temperature, flow velocity and home-entry pipe temperature change is trained through deep learning, an advanced regulation and control time point is calculated in combination with user topological positions and circulating water flow data, the problems that downstream user heat supply is unbalanced and energy loss is increased during low-temperature anti-freezing heat transfer are solved, heat supply is guaranteed to reach the standard, system balance is maintained, and meanwhile the energy consumption is reduced. And heat dissipation is reduced, and energy loss optimization is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of machine learning, in particular to an energy consumption optimization method and system based on deep learning. BACKGROUND

[0002] The heating system heats circulating water through a heat exchange station, and the heated circulating water is delivered to user terminals through a secondary pipe network to release heat, and the cooled circulating water returns to the heat exchange station to be reheated to form a closed-loop heating cycle. As a key infrastructure in winter severe cold regions, it directly ensures the indoor suitable living temperature of residents and maintains the normal life order and the quality of life of residents.

[0003] The topology of the secondary pipe network directly affects the stability of the heating system. The energy consumption of the far-end users increases due to the increased transmission distance of the pipe network. In the late night, shopping malls and other scenarios usually adopt low-temperature anti-freezing operation mode to prevent pipe network cracking and avoid invalid energy consumption. However, when such low-temperature anti-freezing operation users switch to normal heating state, the regulation and change of the circulating water temperature and flow rate in the secondary pipe network will be gradually transmitted along the topology path, which will destroy the original heating balance of the downstream users. In order to maintain the demand, the downstream users will further increase the heating regulation range, forcing the heat exchange station to adjust the output parameters to increase the circulating water temperature and flow rate, thereby intensifying the unorganized emission of heat and indirectly increasing the overall energy consumption. Therefore, there is an urgent need for an energy consumption optimization method and system based on deep learning. SUMMARY

[0004] The purpose of the present application is to provide an energy consumption optimization method and system based on deep learning to solve the problems in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme: an energy consumption optimization method based on deep learning, the energy consumption optimization method comprising the following steps: Step S1, obtaining and analyzing the secondary pipe network data information connected with the heat exchange station and the circulating water flow data information of the secondary pipe network interface to obtain the circulating period of the circulating water in the secondary pipe network; Step S1-1, obtaining the pipe diameter data and the corresponding pipe length data of each pipe section of the secondary pipe network connected with the heat exchange station through the heating completion drawing; calculating the cross-sectional area of each pipe section according to the pipe diameter, and calculating the volume of a single pipe section filled with circulating water combined with the pipe length to obtain the total volume of the circulating water in the secondary pipe network by accumulating the volumes of all pipe sections filled with circulating water. The formula for calculating the cross-sectional area of the pipe diameter is as follows: S i =π×(d i ÷2) 2 ; In the formula, S idenotes the cross-sectional area of the i-th pipe section; d i denotes the pipe diameter of the i-th pipe section; The formula for calculating the volume of circulating water filled by a single pipe section is as follows: V i = S i × L i ; In the formula, V i denotes the volume of circulating water filled by the i-th pipe section; L i denotes the length of the i-th pipe section; The formula for calculating the total volume of circulating water in the secondary pipe network is as follows: ; In the formula, V all denotes the total volume of circulating water in the secondary pipe network; n denotes the number of pipe sections in the secondary pipe network; Step S1-2, obtain the circulating water flow data of the secondary pipe network interface through the flow sensor, the secondary pipe network interface includes the secondary pipe network water supply end interface and the secondary pipe network return water end interface; calculate the average value of the circulating water flow of the secondary pipe network water supply end interface and the secondary pipe network return water end interface; divide the total volume of circulating water in the secondary pipe network obtained in step S1-1 by the average value of the circulating water flow to obtain the circulating period of the circulating water in the secondary pipe network; The formula for calculating the average value of the circulating water flow of the secondary pipe network water supply end interface and the secondary pipe network return water end interface is as follows: Q avg = (Q out + Q in ) ÷ 2; In the formula, Q avg denotes the average value of the circulating water flow of the secondary pipe network water supply end interface and the return water end interface; Q out denotes the circulating water flow of the secondary pipe network water supply end interface; Q in denotes the circulating water flow of the secondary pipe network return water end interface; The formula for calculating the circulating period of the secondary pipe network circulating water is as follows: T = V all ÷ Q avg ; In the formula, T denotes the circulating period of the secondary pipe network circulating water; The total volume of circulating water is calculated by obtaining the diameter and length data of each pipe section of the secondary pipe network through the heating completion drawing, and the circulating water flow at the water supply end and the return end interface of the secondary network is obtained through the flow sensor and the average value is calculated, and then the circulating period of the circulating water in the secondary pipe network is obtained, which provides an accurate time reference for the subsequent steps of heating service area detection, user indoor pipe temperature collection time interval setting, heating portrait construction, etc., to ensure that all subsequent heating-related data collection and analysis are based on a unified time dimension.

[0006] Step S2, detecting the heating service area to which the secondary pipe network belongs according to the circulating period, and topologically positioning and associating storing the users in the heating service area based on the heating completion drawing; Step S2-1, selecting the circulating period obtained in step S1 as a reference to detect the heating service area to which the secondary pipe network belongs, and determining the connection relationship between the users in the heating service area and the secondary pipe network branches and the corresponding pipe sections of each user, wherein the heating service area represents the area heated by the circulating water transported by the secondary pipe network connected to the heat exchange station; Step S2-2, based on the heating completion drawing, extracting the geographic coordinates of each user in the heating service area and the corresponding secondary pipe network branch node number, and positioning the topological position of each user in the secondary pipe network; associating the positioned topological position of the user with the user identification information and storing it in the database; By taking the circulating period as a reference, the connection relationship between the users in the heating service area and the secondary pipe network branches and the corresponding pipe sections are determined, the geographic coordinates of the users and the node numbers of the pipe network branches are extracted based on the heating completion drawing, the topological position of the users is positioned and associated with the user identification information and stored in the database, which clarifies the spatial distribution and connection relationship of the users in the secondary pipe network, and provides accurate spatial topological data support for the subsequent operation of accurately positioning the pipe network branch pipe section corresponding to the low-temperature anti-freezing operation state user, calculating the pipe network path length, and associating the user position with the heating data.

[0007] Step S3, detecting the circulating water temperature of the indoor pipe of each user in the heating service area based on the circulating period, and analyzing the circulating water temperature in the secondary pipe network interface of the heat exchange station to construct a circulating period heating portrait of the heating service area. Step S3-1, based on the cycle period obtained in step S1, the circulating water temperature in the indoor pipe of each user in the heat supply service area is collected at a preset time interval within the cycle period, and the preset time interval is not more than the cycle period; the circulating water temperature of the heat exchange station secondary pipe network water supply end interface is collected at the same time, the density and specific heat capacity data of the circulating water in the secondary pipe network are obtained; the ratio of the circulating water temperature in the indoor pipe of each user to the circulating water temperature of the secondary pipe network water supply end interface is calculated, which is recorded as the single user indoor pipe circulating water temperature ratio; the indoor pipe circulating water heating capacity of each user is calculated, and the single user indoor pipe circulating water heating capacity is equal to the product of the circulating water density, the specific heat capacity, the indoor pipe circulating water flow of the corresponding user and the single user indoor pipe circulating water temperature ratio; the single user indoor pipe circulating water temperature ratios of all users are added and then divided by the total number of users in the heat supply service area, to obtain the average indoor pipe circulating water temperature ratio corresponding to the time interval; the single user indoor pipe circulating water heating capacities of all users are added to obtain the total indoor pipe circulating water heating capacity of the heat supply service area corresponding to the time interval; The single user indoor pipe circulating water heating capacity calculation formula is as follows: ; In the formula, Q j，out represents the indoor pipe circulating water heating capacity of the jth user; p represents the density of the circulating water in the secondary pipe network; c represents the specific heat capacity of the circulating water in the secondary pipe network; Q j，two represents the indoor pipe circulating water flow of the jth user; k j represents the indoor pipe circulating water temperature ratio of the jth user; Step S3-2, record the collection time of each preset time interval, associate the collection time with the total indoor pipe circulating water heating capacity of the heat supply service area corresponding to the time interval, and form a data set of the total indoor pipe circulating water heating capacity of the heat supply service area changing with time within the cycle period; based on the data set, the cycle period heating portrait of the heat supply service area is obtained, which contains the mapping relationship between each collection time and the corresponding total indoor pipe circulating water heating capacity of the heat supply service area within the cycle period, and the cumulative value of the total indoor pipe circulating water heating capacity of the heat supply service area within the cycle period; Based on the cycle period, the user indoor pipe circulating water temperature and the secondary pipe network water supply end interface temperature are collected at a preset time interval, the single user water temperature ratio, the heating capacity, the total heating capacity of the service area and the average water temperature ratio are calculated, the collection time and the total heating capacity are associated to form a data set, and the cycle period heating portrait is constructed, which completely presents the change rule and the cumulative value of the total indoor pipe circulating water heating capacity of the heat supply service area at each time within the cycle period, and provides basic data containing time, temperature and heating capacity for subsequent historical heating data analysis set construction and deep learning model training.

[0008] Step S4, a plurality of historical cycle heat supply images of the heat exchange station are acquired, and a heat supply data analysis set of the heat supply service area is constructed; according to the heat supply data analysis set and a time sequence prediction model, data processing is performed, and a heat supply image of the heat supply service area in the next cycle is predicted, denoted as a predicted cycle heat supply image; Step S4-1, a plurality of cycle heat supply images generated in the historical operation process of the heat exchange station are acquired, and the acquisition time, the total heat supply of the circulating water of the household pipe of the heat supply service area and the cumulative value of the total heat supply in the cycle are extracted in each historical cycle heat supply image; all the extracted historical cycle related data are associated and arranged according to the cycle sequence number, and a heat supply data analysis set of the heat supply service area is constructed; Step S4-2, the constructed heat supply data analysis set of the heat supply service area is input into a time sequence prediction model for data processing, and the time sequence prediction model learns the time variation law and the cumulative value variation law of the total heat supply of the circulating water of the household pipe of the heat supply service area in the historical cycle; a heat supply image of the heat supply service area in the next cycle is output, denoted as a predicted cycle heat supply image, which contains the mapping relationship between each preset acquisition time and the corresponding predicted total heat supply of the circulating water of the household pipe of the heat supply service area in the next cycle, and the predicted cumulative value of the total heat supply of the circulating water of the household pipe of the heat supply service area in the next cycle; The historical plurality of cycle heat supply images of the heat exchange station are acquired, the data such as the acquisition time, the total heat supply and the cumulative value are extracted and arranged into a heat supply data analysis set according to the cycle sequence number, the time sequence prediction model is input to learn the historical heat supply variation law, and the predicted heat supply image of the next cycle is output, which clearly shows the predicted total heat supply and the cumulative value of each preset acquisition time in the next cycle, and provides prospective heat supply demand data basis for the subsequent prediction and adjustment of the regulation parameters of the heat exchange station.

[0009] Step S5, according to the user topology position of the heat supply service area and the real-time flow data information analysis and calculation of the secondary pipe network interface of the heat exchange station, the low-temperature antifreeze advance regulation time of the heat exchange station is obtained when the low-temperature antifreeze operation state of the heat supply service area changes; Step S5-1, retrieve the user identification information in the low-temperature freeze-proof running state in the heating service area from the database, determine the corresponding secondary pipe network branch pipe section of the user in the low-temperature freeze-proof running state in combination with the user topological position calibrated in step S2; extract the time stamp when the household pipe circulating water reaches the standard temperature when the user in the low-temperature freeze-proof running state in the historical cycle period changes from the low-temperature freeze-proof running state to the normal heating state, and the time stamp when the heating station secondary pipe network water supply end interface starts to be controlled in the corresponding historical cycle period, and the control represents the adjustment operation on the circulating water temperature and the circulating water flow rate of the heating station secondary pipe network water supply end interface; Step S5-2, determine the pipe network path length between the secondary pipe network branch pipe section corresponding to the user in the low-temperature freeze-proof running state and the heating station secondary pipe network water supply end interface based on the topological position of the user in the low-temperature freeze-proof running state, calculate the transmission time of the circulating water flowing through the pipe network path length in combination with the real-time flow data of the circulating water of the heating station secondary pipe network interface; subtract the transmission time from the difference between the extracted time stamp when the household pipe circulating water of the user reaches the standard temperature and the starting time stamp of the heating station control, to obtain the low-temperature freeze-proof to heating advance control time of the heating station when the low-temperature freeze-proof running state in the heating service area changes; By retrieving the user identification information in the low-temperature freeze-proof running state, determining the corresponding pipe network branch pipe section in combination with the topological position, extracting the standard reaching time stamp and the control starting time stamp in the historical switching state, calculating the pipe network path transmission time and deriving the low-temperature freeze-proof to heating advance control time, the time parameter accurately quantifies the length of time required to start the control in advance to compensate for the transmission delay of the circulating water, which provides a key time reference basis for subsequent determination of the starting control time point of the heating station and avoids temperature standard reaching delay during heating switching.

[0010] Step S6, perform data training processing on the heating data analysis set of the heating service area through deep learning, map the temperature and flow rate of the circulating water in the secondary pipe network and the temperature change relationship of the household pipe, and obtain the starting control time point of the low-temperature freeze-proof to heating of the heating station in combination with the low-temperature freeze-proof to heating advance control time analysis; Step S6-1, extract the temperature data and flow rate data of the circulating water in the secondary pipe network in the historical cycle period, and the household pipe circulating water temperature data of each user in the heating service area at the corresponding time from the heating data analysis set; combine the user topological position data to arrange the extracted data according to the circulating water temperature, the circulating water flow rate and the household pipe circulating water temperature, and the user topological position, to form a training data set; input the training data set into a pre-set deep learning model for training, and output to obtain a mapping relationship model of the temperature and flow rate of the circulating water in the secondary pipe network and the household pipe circulating water temperature change; Step S6-2, obtaining the initial temperature data of the circulating water in the user's indoor pipe in the heating service area in the low-temperature anti-freezing running state and the target temperature data of the circulating water in the user's indoor pipe in the preset normal heating state, calculating the temperature difference between the two; inputting the temperature difference into the mapping relationship model, combining the topological position data of the corresponding user, and outputting the heating time required for the circulating water in the user's indoor pipe to rise from the initial temperature to the target temperature; subtracting the heating time from the low-temperature anti-freezing to heating advance control time to obtain the time point of the low-temperature anti-freezing to heating start control of the heat exchange station, which represents the specific time when the circulating water in the user's indoor pipe in the low-temperature anti-freezing running state reaches the normal heating state target temperature; predicting the total heating demand in the circulating period heating image, combining the heating demand output by the mapping relationship model, and gradually adjusting the circulating water temperature of the secondary pipe network water supply end interface according to the preset time interval in the circulating period; when adjusting the circulating water flow rate, based on the real-time flow data of the circulating water of the secondary pipe network interface of the heat exchange station and the user topological position, matching the heat delivery demand after temperature adjustment, and synchronously adjusting the circulating water flow rate to maintain the supply and return flow balance; By controlling the circulating water in the secondary pipe network to be in the same temperature and different flow rate running state, and combining the user topological position data, the change amplitude and rate of the circulating water temperature in the indoor pipe under different flow rates can be accurately judged, and the influence law of the flow rate on the temperature transfer efficiency of the indoor pipe can be determined. By controlling the circulating water to be in the same flow rate and different temperature running state, and synchronously combining the user topological position data, the rising or falling amplitude of the circulating water temperature in the indoor pipe under different temperatures can be clearly judged, and the action mechanism of temperature on the heat transfer effect of the indoor pipe can be mastered. The analysis results of the above two control variables provide accurate variable correlation data for the deep learning model training, and further optimize the accuracy of the mapping relationship model of the circulating water temperature, flow rate and indoor pipe temperature change; By extracting the circulating water temperature, flow rate, indoor pipe temperature and user topological position data from the heating data analysis set to construct the training data set, the mapping relationship model among the three is obtained through deep learning model training, the heating time is calculated combining the initial and target temperature difference of the low-temperature anti-freezing user and the topological position, and then the start control time point is obtained. According to this, the circulating water temperature and flow rate of the secondary network water supply end are accurately adjusted, which not only ensures that the low-temperature anti-freezing user reaches the normal heating temperature on time, but also maintains the supply and return flow balance and the stable heating of the downstream user, effectively avoids the additional heat dissipation caused by the imbalance of heating, and realizes the accurate control of energy loss.

[0011] Further, an energy loss optimization system based on deep learning, the energy loss optimization system comprises a circulating period calculation module, a topological calibration storage module, a heating image construction module, a heating image prediction module, an advance control time calculation module and a control start point calculation module. The cycle period calculation module is configured to obtain secondary pipe network data information and secondary pipe network interface circulating water flow data information, and calculate a circulating period of circulating water in the secondary pipe network; the topology calibration storage module is configured to calibrate and store a topology position of a user according to a connection relationship between the user and the secondary pipe network in a heating service area based on the circulating period; the heating portrait construction module is configured to collect a circulating water temperature of a user's indoor pipe and a heat supply end temperature of a secondary network of a heat exchange station based on the circulating period, and construct a circulating period heating portrait of the heating service area; the heating portrait construction module is configured to collect a circulating water temperature of a user's indoor pipe and a heat supply end temperature of a secondary network of a heat exchange station based on the circulating period, and construct a circulating period heating portrait of the heating service area; the advance regulation time calculation module is configured to calculate a low-temperature freeze-proof heat supply advance regulation time in combination with a topology position of a user and real-time flow data of circulating water of a secondary pipe network interface; and the regulation start point calculation module is configured to train a mapping relationship between a circulating water temperature, a flow rate and a change in a temperature of an indoor pipe through deep learning, and calculate a start regulation time point in combination with the advance regulation time. The cycle period calculation module comprises a pipe network flow acquisition unit and a cycle period calculation unit; the pipe network flow acquisition unit is configured to obtain pipe diameter and length data of each pipe section of a secondary pipe network and circulating water flow data of a secondary pipe network interface; and the cycle period calculation unit is configured to calculate a total volume of circulating water in the secondary pipe network and an average value of circulating water flow, and then obtain a circulating period. The topology calibration storage module comprises a service area detection unit and a topology calibration unit; the service area detection unit is configured to determine a connection relationship between a user in a heating service area and a branch of a secondary pipe network and a corresponding pipe network branch pipe section based on a circulating period; and the topology calibration unit is configured to extract a geographic coordinate of the user and a node number of the pipe network branch based on a heating completion drawing, and calibrate a topology position of the user and store the topology position in association with user identification information. The heating portrait construction module comprises a temperature and heat quantity acquisition and calculation unit and a portrait construction unit; the temperature and heat quantity acquisition and calculation unit is configured to acquire temperature data at a preset time interval within a circulating period, and calculate a circulating water temperature ratio of a single user's indoor pipe, a heating quantity and a total heating quantity of the heating service area; and the portrait construction unit is configured to associate an acquisition time with a total heating quantity of circulating water of an indoor pipe of the heating service area, form a data set of a change in the total heating quantity over time within the circulating period, and construct a heating portrait. The heating portrait prediction module comprises a historical data arrangement unit and a predicted portrait generation unit; the historical data arrangement unit is configured to obtain a historical circulating period heating portrait, extract relevant data and arrange the data into a heating data analysis set in association with a circulating period serial number; and the predicted portrait generation unit is configured to input the heating data analysis set into a time series prediction model, learn a historical rule and output a predicted heating portrait of a next circulating period. The advance regulation time calculation module comprises a conversion data extraction unit and a regulation time derivation unit; the conversion data extraction unit is used to call low-temperature anti-freezing operation state user identification information, extract the in-house pipe compliance time stamp and heat exchange station regulation start time stamp when switching from low-temperature anti-freezing to normal heating, and the regulation time derivation unit is used to calculate the transmission time of circulating water flowing through the corresponding pipe network path, and derive the advance regulation time when switching from low-temperature anti-freezing to heating. The regulation start point calculation module comprises a mapping model training unit and a start point derivation unit; the mapping model training unit is used to extract data from the heating data analysis set and associate and arrange the data into a training data set, input a deep learning model to train a mapping relationship model, and the start point derivation unit is used to calculate the difference between the initial temperature and the target temperature of the in-house pipe of the low-temperature anti-freezing user, combine the mapping relationship model and the advance regulation time to derive the regulation time point.

[0012] Compared with the prior art, the present application has the following advantages: 1、The present application calculates the circulation period by obtaining secondary pipe network data and circulating water flow data, combines with the heating completion drawing to calibrate the user topology position and associate and store, clearly defines the time reference and spatial topology relationship of the heating system, solves the problem of lack of unified time dimension and accurate spatial positioning in the prior art of heating data collection and analysis, provides accurate data support for subsequent heating portrait construction and regulation parameter calculation, and improves the accuracy and pertinence of heating data analysis.

[0013] 2、The present application constructs a heating portrait based on the circulation period to collect temperature data, integrates historical portraits to form a data analysis set, and uses a time series prediction model to obtain a predicted heating portrait of the next circulation period, realizes the forward-looking prediction of heating demand, overcomes the limitation of real-time data and lack of advance planning in the prior art of heating regulation, provides a prediction basis for heat exchange station regulation decision, and guarantees the forward-looking and rationality of heating regulation.

[0014] 3、The present application trains a mapping relationship model of circulating water temperature, flow rate and in-house pipe temperature change by deep learning, calculates the advance regulation time and start regulation point in combination with the user topology position and circulating water flow data, accurately adjusts the circulating water parameters, solves the problem of downstream user heating imbalance and increased energy loss when switching from low-temperature anti-freezing to heating, maintains system balance while ensuring user heating compliance, effectively reduces unorganized heat dissipation and energy waste, and improves energy utilization efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 It is a flowchart of the energy loss optimization method based on deep learning of the present application; Figure 2 It is a structural schematic diagram of the energy loss optimization system based on deep learning of the present application. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: As Figure 1 As shown, the present invention provides a technical solution, an energy loss optimization method based on deep learning, the energy loss optimization method comprising the following steps: Step S1: Obtain the secondary pipeline network data information connected to the heat exchange station and the circulating water flow data information of the secondary pipeline network interface, and perform analysis and calculation to obtain the circulation cycle of the circulating water in the secondary pipeline network; Step S1-1: Obtain the pipe diameter data and corresponding pipe length data of each pipe section of the secondary pipe network connected to the heat exchange station through the heating as-built drawings; calculate the cross-sectional area of ​​the corresponding pipe section based on the pipe diameter, and calculate the volume of a single pipe section filled with circulating water based on the length of the pipe section; sum up the volumes of all pipe sections filled with circulating water in the secondary pipe network to obtain the total volume of circulating water in the secondary pipe network. Step S1-2: Obtain circulating water flow data at the secondary pipe network interface using a flow sensor. The secondary pipe network interface includes a secondary pipe network supply end interface and a secondary pipe network return end interface. Calculate the average circulating water flow rate at the secondary pipe network supply end interface and the secondary pipe network return end interface. Divide the total volume of circulating water in the secondary pipe network obtained in step S1-1 by the average circulating water flow rate to obtain the circulation cycle of the circulating water in the secondary pipe network. In practice, the structural parameters of each section of the secondary pipeline network are accurately extracted from the heating as-built drawings. Based on this, the total volume of circulating water is calculated. At the same time, flow sensors are used to obtain the flow data of the supply and return water interfaces of the secondary network in real time. The circulation cycle is determined by the ratio of volume to average flow. It is important to ensure the completeness of the extracted pipeline network section parameters and avoid missing branch sections. At the same time, it is necessary to ensure the stable operation of the flow sensors to obtain reliable flow data, so as to provide a reliable time reference for subsequent steps.

[0018] Step S2: Detect the heating service area of ​​the secondary pipeline network according to the cycle, and perform topology location calibration and topology location association storage for users in the heating service area based on the heating as-built drawings; Step S2-1, taking the cycle period obtained in step S1 as a reference, detecting the heating service area to which the secondary pipe network belongs to determine the connection relationship between the users in the heating service area and the secondary pipe network branches and the pipe network branch pipe sections corresponding to each user, and the heating service area represents the area in which the secondary pipe network connected with the heat exchange station delivers circulating water for heating; Step S2-2, based on the heating completion drawings, extracting the geographic coordinates of each user in the heating service area and the corresponding secondary pipe network branch node number, and calibrating the topological position of each user in the secondary pipe network; associating the calibrated topological position of the user with the user identification information and storing it in the database; In specific implementation, the connection logic between the users in the heating service area and the secondary pipe network branches is determined by detection, and then the geographic coordinates of the users and the corresponding pipe network node numbers are accurately extracted according to the heating completion drawings to complete the topological position calibration and associated storage; it is necessary to check the uniqueness of the corresponding relationship between the users and the pipe network branch pipe sections to ensure the accuracy of the association between the topological position and the user identification information and avoid deviation in subsequent location-based data analysis.

[0019] Step S3, based on the cycle period, detecting the circulating water temperature of the indoor pipe of each user in the heating service area, and combining the circulating water temperature analysis in the secondary pipe network interface of the heat exchange station to construct the cycle period heating portrait of the heating service area; Step S3-1, based on the cycle period obtained in step S1, collecting the circulating water temperature in the indoor pipe of each user in the heating service area at a preset time interval within the cycle period, the preset time interval being not more than the cycle period; at the same time, collecting the circulating water temperature of the water supply end interface of the secondary pipe network of the heat exchange station to obtain the density and specific heat capacity data of the circulating water in the secondary pipe network; calculating the ratio of the circulating water temperature in the indoor pipe of each user to the circulating water temperature of the water supply end interface of the secondary pipe network, which is recorded as the single-user indoor pipe circulating water temperature ratio; calculating the indoor pipe circulating water heating capacity of each user, which is equal to the product of the circulating water density, the specific heat capacity, the indoor pipe circulating water flow of the corresponding user and the single-user indoor pipe circulating water temperature ratio; adding the single-user indoor pipe circulating water temperature ratios of all users and dividing by the total number of users in the heating service area to obtain the average indoor pipe circulating water temperature ratio corresponding to the time interval; adding the indoor pipe circulating water heating capacities of all users to obtain the total indoor pipe circulating water heating capacity of the heating service area corresponding to the time interval; Step S3-2, record the collection time of each preset time interval, associate the collection time with the total heat supply of the circulating water in the heat supply service area, and form a data set of the total heat supply of the circulating water in the heat supply service area in the circulation period; based on the data set, the circulation period heat supply image of the heat supply service area is constructed, which contains the mapping relationship between each collection time and the corresponding total heat supply of the circulating water in the heat supply service area in the circulation period, and the cumulative value of the total heat supply of the circulating water in the heat supply service area in the circulation period; In specific implementation, a reasonable temperature collection interval is set based on the circulation period, the circulating water temperature of the user's indoor pipe and the secondary pipe network water supply end interface is synchronously acquired, the heat supply of a single user and the whole service area is calculated in combination with the circulating water physical parameters, and then the heat supply image is constructed by associating the collection time with the total heat supply; it should be noted that the setting of the collection interval should be able to completely cover the temperature change in the circulation period, and at the same time ensure that the installation position of the temperature collection point is representative, so as to ensure that the calculated heat supply data can truly reflect the heat supply state.

[0020] Step S4, a plurality of circulation period heat supply images of the heat exchange station in the history are acquired, and a heat supply data analysis set of the heat supply service area is constructed; according to the heat supply data analysis set and in combination with a time series prediction model, data processing is performed, and a heat supply image of the heat supply service area in the next circulation period is predicted, which is recorded as a predicted circulation period heat supply image; Step S4-1, a plurality of circulation period heat supply images generated in the historical operation process of the heat exchange station are acquired, and the collection time, the total heat supply of the circulating water in the heat supply service area and the cumulative value of the total heat supply in the circulation period contained in each historical circulation period heat supply image are extracted; all the extracted historical circulation period related data are associated and arranged according to the circulation period serial number, and a heat supply data analysis set of the heat supply service area is constructed; Step S4-2, the constructed heat supply data analysis set of the heat supply service area is input into a time series prediction model for data processing, and the time series prediction model learns the time variation law and the cumulative value variation law of the total heat supply of the circulating water in the heat supply service area in the historical circulation period; a heat supply image of the heat supply service area in the next circulation period is output and recorded as a predicted circulation period heat supply image, which contains the mapping relationship between each preset collection time and the corresponding predicted total heat supply of the circulating water in the heat supply service area in the next circulation period, and the predicted cumulative value of the total heat supply of the circulating water in the heat supply service area in the next circulation period; In specific implementation, key data in the historical cycle period heating image is integrated to form an analysis set, a time series prediction model is used to learn the historical heating change rule, and then a prediction image of the next cycle period is output. It should be noted that continuous and complete historical image data is screened to avoid affecting the model learning effect due to data loss, and attention is paid to the adaptability of the model to the heating rule under different seasons or special weather to ensure the rationality of the prediction result.

[0021] Step S5, according to the user topological position of the heating service area combined with the real-time flow data information analysis and calculation of the secondary pipe network interface of the heat exchange station, the low-temperature freeze-proof conversion heating advance control time of the heat exchange station when the low-temperature freeze-proof running state in the heating service area changes is obtained; Step S5-1, user identification information in the low-temperature freeze-proof running state in the heating service area is called from the database, and the corresponding secondary pipe network branch pipe section of the user in the low-temperature freeze-proof running state is determined combined with the user topological position calibrated in step S2; the time stamp when the indoor pipe circulating water reaches the standard temperature when the user in the low-temperature freeze-proof running state in the historical cycle period changes from the low-temperature freeze-proof running state to the normal heating state, and the time stamp when the heat exchange station secondary pipe network water supply end interface starts to control in the corresponding historical cycle period are extracted from the heating data analysis set, and the control represents the adjustment operation of the circulating water temperature and the circulating water flow rate of the heat exchange station secondary pipe network water supply end interface; Step S5-2, based on the topological position of the user in the low-temperature freeze-proof running state, the pipe network path length between the corresponding secondary pipe network branch pipe section and the heat exchange station secondary pipe network water supply end interface is determined, and the transmission time of the circulating water flowing through the pipe network path length is calculated combined with the real-time flow data of the heat exchange station secondary pipe network interface; the difference between the extracted time stamp when the user indoor pipe circulating water reaches the standard temperature and the heat exchange station control start time stamp is subtracted by the transmission time, and the low-temperature freeze-proof conversion heating advance control time of the heat exchange station when the low-temperature freeze-proof running state in the heating service area changes is obtained; In specific implementation, the low-temperature freeze-proof user information is called from the database, the corresponding pipe network path is determined combined with the topological position, the advance control time is derived by extracting the time stamp of the historical switching state and calculating the transmission time; it should be noted that the accuracy of the historical time stamp extraction is ensured to match the control operation of the corresponding cycle period, and the calculation of the pipe network path length should strictly follow the topological connection relationship to avoid calculation error of the transmission time due to path deviation.

[0022] Step S6, the heating data analysis set of the heating service area is processed by data training through deep learning, the temperature, flow rate and temperature change relationship of the circulating water in the secondary pipe network and the indoor pipe are mapped, and the low-temperature freeze-proof conversion heating advance control time is analyzed to obtain the time point of the low-temperature freeze-proof conversion heating start control of the heat exchange station; Step S6-1: Extract the temperature and flow rate data of circulating water in the secondary pipe network during historical cycles from the heating data analysis set, as well as the corresponding time-based circulating water temperature data of each user's inlet pipe within the heating service area; combine the extracted data with user topology location data and organize them according to circulating water temperature, circulating water flow rate, inlet pipe circulating water temperature, and user topology location to form a training dataset; input the training dataset into a preset deep learning model for training, and output a mapping relationship model between the temperature and flow rate of circulating water in the secondary pipe network and the changes in inlet pipe circulating water temperature; Step S6-2: Obtain the initial temperature data of the circulating water in the user's inlet pipe under low-temperature anti-freeze operation and the target temperature data of the circulating water in the inlet pipe under the preset normal heating state within the heating service area, and calculate the temperature difference between the two; input the temperature difference into the mapping relationship model, and combine it with the topological location data of the corresponding user to output the heating time required for the circulating water in the user's inlet pipe to rise from the initial temperature to the target temperature; subtract the heating time from the low-temperature anti-freeze to heating advance control time to obtain the low-temperature anti-freeze to heating start control time of the heat exchange station, where the low-temperature anti-freeze to heating start control time is represented by the time when the heat exchange station starts to adjust the temperature of the user's inlet pipe. The circulating water temperature and flow rate at the secondary network water supply end interface are adjusted to ensure that the circulating water in the user's inlet pipe, which is in low-temperature antifreeze operation, reaches the target temperature of normal heating at a predetermined time. The total heat demand in the heating profile of the cycle is predicted, and the temperature rise demand output by the mapping relationship model is combined with the temperature rise demand. The circulating water temperature at the secondary network water supply end interface is gradually adjusted according to the predetermined time interval within the cycle. When adjusting the circulating water flow rate, based on the real-time flow data of the circulating water at the secondary network interface of the heat exchange station and the user topology location, the heat delivery demand after temperature adjustment is matched, and the circulating water flow rate is adjusted synchronously to maintain the balance of supply and return flow. In practical implementation, multi-dimensional data is extracted from the heating data analysis set to construct a training set. A deep learning model is used to train the model to obtain the mapping relationship between circulating water parameters and the temperature change of the inlet pipe. Then, the temperature difference of users with low temperature and freeze protection and the topological location are combined to calculate the heating time. Finally, the start-up control time point is determined and the temperature rate is adjusted synchronously. Attention should be paid to the temporal correlation of each parameter in the training data to ensure that the model can accurately capture the dynamic change pattern. At the same time, the temperature rate adjustment should be carried out in a coordinated manner to avoid the imbalance of supply and return flow caused by unilateral adjustment, which will affect the overall heating stability.

[0023] Example 2, as Figure 2 As shown, the present invention provides an energy loss optimization system based on deep learning. The energy loss optimization system includes a cycle calculation module, a topology calibration and storage module, a heating profile construction module, a heating profile prediction module, an advance control time calculation module, and a control start point calculation module. The cycle period calculation module is configured to obtain secondary pipe network data information and secondary pipe network interface circulating water flow data information, and calculate a circulating period of circulating water in the secondary pipe network; the topology calibration storage module is configured to calibrate and store a topology position of a user according to a connection relationship between the user and the secondary pipe network in a heating service area based on the circulating period; the heating portrait construction module is configured to collect a circulating water temperature of a user's household pipe and a heat supply end temperature of a secondary network of a heat exchange station based on the circulating period, and construct a circulating period heating portrait of the heating service area; the heating portrait construction module is configured to collect a circulating water temperature of a user's household pipe and a heat supply end temperature of a secondary network of a heat exchange station based on the circulating period, and construct a circulating period heating portrait of the heating service area; the advance regulation time calculation module is configured to calculate a low-temperature freeze-proof heating advance regulation time in combination with a topology position of a user and real-time flow data of circulating water of a secondary pipe network interface; and the regulation start point calculation module is configured to train a mapping relationship between a circulating water temperature, a flow rate and a household pipe temperature change through deep learning, and calculate a regulation start time point in combination with the advance regulation time. The cycle period calculation module includes a pipe network flow acquisition unit and a cycle period calculation unit; the pipe network flow acquisition unit is configured to obtain pipe diameter and length data of each pipe section of the secondary pipe network and circulating water flow data of a secondary pipe network interface; and the cycle period calculation unit is configured to calculate a total volume of circulating water in the secondary pipe network and an average value of circulating water flow, and further obtain a circulating period. The topology calibration storage module includes a service area detection unit and a topology calibration unit; the service area detection unit is configured to determine a connection relationship between a user in a heating service area and a branch of the secondary pipe network and a corresponding pipe network branch pipe section based on the circulating period; and the topology calibration unit is configured to extract a user geographic coordinate and a pipe network branch node number based on a heating completion drawing, calibrate a topology position of the user and store the topology position in association with user identification information. The heating portrait construction module includes a temperature and heat quantity acquisition and calculation unit and a portrait construction unit; the temperature and heat quantity acquisition and calculation unit is configured to acquire temperature data at a preset time interval within the circulating period, and calculate a single user household pipe circulating water temperature ratio, a heating quantity and a total heating quantity of the heating service area; and the portrait construction unit is configured to associate an acquisition time with a total heating quantity of a household pipe circulating water of the heating service area, form a data set of the total heating quantity changing with time within the circulating period and construct a heating portrait. The heating portrait prediction module includes a historical data arrangement unit and a predicted portrait generation unit; the historical data arrangement unit is configured to obtain a historical circulating period heating portrait, extract relevant data and arrange the data into a heating data analysis set in association with a circulating period serial number; and the predicted portrait generation unit is configured to input the heating data analysis set into a time series prediction model, learn a historical rule and output a predicted heating portrait of a next circulating period. The advance regulation time calculation module comprises a conversion data extraction unit and a regulation time derivation unit; the conversion data extraction unit is used to call low-temperature anti-freezing operation state user identification information, extract in-house pipe compliance time stamp and heat exchange station regulation start time stamp when switching from low-temperature anti-freezing to normal heating; the regulation time derivation unit is used to calculate the transmission time of circulating water flowing through the corresponding pipe network path, and derive the advance regulation time of switching from low-temperature anti-freezing to heating. The regulation start point calculation module comprises a mapping model training unit and a start point derivation unit; the mapping model training unit is used to extract data from the heating data analysis set and associate and arrange the training data set, input the deep learning model to train the mapping relationship model; the start point derivation unit is used to calculate the difference between the initial temperature and the target temperature of the low-temperature anti-freezing user in-house pipe, combine the mapping relationship model and the advance regulation time to derive the regulation time point.

[0024] It is apparent to those skilled in the art that the present application is not limited to the details of the foregoing exemplary embodiments, and that the present application can be implemented in other particular forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, and the scope of the present application should be defined by the appended claims rather than the above description, and it is intended to include all changes falling within the meaning and scope of the equivalent elements of the claims.

Claims

1. A deep learning-based energy loss optimization method, characterized in that: The energy consumption optimization method comprises the following steps: Step S1, obtaining and analyzing the secondary pipe network data information connected with the heat exchange station and the circulating water flow data information of the secondary pipe network interface to obtain the circulating period of the circulating water in the secondary pipe network; Step S2, detecting the heat supply service area to which the secondary pipe network belongs according to the circulating period, and performing topological position calibration and topological position correlation storage of the users in the heat supply service area based on the heat supply completion drawings; Step S3, detecting the circulating water temperature of the indoor pipe of each user in the heat supply service area based on the circulating period, and constructing a circulating period heat supply image of the heat supply service area by combining the circulating water temperature analysis of the secondary pipe network interface of the heat exchange station; Step S4, obtaining a plurality of circulating period heat supply images of the heat exchange station, constructing a heat supply data analysis set of the heat supply service area, and performing data processing on the heat supply data analysis set combined with a time series prediction model to predict a heat supply image of the heat supply service area in the next circulating period, which is recorded as a predicted circulating period heat supply image; Step S5, obtaining the low-temperature freeze-proof advance control time of the heat exchange station when the low-temperature freeze-proof operation state of the heat supply service area changes, by analyzing and calculating the real-time flow data information of the circulating water of the secondary pipe network interface of the heat exchange station based on the topological position of the users in the heat supply service area; Step S6, performing data training processing on the heat supply data analysis set of the heat supply service area by deep learning to map the temperature and flow rate of the circulating water in the secondary pipe network and the temperature change relationship of the indoor pipe, and combining the low-temperature freeze-proof advance control time to obtain the time point of the low-temperature freeze-proof to heat supply start control of the heat exchange station. 2.The energy loss optimization method based on deep learning of claim 1, wherein: The specific steps of step S1 are as follows: Step S1-1, obtaining the pipe diameter data and the corresponding pipe length data of each pipe section of the secondary pipe network connected with the heat exchange station through the heat supply completion drawings; calculating the cross-sectional area of each pipe section according to the pipe diameter, and calculating the volume of a single pipe section filled with circulating water combined with the pipe length to obtain the total volume of the circulating water in the secondary pipe network by accumulating the volumes of all pipe sections filled with circulating water; Step S1-2, obtaining the circulating water flow data of the secondary pipe network interface through a flow sensor, wherein the secondary pipe network interface includes a secondary pipe network water supply end interface and a secondary pipe network return end interface; calculating the average value of the circulating water flow of the secondary pipe network water supply end interface and the secondary pipe network return end interface; dividing the total volume of the circulating water in the secondary pipe network obtained in step S1-1 by the average value of the circulating water flow to obtain the circulating period of the circulating water in the secondary pipe network.

3. The energy loss optimization method based on deep learning according to claim 2, characterized in that: The specific steps of step S2 are as follows: Step S2-1, selecting the circulating period obtained in step S1 as a reference to detect the heat supply service area to which the secondary pipe network belongs, determine the connection relationship between the users in the heat supply service area and the branch pipes of the secondary pipe network, and determine the pipe section corresponding to each user, wherein the heat supply service area represents an area heated by the circulating water delivered by the secondary pipe network connected with the heat exchange station; Step S2-2, based on the heating completion drawings, extracting the geographic coordinates of each user in the heating service area and the corresponding secondary pipe network branch node number, and marking the topological position of each user in the secondary pipe network; associating the marked topological position of the user with the user identification information and storing it in the database.

4. The energy loss optimization method based on deep learning according to claim 3, characterized in that: The specific steps of the step S3 are as follows: Step S3-1, based on the cycle period obtained in step S1, collecting the circulating water temperature in the indoor pipe of each user in the heating service area at a preset time interval within the cycle period, the preset time interval being not more than the cycle period; simultaneously collecting the circulating water temperature of the secondary pipe network water supply end interface of the heat exchange station, obtaining the density and specific heat capacity data of the circulating water in the secondary pipe network; calculating the ratio of the circulating water temperature in the indoor pipe of each user to the circulating water temperature of the secondary pipe network water supply end interface, denoted as the single-user indoor pipe circulating water temperature ratio; calculating the indoor pipe circulating water heating capacity of each user, which is equal to the product of the circulating water density, specific heat capacity, indoor pipe circulating water flow of the corresponding user and single-user indoor pipe circulating water temperature ratio; adding the single-user indoor pipe circulating water temperature ratios of all users and dividing by the total number of users in the heating service area to obtain the average indoor pipe circulating water temperature ratio corresponding to the time interval; adding the single-user indoor pipe circulating water heating capacities of all users to obtain the total indoor pipe circulating water heating capacity of the heating service area corresponding to the time interval; Step S3-2, recording the collection time of each preset time interval, associating the collection time with the total indoor pipe circulating water heating capacity of the heating service area corresponding to the time interval to form a data set of the total indoor pipe circulating water heating capacity of the heating service area changing with time within the cycle period; Based on the data set, a cycle heating portrait of the heating service area is constructed, which includes the mapping relationship between each collection time and the corresponding total indoor pipe circulating water heating capacity of the heating service area within the cycle period, and the cumulative value of the total indoor pipe circulating water heating capacity of the heating service area within the cycle period.

5. The energy loss optimization method based on deep learning according to claim 4, characterized in that: The specific steps of the step S4 are as follows: Step S4-1, obtaining a plurality of cycle heating portraits generated in the historical operation process of the heat exchange station, extracting the collection time, the total indoor pipe circulating water heating capacity of the heating service area and the cumulative value of the total heating capacity within the cycle period included in each historical cycle heating portrait; All the extracted historical cycle related data are associated and sorted according to the cycle number to construct a heating data analysis set of the heating service area; Step S4-2, inputting the constructed heating data analysis set of the heating service area into the time series prediction model for data processing, and learning the time variation law and cumulative value variation law of the total indoor pipe circulating water heating capacity of the heating service area within the historical cycle through the time series prediction model; The output obtains a heat supply image of the heat supply service area in the next cycle period, recorded as a predicted cycle period heat supply image, which includes a mapping relationship between each preset collection time in the next cycle period and a corresponding predicted total heat supply amount of the household pipe circulating water in the heat supply service area, and a predicted cumulative value of the total heat supply amount of the household pipe circulating water in the heat supply service area in the next cycle period.

6. The energy loss optimization method based on deep learning according to claim 5, characterized in that: The specific steps of the step S5 are as follows: Step S5-1, user identification information in the low-temperature freeze-proof operation state in the heat supply service area is called from the database, and the corresponding secondary pipe network branch pipe section of the user in the low-temperature freeze-proof operation state is determined in combination with the user topology position calibrated in the step S2; a time stamp at which the household pipe circulating water of the user in the low-temperature freeze-proof operation state reaches the standard temperature when the user changes from the low-temperature freeze-proof operation state to the normal heat supply state in the historical cycle period is extracted from the heat supply data analysis set, and a time stamp at which the heat exchange station secondary pipe network water supply end interface starts to be controlled in the corresponding historical cycle period is extracted, and the control represents an adjustment operation on the circulating water temperature and the circulating water flow rate of the heat exchange station secondary pipe network water supply end interface; Step S5-2, the pipe network path length between the secondary pipe network branch pipe section corresponding to the user in the low-temperature freeze-proof operation state and the heat exchange station secondary pipe network water supply end interface is determined based on the topology position of the user in the low-temperature freeze-proof operation state, and the transmission time of the circulating water flowing through the pipe network path length is calculated in combination with the real-time flow data of the circulating water of the heat exchange station secondary pipe network interface; The difference between the extracted time stamp at which the household pipe circulating water of the user reaches the standard temperature and the time stamp at which the heat exchange station is controlled is subtracted by the transmission time, so as to obtain the time at which the heat exchange station is controlled in advance when the low-temperature freeze-proof operation state changes in the heat supply service area, recorded as a low-temperature freeze-proof to heat supply advance control time.

7. The energy loss optimization method based on deep learning according to claim 6, characterized in that: The specific steps of the step S6 are as follows: Step S6-1, the temperature data and the flow rate data of the circulating water in the secondary pipe network in the historical cycle period are extracted from the heat supply data analysis set, and the household pipe circulating water temperature data of each user in the heat supply service area at the corresponding time are extracted; the extracted data are associated and arranged according to the circulating water temperature, the circulating water flow rate, the household pipe circulating water temperature, and the user topology position to form a training data set; the training data set is input into a preset deep learning model for training, and a mapping relationship model of the temperature, the flow rate, and the household pipe circulating water temperature change of the circulating water in the secondary pipe network is output; Step S6-2, the initial temperature data of the household pipe circulating water of the user in the low-temperature freeze-proof operation state in the heat supply service area and the target temperature data of the household pipe circulating water in the preset normal heat supply state are obtained, and the temperature difference between the two is calculated; the temperature difference is input into the mapping relationship model, and the topology position data of the corresponding user are combined to output the heating time required for the household pipe circulating water of the user to rise from the initial temperature to the target temperature. The low-temperature anti-freezing heat supply switching time is subtracted from the temperature rising time to obtain a low-temperature anti-freezing heat supply switching starting time point of the heat exchange station, which represents a specific time when the circulating water temperature and flow rate of the secondary pipe network supply end interface are adjusted by the heat exchange station so that the circulating water of the user indoor pipe in the low-temperature anti-freezing operation state reaches the normal heat supply state target temperature at the preset time.

8. A deep learning-based energy loss optimization system applied to the deep learning-based energy loss optimization method of any one of claims 1-7. The energy consumption optimization system comprises a cycle period calculation module, a topology calibration storage module, a heat supply image construction module, a heat supply image prediction module, an advance regulation time calculation module and a regulation starting point calculation module. The cycle period calculation module is configured to acquire secondary pipe network data information and secondary pipe network interface circulating water flow data information, and calculate the cycle period of circulating water in the secondary pipe network; the topology calibration storage module is configured to calibrate and store the user topology position according to the connection relationship between the user and the secondary pipe network in the heat supply service area; and the heat supply image construction module is configured to acquire the circulating water temperature of the user indoor pipe and the secondary network supply end temperature of the heat exchange station based on the cycle period, and construct a cycle period heat supply image of the heat supply service area. The heat supply image construction module is configured to acquire the circulating water temperature of the user indoor pipe and the secondary network supply end temperature of the heat exchange station based on the cycle period, and construct a cycle period heat supply image of the heat supply service area. The advance regulation time calculation module is configured to calculate the low-temperature anti-freezing heat supply switching advance regulation time in combination with the user topology position and the real-time flow data of the secondary pipe network interface circulating water. The regulation starting point calculation module is configured to train the mapping relationship between the circulating water temperature, flow rate and indoor pipe temperature change through deep learning, and calculate the starting regulation time point in combination with the advance regulation time.

9. The energy consumption optimization system based on deep learning according to claim 8, characterized in that: The cycle period calculation module comprises a pipe network flow acquisition unit and a cycle period calculation unit; the pipe network flow acquisition unit is configured to acquire the pipe diameter and length data of each pipe section of the secondary pipe network and the circulating water flow data of the secondary pipe network interface; and the cycle period calculation unit is configured to calculate the total volume of circulating water in the secondary pipe network and the average value of the circulating water flow, and further obtain the cycle period. The topology calibration storage module comprises a service area detection unit and a topology calibration unit; the service area detection unit is configured to determine the connection relationship between the user and the secondary pipe network branch in the heat supply service area and the corresponding pipe network branch pipe section based on the cycle period; and the topology calibration unit is configured to extract the user geographic coordinates and pipe network branch node number based on the heat supply completion drawing, calibrate the user topology position and store the user topology position in association with the user identification information.

10. The energy consumption optimization system based on deep learning according to claim 8, characterized in that: The heat supply image construction module comprises a heat quantity collection and calculation unit and an image construction unit; the heat quantity collection and calculation unit is configured to collect temperature data at preset time intervals in a cycle period, calculate the circulating water temperature ratio of a user's indoor pipe, heat supply quantity, and total heat supply quantity of a heat supply service area; and the image construction unit is configured to associate the collection time and the total heat supply quantity of the circulating water of the indoor pipe of the heat supply service area, form a data set of the total heat supply quantity changing with time in the cycle period, and construct a heat supply image; The heat supply image prediction module comprises a historical data arrangement unit and a predicted image generation unit; the historical data arrangement unit is configured to obtain a historical cycle period heat supply image, extract relevant data, and arrange the data into a heat supply data analysis set according to the cycle period sequence number; and the predicted image generation unit is configured to input the heat supply data analysis set into a time sequence prediction model, learn the historical law, and output a predicted heat supply image of the next cycle period.

11. The energy consumption optimization system based on deep learning according to claim 8, characterized in that: The advance regulation time calculation module comprises a conversion data extraction unit and a regulation time derivation unit; the conversion data extraction unit is configured to call the user identification information in the low-temperature anti-freezing operation state, extract the time stamp of the indoor pipe reaching the standard when the low-temperature anti-freezing is switched to normal heat supply and the time stamp of the regulation start of the heat exchange station; The regulation time derivation unit is configured to calculate the transmission time of the circulating water flowing through the corresponding pipe network path, and derive the advance regulation time when the low-temperature anti-freezing is switched to heat supply; The regulation start point calculation module comprises a mapping model training unit and a start point derivation unit; The mapping model training unit is configured to extract data from the heat supply data analysis set and arrange the data into a training data set, input the data into a deep learning model, and train the model to obtain a mapping relationship model; The start point derivation unit is configured to calculate the difference between the initial temperature and the target temperature of the indoor pipe of the low-temperature anti-freezing user, combine the mapping relationship model and the advance regulation time, and derive the regulation start time point.

Images