A machine learning-based auxiliary analysis method and system for optimizing heating energy efficiency

Abstract

An embodiment of the present application provides a method and system for assisting in analyzing the optimization of heating energy efficiency based on machine learning. First, by analyzing the operation trajectory, scheduling knowledge, and their collaborative data of an exemplary heating operation system, two machine learning networks are trained to estimate collaborative performance knowledge points and evaluate performance respectively. Subsequently, the parameters of these two networks are learned using the data of the learning heating operation system to generate a more optimized heating collaborative analysis network. Through cyclic parameter learning, the network performance is continuously optimized, and finally a collaborative performance knowledge point estimation network is output. This collaborative performance knowledge point estimation network can be used to assist in analyzing the energy efficiency optimization of the target heating operation system, providing effective optimization suggestions, thereby improving the energy efficiency level of the heating system.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular, to a method and system for auxiliary analysis of heating energy efficiency optimization based on machine learning. Background Art

[0002] In the operation and management of heating systems, energy efficiency optimization has always been an important and complex issue. Traditional energy efficiency optimization methods usually rely on experts' experience and intuition, or adjust operating parameters and scheduling strategies through trial and error methods. These methods are not only inefficient but also difficult to adapt to the dynamic changes in the operating environment of heating systems.

[0003] In recent years, with the rapid development of artificial intelligence technology, especially the application of machine learning methods in data analysis and prediction, new solutions have been provided for heating energy efficiency optimization. However, current solutions often only focus on a single operating parameter or scheduling strategy of the heating system and ignore the impact of synergy on energy efficiency. Summary of the Invention

[0004] In order to at least overcome the above deficiencies in the prior art, the purpose of the embodiments of the present application is to provide a method and system for auxiliary analysis of heating energy efficiency optimization based on machine learning.

[0005] According to one aspect of the present application, there is provided a method for auxiliary analysis of heating energy efficiency optimization based on machine learning, the method comprising:

[0006] Obtaining the example heating operation trajectory features, example heating scheduling knowledge features of an example heating operation system, and example synergy data between the example heating operation system and the example heating scheduling knowledge features;

[0007] Loading the example synergy data into a first machine learning network for collaborative performance knowledge point estimation to generate example collaborative performance knowledge points;

[0008] Loading the example heating operation trajectory features and the example heating scheduling knowledge features into a second machine learning network for performance evaluation and analysis to generate example collaborative performance evaluation data;

[0009] Based on the difference between the example collaborative performance evaluation data and the example collaborative performance knowledge points, performing parameter learning on the second machine learning network to generate a basic heating collaborative analysis network;

[0010] Based on the learned heating operation trajectory features, learned heating scheduling knowledge features, learned collaborative data of the learning heating operation system, and the basic heating collaborative analysis network, parameter learning is performed on the first machine learning network to generate a basic collaborative performance knowledge point estimation network; the learned collaborative data is the collaborative data between the learning heating operation system and the learned heating scheduling knowledge features; the learning heating operation system and the exemplary heating operation system are heating operation systems of the same type, the learned heating operation trajectory features and the exemplary heating operation trajectory features are features of the same dimension, the learned heating scheduling knowledge features and the exemplary heating scheduling knowledge features are features of the same dimension, and the learned collaborative data and the exemplary collaborative data are data of the same dimension.

[0011] Perform cyclic parameter learning on the basic heating collaborative analysis network and the basic collaborative performance knowledge point estimation network, and output the basic collaborative performance knowledge point estimation network when the parameter learning is completed as the collaborative performance knowledge point estimation network; the cyclic parameter learning includes cyclic operations: locking the network parameter information of the basic collaborative performance knowledge point estimation network to perform parameter learning on the basic heating collaborative analysis network and locking the network parameter information of the basic heating collaborative analysis network to perform parameter learning on the basic collaborative performance knowledge point estimation network.

[0012] In a possible implementation manner of the first aspect, the method further includes:

[0013] Obtain the target collaborative data of the target heating operation system and the target heating scheduling knowledge features;

[0014] Load the target collaborative data into the collaborative performance knowledge point estimation network for collaborative performance knowledge point estimation to generate target estimated collaborative performance knowledge points;

[0015] Based on the target estimated collaborative performance knowledge points, perform auxiliary analysis on the heating energy efficiency optimization of the target heating operation system to generate an auxiliary analysis result of the heating energy efficiency optimization;

[0016] For example, in a possible implementation manner of the first aspect, the step of performing auxiliary analysis on the heating energy efficiency optimization of the target heating operation system based on the target estimated collaborative performance knowledge points to generate an auxiliary analysis result of the heating energy efficiency optimization includes:

[0017] Generate a corresponding collaborative performance parameter knowledge graph based on the target estimated collaborative performance knowledge points;

[0018] Analyze the collaborative performance parameter knowledge graph to determine the system performance defect point data existing in the target heating operation system;

[0019] Based on the data of system performance defect points existing in the target heating operation system, conduct auxiliary analysis on the energy efficiency optimization of the target heating operation system to generate the results of auxiliary analysis on energy efficiency optimization of heating;

[0020] For example, in a possible implementation manner of the first aspect, the step of generating the corresponding collaborative performance parameter knowledge graph based on the target estimated collaborative performance knowledge points includes:

[0021] Extract key feature data from each target estimated collaborative performance knowledge point;

[0022] Based on the key feature data, use the association rule mining algorithm to analyze the association relationships between the various data nodes associated with the target estimated collaborative performance knowledge points to determine the potential association rules between the data nodes;

[0023] Based on the potential association rules, construct the network structure diagram of the target estimated collaborative performance knowledge points, analyze the corresponding key nodes and paths in the key feature data in the network structure diagram to identify the key factors affecting the energy efficiency performance of the heating operation system;

[0024] According to the key nodes and paths in the network structure diagram, configure the structural framework of the collaborative performance parameter knowledge graph, and the structural framework includes node types, edge types, and attribute definitions;

[0025] After defining the key nodes as the nodes in the collaborative performance parameter knowledge graph, defining the paths between the nodes as edges, and defining the corresponding attributes for the nodes and edges, use a visualization tool to generate the collaborative performance parameter knowledge graph, and the attributes include node types and edge weights;

[0026] For example, in a possible implementation manner of the first aspect, the step of analyzing the collaborative performance parameter knowledge graph to determine the data of system performance defect points existing in the target heating operation system includes:

[0027] By comparing the attribute values of the key nodes with the preset range, identify the abnormal key nodes, and the abnormal key nodes represent the fault points or potential problems in the heating operation system;

[0028] According to the node types of the key nodes and the edge types of the edges, divide the collaborative performance parameter knowledge graph into multiple subgraphs, and each subgraph represents a specific field or functional module in the heating operation system;

[0029] Conduct performance evaluation on each subgraph, and determine the activity of the key nodes, the weight distribution of the edges, and the key paths within the subgraph according to the connection degree and interaction frequency of the key nodes in the subgraph;

[0030] Identify potential defect points in the heating operation system by combining the abnormal key nodes, the activity of the key nodes within the subgraph, the weight distribution of the edges, and the critical path;

[0031] Locate the identified potential defect points by using the critical path and association rules in the collaborative performance parameter knowledge graph, and determine the specific location and influence range of the potential defect points.

[0032] According to one aspect of the embodiments of the present application, a machine learning system is provided. The machine learning system includes a processor and a machine-readable storage medium. Machine-executable instructions are stored in the machine-readable storage medium, and the machine-executable instructions are loaded and executed by the processor to implement the machine learning-based auxiliary analysis method for heating energy efficiency optimization in any of the foregoing possible implementation manners.

[0033] According to one aspect of the embodiments of the present application, a computer program product or a computer program is provided. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the various optional implementation manners of the foregoing three aspects.

[0034] In the technical solutions provided by some embodiments of the present application, the embodiments of the present application achieve precise optimization analysis of heating energy efficiency by combining an exemplary heating operation system with the data characteristics of a learning heating operation system, significantly improving the operation efficiency and energy efficiency level of the heating system. Specifically, first, by obtaining the operation trajectory characteristics, scheduling knowledge characteristics, and corresponding collaborative data of the exemplary heating operation system, a first machine learning network is used to estimate the collaborative performance knowledge points, generating guiding exemplary collaborative performance knowledge points. At the same time, a second machine learning network is used to estimate and analyze the performance of the exemplary heating operation trajectory characteristics and scheduling knowledge characteristics, generating exemplary collaborative performance estimation data, providing a data basis for subsequent parameter learning and network optimization. Based on the comparative analysis between the exemplary collaborative performance estimation data and the exemplary collaborative performance knowledge points, parameter learning is performed on the second machine learning network, generating a basic heating collaborative analysis network that can accurately analyze the performance of the heating system, providing strong support for the subsequent analysis of the learning heating operation system. Further, by learning the operation trajectory characteristics, scheduling knowledge characteristics, and collaborative data of the learning heating operation system, combined with the basic heating collaborative analysis network, parameter learning is performed on the first machine learning network, generating a basic collaborative performance knowledge point estimation network that can accurately estimate the collaborative performance knowledge points of the heating system, providing a scientific basis for heating energy efficiency optimization. Finally, by performing cyclic parameter learning on the basic heating collaborative analysis network and the basic collaborative performance knowledge point estimation network, the performance of the two networks is continuously optimized until the preset convergence requirements are met. The basic collaborative performance knowledge point estimation network after parameter learning is output as a collaborative performance knowledge point estimation network, which can analyze the collaborative performance of the heating system in real time and accurately, providing powerful technical support for the optimization of heating energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained in combination with these drawings without creative efforts.

[0036] Figure 1 It is a flowchart of a method for auxiliary analysis of heating energy efficiency optimization based on machine learning provided by an embodiment of the present application;

[0037] Figure 2 It is a schematic block diagram of the architecture of a machine learning system for implementing the above-mentioned method for auxiliary analysis of heating energy efficiency optimization based on machine learning provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] The following description is provided to enable a person of ordinary skill in the art to implement and combine the present application, and this description is provided in the context of a specific application scenario and its required environment. For a person of ordinary skill in the art, it is obvious that various changes can be made to the disclosed embodiments, and when not departing from the principles and scope of the present application, the general principles defined in the present application can be applied to other embodiments and application scenarios. Therefore, the present application is not limited to the described embodiments, but should be given the broadest scope consistent with the claims.

[0039] Figure 1 FIG. is a schematic flow chart of an auxiliary analysis method for optimizing heating energy efficiency based on machine learning provided by an embodiment of the present application. The auxiliary analysis method for optimizing heating energy efficiency based on machine learning will be introduced in detail below.

[0040] Step S110: Obtain the example heating operation trajectory features, example heating scheduling knowledge features of the example heating operation system, and example collaborative data between the example heating operation system and the example heating scheduling knowledge features.

[0041] Specifically, the example heating operation system is a specific and known heating operation system, which is used to provide data and features as the basis for training and verifying a machine learning model. This example heating operation system may have been operated in an actual environment and has accumulated a large amount of operation data and performance data. For example, a large heating station in a city is an example heating operation system. This heating station may include multiple heat sources, heating pipelines, user interfaces, etc., and has been operating for many years, accumulating a large amount of operation data and user feedback.

[0042] The example heating operation trajectory features describe various data or state information generated during the operation of the example heating operation system, which may include the operation time of the heat source, the temperature change of the heating pipeline, the flow rate of the user interface, etc. For example, in an example heating operation system, some feature segments of the example heating operation trajectory features may include:

[0043] Heat source A starts at 6:00 am and shuts down at 10:00 pm every day.

[0044] The average temperature of heating pipeline B is 50 °C in winter and 30 °C in summer.

[0045] The flow rate of user interface C is 100 cubic meters per hour during peak hours and 50 cubic meters per hour during off-peak hours.

[0046] The example heating scheduling knowledge features describe the strategies, rules, or knowledge used by heating schedulers when scheduling the heating operation system, which may include temperature control strategies, flow rate allocation rules, etc.

[0047] For example, in the exemplary heating operation system, the exemplary heating scheduling knowledge features may include:

[0048] When the outdoor temperature is lower than -5°C, set the temperature of the heating pipeline to 55°C.

[0049] During peak hours, increase the flow rate of user interface C by 20%.

[0050] The exemplary collaborative data describes the collaborative effects or performance manifestations generated by the exemplary heating operation system and the exemplary heating scheduling knowledge features during the actual operation process, and may include one or more parameters among exemplary collaborative transfer monitoring data, exemplary heating scheduling response data, and heating collaborative stability data.

[0051] Thus, in this embodiment, the server first connects to the database of the heating operation system, retrieves relevant data of the exemplary heating operation system (e.g., system A) therefrom, which may include the heating operation trajectory features of system A (such as heat source startup time, heating pipeline temperature change curve, etc.), heating scheduling knowledge features (such as scheduling strategy settings, temperature control target values, etc.), and exemplary collaborative data of system A working in cooperation with the scheduling knowledge features during actual operation. The server stores this data in a local or cloud data warehouse for subsequent analysis.

[0052] Step S120, load the exemplary collaborative data into the first machine learning network for collaborative performance knowledge point estimation to generate exemplary collaborative performance knowledge points, and load the exemplary heating operation trajectory features and the exemplary heating scheduling knowledge features into the second machine learning network for performance manifestation estimation and analysis to generate exemplary collaborative performance estimation data.

[0053] Specifically, the first machine learning network is a pre-trained machine learning model for estimating the collaborative performance knowledge points of the heating operation system based on the exemplary collaborative data, and may specifically be a deep learning network or other types of machine learning algorithms.

[0054] Collaborative performance knowledge point estimation refers to the process of using the first machine learning network to estimate the collaborative performance of the heating operation system, and based on the exemplary collaborative data, generate a prediction or evaluation of the collaborative performance of the heating operation system. For example, given a set of exemplary collaborative data, the first machine learning network can output an estimation result.

[0055] The second machine learning network is another pre-trained machine learning model for estimating the performance of the heating operation system based on the exemplary heating operation trajectory features and exemplary heating scheduling knowledge features. The second machine learning network has different structures and parameters from the first machine learning network. For example, the second machine learning network can be a regression model that receives the exemplary heating operation trajectory features and exemplary heating scheduling knowledge features as inputs and then outputs the estimated performance data.

[0056] The performance estimation and analysis refers to the process of using the second machine learning network to estimate and analyze the performance of the heating operation system, generating predictions or evaluations of the performance of the heating operation system based on the exemplary heating operation trajectory features and exemplary heating scheduling knowledge features. For example, given a set of exemplary heating operation trajectory features and exemplary heating scheduling knowledge features (such as heat source operation time, temperature control strategy, etc.), the second machine learning network can output a performance estimation result.

[0057] That is, in this embodiment, the server loads the exemplary collaborative data into the pre-trained first machine learning network (such as a deep learning model), which is designed to estimate the collaborative performance knowledge points of the heating operation system under specific conditions based on the collaborative data, such as collaborative transfer knowledge points, heating scheduling response knowledge points, and heating collaborative stability knowledge points, etc. At the same time, the server also loads the exemplary heating operation trajectory features and exemplary heating scheduling knowledge features into the second machine learning network, which is used to analyze how these exemplary heating operation trajectory features and exemplary heating scheduling knowledge features affect the performance of the heating system and generate exemplary collaborative performance estimation data, such as collaborative transfer estimation data, heating scheduling response estimation data, and heating collaborative stability estimation data.

[0058] Step S130, based on the difference between the exemplary collaborative performance estimation data and the exemplary collaborative performance knowledge points, perform parameter learning on the second machine learning network to generate a basic heating collaborative analysis network.

[0059] In this embodiment, the server compares the exemplary collaborative performance knowledge points generated by the first machine learning network and the exemplary collaborative performance estimation data generated by the second machine learning network. By calculating the difference between the two (such as the error rate), the server identifies the deficiencies of the second machine learning network in performance estimation. Based on these differences, the server adjusts and optimizes the parameters of the second machine learning network to improve the accuracy of its performance estimation. After multiple iterations and adjustments, the server generates a basic heating collaborative analysis network that can more accurately estimate the performance of the heating operation system.

[0060] Step S140: Based on the learned heating operation trajectory features, learned heating scheduling knowledge features, learned collaborative data of the heating operation system, and the basic heating collaborative analysis network, perform parameter learning on the first machine learning network to generate a basic collaborative performance knowledge point estimation network. The learned collaborative data is the collaborative data between the learned heating operation system and the learned heating scheduling knowledge features. The learned heating operation system and the exemplary heating operation system are of the same type of heating operation system. The learned heating operation trajectory features and the exemplary heating operation trajectory features are features of the same dimension. The learned heating scheduling knowledge features and the exemplary heating scheduling knowledge features are features of the same dimension. The learned collaborative data and the exemplary collaborative data are data of the same dimension.

[0061] In this embodiment, the server obtains the learned heating operation trajectory features, learned heating scheduling knowledge features, and corresponding learned collaborative data of the learned heating operation system (e.g., System B). These data have the same dimension and format as the exemplary data obtained in Step S110. The server loads these data into the basic heating collaborative analysis network to obtain the performance estimation data of System B. At the same time, the server also loads the learned collaborative data into the first machine learning network to obtain the learned collaborative performance knowledge point estimation. Then, the server compares these two results and performs parameter learning on the first machine learning network according to the difference between them to generate a basic collaborative performance knowledge point estimation network, which can better estimate the collaborative performance knowledge points of the learned heating operation system.

[0062] Step S150: Perform cyclic parameter learning on the basic heating collaborative analysis network and the basic collaborative performance knowledge point estimation network, and output the basic collaborative performance knowledge point estimation network when the parameter learning is completed as the collaborative performance knowledge point estimation network. The cyclic parameter learning includes cyclic operations: locking the network parameter information of the basic collaborative performance knowledge point estimation network to perform parameter learning on the basic heating collaborative analysis network and locking the network parameter information of the basic heating collaborative analysis network to perform parameter learning on the basic collaborative performance knowledge point estimation network.

[0063] In this embodiment, the server starts the cyclic parameter learning process. In each cycle, the server first locks the parameters of the basic collaborative performance knowledge point estimation network, and uses the example data and learning data to perform parameter learning on the basic heating collaborative analysis network to improve its estimation accuracy of the performance of the heating operation system. Then, the server locks the parameters of the basic heating collaborative analysis network, and uses the learning data to perform parameter learning on the basic collaborative performance knowledge point estimation network to improve its estimation accuracy of the collaborative performance knowledge points of the heating operation system. This process is repeated until both networks meet the expected performance and convergence criteria. Finally, the server outputs the basic collaborative performance knowledge point estimation network that has completed parameter learning as the collaborative performance knowledge point estimation network for subsequent collaborative performance analysis and energy efficiency optimization assistance of the heating operation system.

[0064] Based on the above steps, the embodiment of the present application realizes the accurate optimization analysis of heating energy efficiency by combining the data characteristics of the example heating operation system and the learning heating operation system, and significantly improves the operation efficiency and energy efficiency level of the heating system. Specifically, first, by obtaining the operation trajectory characteristics, scheduling knowledge characteristics and corresponding collaborative data of the example heating operation system, the first machine learning network is used to estimate the collaborative performance knowledge points, and the guiding example collaborative performance knowledge points are generated. At the same time, through the second machine learning network, the performance estimation analysis of the example heating operation trajectory characteristics and scheduling knowledge characteristics is carried out, and the example collaborative performance estimation data is generated, providing a data basis for subsequent parameter learning and network optimization. Based on the comparative analysis between the example collaborative performance estimation data and the example collaborative performance knowledge points, parameter learning is performed on the second machine learning network to generate the basic heating collaborative analysis network, which can accurately analyze the performance of the heating system and provide strong support for the subsequent analysis of the learning heating operation system. Further, by learning the operation trajectory characteristics, scheduling knowledge characteristics and collaborative data of the learning heating operation system, combined with the basic heating collaborative analysis network, parameter learning is performed on the first machine learning network to generate the basic collaborative performance knowledge point estimation network, which can accurately estimate the collaborative performance knowledge points of the heating system and provide a scientific basis for heating energy efficiency optimization. Finally, through the cyclic parameter learning of the basic heating collaborative analysis network and the basic collaborative performance knowledge point estimation network, the performance of the two networks is continuously optimized until the preset convergence requirements are met. The basic collaborative performance knowledge point estimation network after completing parameter learning is output as the collaborative performance knowledge point estimation network, which can analyze the collaborative performance of the heating system in real time and accurately, providing powerful technical support for the optimization of heating energy efficiency.

[0065] In a possible implementation manner, step S140 may include:

[0066] Step S141: Load the learned heating operation trajectory features and the learned heating scheduling knowledge features into the basic heating collaborative analysis network for performance evaluation and analysis, and generate basic learning collaborative performance data.

[0067] In this embodiment, the server first obtains the learned heating operation trajectory features and the learned heating scheduling knowledge features of the learned heating operation system (e.g., the newly built system B). These features may include the heat source startup time of the new system B, the heating pipeline temperature change curve, the user interface flow data, the temperature control strategy settings, etc.

[0068] Next, the server loads these feature data into the basic heating collaborative analysis network, which is trained based on the data of the exemplary heating operation system and can analyze how the heating operation trajectory features and the heating scheduling knowledge features affect the performance of the heating system.

[0069] When the basic heating collaborative analysis network receives the learned heating operation trajectory features and the learned heating scheduling knowledge features, it can analyze and process these features and generate basic learning collaborative performance data. These basic learning collaborative performance data may include collaborative transfer estimation data, heating scheduling response estimation data, and heating collaborative stability estimation data, which are preliminary estimates of the performance of the new system B under the given features and scheduling knowledge.

[0070] Step S142: Load the learned collaborative data into the first machine learning network for collaborative performance knowledge point estimation, and generate learned estimated collaborative performance knowledge points.

[0071] Meanwhile, the server obtains the learned collaborative data generated during the actual operation of the new system B. These learned collaborative data may include the collaborative transfer monitoring data, heating scheduling response data, and heating collaborative stability data of system B, etc.

[0072] The server loads these learned collaborative data into the first machine learning network, which is designed to estimate the collaborative performance knowledge points of the heating operation system under specific conditions based on the collaborative data. Specifically, the first machine learning network may include multiple sub-networks, such as a collaborative transfer knowledge point estimation network, a heating scheduling response knowledge point estimation network, and a heating collaborative stability knowledge point estimation network, etc.

[0073] After the first machine learning network receives the learning collaboration data, it loads these data into the corresponding sub-networks respectively to estimate the knowledge points of collaboration performance. For example, the collaboration transfer monitoring data will be loaded into the collaboration transfer knowledge point estimation network to generate the learned collaboration transfer knowledge points; the heat supply scheduling response data will be loaded into the heat supply scheduling response knowledge point estimation network to generate the learned heat supply scheduling response knowledge points; the heat supply collaboration stability data will be loaded into the heat supply collaboration stability knowledge point estimation network to generate the learned heat supply collaboration stability knowledge points. Finally, these knowledge points are integrated to form the learned estimated collaboration performance knowledge points.

[0074] Step S143: According to the difference between the learned estimated collaboration performance knowledge points and the basic learned collaboration performance data, perform parameter learning on the first machine learning network to generate the basic collaboration performance knowledge point estimation network.

[0075] Next, the server will compare the difference between the learned estimated collaboration performance knowledge points and the basic learned collaboration performance data. This difference reflects the accuracy of the first machine learning network in estimating the collaboration performance of the new system B, as well as the accuracy of the basic heat supply collaboration analysis network in estimating performance.

[0076] Based on these differences, the server can adjust and optimize the parameters of the first machine learning network. This process can use optimization algorithms such as gradient descent to iteratively update the weights and bias terms of the network, making the network's estimation of the collaboration performance of the new system B more accurate.

[0077] Specifically, the server can calculate the error between the learned estimated collaboration performance knowledge points and the basic learned collaboration performance data, and use these errors as feedback signals to guide the first machine learning network to update its parameters. After multiple iterations and optimizations, the server finally generates the basic collaboration performance knowledge point estimation network. This network is obtained by performing parameter learning on the first machine learning network based on the data and characteristics of the learned heat supply operation system, and can more accurately estimate the collaboration performance knowledge points of the new heat supply operation system.

[0078] Through the above steps, the server uses the data and characteristics of the learned heat supply operation system to perform parameter learning on the first machine learning network, generating the basic collaboration performance knowledge point estimation network. This network can more accurately estimate the collaboration performance knowledge points of the new heat supply operation system, providing strong support for subsequent auxiliary analysis of heat supply energy efficiency optimization.

[0079] In a possible implementation manner, step S150 may include:

[0080] Step S151: Load the example collaboration data into the basic collaborative performance knowledge point estimation network for collaborative performance knowledge point estimation, and generate basic example collaborative performance knowledge points.

[0081] In this embodiment, the server first obtains the previously stored example collaboration data, which describes the collaborative effect of the example heating operation system and scheduling knowledge features in actual operation. The server loads this example collaboration data into the basic collaborative performance knowledge point estimation network, which is obtained by parameter learning of the first machine learning network based on the data of the learning heating operation system before and can estimate the collaborative performance knowledge points of the heating operation system according to the collaboration data.

[0082] When the example collaboration data is processed by the basic collaborative performance knowledge point estimation network, the basic collaborative performance knowledge point estimation network will output basic example collaborative performance knowledge points. These basic example collaborative performance knowledge points include collaborative transfer knowledge points, heating scheduling response knowledge points, and heating collaborative stability knowledge points, etc., which are the estimation results of the collaborative performance of the example heating operation system.

[0083] Step S152: Load the example heating operation trajectory features and the example heating scheduling knowledge features into the basic heating collaboration analysis network for performance estimation analysis, and generate basic example collaborative performance estimation data.

[0084] Immediately afterwards, the server retrieves the example heating operation trajectory features and the example heating scheduling knowledge features from the data warehouse. These features describe the operation situation and scheduling strategy of the example heating operation system. The server loads these features into the basic heating collaboration analysis network, which is obtained by parameter learning based on the data of the example heating operation system before and can analyze how the heating operation trajectory features and scheduling knowledge features affect the performance of the heating system.

[0085] After receiving these features, the basic heating collaboration analysis network will perform analysis and processing, and output basic example collaborative performance estimation data. These basic example collaborative performance estimation data include collaborative transfer estimation data, heating scheduling response estimation data, and heating collaborative stability estimation data, etc., which are the estimation results of the performance of the example heating operation system.

[0086] Step S153: According to the difference between the basic example collaborative performance estimation data and the basic example collaborative performance knowledge points, perform parameter learning on the basic heating collaboration analysis network, and use the network when the parameter learning is completed as the basic heating collaboration analysis network again.

[0087] In this embodiment, the server will then compare the basic example collaborative performance estimation data with the basic example collaborative performance knowledge points. These differences reflect the accuracy of the basic heating collaborative analysis network in estimating performance.

[0088] Based on these differences, the server can adjust and optimize the parameters of the basic heating collaborative analysis network. This process can use optimization algorithms such as gradient descent to iteratively update the weights and bias terms of the network, making the network's estimation of the performance of the heating operation system more accurate. After parameter learning, the server will use the updated network as the basic heating collaborative analysis network again.

[0089] Step S154: Load the learned heating operation trajectory features and the learned heating scheduling knowledge features into the basic heating collaborative analysis network for performance estimation analysis, and generate basic learned collaborative performance data.

[0090] Next, the server obtains the learned heating operation trajectory features and the learned heating scheduling knowledge features of the learned heating operation system. These features describe the operation situation and scheduling strategy of the learned heating operation system. The server loads these features into the basic heating collaborative analysis network that has been updated. After the basic heating collaborative analysis network analyzes and processes these features, it will output basic learned collaborative performance data, which is a preliminary estimate of the performance of the learned heating operation system.

[0091] Step S155: Based on the learned collaborative data and the basic learned collaborative performance data, perform parameter learning on the basic collaborative performance knowledge point estimation network, and use the network when the parameter learning is completed as the basic collaborative performance knowledge point estimation network again.

[0092] In this embodiment, the server simultaneously obtains the learned collaborative data of the learned heating operation system, which describes the collaborative effect of the learned heating operation system in actual operation. The server loads these learned collaborative data and the basic learned collaborative performance data into the basic collaborative performance knowledge point estimation network together.

[0093] The basic collaborative performance knowledge point estimation network estimates the collaborative performance knowledge points of the learned heating operation system based on these data and compares them with the basic learned collaborative performance data. Based on the differences between them, the server adjusts and optimizes the parameters of the basic collaborative performance knowledge point estimation network. This process also uses optimization algorithms such as gradient descent. After parameter learning, the server will use the updated network as the basic collaborative performance knowledge point estimation network again.

[0094] Step S156, loop the step of loading the example collaborative data into the basic collaborative performance knowledge point estimation network for collaborative performance knowledge point estimation to generate basic example collaborative performance knowledge points until, based on the learning collaborative data and the basic learning collaborative performance data, parameter learning is performed on the basic collaborative performance knowledge point estimation network, and the network when the parameter learning is completed is used again as the basic collaborative performance knowledge point estimation network until the network convergence requirement is met.

[0095] The server repeatedly executes steps S151 to S155 to form a process of cyclic parameter learning. In this cycle, the server alternately uses the example collaborative data and the learning collaborative data to perform parameter learning on the two networks to continuously improve their accuracy and generalization ability. The cycle will continue until the performance of the two networks on the validation data reaches the preset convergence requirement.

[0096] Step S157, output the basic heating collaborative analysis network when the parameter learning is completed as the heating collaborative analysis network, and output the basic collaborative performance knowledge point estimation network when the parameter learning is completed as the collaborative performance knowledge point estimation network.

[0097] When the cyclic parameter learning meets the convergence requirement, the server outputs the basic heating collaborative analysis network when the parameter learning is completed as the heating collaborative analysis network, and outputs the basic collaborative performance knowledge point estimation network when the parameter learning is completed as the collaborative performance knowledge point estimation network. These two network models will be used for subsequent collaborative performance analysis and energy efficiency optimization assistance of the heating operation system.

[0098] In a possible implementation manner, the example collaborative data includes one or more parameters among example collaborative transfer monitoring data, example heating scheduling response data, and heating collaborative stability data, the first machine learning network includes one or more estimation networks among a collaborative transfer knowledge point estimation network, a heating scheduling response knowledge point estimation network, and a heating collaborative stability knowledge point estimation network, and the first machine learning network is determined based on the example collaborative data. Step S120 may include:

[0099] Step S121, execute the target collaborative performance knowledge point estimation instruction, where the target collaborative performance knowledge point estimation instruction includes at least one of the first collaborative performance knowledge point estimation, the second collaborative performance knowledge point estimation, and the third collaborative performance knowledge point estimation. Among them, the first collaborative performance knowledge point estimation is to load the example collaborative transfer monitoring data into the collaborative transfer knowledge point estimation network to perform collaborative transfer knowledge point estimation and generate example collaborative transfer knowledge points. The second collaborative performance knowledge point estimation is to load the example heat supply dispatching response data into the heat supply dispatching response knowledge point estimation network to perform heat supply dispatching response knowledge point estimation and generate example heat supply dispatching response knowledge points. The third collaborative performance knowledge point estimation is to load the heat supply collaborative stability data into the heat supply collaborative stability knowledge point estimation network to perform heat supply collaborative stability knowledge point estimation and generate example heat supply collaborative stability knowledge points.

[0100] Step S122, output one or more parameters among the example collaborative transfer knowledge points, the example heat supply dispatching response knowledge points, and the example heat supply collaborative stability knowledge points as the example collaborative performance knowledge points.

[0101] In this embodiment, the server first determines the specific type of the example collaborative data retrieved from the heat supply operation system database. The example collaborative data may include one or more parameters among the example collaborative transfer monitoring data, the example heat supply dispatching response data, and the heat supply collaborative stability data. These data reflect the collaborative performance of the heat supply operation system during actual operation.

[0102] The server identifies that the example collaborative data contains collaborative transfer monitoring data, which is data describing monitoring information such as the fluid flow state and temperature distribution in the heat supply system. The server loads these collaborative transfer monitoring data into the collaborative transfer knowledge point estimation network in the first machine learning network. The collaborative transfer knowledge point estimation network is specifically designed to analyze the collaborative transfer monitoring data and estimate the collaborative transfer knowledge points of the heat supply operation system. Through the processing of the collaborative transfer knowledge point estimation network, the server generates example collaborative transfer knowledge points, which may include the best path of fluid flow, the rationality of temperature distribution, etc.

[0103] The server further identifies that the exemplary collaborative data also includes heat supply dispatching response data, which describes the response speed and accuracy of the heat supply system to dispatching instructions. The server loads this heat supply dispatching response data into the heat supply dispatching response knowledge point estimation network, which can analyze the dispatching response data and estimate the dispatching response knowledge points of the heat supply system. After network processing, the server generates exemplary heat supply dispatching response knowledge points, which may include the optimal time for dispatching response, the adaptability of dispatching strategies, etc.

[0104] If the exemplary collaborative data also includes heat supply collaborative stability data, the server can load this heat supply collaborative stability data into the heat supply collaborative stability knowledge point estimation network, which is responsible for analyzing the stability performance of the heat supply system. Through network processing, the server generates exemplary heat supply collaborative stability knowledge points, which may include the stability performance of the system under different conditions, potential risk factors, etc.

[0105] The server integrates the above-generated exemplary collaborative transfer knowledge points, exemplary heat supply dispatching response knowledge points, and exemplary heat supply collaborative stability knowledge points. These knowledge points together constitute the exemplary collaborative performance knowledge points, comprehensively reflecting the performance of the heat supply operation system in terms of collaborative performance.

[0106] If the exemplary collaborative data includes all three types of data, the server integrates all three types of knowledge points into the exemplary collaborative performance knowledge points. If the exemplary collaborative data only includes some types of data, the server only integrates the corresponding type of knowledge points into the exemplary collaborative performance knowledge points. Finally, the server stores the integrated exemplary collaborative performance knowledge points for subsequent analysis and comparison. These knowledge points will become an important basis for evaluating and optimizing the collaborative performance of the heat supply operation system.

[0107] In a possible implementation manner, the second machine learning network includes one or more estimation networks among a collaborative transfer estimation network, a heat supply dispatching response estimation network, and a heat supply collaborative stability estimation network. The second machine learning network is determined according to the exemplary collaborative data. Step S120 further includes:

[0108] Step S123, execute the target performance estimation instruction, where the target performance estimation instruction includes at least one of a first performance estimation instruction, a second performance estimation instruction, and a third performance estimation instruction. Among them, the first performance estimation instruction is to load the example heating operation trajectory feature and the example heating scheduling knowledge feature into the collaborative transfer estimation network for collaborative transfer estimation to generate an example collaborative transfer result. The second performance estimation instruction is to load the example heating operation trajectory feature and the example heating scheduling knowledge feature into the heating scheduling response estimation network for heating scheduling response estimation to generate an example heating scheduling response result. The third performance estimation instruction is to load the example heating operation trajectory feature and the example heating scheduling knowledge feature into the heating collaborative stability estimation network for heating collaborative stability estimation to generate an example heating collaborative stability result.

[0109] Step S124, output one or more parameters among the example collaborative transfer result, the example heating scheduling response result, and the example heating collaborative stability result as the example collaborative performance estimation data. The example collaborative performance estimation data has the same feature category as the example collaborative performance knowledge points.

[0110] In this embodiment, the server first determines the type of estimation network included in the second machine learning network. In this scenario, the second machine learning network includes one or more of a collaborative transfer estimation network, a heating scheduling response estimation network, and a heating collaborative stability estimation network. These networks are designed and trained based on the characteristics of the example collaborative data and are used to analyze the performance of the heating operation system.

[0111] The server retrieves the example heating operation trajectory feature and the example heating scheduling knowledge feature from the data warehouse. These feature data describe the actual operation situation and scheduling strategy of the example heating operation system and are the input data for performance estimation.

[0112] The server executes the corresponding target performance estimation instruction according to the structure and function of the second machine learning network. These target performance estimation instructions include:

[0113] The first performance estimation instruction:

[0114] The server loads the example heating operation trajectory feature and the example heating scheduling knowledge feature into the collaborative transfer estimation network. The collaborative transfer estimation network analyzes factors such as the trajectory, speed, and efficiency of fluid flow in the heating system based on these feature data for collaborative transfer estimation. Through network processing, the server generates example collaborative transfer results, which may include the best path of fluid flow, flow efficiency evaluation, etc.

[0115] Second performance estimation instruction:

[0116] The server loads the same feature data into the heating dispatch response estimation network. The heating dispatch response estimation network analyzes the response speed and accuracy of the heating system to dispatch instructions, as well as the actual effect of the dispatch strategy. After network processing, the server generates exemplary heating dispatch response results, which may include the time of dispatch response, the adaptability evaluation of the dispatch strategy, etc.

[0117] Third performance estimation instruction (if the second machine learning network includes a heating cooperation stability estimation network): The server loads the feature data into the heating cooperation stability estimation network. This network analyzes the stability performance of the heating system during long-term operation, including the fluctuation of system parameters, failure rate, etc. Through network processing, the server generates exemplary heating cooperation stability results, which may include the stability evaluation of the system, potential risk prediction, etc.

[0118] The server integrates the above-generated exemplary collaborative transfer results, exemplary heating dispatch response results, and (if any) exemplary heating cooperation stability results. These results together constitute the exemplary collaborative performance estimation data, which comprehensively reflect the actual situation of the heating operation system in terms of performance.

[0119] When integrating the data, the server ensures that the feature categories of the exemplary collaborative performance estimation data are the same as those of the exemplary collaborative performance knowledge points for subsequent comparison and analysis. The integrated exemplary collaborative performance estimation data will be stored for subsequent use.

[0120] Through this process, the server comprehensively and accurately analyzes and estimates the performance of the exemplary heating operation system using the second machine learning network, providing strong data support for subsequent optimization and improvement.

[0121] In a possible implementation manner, step S130 may include:

[0122] Step S131, execute the target error calculation instruction. The target error calculation instruction includes one or more parameters of the first error calculation instruction, the second error calculation instruction, and the third error calculation instruction. Among them, the first error calculation instruction determines the first exemplary error parameter based on the difference between the exemplary collaborative transfer result and the exemplary collaborative transfer knowledge point. The second error calculation instruction determines the second exemplary error parameter based on the difference between the exemplary heating dispatch response result and the exemplary heating dispatch response knowledge point. The third error calculation instruction determines the third exemplary error parameter based on the difference between the exemplary heating cooperation stability result and the exemplary heating cooperation stability knowledge point.

[0123] Step S132: Determine the target example error parameter based on one or more of the first example error parameter, the second example error parameter, and the third example error parameter.

[0124] Step S133: Update the network parameter information of the second machine learning network according to the target example error parameter to generate a basic heating collaborative analysis network.

[0125] In this embodiment, the server first executes a target error calculation instruction, which is used to calculate the difference between the example collaborative performance estimation data and the example collaborative performance knowledge points. Since the second machine learning network includes multiple estimation networks (collaborative transfer estimation network, heating scheduling response estimation network, heating collaborative stability estimation network), the target error calculation instruction accordingly includes multiple sub-instructions.

[0126] First error calculation instruction:

[0127] The server first compares the example collaborative transfer result with the example collaborative transfer knowledge points. For example, the example collaborative transfer result output by the collaborative transfer estimation network may show the fluid flow rate and temperature distribution in a specific heating pipeline during a specific period, while the example collaborative transfer knowledge points are the ideal or expected description of this transfer process. The server calculates the difference between the two, such as by calculating indicators such as the deviation and inconsistency of the flow rate and temperature distribution, to obtain the first example error parameter, which quantifies the error of the collaborative transfer estimation network in estimating the fluid transfer performance.

[0128] Second error calculation instruction:

[0129] Next, the server compares the example heating scheduling response result with the example heating scheduling response knowledge points. For example, the heating scheduling response estimation network may output the response time and adjustment range of the heating system to a specific scheduling instruction, while the example heating scheduling response knowledge points are the ideal or expected description of these responses. The server calculates the differences in aspects such as the response time and adjustment range to obtain the second example error parameter, which quantifies the error of the heating scheduling response estimation network in estimating the scheduling response performance.

[0130] Third error calculation instruction (if the second machine learning network includes a heating collaborative stability estimation network):

[0131] Finally, if a heating collaboration stability estimation network exists, the server can compare the exemplary heating collaboration stability results with the exemplary heating collaboration stability knowledge points. For example, the heating collaboration stability estimation network may output a stability assessment of the heating system during long-term operation, while the exemplary heating collaboration stability knowledge point is an ideal or expected description of this stability. The server calculates the differences in the stability assessment, such as the failure rate, parameter fluctuation conditions, etc., to obtain a third exemplary error parameter, which quantifies the error of the heating collaboration stability estimation network in estimating the system stability.

[0132] Based on the first exemplary error parameter, the second exemplary error parameter, and (if any) the third exemplary error parameter, the server comprehensively considers their importance and influence degree to determine a target exemplary error parameter. This process may involve the allocation and adjustment of weights to ensure that the target exemplary error parameter can comprehensively reflect the performance errors of the second machine learning network in multiple aspects.

[0133] Once the target exemplary error parameter is determined, the server uses this target exemplary error parameter to update the network parameter information of the second machine learning network. This process can use optimization algorithms (such as the gradient descent algorithm) to iteratively adjust the weights and bias terms of the network to reduce the target exemplary error parameter. Through multiple iterations and optimizations, the server finally generates a more accurate basic heating collaboration analysis network in terms of performance, which can better estimate the performance of the heating operation system in subsequent analyses and provide strong support for heating energy efficiency optimization.

[0134] In a possible implementation manner, the method further includes: executing one or more parameters among obtaining a first influence coefficient of the exemplary collaboration transfer monitoring data, obtaining a second influence coefficient of the exemplary heating scheduling response data, and obtaining a third influence coefficient of the heating collaboration stability data.

[0135] Step S132 may include:

[0136] Step S1321, executing one or more parameters among determining a first fusion error parameter based on the first influence coefficient and the first exemplary error parameter, determining a second fusion error parameter based on the second influence coefficient and the second exemplary error parameter, and determining a third fusion error parameter based on the third influence coefficient and the third exemplary error parameter.

[0137] Step S1322, determining a target exemplary error parameter based on one or more parameters among the first fusion error parameter, the second fusion error parameter, and the third fusion error parameter.

[0138] In this embodiment, before the server executes the parameter learning for optimizing the second machine learning network, it first needs to obtain the importance degree of different data types on the performance of the heating operation system, that is, the influence coefficient. These influence coefficients represent the weights of different data in evaluating the performance of the heating operation system.

[0139] The server determines the importance of the example collaborative transfer monitoring data in performance evaluation by querying the preset rules or analyzing from historical data. For example, if the collaborative transfer monitoring data has a greater impact on the overall performance of the heating system, then its first influence coefficient will be relatively high. The server stores this coefficient in the memory for subsequent calculations.

[0140] Similarly, the server obtains the influence coefficient of the example heating dispatch response data, that is, the second influence coefficient, which reflects the importance of the heating dispatch response data in evaluating the operation performance of the heating system.

[0141] If the server includes the analysis of heating collaborative stability data, it will also obtain the influence coefficient of the heating collaborative stability data, that is, the third influence coefficient, which represents the weight of the heating collaborative stability data in performance evaluation.

[0142] After determining the influence coefficients of different data, the server starts to determine the fusion error parameter based on these influence coefficients and the example error parameters calculated previously.

[0143] The server multiplies the first influence coefficient by the first example error parameter to obtain the first fusion error parameter, which comprehensively considers the importance of the collaborative transfer monitoring data and the actual error, and is used to more accurately reflect the performance of the collaborative transfer estimation network.

[0144] Similarly, the server multiplies the second influence coefficient by the second example error parameter to calculate the second fusion error parameter, which quantifies the performance error of the heating dispatch response estimation network and considers the importance of the dispatch response data.

[0145] If it includes the analysis of heating collaborative stability data, the server multiplies the third influence coefficient by the third example error parameter to obtain the third fusion error parameter, which reflects the performance error of the heating collaborative stability estimation network and the importance of the stability data.

[0146] Finally, the server determines the target example error parameter based on the calculated fusion error parameters. This process involves weighted summation or other combination methods of different fusion error parameters, depending on the importance of the error parameters and the application scenario.

[0147] For example, the server can directly add three fusion error parameters to obtain the target example error parameter. Or, if the importance of certain data types is higher, the server can assign a greater weight to the corresponding fusion error parameter. In either case, the target example error parameter is a key metric for guiding the parameter learning of the second machine learning network, aiming to improve the network's performance in multiple aspects.

[0148] In a possible implementation manner, the basic heating collaborative analysis network includes one or more of a basic collaborative transfer estimation network, a basic heating scheduling response estimation network, and a basic heating collaborative stability estimation network. Step S141 may include:

[0149] Execute one or more of the following loading instructions for the learned heating operation trajectory feature and the learned heating scheduling knowledge feature:

[0150] The first loading instruction: Load it into the basic collaborative transfer estimation network for collaborative transfer prediction to generate a learned collaborative transfer result.

[0151] The second loading instruction: Load it into the basic heating scheduling response estimation network for heating scheduling response estimation to generate a learned heating scheduling response result.

[0152] The third loading instruction: Load it into the basic heating collaborative stability estimation network for heating collaborative stability estimation to generate a learned heating collaborative stability result.

[0153] Output one or more parameters among the learned collaborative transfer result, the learned heating scheduling response result, and the learned heating collaborative stability result as the learned collaborative performance data.

[0154] In this embodiment, during the performance optimization process of the heating operation system, the server has trained and optimized the basic heating collaborative analysis network through a series of steps. The basic heating collaborative analysis network includes one or more of a basic collaborative transfer estimation network, a basic heating scheduling response estimation network, and a basic heating collaborative stability estimation network. Next, the server will use this network to perform a performance estimation analysis on the newly acquired data of the heating operation system to be learned.

[0155] The server first obtains the to-be-learned heating operation trajectory feature and the to-be-learned heating scheduling knowledge feature of the heating operation system to be learned (for example, the newly built system B). These feature data describe key information such as the heat source startup time, heating pipeline temperature change, user interface flow rate, and temperature control strategy in the actual operation of the new system B.

[0156] The server then loads this data to be learned into the basic heating collaborative analysis network. Since this network contains multiple sub-networks, the server can execute different loading instructions for the data to be learned.

[0157] The server loads the characteristics of the heating operation trajectory to be learned and the characteristics of the heating scheduling knowledge to be learned into the basic collaborative flow estimation network. Based on these characteristic data, the basic collaborative flow estimation network analyzes the trajectory, speed, efficiency, etc. of the fluid flow in the heating system to perform collaborative flow prediction. After being processed by the network, the server obtains the learned collaborative flow results, which may include the predicted fluid flow path, flow velocity distribution, etc.

[0158] Furthermore, the server loads the same characteristic data into the basic heating scheduling response estimation network. The basic heating scheduling response estimation network analyzes the response speed and accuracy of the heating system to the scheduling instructions and predicts the response of the system to the scheduling strategy. Through network processing, the server generates the learned heating scheduling response results, such as the predicted response time, adjustment range, etc.

[0159] Furthermore, if the basic heating collaborative analysis network includes the basic heating collaborative stability estimation network, the server can load the characteristic data into this network. The basic heating collaborative stability estimation network analyzes the stability performance of the heating system during long-term operation and predicts the stability state of the system. After being processed by the network, the server obtains the learned heating collaborative stability results, such as the predicted system stability evaluation, potential risk prediction, etc.

[0160] The server integrates the above-generated learned collaborative flow results, learned heating scheduling response results, and (if any) learned heating collaborative stability results. These results together constitute the learned collaborative performance data, comprehensively reflecting the prediction of the performance of the heating operation system to be learned.

[0161] When integrating the data, the server ensures that the format and structure of the learned collaborative performance data are consistent with the previously generated exemplary collaborative performance estimation data for subsequent comparison and analysis. The integrated learned collaborative performance data will be stored for subsequent use.

[0162] Through this process, the server uses the basic heating collaborative analysis network to comprehensively and accurately predict and analyze the performance of the heating operation system to be learned, providing strong data support for subsequent optimization and improvement.

[0163] In a possible implementation manner, the learned collaborative data includes at least one of learned collaborative flow monitoring data, learned heating scheduling response data, and learned stable operation duration data, and the step S142 may include:

[0164] Step S1421, execute the learning knowledge point estimation instruction. The learning knowledge point estimation instruction includes one or more parameters among the first knowledge point estimation instruction, the second knowledge point estimation instruction, and the third knowledge point estimation instruction. Among them, the first knowledge point estimation instruction is to load the learning collaborative transfer monitoring data into the collaborative transfer knowledge point estimation network, perform collaborative transfer knowledge point estimation, and generate learning collaborative transfer knowledge points. The second knowledge point estimation instruction is to load the learning heating dispatch response data into the heating dispatch response knowledge point estimation network, perform heating dispatch response knowledge point estimation, and generate learning heating dispatch response knowledge points. The third knowledge point estimation instruction is to load the learning stable operation duration data into the heating collaborative stability knowledge point estimation network, perform heating collaborative stability knowledge point estimation, and generate learning heating collaborative stability knowledge points.

[0165] Step S1422, output one or more parameters among the learning collaborative transfer knowledge points, the learning heating dispatch response knowledge points, and the learning heating collaborative stability knowledge points as the learning estimated collaborative performance knowledge points.

[0166] In this embodiment, during the performance analysis and optimization of the heating operation system, the server has trained multiple machine learning networks, including the collaborative transfer knowledge point estimation network, the heating dispatch response knowledge point estimation network, and the heating collaborative stability knowledge point estimation network. Next, the server will use these networks to process the newly acquired learning collaborative data of the heating operation system to be learned to estimate its collaborative performance knowledge points.

[0167] The server first obtains the learning collaborative data of the heating operation system to be learned (for example, the newly built system B). This data includes one or more of the learning collaborative transfer monitoring data, the learning heating dispatch response data, and the learning stable operation duration data. This data reflects the collaborative performance of the heating station during actual operation.

[0168] The server executes the corresponding learning knowledge point estimation instruction according to the type and content of the learning collaborative data.

[0169] If the learning collaborative data contains the learning collaborative transfer monitoring data, the server loads this part of the data into the collaborative transfer knowledge point estimation network. The collaborative transfer knowledge point estimation network analyzes the monitoring data of the fluid flow in the heating system, such as the flow rate, temperature distribution, etc., according to the learning collaborative transfer monitoring data. The network processes these data to estimate the collaborative transfer knowledge points of the heating system, that is, the best path, efficiency, etc. of the fluid flow, and generates the learning collaborative transfer knowledge points.

[0170] If the learning collaboration data includes learning heat supply scheduling response data, the server loads this part of the data into the heat supply scheduling response knowledge point estimation network. The heat supply scheduling response knowledge point estimation network analyzes the response data of the heat supply system to scheduling instructions, such as response time, adjustment range, etc. Based on these data, the network estimates the heat supply scheduling response knowledge points of the heat supply system, that is, the effectiveness of the scheduling strategy, response speed, etc., and generates learning heat supply scheduling response knowledge points.

[0171] If the learning collaboration data includes learning stable operation duration data, the server loads this part of the data into the heat supply collaboration stability knowledge point estimation network. The heat supply collaboration stability knowledge point estimation network analyzes the stable operation duration data of the heat supply system and evaluates the stability performance of the system under different conditions. Through these data, the network estimates the heat supply collaboration stability knowledge points of the heat supply system, that is, the stability level of the system, potential risks, etc., and generates learning heat supply collaboration stability knowledge points.

[0172] The server integrates the generated learning collaboration transfer knowledge points, learning heat supply scheduling response knowledge points, and (if any) learning heat supply collaboration stability knowledge points. These knowledge points together constitute the learning estimated collaboration performance knowledge points, which comprehensively reflect the estimated situation of the heat supply operation system to be learned in terms of collaboration performance.

[0173] When integrating the data, the server ensures that the format and structure of the learning estimated collaboration performance knowledge points are consistent with the previously generated example collaboration performance knowledge points for subsequent comparison and analysis. The integrated learning estimated collaboration performance knowledge points will be stored for subsequent use.

[0174] Through this process, the server processes and analyzes the learning collaboration data of the heat supply operation system to be learned using multiple machine learning networks, generates the learning estimated collaboration performance knowledge points, and provides strong data support for the performance optimization of the heat supply operation system.

[0175] In a possible implementation manner, step S143 may include:

[0176] Step S1431, execute the target learning error calculation instruction. The target learning error calculation instruction includes one or more parameters among the first learning error calculation instruction, the second learning error calculation instruction, and the third learning error calculation instruction. Among them, the first learning error calculation instruction determines the first learning error parameter based on the difference between the learning collaborative transfer result and the learning collaborative transfer knowledge point. The first learning error calculation instruction determines the second learning error parameter based on the difference between the learning heat supply scheduling response result and the learning heat supply scheduling response knowledge point. The third learning error calculation instruction determines the third learning error parameter based on the difference between the learning heat supply collaborative stability result and the learning heat supply collaborative stability knowledge point.

[0177] Step S1432, determine the target learning error parameter based on one or more parameters among the first learning error parameter, the second learning error parameter, and the third learning error parameter.

[0178] Step S1433, perform parameter learning on the first machine learning network according to the target learning error parameter to generate the basic collaborative performance knowledge point estimation network.

[0179] In this embodiment, during the performance analysis and optimization of the heat supply operation system, the server has obtained learning collaborative data through the previous steps and generated learning estimated collaborative performance knowledge points and basic learning collaborative performance data. Next, the server will perform parameter learning on the first machine learning network based on the difference between the two to generate the basic collaborative performance knowledge point estimation network.

[0180] The server first executes the target learning error calculation instruction, and the purpose of this target learning error calculation instruction is to quantify the difference between the learning estimated collaborative performance knowledge points and the basic learning collaborative performance data, and this difference reflects the accuracy of the first machine learning network in estimating the collaborative performance knowledge points.

[0181] The server compares the learning collaborative transfer result (i.e., the fluid flow situation of the heat supply system predicted by the basic heat supply collaborative analysis network) with the learning collaborative transfer knowledge point (i.e., the collaborative transfer knowledge point estimated by the first machine learning network). The server calculates the difference between the two, such as the deviation in aspects such as flow rate and temperature distribution, so as to determine the first learning error parameter, and this parameter quantifies the error of the collaborative transfer estimation network in predicting the fluid transfer performance.

[0182] The server compares the learned heating scheduling response result (i.e., the response of the heating system to the scheduling instruction predicted by the basic heating collaboration analysis network) with the learned heating scheduling response knowledge points (i.e., the scheduling response knowledge points estimated by the first machine learning network). The server calculates the difference between the two, such as the deviation in aspects such as response time and adjustment range, so as to determine the second learning error parameter, which quantifies the error of the heating scheduling response estimation network in predicting the scheduling response performance.

[0183] The server compares the learned heating collaboration stability result (i.e., the stability of the heating system predicted by the basic heating collaboration analysis network) with the learned heating collaboration stability knowledge points (i.e., the collaboration stability knowledge points estimated by the first machine learning network). The server calculates the difference between the two, such as the deviation in aspects such as stability assessment and potential risk prediction, so as to determine the third learning error parameter, which quantifies the error of the heating collaboration stability estimation network in predicting the system stability.

[0184] After calculating the first learning error parameter, the second learning error parameter, and (if any) the third learning error parameter, the server comprehensively considers these parameters to determine a target learning error parameter. This process may involve the allocation and adjustment of weights to ensure that the target learning error parameter can comprehensively reflect the performance error of the first machine learning network in multiple aspects.

[0185] Finally, the server performs parameter learning on the first machine learning network based on the target learning error parameter. This process usually uses an optimization algorithm (such as the gradient descent algorithm) to iteratively adjust the weights and bias terms of the network to reduce the target learning error parameter. Through multiple iterations and optimizations, the server finally generates a more accurate basic collaboration performance knowledge point estimation network, which can better estimate the collaboration performance knowledge points of the heating operation system in subsequent analysis and provide strong support for heating energy efficiency optimization.

[0186] In a possible implementation manner, the method further includes:

[0187] Step S160, obtaining the target collaboration data of the target heating operation system and the target heating scheduling knowledge features.

[0188] In this embodiment, during the optimization process of the heating operation system, the server has trained and optimized multiple machine learning networks through the previous steps. Next, the server will use these networks to perform performance analysis and energy efficiency optimization assistance on the new target heating operation system.

[0189] The server first connects to the database or data acquisition system of the target heating operation system, obtains the heating operation trajectory characteristics and target heating scheduling knowledge characteristics of the system, and obtains the target collaborative data of the target heating operation system and the target heating scheduling knowledge characteristics.

[0190] Step S170: Load the target collaborative data into the collaborative performance knowledge point estimation network for collaborative performance knowledge point estimation to generate target estimated collaborative performance knowledge points.

[0191] In this embodiment, the server loads the obtained target collaborative data into the previously trained collaborative performance knowledge point estimation network. This collaborative performance knowledge point estimation network has been trained and optimized through a large number of example data and data to be learned, and can accurately estimate the collaborative performance knowledge points of the heating operation system.

[0192] After loading the target collaborative data, the collaborative performance knowledge point estimation network will analyze and process it to generate target estimated collaborative performance knowledge points. These target estimated collaborative performance knowledge points include collaborative transfer knowledge points, heating scheduling response knowledge points, and heating collaborative stability knowledge points of heating station G under specific conditions, etc. These knowledge points comprehensively reflect the collaborative performance of heating station G in actual operation.

[0193] Step S180: Based on the target estimated collaborative performance knowledge points, perform auxiliary analysis on the energy efficiency optimization of the target heating operation system to generate an auxiliary analysis result of energy efficiency optimization for heating.

[0194] In this embodiment, after obtaining the target estimated collaborative performance knowledge points, the server will perform auxiliary analysis on the energy efficiency optimization of the target heating operation system based on these knowledge points. This process may include the following steps:

[0195] The server first analyzes the collaborative performance bottlenecks reflected in the target estimated collaborative performance knowledge points. For example, if it is found that there are obvious deficiencies in the collaborative transfer of heating station G, it may mean that the layout of the heating pipeline is unreasonable or the control of the fluid flow rate is improper.

[0196] Based on the analysis results of the collaborative performance bottlenecks, the server further identifies the potential for energy efficiency optimization of heating station G. For example, by adjusting the layout of the heating pipeline, optimizing the fluid flow rate control strategy, etc., it may be possible to significantly improve the heating efficiency of the heating station.

[0197] The server generates specific optimization suggestions according to the analysis results of the energy efficiency optimization potential. These suggestions may include the scheme for improving the layout of the heating pipeline, the parameter settings for optimizing the fluid flow rate control strategy, etc. These suggestions are aimed at helping heating station G improve the heating efficiency and reduce the energy consumption cost.

[0198] Finally, the server integrates the above analysis process and results into an auxiliary analysis report on heating energy efficiency optimization. This report details the current situation, existing problems, and potential optimization solutions of heating station G in terms of collaborative performance. The management personnel of heating station G can formulate specific energy efficiency optimization measures based on this report.

[0199] For example, in a possible implementation manner, step S180 includes:

[0200] Step S181, generating a corresponding collaborative performance parameter knowledge graph based on the target estimated collaborative performance knowledge points.

[0201] In this embodiment, during the optimization process of the heating operation system, the server has obtained the target collaborative data of the target heating operation system through previous steps, and generated the target estimated collaborative performance knowledge points using the collaborative performance knowledge point estimation network. Next, the server will perform auxiliary analysis on heating energy efficiency optimization based on these knowledge points and generate corresponding analysis results.

[0202] The server first starts to construct a collaborative performance parameter knowledge graph according to the parameters in the target estimated collaborative performance knowledge points, such as collaborative transfer knowledge points, heating dispatch response knowledge points, heating collaborative stability knowledge points, etc.

[0203] For example, the server extracts key parameters from the target estimated collaborative performance knowledge points, such as uneven temperature distribution in the heating pipeline, unreasonable fluid flow rate control strategy, too long heating dispatch response time, etc. Then, the server uses a knowledge graph construction tool to take these key parameters as nodes and their association relationships as edges to construct a collaborative performance parameter knowledge graph. In this graph, the nodes may show mutual influence, causal relationships, etc. through different edges.

[0204] Step S182, analyzing the collaborative performance parameter knowledge graph to determine the system performance defect point data existing in the target heating operation system.

[0205] In this embodiment, the server then conducts in-depth analysis on the generated collaborative performance parameter knowledge graph. This analysis process can include multiple aspects:

[0206] Abnormal node identification: The server identifies abnormal nodes by comparing the attribute values (such as temperature distribution range, flow rate control parameters, etc.) of each node in the graph with the preset normal range or standard value. These abnormal nodes may represent fault points or potential problems in the heating operation system.

[0207] Sub - graph partitioning: The server divides the knowledge graph into multiple sub - graphs according to the types of nodes (such as heating pipelines, heat sources, user interfaces, etc.) and the types of edges (such as fluid flow relationships, temperature transfer relationships, etc.). Each sub - graph represents a specific domain or functional module in the heating operation system.

[0208] Performance evaluation: The server conducts performance evaluation on each sub - graph. By analyzing indicators such as the connectivity degree and interaction frequency of key nodes in the sub - graph, it determines the activity of key nodes within the sub - graph, the weight distribution of edges, and the critical path.

[0209] Defect point identification: Combining information such as abnormal nodes, the activity of key nodes within the sub - graph, the weight distribution of edges, and the critical path, the server can identify potential defect points in the heating operation system. These defect points can be the main reasons for system performance degradation and increased energy consumption.

[0210] Step S183: Based on the system performance defect point data existing in the target heating operation system, conduct auxiliary analysis on the energy efficiency optimization of the target heating operation system, and generate an auxiliary analysis result for energy efficiency optimization of heating.

[0211] After determining the system performance defect point data, the server starts to conduct in - depth analysis on these defect points and generates corresponding auxiliary analysis results for energy efficiency optimization of heating.

[0212] Defect point location: The server uses the critical path and association rules in the knowledge graph to accurately locate the identified potential defect points, and determine their specific locations and influence ranges.

[0213] Optimization strategy suggestions: Based on the location results of the defect points, the server generates specific optimization strategy suggestions. These suggestions can include improving the layout of heating pipelines, adjusting the fluid flow rate control strategy, optimizing the heating scheduling response mechanism, etc. These suggestions aim to improve the collaborative performance of the heating operation system and reduce energy consumption costs.

[0214] Generate an auxiliary analysis report for optimization: Finally, the server integrates the above - mentioned analysis process and results into an auxiliary analysis report for energy efficiency optimization of heating. This report details the performance defect points existing in the heating operation system, optimization strategy suggestions, and expected optimization effects. The management personnel of the heating operation system can formulate specific energy efficiency optimization measures based on this report.

[0215] For example, in a possible implementation manner, step S181 includes:

[0216] Step S1811: Extract key feature data from each target estimated collaborative performance knowledge point.

[0217] In this embodiment, during the optimization analysis of the heating operation system, the server has generated target estimated collaborative performance knowledge points through the previous steps. Next, the server will generate a collaborative performance parameter knowledge graph based on these knowledge points to more intuitively analyze and identify the key factors affecting the energy efficiency performance of the heating operation system.

[0218] The server first traverses the set of target estimated collaborative performance knowledge points and extracts the key feature data from them. These key feature data can include temperature distribution data of heating pipelines, fluid flow velocity data, heat source operation efficiency data, user interface flow data, etc. These data are the basis for analyzing the performance of the heating operation system.

[0219] For example, the server extracts the temperature distribution data of heating pipeline A and finds that its temperature fluctuates greatly during certain time periods; at the same time, it also extracts the flow velocity data of the fluid in heating pipeline B and finds that the flow velocity is unstable. These data may all affect the energy efficiency performance of the heating system.

[0220] Step S1812: Based on the key feature data, use the association rule mining algorithm to analyze the association relationships between the various data nodes associated with the target estimated collaborative performance knowledge points to determine the potential association rules between the various data nodes.

[0221] The server then uses an association rule mining algorithm (such as the Apriori algorithm or the FP-Growth algorithm) to analyze the association relationships between the extracted key feature data. The association rule mining algorithm can discover the potential association rules between data items, that is, which data items often appear together or have a certain dependency relationship.

[0222] In this example, the server runs the association rule mining algorithm to analyze the association relationship between the temperature distribution data of heating pipeline A and the flow velocity data of the fluid in heating pipeline B. The algorithm may find that when the temperature of heating pipeline A fluctuates greatly, the flow velocity of the fluid in heating pipeline B also tends to be unstable. This forms a potential association rule.

[0223] Step S1813: Based on the potential association rules, construct a network structure diagram of the target estimated collaborative performance knowledge points, analyze the corresponding key nodes and paths in the key feature data in the network structure diagram to identify the key factors affecting the energy efficiency performance of the heating operation system.

[0224] In this embodiment, the server constructs a network structure diagram of the target estimated collaborative performance knowledge points according to the potential association rules obtained by the association rule mining algorithm. In this network structure diagram, each key feature data corresponds to a node, and the association relationship between the nodes is represented by an edge. The weight of the edge can be set according to the confidence or support of the association rule.

[0225] After constructing the network structure diagram, the server analyzes the key nodes and paths in the diagram. Key nodes are usually those with high centrality or connectivity in the diagram, and they often represent the key factors affecting the energy efficiency performance of the heating system. Key paths are the paths formed by the edges connecting key nodes, revealing the mutual influence relationships between key factors.

[0226] In this example, the server discovers that Heating Pipe A and Heating Pipe B have high centrality in the network structure diagram, and there is an edge with a high weight between them. This indicates that the temperature fluctuation of Heating Pipe A and the unstable flow rate of Heating Pipe B are the key factors affecting the energy efficiency performance of the heating system, and there is a mutual influence relationship between them.

[0227] Step S1814, according to the key nodes and paths in the network structure diagram, configure the structural framework of the collaborative performance parameter knowledge graph, and the structural framework includes node types, edge types, and attribute definitions.

[0228] After identifying the key nodes and paths, the server starts to configure the structural framework of the collaborative performance parameter knowledge graph, which includes node types, edge types, attribute definitions, etc.

[0229] In this example, the server defines Heating Pipe A and Heating Pipe B as the "heating pipe" type in the node type; defines their association relationship as the "temperature - flow rate association" type in the edge type; and defines corresponding attributes for the nodes and edges. For example, node attributes include pipe name, length, etc., and edge attributes include confidence and support of the association rule, etc.

[0230] Step S1815, after defining the key nodes as the nodes in the collaborative performance parameter knowledge graph, defining the paths between the nodes as edges, and defining corresponding attributes for the nodes and edges, use a visualization tool to generate the collaborative performance parameter knowledge graph, and the attributes include node type and edge weight.

[0231] Finally, the server uses a visualization tool (such as Gephi, D3.js, etc.) to present the configured collaborative performance parameter knowledge graph. In the visualization graph, key nodes and key paths will be highlighted so that users can quickly identify the key factors affecting the energy efficiency performance of the heating system and their mutual relationships.

[0232] Through this process, the server generates an intuitive and easy - to - understand collaborative performance parameter knowledge graph, providing strong data support and analysis tools for the energy efficiency optimization of the heating operation system.

[0233] For example, in a possible implementation manner, step S182 includes:

[0234] Step S1821: By comparing the attribute values of the key nodes with the differences in the preset ranges, identify the abnormal key nodes, where the abnormal key nodes represent the fault points or potential problems in the heating operation system.

[0235] In this embodiment, after the server generates the collaborative performance parameter knowledge graph, it is next necessary to deeply analyze this graph to identify the system performance defect points existing in the target heating operation system. The following are detailed scenario examples for each analysis step.

[0236] The server will first traverse all the key nodes in the collaborative performance parameter knowledge graph and obtain the attribute values of these nodes (such as temperature, flow rate, pressure, etc.). Then, the server can compare these attribute values with the preset normal ranges. If the attribute value of a certain node exceeds the normal range, then the server will identify it as an abnormal key node.

[0237] For example, among the key nodes of heating pipeline A, the server finds that the temperature attribute value is 85°C, while the preset normal temperature range is 60°C to 75°C. Since 85°C exceeds the normal range, the server identifies heating pipeline A as an abnormal key node, which may mean that there is an overheating problem with heating pipeline A.

[0238] Step S1822: According to the node types of the key nodes and the edge types of the edges, divide the collaborative performance parameter knowledge graph into multiple subgraphs, where each subgraph represents a specific domain or functional module in the heating operation system.

[0239] Next, the server will divide the collaborative performance parameter knowledge graph into multiple subgraphs according to the node types of the key nodes and the edge types of the edges. Each subgraph represents a specific domain or functional module in the heating operation system.

[0240] For example, the server can divide all the key nodes and edges related to the heat source into a subgraph named "Heat Source Management"; divide all the key nodes and edges related to the heating pipelines into the "Heating Pipeline Network" subgraph; divide all the key nodes and edges related to the user interface into the "User Interface Service" subgraph.

[0241] Step S1823: Perform performance evaluation on each of the subgraphs, and determine the activity of the key nodes, the weight distribution of the edges, and the critical path within the subgraph according to the connection degree and interaction frequency of the key nodes in the subgraph.

[0242] The server can perform performance evaluation on each subgraph, and this evaluation process mainly includes calculating the activity of the key nodes within the subgraph, the weight distribution of the edges, and identifying the critical path.

[0243] Taking the "heating pipeline network" sub - graph as an example, the server can analyze the connection degree and interaction frequency of each key node in this sub - graph to determine their importance in the heating pipeline network. Nodes with a high connection degree tend to have a higher activity level and have a greater impact on the performance of the heating system. At the same time, the server will also identify key paths based on the weight distribution of the edges, and these paths usually represent the main channels of fluid flow in the heating pipeline network.

[0244] Step S1824, combining the abnormal key nodes, the activity levels of the key nodes within the sub - graph, the weight distribution of the edges, and the key paths, identify potential defect points in the heating operation system.

[0245] After completing the performance evaluation of the sub - graph, the server can combine information such as abnormal key nodes, the activity levels of key nodes within the sub - graph, the weight distribution of the edges, and the key paths to identify potential defect points in the heating operation system.

[0246] Continuing with the "heating pipeline network" sub - graph as an example, the server discovers that heating pipeline A is not only an abnormal key node (abnormal temperature), but also has a high connection degree in the sub - graph, indicating that it has an important impact on the performance of the heating system. At the same time, the server also finds that the edge weight between heating pipeline A and heating pipeline B is large, which means that the fluid flow between them also has a relatively significant impact on the heating system. Based on this information, the server can preliminarily judge that there may be potential defect points in heating pipeline A and its connection with heating pipeline B.

[0247] Step S1825, using the key paths and association rules in the collaborative performance parameter knowledge graph, locate the identified potential defect points to determine the specific location and influence range of the potential defect points.

[0248] Finally, the server can use the key paths and association rules in the collaborative performance parameter knowledge graph to more accurately locate the identified potential defect points and determine their specific location and influence range.

[0249] In the above example, the server can further analyze the key paths and association rules between heating pipeline A and heating pipeline B. By tracking the flow paths of the fluid in these two pipelines and their mutual influence relationships, the server can more accurately locate the specific location of the potential defect point (such as a specific section of heating pipeline A or its connection with heating pipeline B), and evaluate the range of its possible impact on the performance of the heating system (such as which user interfaces' heating effects are affected and how much energy loss is caused, etc.).

[0250] Figure 2The figure shows the hardware structure diagram of the machine learning system 100 provided by the embodiments of the present application for implementing the above-mentioned machine learning-based auxiliary analysis method for heating energy efficiency optimization, as Figure 2 shown, the machine learning system 100 may include a processor 110, a machine-readable storage medium 120, a bus 130, and a communication unit 140.

[0251] In an alternative embodiment, the machine learning system 100 may be a single server or a server group. The server group may be centralized or distributed (for example, the machine learning system 100 may be a distributed system). In an alternative embodiment, the machine learning system 100 may be local or remote. For example, the machine learning system 100 may access the information and / or data stored in the machine-readable storage medium 120 via a network. For another example, the machine learning system 100 may be directly connected to the machine-readable storage medium 120 to access the stored information and / or data. In an alternative embodiment, the machine learning system 100 may be implemented on a cloud platform. By way of example only, the cloud platform may include a private cloud, a public cloud, an aggregated cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, etc. or any integration thereof.

[0252] The machine-readable storage medium 120 may store data and / or instructions. In an alternative embodiment, the machine-readable storage medium 120 may store data obtained from an external terminal. In an alternative embodiment, the machine-readable storage medium 120 may store the data and / or instructions used by the machine learning system 100 to execute or use to complete the exemplary methods described in the present application. In an alternative embodiment, the machine-readable storage medium 120 may include a mass storage device, a removable storage device, a volatile read / write memory, a read-only memory, etc. or any integration thereof. Exemplary mass storage devices may include magnetic disks, optical disks, solid-state disks, etc. Exemplary removable storage devices may include flash drives, floppy disks, optical disks, memory cards, compact disks, magnetic tapes, etc.

[0253] In a specific implementation process, multiple processors 110 execute the computer-executable instructions stored in the machine-readable storage medium 120, so that the processors 110 can execute the machine learning-based auxiliary analysis method for heating energy efficiency optimization in the above method embodiments. The processors 110, the machine-readable storage medium 120, and the communication unit 140 are connected through the bus 130, and the processors 110 can be used to describe the transceiver actions of the communication unit 140.

[0254] For the specific implementation process of the processor 110, reference may be made to the various method embodiments executed by the above-mentioned machine learning system 100. The implementation principles and technical effects are similar, and will not be elaborated herein in this embodiment.

[0255] In addition, an embodiment of the present application further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above-mentioned auxiliary analysis method for optimizing heating energy efficiency based on machine learning is implemented.

[0256] Similarly, it should be noted that, in order to simplify the description of the present application disclosure and thus help the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, sometimes multiple features are merged into one embodiment, drawing or description thereof.

Claims

1. An auxiliary analysis method for optimizing heating energy efficiency based on machine learning, characterized in that, The method includes: Obtaining the example heating operation trajectory characteristics, example heating scheduling knowledge characteristics of the example heating operation system, and the example collaborative data between the example heating operation system and the example heating scheduling knowledge characteristics; Loading the example collaborative data into a first machine learning network for collaborative performance knowledge point estimation to generate example collaborative performance knowledge points; Loading the example heating operation trajectory characteristics and the example heating scheduling knowledge characteristics into a second machine learning network for performance estimation analysis to generate example collaborative performance estimation data; Performing parameter learning on the second machine learning network according to the difference between the example collaborative performance estimation data and the example collaborative performance knowledge points to generate a basic heating collaborative analysis network; Performing parameter learning on the first machine learning network according to the learned heating operation trajectory characteristics, learned heating scheduling knowledge characteristics, learned collaborative data of the learned heating operation system, and the basic heating collaborative analysis network to generate a basic collaborative performance knowledge point estimation network; the learned collaborative data is the collaborative data between the learned heating operation system and the learned heating scheduling knowledge characteristics; the learned heating operation system and the example heating operation system are heating operation systems of the same type, the learned heating operation trajectory characteristics and the example heating operation trajectory characteristics are features of the same dimension, the learned heating scheduling knowledge characteristics and the example heating scheduling knowledge characteristics are features of the same dimension, and the learned collaborative data and the example collaborative data are data of the same dimension; Performing cyclic parameter learning on the basic heating collaborative analysis network and the basic collaborative performance knowledge point estimation network, and outputting the basic collaborative performance knowledge point estimation network when the parameter learning is completed as a collaborative performance knowledge point estimation network, where the collaborative performance knowledge point estimation network is used to perform auxiliary analysis on the heating energy efficiency optimization of the target heating operation system to generate a heating energy efficiency optimization auxiliary analysis result; the cyclic parameter learning includes cyclic operations: locking the network parameter information of the basic collaborative performance knowledge point estimation network to perform parameter learning on the basic heating collaborative analysis network and locking the network parameter information of the basic heating collaborative analysis network to perform parameter learning on the basic collaborative performance knowledge point estimation network.

2. The auxiliary analysis method for optimizing heating energy efficiency based on machine learning according to claim 1, wherein The performing parameter learning on the first machine learning network according to the learned heating operation trajectory characteristics, learned heating scheduling knowledge characteristics, learned collaborative data of the learned heating operation system, and the basic heating collaborative analysis network to generate a basic collaborative performance knowledge point estimation network includes: Loading the learned heating operation trajectory characteristics and the learned heating scheduling knowledge characteristics into the basic heating collaborative analysis network for performance estimation analysis to generate basic learned collaborative performance data; Loading the learned collaborative data into the first machine learning network for collaborative performance knowledge point estimation to generate learned estimated collaborative performance knowledge points; Based on the difference between the learned estimation collaborative performance knowledge points and the basic learning collaborative performance data, parameter learning is performed on the first machine learning network to generate the basic collaborative performance knowledge point estimation network.

3. The auxiliary analysis method for optimizing heating energy efficiency based on machine learning according to claim 2, characterized in that The cyclic parameter learning of the basic heat supply collaborative analysis network and the basic collaborative performance knowledge point estimation network, and outputting the basic collaborative performance knowledge point estimation network when the parameter learning is completed as the collaborative performance knowledge point estimation network, includes: Loading the example collaborative data into the basic collaborative performance knowledge point estimation network for collaborative performance knowledge point estimation to generate basic example collaborative performance knowledge points; Loading the example heat supply operation trajectory features and the example heat supply scheduling knowledge features into the basic heat supply collaborative analysis network for performance estimation and analysis to generate basic example collaborative performance estimation data; Based on the difference between the basic example collaborative performance estimation data and the basic example collaborative performance knowledge points, parameter learning is performed on the basic heat supply collaborative analysis network, and the network when the parameter learning is completed is used again as the basic heat supply collaborative analysis network; Loading the learned heat supply operation trajectory features and the learned heat supply scheduling knowledge features into the basic heat supply collaborative analysis network for performance estimation and analysis to generate basic learned collaborative performance data; Based on the learned collaborative data and the basic learned collaborative performance data, parameter learning is performed on the basic collaborative performance knowledge point estimation network, and the network when the parameter learning is completed is used again as the basic collaborative performance knowledge point estimation network; Repeating the steps of loading the example collaborative data into the basic collaborative performance knowledge point estimation network for collaborative performance knowledge point estimation to generate basic example collaborative performance knowledge points, to performing parameter learning on the basic collaborative performance knowledge point estimation network based on the learned collaborative data and the basic learned collaborative performance data, and using the network when the parameter learning is completed again as the basic collaborative performance knowledge point estimation network until the network convergence requirement is met; Outputting the basic heat supply collaborative analysis network when the parameter learning is completed as the heat supply collaborative analysis network, and outputting the basic collaborative performance knowledge point estimation network when the parameter learning is completed as the collaborative performance knowledge point estimation network.

4. The machine learning-based auxiliary analysis method for optimizing heating energy efficiency according to claim 1, wherein The example collaborative data includes one or more parameters among example collaborative transfer monitoring data, example heat supply scheduling response data, and heat supply collaborative stability data. The first machine learning network includes one or more estimation networks among a collaborative transfer knowledge point estimation network, a heat supply scheduling response knowledge point estimation network, and a heat supply collaborative stability knowledge point estimation network. The first machine learning network is determined based on the example collaborative data. The loading the example collaborative data into the first machine learning network for collaborative performance knowledge point estimation to generate example collaborative performance knowledge points includes: Execute the target collaborative performance knowledge point estimation instruction, where the target collaborative performance knowledge point estimation instruction includes at least one of the first collaborative performance knowledge point estimation, the second collaborative performance knowledge point estimation, and the third collaborative performance knowledge point estimation; wherein, the first collaborative performance knowledge point estimation is to load the example collaborative transfer monitoring data into the collaborative transfer knowledge point estimation network, perform collaborative transfer knowledge point estimation, and generate example collaborative transfer knowledge points; the second collaborative performance knowledge point estimation is to load the example heat supply dispatching response data into the heat supply dispatching response knowledge point estimation network, perform heat supply dispatching response knowledge point estimation, and generate example heat supply dispatching response knowledge points; the third collaborative performance knowledge point estimation is to load the heat supply collaborative stability data into the heat supply collaborative stability knowledge point estimation network, perform heat supply collaborative stability knowledge point estimation, and generate example heat supply collaborative stability knowledge points; Output one or more parameters among the example collaborative transfer knowledge points, the example heat supply dispatching response knowledge points, and the example heat supply collaborative stability knowledge points as the example collaborative performance knowledge points.

5. The auxiliary analysis method for optimizing heating energy efficiency based on machine learning according to claim 4, characterized in that The second machine learning network includes one or more estimation networks among the collaborative transfer estimation network, the heat supply dispatching response estimation network, and the heat supply collaborative stability estimation network. The second machine learning network is determined based on the example collaborative data. Loading the example heat supply operation trajectory features and the example heat supply dispatching knowledge features into the second machine learning network for performance estimation and analysis to generate example collaborative performance estimation data includes: Execute the target performance estimation instruction, where the target performance estimation instruction includes at least one of the first performance estimation instruction, the second performance estimation instruction, and the third performance estimation instruction; wherein, the first performance estimation instruction is to load the example heat supply operation trajectory features and the example heat supply dispatching knowledge features into the collaborative transfer estimation network for collaborative transfer estimation, and generate example collaborative transfer results; the second performance estimation instruction is to load the example heat supply operation trajectory features and the example heat supply dispatching knowledge features into the heat supply dispatching response estimation network for heat supply dispatching response estimation, and generate example heat supply dispatching response results; the third performance estimation instruction is to load the example heat supply operation trajectory features and the example heat supply dispatching knowledge features into the heat supply collaborative stability estimation network for heat supply collaborative stability estimation, and generate example heat supply collaborative stability results; Output one or more parameters among the example collaborative transfer results, the example heat supply dispatching response results, and the example heat supply collaborative stability results as the example collaborative performance estimation data; the example collaborative performance estimation data has the same feature category as the example collaborative performance knowledge points; Among them, parameter learning is performed on the second machine learning network according to the difference between the example collaborative performance estimation data and the example collaborative performance knowledge points to generate a basic heat supply collaborative analysis network, including: Execute the target error calculation instruction; the target error calculation instruction includes one or more parameters among the first error calculation instruction, the second error calculation instruction, and the third error calculation instruction; wherein, the first error calculation instruction is to determine the first example error parameter according to the difference between the example collaborative transfer result and the example collaborative transfer knowledge point; the second error calculation instruction is to determine the second example error parameter according to the difference between the example heat supply dispatching response result and the example heat supply dispatching response knowledge point; the third error calculation instruction is to determine the third example error parameter according to the difference between the example heat supply collaborative stability result and the example heat supply collaborative stability knowledge point; Determine the target example error parameter according to one or more parameters among the first example error parameter, the second example error parameter, and the third example error parameter; Update the network parameter information of the second machine learning network according to the target example error parameter to generate a basic heat supply collaborative analysis network.

6. The machine learning-based auxiliary analysis method for optimizing heating energy efficiency according to claim 5, characterized in that The method further includes: Execute one or more parameters among the first influence coefficient for obtaining the example collaborative transfer monitoring data, the second influence coefficient for obtaining the example heat supply dispatching response data, and the third influence coefficient for obtaining the heat supply collaborative stability data; The determining the target example error parameter according to one or more parameters among the first example error parameter, the second example error parameter, and the third example error parameter includes: Execute one or more parameters among determining the first fusion error parameter based on the first influence coefficient and the first example error parameter, determining the second fusion error parameter based on the second influence coefficient and the second example error parameter, and determining the third fusion error parameter based on the third influence coefficient and the third example error parameter; Determine the target example error parameter based on one or more parameters among the first fusion error parameter, the second fusion error parameter, and the third fusion error parameter.

7. The auxiliary analysis method for optimizing heating energy efficiency based on machine learning according to claim 2, characterized in that The basic heat supply collaborative analysis network includes one or more networks among the basic collaborative transfer estimation network, the basic heat supply dispatching response estimation network, and the basic heat supply collaborative stability estimation network. The loading the learned heat supply operation trajectory feature and the learned heat supply dispatching knowledge feature into the basic heat supply collaborative analysis network for performance estimation analysis to generate the basic learning collaborative performance data includes: Execute the following one or more loading instructions for the learned heat supply operation trajectory feature and the learned heat supply dispatching knowledge feature: The first loading instruction: Load into the basic collaborative transfer estimation network for collaborative transfer prediction to generate a learned collaborative transfer result; The second loading instruction: Load into the basic heat supply dispatching response estimation network for heat supply dispatching response estimation to generate a learned heat supply dispatching response result; The third loading instruction: Load into the basic heat supply collaborative stability estimation network for heat supply collaborative stability estimation to generate a learned heat supply collaborative stability result; Output one or more parameters among the learning collaborative transfer result, the learning heat supply scheduling response result, and the learning heat supply collaborative stability result as the learning collaborative performance data; Among them, the learning collaborative data includes at least one of learning collaborative transfer monitoring data, learning heat supply scheduling response data, and learning stable operation duration data. Loading the learning collaborative data into the first machine learning network for collaborative performance knowledge point estimation to generate learning estimated collaborative performance knowledge points includes: Execute the learning knowledge point estimation instruction; the learning knowledge point estimation instruction includes one or more parameters among the first knowledge point estimation instruction, the second knowledge point estimation instruction, and the third knowledge point estimation instruction; among them, the first knowledge point estimation instruction is to load the learning collaborative transfer monitoring data into the collaborative transfer knowledge point estimation network for collaborative transfer knowledge point estimation to generate learning collaborative transfer knowledge points; the second knowledge point estimation instruction is to load the learning heat supply scheduling response data into the heat supply scheduling response knowledge point estimation network for heat supply scheduling response knowledge point estimation to generate learning heat supply scheduling response knowledge points; the third knowledge point estimation instruction is to load the learning stable operation duration data into the heat supply collaborative stability knowledge point estimation network for heat supply collaborative stability knowledge point estimation to generate learning heat supply collaborative stability knowledge points; Output one or more parameters among the learning collaborative transfer knowledge points, the learning heat supply scheduling response knowledge points, and the learning heat supply collaborative stability knowledge points as the learning estimated collaborative performance knowledge points.

8. The auxiliary analysis method for optimizing heating energy efficiency based on machine learning according to claim 7, characterized in that Based on the difference between the learning estimated collaborative performance knowledge points and the basic learning collaborative performance data, performing parameter learning on the first machine learning network to generate the basic collaborative performance knowledge point estimation network includes: Execute the target learning error calculation instruction; the target learning error calculation instruction includes one or more parameters among the first learning error calculation instruction, the second learning error calculation instruction, and the third learning error calculation instruction; among them, the first learning error calculation instruction is to determine the first learning error parameter based on the difference between the learning collaborative transfer result and the learning collaborative transfer knowledge points; the first learning error calculation instruction is to determine the second learning error parameter based on the difference between the learning heat supply scheduling response result and the learning heat supply scheduling response knowledge points; the third learning error calculation instruction is to determine the third learning error parameter based on the difference between the learning heat supply collaborative stability result and the learning heat supply collaborative stability knowledge points; Determine the target learning error parameter based on one or more parameters among the first learning error parameter, the second learning error parameter, and the third learning error parameter; Perform parameter learning on the first machine learning network based on the target learning error parameter to generate the basic collaborative performance knowledge point estimation network.

9. The machine learning-based auxiliary analysis method for optimizing heating energy efficiency according to any one of claims 1-8, characterized in that The method further includes: Obtain the target collaborative data of the target heat supply operation system and the target heat supply scheduling knowledge features; Loading the target collaborative data into a collaborative performance knowledge point estimation network for estimating collaborative performance knowledge points to generate target estimated collaborative performance knowledge points; Based on the target estimated collaborative performance knowledge points, performing auxiliary analysis on the target heating operation system for heating energy efficiency optimization to generate an auxiliary analysis result of heating energy efficiency optimization.

10. A machine learning system, characterized in that, The machine learning system includes a processor and a machine-readable storage medium, and machine-executable instructions are stored in the machine-readable storage medium. The machine-executable instructions are loaded and executed by the processor to implement the method for auxiliary analysis of heating energy efficiency optimization based on machine learning according to any one of claims 1-9.

Images