A method and device for analyzing heating system failure based on artificial intelligence

Abstract

The present invention provides a method and device for analyzing a heating system fault based on artificial intelligence, and relates to the field of data processing technology. The method comprises: fusing environmental dynamic data with a digital twin model to establish a dynamic digital twin model; analyzing the fault type based on the dynamic digital twin model, wherein the dynamic digital twin model identifies the fault type by comparing the normal operating data, the fault operating data mode, and the influence of the environmental dynamic data; predicting the fault development trend and influence according to the fault type; and estimating the risk level according to the fault development trend and influence. By establishing a digital twin model and comparing it with the actual operating data, the present invention can more accurately detect abnormal parameters of the heating system and realize rapid positioning of the fault.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and device for analyzing failures of a heating system based on artificial intelligence. Background Art

[0002] With the widespread application of building heating systems, their safe and stable operation is of great significance to ensuring the use function and living comfort of buildings. However, traditional heating systems have problems such as insufficient utilization of operating data and delayed fault detection response, which makes it difficult to quickly and accurately locate system faults and conduct risk assessment. Currently, the commonly used heating system fault analysis methods mainly include:

[0003] 1. Rely on the operator's experience and on-site inspection to locate the fault. This method relies on personal experience and has low accuracy and efficiency.

[0004] 2. Use physical or mathematical modeling to analyze faults, but traditional models cannot integrate various influencing factors, and the simulation of the system is not accurate enough. The fault type is judged based on the data collected by the sensor, but it is difficult to consider the impact of environmental dynamic factors, and the prediction of fault development trends is insufficient; rely on historical statistical data for risk assessment, but it is difficult to adapt to the impact of environmental changes on risks. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a heating system fault analysis method and device based on artificial intelligence. By establishing a digital twin model and comparing it with the actual operation data, it is possible to more accurately detect abnormal parameters of the heating system and quickly locate the fault.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0007] In a first aspect, a heating system fault analysis method based on artificial intelligence, the method comprising:

[0008] Obtaining the operating data of the heating system;

[0009] Build a digital twin model of the heating system based on operating data;

[0010] When the digital twin model detects abnormal heating system parameters, it obtains environmental dynamic data;

[0011] Integrate environmental dynamic data with the digital twin model to establish a dynamic digital twin model;

[0012] Analyze fault types based on the dynamic digital twin model. The dynamic digital twin model identifies fault types by comparing the patterns of normal operation data, fault operation data, and the impact of environmental dynamic data.

[0013] According to the fault type, predict the fault development trend and impact;

[0014] Estimate the risk level based on fault development trends and impacts.

[0015] Furthermore, a digital twin model of the heating system is established based on the operating data, including:

[0016] Obtaining the structural data of the heating system and the historical operation data of the heating system;

[0017] Construct a physical model of the heating system based on the structural data and historical operation data of the heating system;

[0018] According to the physical model, a mathematical model of the heating system is established;

[0019] By using historical operation data as input signals, unknown parameters in the mathematical model are adjusted so that the mathematical model outputs historical data that meets the requirements;

[0020] Input historical data that meets the requirements into the physical model and mathematical model to form a digital twin model of the heating system.

[0021] Furthermore, when the digital twin model detects abnormal heating system parameters, it obtains environmental dynamic data, including:

[0022] When the digital twin model simulates the operation of the heating system, it monitors the key parameters of the heating system and compares the key parameters with the set normal operating range;

[0023] If the monitored key parameters are ≥ normal range, the heating system is judged to be faulty or abnormal, and the heating system automatically obtains relevant environmental dynamic data.

[0024] Furthermore, the environmental dynamic data is integrated with the digital twin model to establish a dynamic digital twin model, including:

[0025] Preprocessing the acquired environmental dynamic data to obtain preprocessed data;

[0026] Analyze the correlation points between the parameters in the preprocessed data and the digital twin model, and determine the submodules related to the environmental dynamic data;

[0027] Add an input interface that matches the environmental dynamic data parameters in the framework of the digital twin model;

[0028] Design a dynamic data-driven mechanism to activate the update calculation of the submodule when the preprocessed data is input through the newly added interface;

[0029] Call the newly added dynamic data input interface in the model framework and activate the dynamic driving mechanism to generate a dynamic digital twin model.

[0030] Furthermore, based on the dynamic digital twin model, the fault type is analyzed. The dynamic digital twin model identifies the fault type by comparing the normal operation data, the fault operation data mode and the impact of the environmental dynamic data, including:

[0031] Within the framework of the dynamic digital twin model, establish a normal operation data module, a fault operation data module, and an environmental dynamic data module;

[0032] Inputting normal operation data, fault operation data and environmental dynamic data into the operation data module, the fault operation data module and the environmental dynamic data module respectively;

[0033] Compare the pattern features of the operation data module, the fault operation data module and the environmental dynamic data module respectively to obtain the difference between the fault operation data and the normal operation data;

[0034] According to the impact of environmental dynamic data, analyze the causes of the differences and determine the type of fault that causes the differences.

[0035] Furthermore, according to the fault type, the fault development trend and impact are predicted, including:

[0036] Build a knowledge base of fault types;

[0037] After determining the fault type, use the fault type as a keyword to query the knowledge base to obtain the development law information of the fault type;

[0038] Use the historical cases of type failures obtained from the knowledge base to train the prediction model of failure development to obtain the final prediction model;

[0039] Input the parameters of the current fault into the final prediction model to obtain the fault development path;

[0040] Analyze the impact scope of the knowledge base and the fault development path to obtain analysis results;

[0041] According to the influence of environmental dynamic data on the analysis results, corrections are made to obtain the final impact prediction.

[0042] Furthermore, based on the fault development trend and impact, the risk level is estimated, including:

[0043] Construct a fault risk assessment model based on the fault development speed, impact scope and environmental dynamic factors;

[0044] Preset the determination thresholds of different risk levels in the fault risk assessment model;

[0045] Calculate the comprehensive risk value based on the fault development trend and fault risk assessment model;

[0046] According to the comprehensive risk value and the preset threshold, a preliminary risk level is obtained;

[0047] According to environmental dynamic factors, the preliminary risk level is adjusted to obtain the adjusted result;

[0048] Based on the adjusted results, the final risk level is obtained.

[0049] In a second aspect, a heating system fault analysis device based on artificial intelligence comprises:

[0050] The acquisition module is used to acquire the operating data of the heating system; establish a digital twin model of the heating system based on the operating data; acquire environmental dynamic data when the digital twin model detects that the parameters of the heating system are abnormal; fuse the environmental dynamic data with the digital twin model to establish a dynamic digital twin model;

[0051] The processing module is used to analyze the fault type based on the dynamic digital twin model. The dynamic digital twin model identifies the fault type by comparing the patterns of normal operation data, fault operation data and the influence of environmental dynamic data; predicts the fault development trend and impact based on the fault type; and estimates the risk level based on the fault development trend and impact.

[0052] According to a third aspect, a computing device includes:

[0053] one or more processors;

[0054] The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the above method.

[0055] In a fifth aspect, a computer-readable storage medium stores a program, and the program implements the above method when executed by a processor.

[0056] The above solution of the present invention includes at least the following beneficial effects:

[0057] The above-mentioned scheme of the present invention, by establishing a digital twin model and comparing it with the actual operation data, can more accurately detect abnormal parameters of the heating system and realize rapid positioning of faults; by fusing the environmental dynamic data with the digital twin model to establish a dynamic digital twin model, it can consider the impact of environmental factors on faults and realize dynamic analysis of fault types; based on the dynamic digital twin model, by comparing the patterns of normal operation data, fault operation data and the impact of environmental dynamic data, it can predict the development trend and impact of faults and take corresponding maintenance and processing measures in advance; according to the development trend and impact of faults, combined with the risk assessment model, it can estimate the risk level of heating system failures and help decision makers assess the impact of faults on system operation and users; through rapid fault location, dynamic fault analysis and risk assessment, it is possible to take timely measures to improve the reliability and safety of the heating system and reduce the impact of system failures on users. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart of a heating system fault analysis method based on artificial intelligence provided in an embodiment of the present invention.

[0059] Figure 2 It is a schematic diagram of a heating system fault analysis device based on artificial intelligence provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0061] like Figure 1 As shown, an embodiment of the present invention provides a heating system fault analysis method based on artificial intelligence, the method comprising:

[0062] Step 11, obtaining operation data of the heating system;

[0063] Step 12: Establish a digital twin model of the heating system based on the operating data;

[0064] Step 13, when the digital twin model detects that the heating system parameters are abnormal, obtain environmental dynamic data;

[0065] Step 14, integrating the environmental dynamic data with the digital twin model to establish a dynamic digital twin model;

[0066] Step 15: Analyze the fault type based on the dynamic digital twin model. The dynamic digital twin model identifies the fault type by comparing the patterns of normal operation data, fault operation data, and the influence of environmental dynamic data;

[0067] Step 16, predicting the fault development trend and impact according to the fault type;

[0068] Step 17: Estimate the risk level based on the fault development trend and impact.

[0069] In an embodiment of the present invention, by establishing a digital twin model and comparing it with the actual operation data, it is possible to more accurately detect abnormal parameters of the heating system and quickly locate the fault; by fusing the environmental dynamic data with the digital twin model to establish a dynamic digital twin model, it is possible to consider the impact of environmental factors on the fault and realize dynamic analysis of the fault type; based on the dynamic digital twin model, by comparing the patterns of normal operation data, fault operation data and the impact of environmental dynamic data, it is possible to predict the development trend and impact of the fault and take corresponding maintenance and processing measures in advance; according to the development trend and impact of the fault, combined with the risk assessment model, it is possible to estimate the risk level of the heating system failure, and help decision makers assess the impact of the fault on the system operation and users; through rapid fault location, dynamic fault analysis and risk assessment, timely measures can be taken to improve the reliability and safety of the heating system and reduce the impact of system failures on users.

[0070] In a preferred embodiment of the present invention, the above step 12 may include:

[0071] Step 121, obtaining structural data of the heating system and historical operation data of the heating system;

[0072] Step 122, constructing a physical model of the heating system according to the structural data and historical operation data of the heating system;

[0073] Step 123, establishing a mathematical model of the heating system according to the physical model;

[0074] Step 124, using the historical operation data as an input signal, adjusting the unknown parameters in the mathematical model so that the mathematical model outputs historical data that meets the requirements;

[0075] Step 125, inputting the historical data that meets the requirements into the physical model and the mathematical model to form a digital twin model of the heating system.

[0076] In an embodiment of the present invention, a physical model is constructed by structural data to accurately reflect the physical characteristics of the system. A mathematical model is established on the basis of the physical model to enable the model to have prediction and simulation capabilities. The parameters of the mathematical model are calibrated using historical operation data to make the model output consistent with the actual system. Multi-source data drives the physical model and the mathematical model to form a digital twin model that reflects the actual operating characteristics of the system. By refining the steps of digital twin modeling, it is beneficial to improve the accuracy and interpretability of the model, making subsequent model-based fault detection and prediction more accurate and reliable, thereby improving the maintainability and safety of the system.

[0077] In another preferred embodiment of the present invention, in the above step 121, the structural data includes a pipe network layout diagram, equipment parameter table, pipe material parameters, etc. of the heating system, and these data can be obtained from sources such as system design drawings and equipment manuals. Historical operation data refers to the operating parameters of the system over a period of time in the past, such as data sequences such as temperature, pressure, flow, etc. These data can be obtained from databases, logs, etc. The structural data reflects the static information of the system such as connection, composition, and materials, and the operating data reflects the dynamic working status of the system. The two types of data complement each other and are both very important for establishing a digital twin model. The structural data provides the basic framework, and the operating data provides the basis for parameter adjustment.

[0078] In another preferred embodiment of the present invention, in the above step 122, based on the structural data, the components such as pipes, pipe fittings, equipment and their connection relationships in the heating system are determined to establish the physical network structure of the system; the physical parameters of each component, such as the length, diameter, wall thickness, material of the pipe, the model and specifications of the equipment and other parameters are determined; model blocks matching the actual components are obtained, and they are assembled according to the structural connection relationships to form a physical model network of the heating system; the parameters of the components in the model are adjusted to be consistent with the component parameters in the actual system; the temperature, pressure and other parameters in the historical operation data are obtained as the model boundary conditions and initialization conditions; the physical model is run so that its output results conform to the historical operation data under the given boundary and initialization conditions; the above steps are repeated until the error between the model output and the actual system operation data meets the requirements, and finally a physical model that can reflect the physical characteristics of the actual heating system is obtained.

[0079] In another preferred embodiment of the present invention, in the above step 123, a pipeline flow continuity equation is established based on the physical model. Momentum equation And the energy equation in, represents the partial derivative with respect to time, ρ represents the fluid density, represents the Reynolds mean of the velocity vector, represents the gradient operator, represents the tensor product, μ represents the dynamic viscosity coefficient, represents the pressure gradient, represents the velocity gradient, T represents the transposed sign, represents the stress tensor, g represents the acceleration of gravity, and φ is the viscous heat dissipation term. In the above step 124, the historical operation data of the heating system in the past period of time, such as time series data such as temperature, pressure, and flow, are obtained; the historical operation data are input into the established mathematical model in chronological order; the mathematical model is run to obtain the output response results of each variable; the output results of the mathematical model are compared with the corresponding historical operation data. If there is a difference between the two, an unknown parameter in the mathematical model is selected for adjustment, and the mathematical model is repeatedly run until its output response curve matches the historical operation data, and other unknown parameters are continuously selected for adjustment so that the model output of each variable can meet the requirements of being consistent with the historical data.

[0080] In another preferred embodiment of the present invention, in the above step 125, the mathematical model after parameter adjustment is combined with the constructed physical model, and the preprocessed historical operation data is input into the physical model and the mathematical model at the same time; the combined model is run to obtain the simulated response of the system under given historical conditions, and the model output results are compared with the actual historical operation data. If there is a difference between the two, the relevant parameters in the physical model are adjusted; the above steps are repeated until the model output is basically consistent with the historical data, and finally a digital twin model containing an accurate physical model and a mathematical model is obtained.

[0081] In a preferred embodiment of the present invention, the above step 13 may include:

[0082] Step 131, when the digital twin model simulates the operation of the heating system, monitors the key parameters of the heating system and compares the key parameters with the set normal operating range;

[0083] Step 132: If the monitored key parameter is greater than or equal to the normal range, it is determined that the heating system has a fault or an abnormality, and the heating system automatically obtains relevant environmental dynamic data.

[0084] In the embodiment of the present invention, the key parameters are monitored in real time through the digital twin model. Once they exceed the normal range, the abnormality can be quickly located. If the parameters exceed the normal range, it is judged as a fault. The environmental data is obtained in real time without waiting for manual confirmation, which shortens the response time. Compared with simply judging the fault based on the parameter threshold, the digital twin model monitoring and automatic acquisition of environmental data can quickly locate the fault, capture the impact of environmental changes on the fault in real time, improve the efficiency of subsequent fault diagnosis and processing, reduce maintenance costs, and improve system reliability.

[0085] In another preferred embodiment of the present invention, in the above step 131, in the digital twin model, the key parameters of the heating system, such as flow, pressure, temperature, etc., are determined; according to the preset conditions, the normal operating range of these key parameters is determined; during the operation of the model, the simulation value of the key parameter is obtained in real time, and the simulation value is compared with the normal range to determine whether it exceeds the range. If it exceeds the range, the parameter name, simulation value, exceeded value and other information are recorded, and the moment of exceeding the range is determined as the abnormal moment detected by the model. In the above step 132, the judgment module receives the key parameter monitoring results from the digital twin model, analyzes the monitoring results, and if any key parameter exceeds the normal range, it is determined that the system has a fault. The fault judgment result triggers the environmental data acquisition module, and the environmental data acquisition module calls the interface of the meteorological station or environmental sensor to obtain real-time environmental dynamic data related to the fault, such as temperature, humidity, wind speed, etc.

[0086] In a preferred embodiment of the present invention, the above step 14 may include:

[0087] Step 141, preprocessing the acquired environmental dynamic data to obtain preprocessed data;

[0088] Step 142, analyzing the correlation points between the parameters in the preprocessed data and the digital twin model, and determining the submodules related to the environmental dynamic data;

[0089] Step 143, adding an input interface matching the environmental dynamic data parameters in the framework of the digital twin model;

[0090] Step 144, design a dynamic data-driven mechanism to activate the update calculation of the submodule when the preprocessed data is input through the newly added interface;

[0091] Step 145, calling the newly added dynamic data input interface in the model framework and activating the dynamic driving mechanism to generate a dynamic digital twin model.

[0092] In the embodiment of the present invention, a special interface is added to enable dynamic data to directly drive the model, and a driving mechanism is designed to enable data to directly activate the update of the submodule, and the dynamic update of the model is realized by calling the interface. By designing the dynamic driving process of the model, environmental data can be effectively utilized, so that the digital twin model can dynamically respond to external changes, improve the effect of model prediction of faults, and enhance the adaptability of the system.

[0093] In another preferred embodiment of the present invention, in the above step 141, by The acquired environmental dynamic data is processed to obtain a processed signal X, where w(t) is the weight function of the original signal x(t), X(a, b) is the signal after wavelet transformation, a is the scale parameter, b is the translation parameter, ψ is the mother wavelet function, and t is the time variable, where w(t)=R(t, t0)=E[x(t0+t)x(t0)], R(t, t0) is the autocorrelation function of the original signal at time t0, R is the sign of the autocorrelation function, t0 is the reference time in the autocorrelation function, E is the expectation or mean operator, x is the sign of the original signal, x(t0) is the value of the original signal at time t0, and x(t0+t) is the value of the original signal at time t0+t.

[0094] In another preferred embodiment of the present invention, in the above step 142, a parameter association metric is constructed based on the preprocessed data. Among them, G ij represents the association measure between parameters i and j, T is the time length, is the mean of parameter i, w t is the time weight. According to the parameter association metric G ij , establish the weighted association matrix R = W r ×W c ×G, where W r is the row weight matrix, W c is the column weight matrix, the row weight matrix W r =diag(w r1 , w r2 ,…,w rn ), where w ri is the row weight coefficient of the i-th parameter, r represents the row, n represents the number of parameters, and c represents the column; the column weight matrix W c =diag(w c1 , w c2 , ..., w cn ), where w ci is the column weight coefficient of the i-th parameter, G is the original incidence matrix, where

[0095] In another preferred embodiment of the present invention, in the above step 143, the input interface type and format that needs to be added are determined according to the association between the above-determined environmental dynamic data parameters and the digital twin model; in the software framework of the digital twin model, input ports or interface functions that match these parameters are added; the name, format, accepted parameter type and other attributes of the interface are set to be consistent with the definition of the environmental dynamic data parameters; the interface function is responsible for receiving real-time input from environmental data and processing it into a standard format used internally in the model; the data input by the interface will update the relevant variables and parameters in the model in real time; the interface function is tested to ensure that the dynamic environmental data can be smoothly input into the digital twin model through the interface; the interface-related technical documents are recorded, including the name, format, parameters, etc., to facilitate subsequent use and maintenance; and finally a digital twin model framework that can interact with environmental dynamic data in real time is obtained.

[0096] In another preferred embodiment of the present invention, in the above step 144, the framework of the digital twin model is analyzed to determine the sub-modules that need to be connected to the dynamic data drive; a driving interface is designed in these sub-modules to receive data from the newly added input interface; when dynamic data is input through the newly added interface, the driving interface will activate the flag signal; after the sub-module program detects the flag signal, it calls the internal update function; the update function includes the processes of reading dynamic data, updating the internal state variables of the module, and re-running the module logic; through this mechanism, the input of dynamic data can directly drive the update calculation of the sub-module; repeat the above process and expand it to multiple sub-modules of the digital twin model; finally, when external dynamic data is input, the update calculation inside the model is automatically activated.

[0097] In another preferred embodiment of the present invention, in the above step 145, in the running module of the digital twin model, the newly added dynamic data input interface is called; the external environmental dynamic data is input into the model through the interface; the input data is automatically passed to the driving interface of the relevant sub-module; the driving interface of the sub-module detects the new data and activates the flag signal; the flag signal triggers the update function of the sub-module to run; the update function updates the sub-module status and completes a dynamically driven cyclic calculation; the model running module will periodically call the interface to obtain the latest environmental data; the sub-module will be driven by the environmental data to realize dynamic update of the model, and finally obtain a digital twin model that can actively respond to external dynamic data, which can better simulate the operation of the system in a complex dynamic environment.

[0098] In a preferred embodiment of the present invention, the above step 15 may include:

[0099] Step 151, within the framework of the dynamic digital twin model, establish a normal operation data module, a fault operation data module and an environmental dynamic data module;

[0100] Step 152, inputting the normal operation data, the fault operation data and the environmental dynamic data into the operation data module, the fault operation data module and the environmental dynamic data module respectively;

[0101] Step 153, respectively comparing the pattern features of the operation data module, the fault operation data module and the environment dynamic data module to obtain the difference between the fault operation data and the normal operation data;

[0102] Step 154 ​​, analyzing the cause of the difference according to the influence of the environmental dynamic data, and determining the type of fault causing the difference.

[0103] In the embodiment of the present invention, independent modules are established to store different types of data, so that the comparison between data is clearer. By comparing the data patterns of different modules, the abnormal patterns related to the fault can be effectively identified. By comprehensively considering the dynamic factors of the environment, the errors caused by the environment can be eliminated and the cause of the fault can be accurately located. Compared with simple data comparison, the introduction of independent data modules and consideration of environmental factors can improve the accuracy of fault diagnosis, reduce the interference caused by the external environment, enable maintenance personnel to locate faults more quickly and accurately, and improve the reliability and safety of the system.

[0104] In another preferred embodiment of the present invention, in the above steps 151 and 152, a normal operation data module, a fault operation data module and an environmental dynamic data module are set within the software framework of the dynamic digital twin model; the normal operation data module is used to store the historical operation data of the system under fault-free conditions; the fault operation data module is used to store the operation data of the system under known fault conditions; the environmental dynamic data module is used to store external environmental dynamic data related to the system operation; the three modules are connected through a standardized interface to facilitate subsequent data interaction and comparative analysis, and data reading and storage functions are written for each module to realize the collection and organization of various types of data, set module permissions, and only allow internal access to the model to prevent external illegal tampering with data, and test whether the module functions meet expectations and can correctly store and provide data of the corresponding type, so as to finally obtain an independent data module.

[0105] In another preferred embodiment of the present invention, in the above step 153, the parameter time series in the normal operation data module is analyzed to extract the normal mode feature X; the fault operation data module is identified and processed to extract the fault mode feature Y; and a matching model is established. Among them, T represents the length of the time series, t represents the time index, and S ij represents the matching score of data sequences i and j, x it represents the value of sequence i at time t, x jt represents the value of time series j at time t, w tis the time series weight; according to the matching model, Calculate the similarity between the normal mode feature X and the fault mode feature Y. W is the matching weight matrix, n represents the number of time series in the normal mode feature X, m represents the number of time series in the fault mode feature Y, and W ij Represents the weight corresponding to the match between time series i and time series j; therefore, if the SIM(X, Y) value is too low, it means that there is a significant difference between the normal mode feature X and the fault mode feature Y. According to the difference, the data difference between the fault operation and the normal operation is obtained.

[0106] In another preferred embodiment of the present invention, in the above step 154, based on the difference results obtained, the temperature, humidity and other parameter data in the environmental dynamic data module are checked, and the change time series of the environmental data and the state parameters are compared to determine whether the environmental change is related to the state change. If there is a certain correlation, it is determined that the environmental change is one of the causes of the difference. According to the range of parameter value changes caused by environmental factors, the most likely fault type is determined by referring to the fault mode library. If the environmental factors cannot explain all the differences, there are other causes of the fault, and the diagnosis process needs to be adjusted for further judgment. Finally, the state parameter differences and the influence of environmental factors are combined to determine the fault type that caused the difference. If it cannot be determined, it is marked as unknown cause and more diagnostic information needs to be obtained.

[0107] In a preferred embodiment of the present invention, the above step 16 may include:

[0108] Step 161, constructing a fault type knowledge base;

[0109] Step 162, after determining the fault type, query the knowledge base using the fault type as a keyword to obtain the development law information of the fault type;

[0110] Step 163, using the historical cases of type failures obtained from the knowledge base, training a prediction model for failure development to obtain a final prediction model;

[0111] Step 164, inputting the parameters of the current fault into the final prediction model to obtain the fault development path;

[0112] Step 165, analyzing the impact scope of the knowledge base and the fault development path to obtain an analysis result;

[0113] Step 166, make corrections based on the impact of the environmental dynamic data on the analysis results to obtain a final impact prediction.

[0114] In the embodiment of the present invention, the present invention can make full use of the knowledge base and environmental data to make the prediction of the development trend of the fault more intelligent and dynamic, and comprehensively judge the impact range to provide a basis for subsequent processing decisions and improve system reliability.

[0115] In another preferred embodiment of the present invention, the above-mentioned step 161 may include: obtaining historical fault case data, obtaining case data of different types of faults from the operation log and maintenance records of the heating system, preprocessing the case data, and extracting fault-related characteristic parameters to build a structured case library; determining the hierarchical relationship of fault types, and building a conceptual system corresponding to the fault types; labeling the case data, semantically labeling the case data with the fault type, and labeling the fault type corresponding to each case; storing the preprocessed and labeled case data into the knowledge base, and providing a query interface for subsequent steps to call.

[0116] In another preferred embodiment of the present invention, the above-mentioned step 162 may include: inputting the currently determined fault type, taking the fault type as a query keyword, searching for case data containing the keyword in the knowledge base constructed in step 161, extracting a case data subset related to the keyword from the search results, performing statistical analysis on the case data subset to obtain a probability distribution of fault development, analyzing the changing trends and laws of fault parameters at different stages, constructing a state transition diagram model of fault development, and extracting development law knowledge from the state transition diagram; formatting and structuring the extracted development law knowledge, organizing it into law information, and returning the law information as a query result.

[0117] In another preferred embodiment of the present invention, the above step 163 may include: establishing a prediction model for predicting the development trend of heating system faults; extracting fault cases from the historical operation log of the heating system, obtaining state parameters related to the faults as features, and annotations of the fault development as labels; performing data preprocessing on the fault cases and extracting features to construct samples; selecting a model structure for processing time series prediction tasks, and setting training parameters, using the preprocessed sample data to train the prediction model, and minimizing the loss function through iteration; evaluating the model performance on an independent test set, and if poor, adjusting the prediction model and retraining, and repeating the above steps until a prediction model with good performance is trained.

[0118] In another preferred embodiment of the present invention, the above step 164 may include: inputting the state data of the new fault into the prediction model to predict the fault development trend, converting the state sequence into a readable fault development path, mapping each state to the semantic description of the corresponding fault stage, constructing a text or chart description of the fault development path, and returning the result describing the current fault development path. The above step 165 may include: querying the historical cases related to the current fault type in the knowledge base to obtain its impact range data; analyzing the impact range statistics of the historical cases to determine the distribution probability of different types of impacts; querying the impact range knowledge graph model of the current fault type in the knowledge base; loading the development path result of the current fault; marking the state transition points that may have an impact on the development path; combining the transition points, the impact range distribution and the knowledge graph to analyze the impact range of the current fault; generating an impact range analysis report as a result. The above step 166 may include: obtaining real-time environmental dynamic data related to the current fault, processing the dynamic data, and extracting relevant influencing factors; analyzing how the dynamic factors affect the range analysis results, correcting the analysis results, narrowing or expanding the impact range, and adjusting the probability of the impact; comprehensively considering the correction of dynamic factors to obtain the final impact prediction.

[0119] In a preferred embodiment of the present invention, the above step 17 may include:

[0120] Step 171, constructing a fault risk assessment model according to the fault development speed, impact scope and environmental dynamic factors;

[0121] Step 172, presetting determination thresholds of different risk levels in the fault risk assessment model;

[0122] Step 173, calculating a comprehensive risk value according to the fault development trend and the fault risk assessment model;

[0123] Step 174, obtaining a preliminary risk level according to the comprehensive risk value and a preset threshold;

[0124] Step 175, adjusting the preliminary risk level according to the environmental dynamic factors to obtain an adjusted result;

[0125] Step 176, obtaining a final risk level based on the adjusted result.

[0126] In an embodiment of the present invention, constructing a risk assessment model can systematically consider multiple factors and conduct quantitative risk analysis; presetting thresholds can clarify risk standards at different levels, making the assessment results more meaningful for reference; calculating a comprehensive risk value can integrate multiple factors and conduct a comprehensive assessment; obtaining a preliminary risk level and making corrections can combine quantitative analysis and qualitative judgment, consider environmental dynamic factors to make the results more in line with actual conditions, and make assessments and corrections to make the final results more accurate and reliable.

[0127] In another preferred embodiment of the present invention, sensors are used to collect parameters such as real-time temperature, pressure, and flow, characterize the speed of fault development, analyze the system topology, determine the number of devices within the impact range, obtain environmental dynamic data such as wind speed and humidity from the meteorological station, and build a risk assessment model based on a neural network. The input layer receives data of various factors, labels the risk level of historical fault cases, and trains the risk assessment model parameters to fit the data; the high risk value interval is determined to be 0.7-1, the medium risk is 0.3-0.7, and the low risk is 0-0.3; the threshold nodes of the corresponding interval are set in the output layer of the risk assessment model for classifying the risk level; the state sequence of the fault development trend is input into the risk assessment model; the hidden layer integrates the information of various factors and outputs the comprehensive risk value result; the calculation result is a probability value within the range of 0 to 1; the relationship between the risk value and the threshold is compared, and the preliminary risk level result is determined according to the interval, and the preliminary result and process data are recorded for tracing. Determine whether the environmental dynamic factors are abnormal, analyze their impact on the risk, increase or decrease the preliminary level according to the impact relationship, and record the correction basis to explain the results. The final risk level is determined by the revised results, and an assessment report is generated, detailing the analysis process, results and data basis, providing a report for decision-making reference and reducing risks.

[0128] Among them, the environmental dynamic data such as wind speed and humidity are obtained from the weather station to build a risk assessment model based on neural network, including:

[0129] Query the real-time data interface of the weather station to obtain monitoring data of parameters such as wind speed, humidity, temperature, and air pressure; clean the meteorological data, normalize the cleaned data, and map it to the range of 0-1; construct a risk assessment neural network consisting of an input layer, 3 hidden layers, and an output layer. The number of nodes in the input layer is the number of meteorological parameters. The number of nodes in the first hidden layer is set to twice that of the input layer; use the ReLU activation function and add a Dropout layer to prevent overfitting. The output layer contains one node, gives the normalized risk assessment value, collects historical fault cases, and marks the risk level as a sample label; trains the network model, adjusts the parameters, and evaluates the model to ensure that it can effectively fit the case target, saves the optimized model and parameters, and prepares to receive meteorological data input in real time.

[0130] like Figure 2As shown, an embodiment of the present invention further provides a heating system fault analysis device 20 based on artificial intelligence, comprising:

[0131] The acquisition module 21 is used to acquire the operation data of the heating system; establish a digital twin model of the heating system according to the operation data; acquire environmental dynamic data when the digital twin model detects that the parameters of the heating system are abnormal; and fuse the environmental dynamic data with the digital twin model to establish a dynamic digital twin model;

[0132] The processing module 22 is used to analyze the fault type based on the dynamic digital twin model. The dynamic digital twin model identifies the fault type by comparing the patterns of normal operation data, fault operation data and the influence of environmental dynamic data; predicts the fault development trend and impact based on the fault type; and estimates the risk level based on the fault development trend and impact.

[0133] Optionally, a digital twin model of the heating system is established based on the operating data, including:

[0134] Obtaining the structural data of the heating system and the historical operation data of the heating system;

[0135] Construct a physical model of the heating system based on the structural data and historical operation data of the heating system;

[0136] According to the physical model, a mathematical model of the heating system is established;

[0137] By using historical operation data as input signals, unknown parameters in the mathematical model are adjusted so that the mathematical model outputs historical data that meets the requirements;

[0138] Input historical data that meets the requirements into the physical model and mathematical model to form a digital twin model of the heating system.

[0139] Optionally, when the digital twin model detects abnormal heating system parameters, it obtains environmental dynamic data, including:

[0140] When the digital twin model simulates the operation of the heating system, it monitors the key parameters of the heating system and compares the key parameters with the set normal operating range;

[0141] If the monitored key parameters are ≥ normal range, the heating system is judged to be faulty or abnormal, and the heating system automatically obtains relevant environmental dynamic data.

[0142] Optionally, the environmental dynamic data is integrated with the digital twin model to establish a dynamic digital twin model, including:

[0143] Preprocessing the acquired environmental dynamic data to obtain preprocessed data;

[0144] Analyze the correlation points between the parameters in the preprocessed data and the digital twin model, and determine the submodules related to the environmental dynamic data;

[0145] Add an input interface that matches the environmental dynamic data parameters in the framework of the digital twin model;

[0146] Design a dynamic data-driven mechanism to activate the update calculation of the submodule when the preprocessed data is input through the newly added interface;

[0147] Call the newly added dynamic data input interface in the model framework and activate the dynamic driving mechanism to generate a dynamic digital twin model.

[0148] Optionally, the fault type is analyzed based on a dynamic digital twin model. The dynamic digital twin model identifies the fault type by comparing the patterns of normal operation data, fault operation data, and the impact of environmental dynamic data, including:

[0149] Within the framework of the dynamic digital twin model, establish a normal operation data module, a fault operation data module, and an environmental dynamic data module;

[0150] Inputting normal operation data, fault operation data and environmental dynamic data into the operation data module, the fault operation data module and the environmental dynamic data module respectively;

[0151] Compare the pattern features of the operation data module, the fault operation data module and the environmental dynamic data module respectively to obtain the difference between the fault operation data and the normal operation data;

[0152] According to the impact of environmental dynamic data, analyze the causes of the differences and determine the type of fault that causes the differences.

[0153] Optionally, based on the fault type, predict the fault development trend and impact, including:

[0154] Build a knowledge base of fault types;

[0155] After determining the fault type, use the fault type as a keyword to query the knowledge base to obtain the development law information of the fault type;

[0156] Use the historical cases of type failures obtained from the knowledge base to train the prediction model of failure development to obtain the final prediction model;

[0157] Input the parameters of the current fault into the final prediction model to obtain the fault development path;

[0158] Analyze the impact scope of the knowledge base and the fault development path to obtain analysis results;

[0159] According to the influence of environmental dynamic data on the analysis results, corrections are made to obtain the final impact prediction.

[0160] Optionally, estimate the risk level based on the fault development trend and impact, including:

[0161] Construct a fault risk assessment model based on the fault development speed, impact scope and environmental dynamic factors;

[0162] Preset the determination thresholds of different risk levels in the fault risk assessment model;

[0163] Calculate the comprehensive risk value based on the fault development trend and fault risk assessment model;

[0164] According to the comprehensive risk value and the preset threshold, a preliminary risk level is obtained;

[0165] According to environmental dynamic factors, the preliminary risk level is adjusted to obtain the adjusted result;

[0166] Based on the adjusted results, the final risk level is obtained.

[0167] It should be noted that the device is a device corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0168] The embodiment of the present invention further provides a computing device, comprising: a processor, a memory storing a computer program, wherein when the computer program is executed by the processor, the method described above is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0169] The embodiment of the present invention also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method described above. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0170] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0171] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0172] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0173] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0174] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0175] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0176] In addition, it should be noted that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0177] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code that implements the method or device. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0178] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A heating system fault analysis method based on artificial intelligence, characterized in that: The method comprises: Obtaining the operating data of the heating system; Build a digital twin model of the heating system based on operating data; When the digital twin model detects abnormal heating system parameters, it obtains environmental dynamic data; Integrate environmental dynamic data with the digital twin model to establish a dynamic digital twin model; Analyze fault types based on the dynamic digital twin model. The dynamic digital twin model identifies fault types by comparing the patterns of normal operation data, fault operation data, and the impact of environmental dynamic data. According to the fault type, predict the fault development trend and impact; Estimate the risk level based on the fault development trend and impact; Based on the operating data, a digital twin model of the heating system is established, including: Obtaining the structural data of the heating system and the historical operation data of the heating system; Construct a physical model of the heating system based on the structural data and historical operation data of the heating system; According to the physical model, a mathematical model of the heating system is established; By using historical operation data as input signals, unknown parameters in the mathematical model are adjusted so that the mathematical model outputs historical data that meets the requirements; Input the historical data that meets the requirements into the physical model and mathematical model to form a digital twin model of the heating system; When the digital twin model detects abnormal heating system parameters, it obtains environmental dynamic data, including: When the digital twin model simulates the operation of the heating system, it monitors the key parameters of the heating system and compares the key parameters with the set normal operating range; If the monitored key parameters are ≥ normal range, the heating system is judged to be faulty or abnormal, and the heating system automatically obtains relevant environmental dynamic data; Integrate environmental dynamic data with the digital twin model to establish a dynamic digital twin model, including: Preprocessing the acquired environmental dynamic data to obtain preprocessed data, specifically including: Process the acquired environmental dynamic data to obtain a processed signal X, where w(t) is the weight function of the original signal x(t), X(a,b) is the signal after wavelet transformation, a is the scale parameter, b is the translation parameter, ψ is the mother wavelet function, and t is the time variable, where w(t)=R(t,t0)=E[x(r0+t)x(t0)], R(t,t0) is the autocorrelation function of the original signal at time t0, R is the sign of the autocorrelation function, t0 is the reference time in the autocorrelation function, E is the expectation or mean operator, x is the sign of the original signal, x(t0) is the value of the original signal at time t0, and x(t0+t) is the value of the original signal at time t0+t; Analyze the correlation points between the parameters in the preprocessed data and the digital twin model, and determine the submodules related to the environmental dynamic data, including: constructing parameter correlation metrics based on the preprocessed data Among them, G ij represents the association measure between parameters i and j, T is the time length, is the mean of parameter i, w t is the time weight; according to the parameter association metric G ij , establish the weighted association matrix R = W r ×W c ×G, where W r is the row weight matrix, W c is the column weight matrix, the row weight matrix W r =diag(w r1 , w r2 ,…,w rn ), where w ri is the row weight coefficient of the i-th parameter, r represents the row, n represents the number of parameters, and c represents the column; the column weight matrix W c =diag(w c1 , w c2 , ..., w cn ), where w ci is the column weight coefficient of the i-th parameter, G is the original incidence matrix, where Add an input interface that matches the environmental dynamic data parameters in the framework of the digital twin model; Design a dynamic data-driven mechanism to activate the update calculation of the submodule when the preprocessed data is input through the newly added interface; Call the newly added dynamic data input interface in the model framework to activate the dynamic driving mechanism to generate a dynamic digital twin model; Analyze fault types based on the dynamic digital twin model. The dynamic digital twin model identifies fault types by comparing the patterns of normal operation data, fault operation data, and the impact of environmental dynamic data, including: Within the framework of the dynamic digital twin model, establish a normal operation data module, a fault operation data module, and an environmental dynamic data module; Inputting normal operation data, fault operation data and environmental dynamic data into the operation data module, the fault operation data module and the environmental dynamic data module respectively; Compare the pattern features of the operation data module, the fault operation data module and the environmental dynamic data module respectively to obtain the difference between the fault operation data and the normal operation data; According to the impact of environmental dynamic data, analyze the causes of the difference and determine the type of fault that causes the difference; According to the fault type, predict the fault development trend and impact, including: Build a knowledge base of fault types; After determining the fault type, use the fault type as a keyword to query the knowledge base to obtain the development law information of the fault type; Use the historical cases of type failures obtained from the knowledge base to train the prediction model of failure development to obtain the final prediction model; Input the parameters of the current fault into the final prediction model to obtain the fault development path; Analyze the impact scope of the knowledge base and the fault development path to obtain analysis results; According to the impact of environmental dynamic data on the analysis results, corrections are made to obtain the final impact prediction; Estimate the risk level based on the fault development trend and impact, including: Construct a fault risk assessment model based on the fault development speed, impact scope and environmental dynamic factors; Preset the determination thresholds of different risk levels in the fault risk assessment model; Calculate the comprehensive risk value based on the fault development trend and fault risk assessment model; According to the comprehensive risk value and the preset threshold, a preliminary risk level is obtained; According to environmental dynamic factors, the preliminary risk level is adjusted to obtain the adjusted result; Based on the adjusted results, the final risk level is obtained.

2. A heating system fault analysis device based on artificial intelligence, implementing the method according to claim 1, characterized in that: include: An acquisition module, used for acquiring operation data of a heating system; Build a digital twin model of the heating system based on operating data; When the digital twin model detects abnormal heating system parameters, it obtains environmental dynamic data; Integrate environmental dynamic data with the digital twin model to establish a dynamic digital twin model; A processing module for analyzing the fault type based on a dynamic digital twin model. The dynamic digital twin model identifies the fault type by comparing the patterns of normal operation data, fault operation data, and the impact of environmental dynamic data; According to the fault type, predict the fault development trend and impact; Estimate the risk level based on fault development trends and impacts.

3. A computing device, characterized in that: include: one or more processors; A storage device, used for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method as claimed in claim 1.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which implements the method according to claim 1 when executed by a processor.

Images