CN-121997083-A - Electric fireplace fault self-diagnosis method and system

CN121997083ACN 121997083 ACN121997083 ACN 121997083ACN-121997083-A

Abstract

The invention relates to the technical field of intelligent household equipment and operation and maintenance of the Internet of things, and discloses a self-diagnosis method and a self-diagnosis system for faults of an electric fireplace, wherein the method comprises the steps of obtaining multidimensional sensing data and a history record; the method comprises the steps of carrying out feature extraction and clustering calculation according to multi-dimensional sensing data to obtain a difference feature set, inputting the difference feature set into a pre-built causal model to carry out deduction, extracting to obtain a potential abnormal link, merging to obtain a state matrix according to the potential abnormal link and the history record, constructing a difference directed graph based on the state matrix, locating an abnormal source, inputting a state vector corresponding to the abnormal source into a pre-built Bayesian network to carry out deduction, and determining a fault starting point. The method can realize the accurate positioning of the fault root cause of the electric fireplace.

Inventors

YANG PENG
ZHU XIAOLIN
CHEN MIANFENG

Assignees

勃格科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. A method for self-diagnosing faults of an electric fireplace, comprising: Acquiring multidimensional sensing data, and performing feature extraction on the multidimensional sensing data to obtain a response sequence set; performing cluster analysis according to the response sequence set to obtain a difference characteristic set; Inputting the difference feature set into a pre-constructed causal model for deduction to obtain an offset sequence, and analyzing the degree of deviation according to the offset sequence to obtain a potential abnormal link; Acquiring a history record, carrying out delay comparison on the history record and the potential abnormal link to obtain an offset triggering sequence, fusing the response sequence set and the history record according to the offset triggering sequence to obtain a state matrix, and constructing a differential directed graph based on the state matrix; Extracting directed edges with weights exceeding a preset weight threshold value from the difference directed graph, splicing to obtain a reaction path, calculating the degree of entry of nodes in the reaction path to obtain an degree value, and positioning the nodes with the degree value of zero as abnormal sources; And extracting a state vector corresponding to the abnormal source, inputting the state vector into a pre-constructed Bayesian network to be deduced to obtain a probability path, comparing the probability path with the reaction path to obtain a consistency score, and determining the abnormal source with the consistency score higher than a preset consistency threshold as a fault starting point.
2. The method for self-diagnosing a fault in an electric fireplace of claim 1, wherein the feature extraction of the multidimensional sensing data to obtain a response sequence set comprises: The multi-dimensional sensing data are aligned by using a time stamp to obtain fusion data, wherein the multi-dimensional sensing data comprise temperature data, current data and voltage data; extracting features of the fusion data to obtain an initial feature set, and complementing the initial feature set by an interpolation method to obtain a complete feature set; Performing time sequence analysis on the complete feature set to obtain a state mapping matrix; And calculating state switching delay time according to the state mapping matrix to obtain the response sequence set.
3. The method for self-diagnosing a fault in an electric fireplace according to claim 1, wherein the performing cluster analysis according to the response sequence set to obtain a difference feature set includes: carrying out dimensionless treatment on each sequence in the response sequence set to obtain a standardized sequence set, and calculating correlation coefficients between every two sequences in the standardized sequence set by using a cross-correlation function to obtain an intensity value; extracting sequence pairs with the intensity values exceeding a preset association threshold value, and polymerizing to obtain an interaction set; Calculating time lag difference values of each sequence pair in the interaction set as time lag correlation values, and constructing an expansion interaction matrix according to the time lag correlation values; Grouping sequences with similar time lag correlation values in the extended interaction matrix by using a clustering algorithm to obtain a difference response cluster; and extracting the deviation vector of the difference response cluster, and combining the deviation vector to obtain a difference feature set.
4. The method for self-diagnosing a fault of an electric fireplace according to claim 1, wherein the step of analyzing the deviation according to the deviation sequence to obtain a potential abnormal link includes: analyzing the offset sequence to obtain delay time, comparing the delay time with a preset reference delay interval to obtain an exceeding part, and calculating the ratio of the exceeding part to the upper limit value of the reference delay interval to obtain the deviation degree; Comparing the deviation degree with a preset deviation threshold, extracting parameters with the numerical value larger than the preset deviation threshold to construct a potential fault set, and positioning parameters in the potential fault set as potential abnormal links.
5. The method for self-diagnosing a fault in an electric fireplace according to claim 1, wherein the step of obtaining an offset triggering sequence by comparing the history with the potential abnormal links includes: analyzing the history record, extracting delay characteristic fragments of the potential abnormal links, and matching the delay characteristic fragments with a preset history abnormal section to obtain a matched section; Comparing the matching section with the response sequence set by using a dynamic time warping algorithm, and calculating to obtain an optimal alignment path; And analyzing the time mapping relation in the optimal alignment path, extracting the time offset of the response sequence set relative to the matching section, and sequencing according to the time offset to obtain an offset triggering sequence.
6. The method for self-diagnosing a fault in an electric fireplace according to claim 1, wherein the fusing the response sequence set and the history record according to the offset triggering sequence to obtain a state matrix, and constructing a differential directional map based on the state matrix comprises: Extracting the response sequence set and the history record in a time window corresponding to the offset triggering sequence, and arranging and corresponding the response sequence set and the history record according to a parameter type to obtain a state matrix; Calculating the deviation value of the real-time data and the historical data in the state matrix, taking the parameter nodes as the graph vertexes, taking the time sequence association direction among the parameters as the directed edges, and constructing the deviation value as the directed edge weight to obtain the differential directed graph.
7. The method for self-diagnosing faults of an electric fireplace according to claim 1, wherein the steps of extracting the directed edges with weights exceeding a preset weight threshold in the differential directed graph, splicing to obtain a reaction path, calculating the degree of entry of the nodes in the reaction path to obtain a degree value, and positioning the nodes with the degree value of zero as abnormal sources comprise the steps of: Traversing all directed edges in the difference directed graph, screening out edges with weights lower than a preset weight threshold, and extracting the remaining directed edges as high-risk propagation links; splicing the high-risk propagation links end to end according to the time sequence to obtain a reaction path; counting the number of leading associated edges of each node in the reaction path to obtain an ingress value, and extracting a starting node with zero ingress value to mark the starting node as an abnormal source.
8. The method according to claim 1, wherein the step of inputting the state vector into a pre-constructed bayesian network to be deduced to obtain a probability path, comparing the probability path with the reaction path to obtain a consistency score, and determining an abnormal source with the consistency score higher than a preset consistency threshold as a fault starting point comprises the steps of: Inputting the state vector as an evidence node into a pre-constructed Bayesian network diagram, and carrying out probability deduction to update the node state to obtain posterior probability distribution of each associated node; calculating the conditional dependence probability of each node in the posterior probability distribution, and extracting the branch with the highest cumulative probability to obtain a probability path; And calculating the node overlapping rate of the probability path and the reaction path to obtain a consistency score, and extracting an abnormal source of which the consistency score exceeds a preset consistency threshold value to determine as a fault starting point.
9. The method of claim 1, further comprising, after said determining an anomaly source having a consistency score above a preset consistency threshold as a failure origin: Extracting a downstream node linkage relation positioned at the fault starting point in the reaction path; Smoothing the state parameters of the fault starting point to generate a reference fluctuation sequence, and carrying out mapping deduction on the reference fluctuation sequence to the downstream nodes along the node linkage relation by utilizing a preset adjustment vector to obtain the predicted track of each node; Extracting a propagation branch with a parameter deviation degree larger than a preset linkage threshold value from the predicted track, merging the propagation branch with the reaction path, performing topology reconstruction to generate a reconstruction network, and outputting the reconstruction network as a tracing chain.
10. An electric fireplace fault self-diagnosis system, comprising: the data processing module is used for acquiring multidimensional sensing data, and extracting characteristics of the multidimensional sensing data to obtain a response sequence set; The characteristic difference module is used for carrying out cluster analysis according to the response sequence set to obtain a difference characteristic set; The deviation detection module is used for inputting the difference feature set into a pre-constructed causal model to be deduced to obtain a deviation sequence, and carrying out deviation analysis according to the deviation sequence to obtain a potential abnormal link; the map construction module is used for acquiring a history record, carrying out delay comparison on the history record and the potential abnormal link to obtain an offset triggering sequence, fusing the response sequence set and the history record according to the offset triggering sequence to obtain a state matrix, and constructing a differential directed map based on the state matrix; The source positioning module is used for extracting directed edges with weights exceeding a preset weight threshold in the differential directed graph, splicing the directed edges to obtain a reaction path, calculating the degree of incidence of nodes in the reaction path to obtain an degree value, and positioning the nodes with the degree value of zero as abnormal sources; The tracing output module is used for extracting a state vector corresponding to the abnormal source, inputting the state vector into a pre-constructed Bayesian network to be deduced to obtain a probability path, comparing the probability path with the reaction path to obtain a consistency score, and determining the abnormal source with the consistency score higher than a preset consistency threshold as a fault starting point.

Description

Electric fireplace fault self-diagnosis method and system Technical Field The invention relates to the technical field of intelligent household equipment and operation and maintenance of the Internet of things, in particular to a fault self-diagnosis method and system of an electric fireplace. Background In the field of modern intelligent household equipment, an electric fireplace is used as comprehensive equipment integrating heating and decorative light effects, and the internal operation of the electric fireplace relates to complex interaction of multiple physical fields such as electric energy conversion, heat transfer, light effect display and the like. Along with the increasingly high integration and intellectualization of the functions of the equipment, the reliability monitoring of the running state of the equipment gradually evolves towards the direction of fault prediction and health management so as to realize accurate abnormal monitoring and maintenance before serious damage or potential safety hazard of the equipment occurs. Currently, for fault diagnosis of electric fireplaces and the like, in one prior art, a hard threshold judgment mechanism (such as over-temperature alarm or current overload outage) of a single sensor is generally relied on, or simple comparison is performed by adopting section data based on static rules. However, in the actual dynamic operation process of an electric fireplace, high dynamic coupling and significant time response differences exist among multiple parameters such as electric energy input, heat output, light effect presentation and the like. For example, a minor abnormality of the control circuit may first cause a hidden fluctuation of the electric energy input, and after a certain thermal inertia delay, the heat output is shifted, thereby causing a linkage instability of the light efficiency module. In the prior art, dynamic interaction rules and deep causal topological relations of multidimensional parameters in time sequence are fractured only by means of isolated static threshold values or single-dimensional surface data. When equipment fails to cause multi-parameter linkage mutation, the existing diagnosis system can only capture the most obvious appearance characteristics at the tail end of a propagation chain, and cannot clear the time sequence and space propagation venation of parameter abnormality. Therefore, the technical problems that it is difficult to precisely trace the abnormal chain reaction path and effectively locate the initial source of the fault exist in the prior art. Disclosure of Invention The invention provides a fault self-diagnosis method and system of an electric fireplace, which are used for solving the technical problems that in the prior art, an abnormal chain reaction path is difficult to trace accurately and an initial source of a fault is effectively positioned. In order to solve the above technical problems, the present invention provides a fault self-diagnosis method for an electric fireplace, comprising: Acquiring multidimensional sensing data, and performing feature extraction on the multidimensional sensing data to obtain a response sequence set; performing cluster analysis according to the response sequence set to obtain a difference characteristic set; Inputting the difference feature set into a pre-constructed causal model for deduction to obtain an offset sequence, and analyzing the degree of deviation according to the offset sequence to obtain a potential abnormal link; Acquiring a history record, carrying out delay comparison on the history record and the potential abnormal link to obtain an offset triggering sequence, fusing the response sequence set and the history record according to the offset triggering sequence to obtain a state matrix, and constructing a differential directed graph based on the state matrix; Extracting directed edges with weights exceeding a preset weight threshold value from the difference directed graph, splicing to obtain a reaction path, calculating the degree of entry of nodes in the reaction path to obtain an degree value, and positioning the nodes with the degree value of zero as abnormal sources; And extracting a state vector corresponding to the abnormal source, inputting the state vector into a pre-constructed Bayesian network to be deduced to obtain a probability path, comparing the probability path with the reaction path to obtain a consistency score, and determining the abnormal source with the consistency score higher than a preset consistency threshold as a fault starting point. In a second aspect, the present invention provides a fault self-diagnostic system for an electric fireplace comprising: the data processing module is used for acquiring multidimensional sensing data, and extracting characteristics of the multidimensional sensing data to obtain a response sequence set; The characteristic difference module is used for carrying out cluster analysis according to the respons