Solving method and device based on multiple observation aggregation fault diagnosis and medium

Abstract

The invention belongs to the technical field of circuit fault diagnosis, and particularly relates to a solving method and device based on multiple observation aggregation fault diagnosis and a medium. In order to solve the problems of one-sided single observation information, redundant diagnosis results and the like, multiple times of fault observation are performed on a circuit system, and multiple independent diagnosis results are generated and integrated into a diagnosis set; establishing a mapping relation between a diagnosis set and a unique keyword by using a map function of the STL, and converting the component combination into an integer identifier to simplify operation; calling a BWSS algorithm to perform collision calculation on the mapped keywords to obtain a candidate solution set covering all observation scenes; after the keyword solution set is converted to an original diagnosis form, deleting a superset according to a redundancy rule to form a simplified intermediate diagnosis solution set; and finally, through a branch and bound minimization algorithm, according to potential classification and ascending order arrangement of the diagnosis sets, high-order supersets are deleted layer by layer by adopting a heuristic strategy, and finally, a minimum diagnosis solution set of all faults is output.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of circuit fault diagnosis, and particularly relates to a solving method, device and medium based on multiple observation aggregation fault diagnosis. BACKGROUND

[0002] In the field of circuit fault diagnosis, accurately and efficiently locating fault components is the core requirement to ensure stable operation of the system. However, existing fault diagnosis schemes have many technical problems, which are difficult to meet the diagnosis needs of complex circuit systems. The specific problems are as follows: Existing technologies rely on single diagnosis algorithm observation to obtain fault information, and can only generate a single fault diagnosis set. Since circuit system faults have multiple causes and correlation characteristics, single observation is easily affected by noise interference and observation angle limitations, resulting in biased diagnosis results. For example, only through single input and output signal analysis, some hidden fault components may be missed, or normal components may be mistakenly judged as faults, ultimately causing inaccurate fault location, increasing subsequent debugging costs and misjudgment risks.

[0003] Some schemes attempt to improve the comprehensiveness of diagnosis through multiple observations, but lack a standardized multi-observation data processing mechanism. On the one hand, the diagnosis results generated by multiple observations are not uniform in format, such as inconsistent component fault state descriptions and fragmented diagnosis set structures, making it difficult to directly aggregate and analyze. On the other hand, repeated or similar diagnosis results are not effectively removed, resulting in a large amount of redundant data in subsequent calculations, which not only occupies additional storage resources, but also significantly increases algorithm running time, reducing diagnosis efficiency.

[0004] The core goal of fault diagnosis is to obtain a minimal fault set (i.e., the least component combination that can explain all fault phenomena), but existing minimization methods have significant defects. Most schemes use a full two-by-two comparison strategy, which requires subset relationship verification for all candidate diagnosis solutions, resulting in exponential growth of time complexity (e.g., when the number of candidate solutions is 12, dozens of comparisons are needed). Moreover, they do not combine fault probability characteristics (e.g., when component faults are independent of each other, the fault probability of a set with fewer elements is higher) for targeted screening, resulting in a large number of diagnosis solutions with chaotic priority, making it difficult for operators to quickly locate high-probability fault components, and making it difficult to implement engineering.

[0005] Existing schemes often use single technologies such as meet-set algorithm and minimization algorithm in isolation, without forming a complete technical chain of data input - preprocessing - calculation - screening: for example, some schemes only focus on the integrity of meet-set calculation, but ignore the impact of redundant solutions on subsequent minimization; or only optimize the efficiency of minimization, but do not solve the problem of standardized input of multi-observation data. This fragmented technical design leads to poor connection of the overall diagnosis process, making it difficult to balance diagnosis accuracy and efficiency, and unable to meet the diagnosis needs of complex circuit systems.

[0006] The prior art with application number CN202410256632.8 discloses a simulation circuit fault diagnosis method based on matrix model parameter identification, including the following steps: Pre-measurement stage: Step one, number the diagnosed device and set its fault state; Step two, sample the output signal of each fault state of each diagnosed device in time sequence, form an ordered matrix, and use Laplace operator to reduce dimension and sharpen, obtain the trace and spectral radius of the matrix as fault features; Step three, least square fitting is performed on the numerical value of the obtained discrete fault features to obtain a continuous fitting equation; Diagnosis stage: Step four, input the same excitation to the fault circuit as in the pre-measurement stage to obtain the output response sequence of the circuit under each fault state, form an output response matrix, and obtain the measurement value of the trace and the measurement value of the spectral radius after dimension reduction by the Laplace operator; Step five, first substitute the measurement value of the trace into the fitting trace equation of each fault device to obtain a series of solutions, and then substitute these solutions as independent variables into the fitting equation of the spectral radius of the corresponding device again to obtain a series of spectral radius calculation values; Step six, find the one closest to the measurement value of the spectral radius in the series of spectral radius calculation values, if the solution of the corresponding measurement value of the trace is within the tolerance range of the device, it is determined that the circuit is fault-free; otherwise, the device corresponding to this spectral radius calculation value is the fault element, and fault location is realized; Step seven, after fault location, substitute the spectral radius measurement value back into the fitting equation of the spectral radius of the located fault device to obtain the fault device parameter value calculated by the spectral radius, and take the average of the fault device value calculated by the trace before as the final identification parameter of the fault device.

[0007] Therefore, in the fault diagnosis determination solution set method, there is only a single observation information minimization solution, the final diagnosis accuracy is insufficient, the multi-observation data integration is difficult, and the large amount of data after integration leads to high calculation redundancy, the large amount of data obtained by multi-observation is inefficient in minimization screening, and the engineering practicability is poor, and the algorithm cooperativity is insufficient. SUMMARY

[0008] To solve the above problems, the application provides a solving method, system, equipment and medium based on multiple observation aggregation fault diagnosis, the application generates multiple independent diagnosis results and integrates them into a diagnosis set by observing the circuit system multiple times; the map function of STL is used to establish the mapping relationship between the diagnosis set and the unique keyword, and the component combination is converted into an integer identifier to simplify the operation; the BWSS algorithm is called to calculate the collision of the mapped keyword, and a candidate solution set covering all observation scenarios is obtained; after the keyword solution set is converted back to the original diagnosis form, the superset is deleted according to the redundancy rule to form a simplified intermediate diagnosis solution set; finally, the branch and bound minimization algorithm is used to sort and arrange the diagnosis set in ascending order according to the potential, and the heuristic strategy is used to delete the high-order superset layer by layer, and finally the minimal diagnosis solution set that can explain all faults is output.

[0009] To achieve the above purpose, the application adopts the following technical solutions: The first aspect of the application provides a solving method based on multiple observation aggregation fault diagnosis, including the following steps: Step S1, multiple fault diagnoses are obtained by observing the fault multiple times using a diagnosis algorithm, and multiple observation diagnosis results are combined to form a first diagnosis set; the specific steps are as follows: Step S11, the diagnosis algorithm is called to observe multiple times to obtain multiple observation fault diagnosis sets Diag1, Diag2, …, Diag i Each diagnosis set Diag i contains multiple diagnoses, and each diagnosis contains at least one component; Step S12, the multiple diagnosis sets are combined to form a new first diagnosis set Diag={Diag1, Diag2, …, Diag i}.

[0010] Step S2, the first diagnosis set is mapped to a second diagnosis set using a keyword, the second diagnosis set is processed using the BWSS algorithm to obtain a third diagnosis set after collision, and the redundant solutions of the third diagnosis set after collision are identified and deleted to obtain a fourth diagnosis set after removing the redundancy; the specific steps are as follows: Step S21, the map function of the standard template library STL is used to map the components of the first diagnosis set Diag, and all diagnoses are mapped into a unique keyword to form a second diagnosis set {Diag1 ’ ’ , Diag2 ’ , …, Diag i ’}; Step S22, the BWSS algorithm is used to process the second diagnostic set to diagnose the collision of the diagnostic set, and a third diagnostic set MHS (Diag1 ’ , Diag2 ’ , …, Diag i ’ )={Diag1'', Diag2'', …, Diag i ''} is obtained after the collision. Step S23, the redundant rules are removed from the collided diagnostic set MHS, the redundant diagnoses in each diagnostic set Diag i are deleted, and the diagnosis is added to the fourth diagnostic set MHS', and the diagnosis containing the key of the third diagnostic set MHS is deleted, and the remaining key diagnosis in the third diagnostic set MHS is reflected into a component diagnosis and added to the fourth diagnostic set MHS'.

[0011] Step S3, the fourth diagnostic set after removing the redundancy is processed by using a branch and bound minimization strategy to obtain a minimum diagnostic solution set of multiple observation diagnosis results. The specific steps are as follows: Step S31, the fourth diagnostic set MHS' is processed by using a branch and bound minimization strategy to minimize the diagnosis aggregation, and the fourth diagnostic set MHS' is first decomposed according to the potential to obtain the diagnosis corresponding to each potential, and the potentials are sorted in ascending order according to the size of the potential. Step S32, the sorted diagnostic set is processed by using a heuristic strategy to identify and delete the superset, and finally a minimum diagnostic solution set is obtained.

[0012] The second aspect of the present application provides an electronic device, comprising a memory 102, a processor 101, a display module 103, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to realize the steps of any one of the foregoing solving methods of the multiple observation aggregation fault diagnosis.

[0013] The third aspect of the present application provides a readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to realize the steps of any one of the foregoing solving methods of the multiple observation aggregation fault diagnosis.

[0014] The present application has the advantages that: multiple observation fault formation diagnosis set is formed, the diagnosis accuracy is improved and the misjudgment risk is reduced by aggregating the diagnosis set; the one-sided diagnosis result of single observation is integrated into a standardized diagnosis set (Diag={Diag1, Diag2, …, Diag i}, the use of multiple observation cross-validation covers more fault scenarios, which can effectively exclude noise interference and implicit fault omission in single observation, ensure that the final diagnosis solution can explain all fault phenomena, and significantly reduce the debugging cost caused by misjudgment; the application innovatively introduces the map function of C++ STL to realize the mapping of component set to unique key, such as {x2, x4} mapped to 2, which converts complex set comparison into efficient integer comparison, uses the red-black tree structure of map to realize O (log n) search and insertion efficiency, and greatly reduces the redundancy of multiple observation data; at the same time, through the preprocessing stage deletion superset strategy, the invalid candidate solution is reduced in advance before minimization calculation, the subsequent minimization calculation amount is reduced, and the running time of the whole diagnosis process is shortened; the branch and bound minimization strategy has a breakthrough efficiency advantage: by grouping in ascending order of diagnosis potential, it is only necessary to judge whether the solution with large potential is a superset of the solution with small potential, without bidirectional comparison, and the time complexity is reduced from exponential to nearly linear; combined with heuristic screening rules, the comparison logic is further simplified, for example, when processing 12 candidate solutions, only 7 core comparisons are needed to get the minimum diagnosis solution set. The application forms a complete technical chain of multiple observation of fault - preprocessing (mapping + BWSS collision + redundancy deletion) - branch and bound minimization, and each link is cooperatively adapted: the standardized diagnosis set provides a unified data basis for subsequent processing, the map mapping and superset deletion reduce the burden of collision calculation, and the branch and bound strategy improves the efficiency of minimization. This whole-chain optimization design ensures smooth connection of the diagnosis process and guarantees the accuracy of diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the present application, the drawings needed to be used in the description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0016] Figure 1 The flowchart of the method of the present application is shown in the figure. Figure 2 The component mapping flowchart of the method of the present application is shown in the figure. Figure 3 The branch and bound minimization flowchart of the method of the present application is shown in the figure. Figure 4 The schematic diagram of the device structure of the present application is shown in the figure.

[0017] Among them, 101 is a processor, 102 is a memory, and 103 is a display module. DETAILED DESCRIPTION

[0018] With reference to the accompanying drawings: the technical solutions in the embodiments of the present application will be described clearly and completely, obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0020] Embodiment 1, a solving method based on multiple observation aggregation fault diagnosis as shown in the following steps: Figure 1 Step S1, through the observation of multiple diagnosis algorithms for faults, a plurality of fault diagnoses are obtained, and the multiple observation diagnosis results form a first diagnosis set; the specific steps are as follows: Step S11, call the diagnosis algorithm for multiple observations to obtain a plurality of observation fault diagnosis sets Diag1, Diag2, …, Diag i . i Each diagnosis Diag i contains a plurality of components, assuming that the system to be diagnosed has m components, then the diagnosis d m = {x1, x2, …, x i}. Step S12, combine the multiple diagnosis sets to form a first diagnosis set Diag = {Diag1, Diag2, …, Diag ’}.

[0021] Step S2, perform key mapping on the first diagnosis set to form a second diagnosis set, perform BWSS algorithm on the second diagnosis set to obtain a third diagnosis set after collision, and perform redundant solution identification and deletion on the third diagnosis set after collision to obtain a fourth diagnosis set after removing redundancy; The specific steps are as follows: Step S21, use the map function of the table needle template library STL to map all diagnosis results of the first diagnosis set, map all diagnoses to a unique key, form new diagnoses Diag1 ’ i , Diag2 ’ , …, Diag i ; combine all new diagnoses to obtain a second diagnosis set; the component mapping rule is: ’ Step S211, the diagnostic set {Diag1, Diag2, Diag i} corresponding to multiple observations is converted into DiagMap<map, int>, DiagMap is traversed to obtain all diagnostic groups Diag1, Diag2, Diag i， , at this time the keyword nextKey=1; Step S212, each diagnostic group is traversed to obtain the set S1, S2, … S n in each diagnostic group. Step S213, according to the set S1, S2, … S n , the keyword mapping of each set is performed. Step S2131, if S n is in DiagMap, the set is skipped and the keyword mapping is not performed, nextKey is unchanged. Step S2132, if S n is not in DiagMap, the keyword of S n is nextKey, and then nextKey=nextKey+1 is executed.

[0022] Step S21321, if there is a set S n in the current diagnostic group, step S213 is executed. Step S21322, if there is no set S n in the current diagnostic group, it is determined whether there are more diagnostic groups to be mapped. Step S213221, if there are more diagnostic groups to be mapped, step S212 is executed. Step S213222, if all diagnostic groups are mapped, the keyword mapping is ended.

[0023] Step S22, the second diagnostic set is processed using the BWSS algorithm to obtain the collision of the diagnostic set, and the third diagnostic set MHS (Diag1 ’ ’ , Diag2 ’ , …, Diag i ’ )={Diag1’’, Diag2’’,…, Diag i ’’} after collision. Step S23, the third diagnostic set MHS after collision is processed by removing redundant rules to delete each diagnostic set Diag idiagnoses in the third diagnosis set MHS are removed, and the remaining key diagnoses in the third diagnosis set MHS are reflected into component diagnoses and added to the fourth diagnosis set MHS'; the redundancy removal rule: if a diagnosis, its subset appears in all diagnoses, then the diagnosis is removed from the diagnosis set Diag i and added to the new diagnosis set; Step S3, the fourth diagnosis set after removing the redundancy is processed by using the branch and bound minimization strategy to obtain the minimal diagnosis solution set of the multiple observation diagnosis results; the specific steps are as follows: Step S31, the diagnosis aggregation of the fourth diagnosis set MHS' is minimized by using the branch and bound minimization strategy, and the potential of each diagnosis set in the fourth diagnosis set MHS' is the number of components, first, the diagnosis set MHS' is decomposed according to the potential to obtain the diagnosis corresponding to each potential, and the potentials are sorted in ascending order according to the size of the potential; Step S32, the sorted diagnosis set is processed by using the heuristic strategy to identify and delete the superset, and finally the minimal diagnosis solution set is obtained; the heuristic strategy is: The superset is defined as: if a diagnosis, its subset appears in all diagnoses, then the diagnosis is an aggregated diagnosis, but the set formed by adding any component to it is a superset.

[0024] From the smallest potential group, the minimal solution is first stored in the new diagnosis solution set Diag''', and then the set with a potential greater than the minimal solution is judged: if there is a superset that can cover the minimal solution, the superset is deleted; if not, the original set is added to Diag'''. When a new minimal solution is added, if at least one component in the minimal solution of the diagnosis set Diag''' is not in the added solution, the diagnosis set to be added can be added to Diag'''.

[0025] Embodiment 2, a solving method based on multiple observation aggregated fault diagnosis, system observation OBS includes the input in and the output out of the system, when the system to be diagnosed receives the i-th signal in i , the system will give the corresponding actual output value out i , at the same time, according to the logical relationship of the system, the theoretical output value out i ' of the system working normally is obtained, when the theoretical value and the actual value appear obvious deviation, it is determined that at least one component in the system fails, if there is no obvious deviation, this observation is invalid observation for the system with fault. For each observation OBS i , the diagnosis algorithm is called to obtain all possible fault diagnosis sets Diag i ={d1, d2, …, di}, for each diagnosis d contains several components, assuming that the system to be diagnosed has m components, then the diagnosis d = { x1, x2,..., xm}, x = 0 or 1, 0 represents that the component is not in the diagnosis d, 1 represents that it exists. For convenience, only the value of x = 1 appears in the expression of d. m}, x = 0 or 1, 0 represents that the component is not in the diagnosis d, 1 represents that it exists. For convenience, only the value of x = 1 appears in the expression of d.

[0026] Suppose that the present application obtains 3 abnormal observation results Diag = { Diag1, Diag2, Diag3}, Diag1 = { { x1}, { x2, x4}, { x3, x7}}, Diag2 = { { x1}, { x2, x5}} and Diag3 = { { x1, x7}, { x2, x7}, { x3, x6}, { x8, x9}}.

[0027] It can be seen that a single observation cannot accurately find all the fault components, and if debugging is performed one by one, when the system size is large, not only a lot of time will be wasted, but also some misjudgments may be caused. The present application proposes an aggregation diagnosis algorithm based on multiple observations.

[0028] Firstly, the present application calls the map function of the STL module in C++ to map all the diagnosis results Diag, and gives a unique key to all the diagnoses. The mapping rule flow chart is as shown in Figure 2 , and the specific rule is: convert the diagnosis set Diag corresponding to multiple observations into DiagMap < map, int>, traverse DiagMap to obtain all diagnosis groups Diag1, Diag2, Diag3; at this time, the key nextKey = 1; traverse each diagnosis group to obtain the set S1, S2,... S n in each diagnosis group; according to the set S1, S2,... S n , map the key of each set; if S n is in DiagMap, skip this set and do not map the key, nextKey remains unchanged; if S n is not in DiagMap, the key of S n is nextKey, and then nextKey = nextKey + 1. This method efficiently utilizes the red-black tree implementation of map to ensure O (log n) time complexity of search and insertion, and is suitable for the scene of dynamically allocating unique identification. The final key mapping result is: { x1} = 1, { x2, x4} = 2, { x3, x7} = 3, { x2, x5} = 4, { x1, x7} = 5, { x2, x7} = 6, { x3, x6} = 7, { x8, x9} = 8.

[0029] Diag1' = {1, 2, 3}, Diag2' = {1, 4} and Diag3' = {5, 6, 7, 8}.

[0030] The BWSS algorithm is called to find a set of candidate key sets, which satisfy that the set has at least one common element with each observed diagnosis set (i.e. collision, ensuring that all observed scenarios are covered), and there is no redundant element in the set. A total of 12 results of the diagnosis are obtained after collision and mapping: MHS({Diag1', Diag2', Diag3'}) = {{1, 5}, {1, 6}, {1, 7}, {1, 8}, {2, 4, 5}, {2, 4, 6}, {2, 4, 7}, {2, 4, 8}, {3, 4, 5}, {3, 4, 6}, {3, 4, 7}, {3, 4, 8}}.

[0031] According to the mapping table, the minimum collision set result is written in the form of diagnosis, and the aggregated diagnosis result is obtained: {1, 5} = {x1, x7}; {1, 6} = {x1, x2, x7}; {1, 7} = {x1, x3, x6}; {1, 8} = {x1, x8, x9}; {2, 4, 5} = {x1, x2, x4, x5, x7}; {2, 4, 6} = {x2, x4, x5, x7}; {2, 4, 7} = {x2, x3, x4, x5, x6}; {2, 4, 8} = {x2, x4, x5, x8, x9}; {3, 4, 5} = {x1, x2, x3, x5, x7}; {3, 4, 6} = {x2, x3, x5, x7}; {3, 4, 7} = {x2, x3, x5, x6, x7}; {3, 4, 8} = {x2, x3, x5, x7, x8, x9}; For the above results, the present application further needs further screening, i.e. minimization of the solution set. Because statistically, {x1, x7} and {x1, x2, x7} are two solutions, the former is more likely to fail, and if the failures of all components are independent of each other, the prior probability of failure of each component is assumed to be 10 -3 , then the possibility of failure of the former is one thousand times that of the latter, so the present application needs to optimize and delete the diagnosis set after collision; According to the above analysis, it can be seen that there are more redundant solutions, and the application optimizes the above algorithm to avoid the generation of partial redundant diagnosis, thereby reducing the time of the subset detection process. The redundant rule is: if a diagnosis, its subset appears in all diagnoses, then delete the diagnosis from the diagnosis set Diag, and add the diagnosis to the new diagnosis.

[0032] For the diagnosis d={x1, x7}, its subset {x1}, {x1, x7} appears in all diagnoses (Diag1, Diag2, Diag3), then d={x1, x7} is an aggregate diagnosis (when the components in d fail, it satisfies all observations), and the redundancy needs to be removed. After removing the redundancy, Diag1''={1, 2, 3}, Diag2''={1, 4} and Diag3''={6, 7, 8}. The diagnosis {x1, x7} is deleted from Diag3''.

[0033] The diagnosis set MHS({Diag1'', Diag2'', Diag3''}) after removing the redundancy is {{1, 6}, {1, 7}, {1, 8}, {2, 4, 6}, {2, 4, 7}, {2, 4, 8}, {3, 4, 6}, {3, 4, 7}, {3, 4, 8}}, and the corresponding diagnosis form is obtained according to the mapping relationship of the key: {1, 6}={x1, x2, x7}; {1, 7}={x1, x3, x6}; {1, 8}={x1, x8, x9}; {2, 4, 6}={x2, x4, x5, x7}; {2, 4, 7}={x2, x3, x4, x5, x6}; {2, 4, 8}={x2, x4, x5, x8, x9}; {3, 4, 6}={x2, x3, x5, x7}; {3, 4, 7}={x2, x3, x5, x6, x7}; {3, 4, 8}={x2, x3, x5, x7, x8, x9}; The running efficiency of the minimal hitting set algorithm is improved, and the running efficiency of the aggregate diagnosis is also improved.

[0034] Minimization is the core bottleneck that hinders the solving efficiency of aggregate diagnosis, and its running time accounts for a large part of the whole process. The application proposes a branch and bound minimization strategy, which realizes efficiency breakthrough through innovative sorting logic and screening rules. The specific process is as follows Figure 3As shown, firstly, all diagnoses are sorted according to the size of the potential (i.e. the number of components contained), and branching is performed according to the potential; then, layer-by-layer minimization screening is performed by using a heuristic strategy, and the bound is completed by using the heuristic strategy for superset deletion and cross-potential layer-by-layer screening. Finally, the minimal diagnosis solution set is obtained.

[0035] It is assumed that all diagnoses are grouped according to the size of the potential, and the potential of each group of diagnoses is the same.

[0036] The present application performs a minimization operation on each group of diagnoses, i.e. it is judged whether there is the same solution.

[0037] The set of the potential of the diagnosis solution set after removing the redundant solution set is Card={2, 3, 4, 5, 6}, the diagnosis solution set is grouped according to the potential and sorted in ascending order, and the result is as follows: When the potential is 2, C2={{x1, x7}}; When the potential is 3, C3={{x1, x2, x7}, {x1, x3, x6}, {x1, x8, x9}}; When the potential is 4, C4={{x2, x4, x5, x7}, {x2, x3, x5, x7}}; When the potential is 5, C5={{x2, x3, x4, x5, x6}, {x2, x4, x5, x8, x9}, {x2, x3, x5, x6, x7}}; When the potential is 6, C6={{x2, x3, x5, x7, x8, x9}}.

[0038] After sorting, self-checking of diagnoses with the same potential is performed first, and the comparison with the subset is different, because the potential of each group of diagnoses is the same, and it is only necessary to judge that there is one component in the solution set, i.e. the judgment is ended. According to the size of the potential, ascending minimization is performed.

[0039] After self-checking, the diagnosis with a small potential may be a subset of the diagnosis with a large potential, and the diagnosis with a large potential cannot be a subset of the diagnosis with a small potential, so it is only necessary to compare whether the diagnosis with a large potential is a superset.

[0040] Firstly, the diagnoses in the group with the smallest potential are all minimal, and are stored in a new container Diag={{x1, x7}}; then, the next group is judged, because {x1, x2, x7} is a superset of {x1, x7}, and x7 is not in {x1, x3, x6}, {x1, x8, x9}, so Diag={{x1, x7}, {x1, x3, x6}, {x1, x8, x9}}. It can be seen that when a minimal solution is newly added, the heuristic strategy is: it is not necessary to judge whether they and the solution in the container are subsets of each other, and it is only necessary to judge that the minimal solution in the container has at least one component not in the solution to be added. The operation will save half. The final fault diagnosis solution set is: {x1, x7}, {x1, x3, x6}, {x1, x8, x9}, {x2, x4, x5, x7}, {x2, x3, x4, x5, x6}, {x2, x4, x5, x8, x9}, {x2, x3, x5, x7}.

[0041] The present application improves the diagnostic accuracy and reduces the risk of misjudgment by observing the aggregation of core logic multiple times. The present application integrates the one-sided diagnostic results of single observation into a standardized diagnostic set (Diag = {d1, d2, …, d i}, covers more fault scenarios by using multi-observation cross-validation, effectively eliminates noise interference and hidden fault omission in single observation, ensures that the final diagnostic solution can explain all fault phenomena, and significantly reduces the debugging cost caused by misjudgment. The present application innovatively introduces the map function of C++STL to realize the mapping of component set-unique key, such as {x2, x4} mapped to 2, converts the complex set comparison into efficient integer comparison, uses the map red-black tree structure to realize the O (log n) search and insertion efficiency, and greatly reduces the redundancy of multi-observation data. At the same time, through the preprocessing stage deletion superset strategy, the invalid candidate solution is reduced in advance before the minimization calculation, the subsequent minimization calculation amount is reduced, and the running time of the overall diagnostic process is shortened. The branch and bound minimization strategy has a breakthrough efficiency advantage: by grouping in ascending order of diagnostic potential, it only needs to judge whether the potential large solution is a superset of the potential small solution, without bidirectional comparison, and the time complexity is reduced from exponential to nearly linear. Combined with heuristic screening rules, the comparison logic is further simplified, for example, when processing 12 candidate solutions, only 7 core comparisons are needed to get the minimal diagnostic solution set. The present application forms a complete technical chain of multiple observation of fault-preprocessing (mapping + BWSS collision + redundancy deletion)-branch and bound minimization, and each link is cooperatively adapted: the standardized diagnostic set provides a unified data basis for subsequent processing, the map mapping and superset deletion reduce the burden of collision calculation, and the branch and bound strategy improves the efficiency of minimization. This whole-chain optimization design ensures smooth connection of the diagnostic process and guarantees the accuracy of diagnosis.

[0042] Embodiment 3, as shown in Figure 4 Embodiment 3, as shown in

[0043] Embodiment 4, a computer readable storage medium, having a computer program stored thereon, the program being executed by a processor to implement the steps of the solving method based on multiple observation aggregation fault diagnosis in embodiment 1.

[0044] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage and so forth) embodying computer-readable program code.

[0045] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 means for carrying out each of the

[0046] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 means for carrying out each of the

[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 means for carrying out each of the

[0048] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing relevant hardware, and the program can be stored in a computer readable storage medium. When the program is executed, the program can include the processes of the above-mentioned embodiment methods. The storage medium can be a magnetic disc, an optical disc, a Read-Only Memory (ROM) or a Random Access Memory (RAM) and the like.

[0049] The above description of disclosed embodiments enables one of ordinary skill in the art to make or use the application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A solving method based on multiple observation aggregation fault diagnosis, characterized in that: The method comprises the following steps: [1] Step S1, obtaining multiple fault diagnoses by observing multiple diagnosis algorithms of the fault, forming a first diagnosis set with the multiple observation diagnosis results; Step S2, performing keyword mapping on the first diagnosis set to form a second diagnosis set, performing BWSS algorithm processing on the second diagnosis set to obtain a third diagnosis set after collision, and performing redundant solution identification and deletion on the third diagnosis set after collision to obtain a fourth diagnosis set after removing redundancy; Step S3, processing the fourth diagnosis set after removing redundancy by using a branch and bound minimization strategy to obtain a minimal diagnosis solution set of the multiple observation diagnosis results.

2. The method of claim 1, wherein: The specific steps of S1 are as follows:[2] Step S11, calling the diagnostic algorithm to make multiple observations to obtain multiple observation fault diagnosis sets Diag1, Diag2, …, Diag i Each diagnostic set Diag i Contains multiple diagnoses, each diagnosis contains at least one component; Step S12: Combine the multiple fault diagnosis sets to form the first diagnosis set Diag = {Diag1, Diag2, ..., Diag...} i } 3. The method of claim 2, wherein: The specific steps of S2 are as follows:[3] Step S21, using the map function of the standard template library STL to component map the first diagnostic set Diag, mapping all the diagnostics to a unique key, forming the second diagnostic set {Diag1 ’ ’ , Diag2 ’ , …, Diag i ’} Step S22, the BWSS algorithm is used to process the second diagnostic set to diagnose the collision of the diagnostic set, and a third diagnostic set MHS (Diag1 ’ , Diag2 ’ , …, Diag i ’ ) after collision is obtained. i ’’} Step S23, the diagnosis set MHS after the collision is processed by removing the redundant rules, deleting each diagnosis set Diag i diagnosis in the third diagnosis set MHS, the remaining key diagnosis in the third diagnosis set MHS is reflected into the component diagnosis and added to the fourth diagnosis set MHS'.

4. The solving method based on multiple observation aggregated fault diagnosis according to claim 3, characterized in that: The specific steps of S3 are as follows:[4] Step S31, performing minimal diagnosis aggregation on the fourth diagnosis set MHS' by using a branch and bound minimization strategy, first decomposing the fourth diagnosis set MHS' according to the potential to obtain diagnoses corresponding to each potential, and sorting the diagnoses in ascending order according to the size of the potential; Step S32, identifying and deleting the superset of the sorted diagnosis set by using a heuristic strategy, and finally obtaining a minimal diagnosis solution set.

5. The method of claim 3, wherein: The component mapping rule in step S21 is as follows:[5] Step S211, the first diagnostic set {Diag1, Diag2, …, Diag i} corresponding to multiple observations is converted into DiagMap<map, int>, at this time the key nextKey=1, DiagMap is traversed to obtain all diagnostic groups Diag1, Diag2, …, Diag i， ; Step S212, traversing each diagnosis group to obtain the set S1, S2, … S n ; Step S213, according to the set S1, S2, … S n For each set S n Perform mapping of keyword words; Step S2131, if S n In DiagMap, this set is skipped, no mapping of the key is performed, and nextKey is unchanged. Step S2132, if S n not in DiagMap then S n The key is nextKey, and then nextKey = nextKey + 1 is executed. Step S21321, if there is a set S in the current diagnosis group n Step S213 is executed. Step S21322, if the current diagnosis group is not set S n , it is determined whether there are more diagnosis groups to be mapped; Step S213221, if there are more diagnosis groups that have not been mapped, step S212 is executed; Step S213222, if all diagnosis groups are mapped as keywords, the keyword mapping is ended.

6. The solving method based on multiple observation aggregated fault diagnosis according to claim 3 or 5, characterized in that: The redundancy removal rule in step S24 is as follows: if a diagnosis and its subsets appear in all diagnoses, the diagnosis is deleted from the diagnosis set, and the diagnosis is added to a new diagnosis set.

7. The method for solving the aggregated fault diagnosis based on multiple observations according to claim 4, wherein the superset is defined as follows: if a diagnosis and its subsets appear in all diagnoses, the diagnosis is an aggregated diagnosis, but the set formed by adding any component to the diagnosis is a superset.[6] 8. The method of claim 4, wherein: The potential is the number of components in each diagnosis set of the diagnosis set MHS', and the heuristic strategy is as follows:[7] Starting from the smallest potential group of the diagnosis set MHS', the minimal solution is first stored in the new diagnosis solution set Diag''', and then the set with a potential greater than the minimal solution is judged: if there is a superset that can cover the minimal solution, the superset is deleted; if not, the original set is added to Diag'''; when a new minimal solution is added, if at least one component in the minimal solution of the diagnosis solution set Diag''' is not in the added solution, the diagnosis set to be added is added to Diag'''.

9. An electronic device comprising a memory (102), a processor (101), a display module (103) and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the steps of the method for solving the aggregated fault diagnosis based on multiple observations according to any one of claims 1 to 8 when executing the program.

10. A readable storage medium, having stored thereon a computer program, characterized in that, The computer program implements the steps of the method for solving the aggregated fault diagnosis based on multiple observations according to any one of claims 1 to 8 when executed by the processor.

Images