Fault diagnosis subsystem design method and device for equipment intelligent operation and maintenance decision

Abstract

The invention discloses an equipment intelligent operation and maintenance decision-oriented fault diagnosis subsystem design method and device, and relates to the technical field of fault diagnosis, and the method comprises the steps: obtaining the related data of target equipment, and obtaining a plurality of to-be-diagnosed fault modes of the target equipment based on the related data; constructing a plurality of candidate diagnosis algorithm combination schemes based on each to-be-diagnosed fault mode; optimizing the plurality of candidate diagnosis algorithm combination schemes based on an optimization algorithm to obtain a plurality of diagnosis algorithm combination schemes meeting a critical time index; and evaluating the diagnosis precision of the plurality of diagnosis algorithm combination schemes according to the evaluation model, and obtaining an optimal diagnosis algorithm combination scheme based on the diagnosis precision. Through the optimized diagnosis algorithm combination scheme, it is ensured that each related diagnosis algorithm can make a diagnosis conclusion within a critical time index, the capacity of capturing and analyzing early fault signals is improved, the faults are effectively prevented from being diffused in a propagation chain in equipment, and therefore early diagnosis of the faults is achieved.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and more specifically to a fault diagnosis subsystem design method and apparatus for intelligent equipment operation and maintenance decision-making. Background Technology

[0002] Currently, the fault diagnosis subsystems of active-duty equipment health management systems generally suffer from untimely diagnosis and low accuracy when dealing with equipment with increasing complexity and functional coupling. With the development of modern technology, the close interconnections between different components within equipment have formed complex fault propagation chains, making it possible for minor faults to quickly affect critical components or subsystems, thereby leading to the failure of important missions.

[0003] However, current fault diagnosis subsystems fail to fully utilize this information to achieve earlier and more accurate fault identification. On the one hand, they struggle to capture and analyze early fault signals in a timely manner to prevent faults from spreading to a wider range of systems through internal propagation chains. On the other hand, lower diagnostic accuracy limits the possibility of shifting from scheduled maintenance to condition-based maintenance, which not only wastes maintenance resources but may also lead to over-maintenance and unnecessary entry of equipment into maintenance procedures due to false fault reports, affecting the execution of normal tasks.

[0004] Therefore, improving the speed and accuracy of fault diagnosis, while simultaneously achieving real-time monitoring and assessment of equipment health status, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a fault diagnosis subsystem design method and apparatus for intelligent operation and maintenance decision-making of equipment, which overcomes the above-mentioned defects.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A design method for a fault diagnosis subsystem oriented towards intelligent equipment operation and maintenance decision-making, comprising the following steps:

[0008] Acquire relevant data of the target equipment, and obtain multiple fault modes to be diagnosed of the target equipment based on the relevant data;

[0009] Based on each of the aforementioned fault modes to be diagnosed, construct multiple candidate diagnostic algorithm combination schemes;

[0010] The candidate diagnostic algorithm combination schemes are optimized based on the optimization algorithm to obtain multiple diagnostic algorithm combination schemes that meet the critical time index.

[0011] The diagnostic accuracy of multiple diagnostic algorithm combinations is evaluated using an evaluation model, and the optimal diagnostic algorithm combination is obtained based on the diagnostic accuracy.

[0012] Optionally, the step of obtaining the fault mode to be diagnosed is as follows:

[0013] Based on the relevant data, the failure mode type of the target equipment is extracted;

[0014] Analyze the failure propagation chain for each failure mode type using reliability engineering methods;

[0015] Multiple fault modes to be diagnosed for the target equipment are determined based on the fault propagation chain of each fault mode type.

[0016] Optionally, the step of obtaining the candidate diagnostic algorithm combination scheme is as follows:

[0017] Based on the relevant data of the target equipment, relevant diagnostic algorithms and data sets for each of the fault modes to be diagnosed are extracted, namely the first data set;

[0018] Based on the aforementioned diagnostic algorithms, a combination scheme of multiple candidate diagnostic algorithms is constructed to satisfy the system's computational resource constraints.

[0019] Optionally, the step of obtaining the diagnostic algorithm combination scheme is as follows:

[0020] Step 31: Obtain the sampling frequency, input data length, and first data set of each candidate diagnostic algorithm in the candidate diagnostic algorithm combination scheme;

[0021] Step 32: Divide the first dataset according to a preset segmentation ratio to obtain a first training dataset and a first test dataset;

[0022] Step 33: Extract data samples from the first training dataset and the first test dataset based on the input data length, and construct the second training dataset and the second test dataset respectively;

[0023] Step 34: Train the candidate diagnostic algorithm based on the second training dataset, and evaluate the first diagnostic accuracy of the trained candidate diagnostic algorithm at the input signal length based on the test dataset;

[0024] Step 35: Update the first diagnostic accuracy to the diagnostic accuracy set, determine the updated input data length according to the diagnostic accuracy set, jump to step 33, until the iteration stop condition is reached, and obtain the optimal input data length and the optimal diagnostic accuracy estimate.

[0025] Step 36: Construct an updated diagnostic algorithm based on the optimal input data length and the optimal diagnostic accuracy estimate, and replace the candidate diagnostic algorithm with the updated diagnostic algorithm to form the diagnostic algorithm combination scheme.

[0026] Optionally, the step of obtaining the updated input data length is as follows: the maximum value in the diagnostic accuracy set is used as the updated diagnostic accuracy estimate of the diagnostic algorithm, and the position number of the updated diagnostic accuracy estimate in the diagnostic accuracy set is the updated input data length.

[0027] Optionally, the step of obtaining the optimal combination of diagnostic algorithms is as follows:

[0028] A second dataset is constructed based on multiple combinations of the aforementioned diagnostic algorithms;

[0029] The second dataset is input into the evaluation model for training and learning to obtain multiple diagnostic accuracies;

[0030] The optimal combination of diagnostic algorithms is obtained based on multiple diagnostic accuracies.

[0031] In one embodiment, the second data set includes the diagnostic results and data sample labels of each diagnostic algorithm in the diagnostic algorithm combination scheme.

[0032] A fault diagnosis subsystem design device for intelligent equipment operation and maintenance decision-making includes:

[0033] A graphical interface module is used to obtain the system design parameters and display the design results;

[0034] The system design module is used to execute the design method based on the design parameters to generate the design results.

[0035] As can be seen from the above technical solution, the present invention discloses a fault diagnosis subsystem design method and device for intelligent equipment operation and maintenance decision-making, which has the following advantages compared with the prior art:

[0036] Improve fault diagnosis speed: Through the optimized combination of diagnostic algorithms, it is ensured that each relevant diagnostic algorithm can make a diagnostic conclusion within the critical time index. This not only improves the ability to capture and analyze early fault signals, but also effectively prevents the fault from spreading in the propagation chain within the equipment, thereby achieving early fault diagnosis.

[0037] Enhancing diagnostic accuracy: By introducing ensemble learning technology to integrate the capabilities of different diagnostic algorithms, diagnostic accuracy assessment is achieved on a case-by-case basis. This measure significantly improves the accuracy of fault diagnosis, reduces false alarms, enables condition-based maintenance, avoids unnecessary maintenance procedures, and improves equipment availability and mission execution efficiency.

[0038] Real-time health monitoring and assessment: The improved fault diagnosis subsystem can monitor the health status of equipment in real time and provide immediate and accurate assessment results. This is especially important for modern equipment with high complexity and strong functional coupling, as it can help maintenance personnel understand the equipment status in a timely manner, formulate scientific and reasonable maintenance strategies, and reduce unplanned downtime. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of the method flow provided by the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] One embodiment of the present invention discloses a design method for a fault diagnosis subsystem oriented towards intelligent equipment operation and maintenance decision-making, such as... Figure 1 As shown, the specific steps are as follows:

[0043] Step 1: Obtain relevant data of the target equipment, and based on the relevant data, obtain multiple fault modes of the target equipment to be diagnosed;

[0044] Step 2: Construct multiple candidate diagnostic algorithm combination schemes based on each fault mode to be diagnosed;

[0045] Step 3: Optimize multiple candidate diagnostic algorithm combination schemes based on optimization algorithms to obtain multiple diagnostic algorithm combination schemes that meet the critical time index;

[0046] Step 4: Evaluate the diagnostic accuracy of multiple diagnostic algorithm combinations based on the evaluation model, and obtain the optimal diagnostic algorithm combination based on the diagnostic accuracy.

[0047] In one embodiment, the step of obtaining the fault mode to be diagnosed is as follows:

[0048] Step 11: Extract the failure mode type of the target equipment based on relevant data;

[0049] Step 12: Analyze the failure propagation chain for each failure mode type using reliability engineering methods;

[0050] Step 13: Determine multiple fault modes to be diagnosed for the target equipment based on the fault propagation chain of each fault mode type.

[0051] Furthermore, the failure modes to be diagnosed for the target equipment are identified through the following steps: First, relevant data on the target equipment are collected, including design materials, quality analysis materials, and quality problem records and maintenance manuals of similar products. Then, based on the relevant data, the failure mode types of the target equipment are extracted. Finally, FMEA and FTA methods are used to analyze the possible failure propagation chains of each failure mode, clarify the possible mission impact and harmful consequences of the failure mode, and determine the failure modes to be diagnosed for the target equipment accordingly.

[0052] In one embodiment, the step of obtaining the candidate diagnostic algorithm combination scheme is as follows:

[0053] Step 21: Extract relevant diagnostic algorithms and datasets for each fault mode to be diagnosed based on the relevant data of the target equipment, i.e., the first dataset;

[0054] Step 22: Construct a combination of multiple candidate diagnostic algorithms that satisfy the system's computational resource constraints based on relevant diagnostic algorithms.

[0055] Furthermore, taking the fault mode to be diagnosed as a unit, multiple different candidate diagnostic algorithm combination schemes are formulated. The specific steps are as follows: First, taking the fault mode to be diagnosed as a unit, relevant diagnostic algorithms targeting that fault mode are collected. and the corresponding training data set for the algorithm. Then, based on the computational resource design constraints and requirements of the equipment health management system, and combined with the design experience of similar equipment health management system fault diagnosis subsystems, several different candidate diagnostic algorithm combination schemes were constructed, forming a set of candidate diagnostic algorithm combination schemes. This means that the set contains Different combinations of candidate diagnostic algorithms, each of the following combinations All consist of multiple different types of diagnostic algorithms. Composition, which can be represented as .

[0056] In one embodiment, the step of obtaining the diagnostic algorithm combination scheme is as follows:

[0057] Step 31: Obtain the sampling frequency, input data length, and first data set of each candidate diagnostic algorithm in the candidate diagnostic algorithm combination scheme;

[0058] Step 32: Divide the first dataset according to the preset splitting ratio to obtain the first training dataset and the first test dataset;

[0059] Step 33: Extract data samples from the first training dataset and the first test dataset based on the length of the input data, and construct the second training dataset and the second test dataset.

[0060] Step 34: Train the candidate diagnostic algorithm based on the second training dataset, and evaluate the first diagnostic accuracy of the trained candidate diagnostic algorithm with respect to the input signal length based on the test dataset;

[0061] Step 35: Update the first diagnostic accuracy to the diagnostic accuracy set, determine the length of the updated input data based on the diagnostic accuracy set, and jump to step 33 until the iteration stop condition is met to obtain the optimal input data length and the optimal diagnostic accuracy estimate.

[0062] Step 36: Construct an updated diagnostic algorithm based on the optimal input data length and the optimal diagnostic accuracy estimate. Replace the candidate diagnostic algorithms with the updated diagnostic algorithm to form a diagnostic algorithm combination scheme.

[0063] In one embodiment, the step of obtaining the updated input data length is as follows: the maximum value in the diagnostic accuracy set is used as the updated diagnostic accuracy estimate of the updated diagnostic algorithm, and the position number of the updated diagnostic accuracy estimate in the diagnostic accuracy set is the updated input data length.

[0064] Furthermore, a candidate diagnostic algorithm combination scheme Each candidate diagnostic algorithm Based on this, an optimization algorithm was designed to meet the critical time requirement. The iterative steps of the optimization algorithm are shown in Table 1. The newly designed diagnostic algorithm was then used. Replacement candidate diagnostic algorithm combination scheme Candidate diagnostic algorithms in To obtain a combination of diagnostic algorithms, the specific steps are as follows: First, determine each candidate diagnostic algorithm. Data collection frequency in the corresponding dataset Based on candidate diagnostic algorithm Length of input data required for diagnosis and data sampling frequency It can be done through formula Estimating the diagnostic algorithm Time taken to reach a diagnostic conclusion Then, an optimization algorithm is executed for each candidate diagnostic algorithm. Improve the design to obtain performance indicators at critical time. It has the diagnostic algorithm with the highest diagnostic accuracy. , will be The definition of the set of triples, i.e., algorithm rules and The same applies, the length of the input data is determined by... It is confirmed that it has a high diagnostic accuracy. Finally, a diagnostic algorithm is used. Replacement candidate diagnostic algorithm combination scheme Candidate diagnostic algorithms in This allows for the updating of the candidate diagnostic algorithm combination scheme.

[0065] Table 1

[0066] In one embodiment, the step of obtaining the optimal combination of diagnostic algorithms is as follows:

[0067] Step 41: Construct a second dataset based on a combination of multiple diagnostic algorithms;

[0068] Step 42: Input the second dataset into the evaluation model for training and learning to obtain multiple diagnostic accuracies;

[0069] Step 43: Obtain the optimal combination of diagnostic algorithms based on multiple diagnostic accuracies.

[0070] In one embodiment, the second dataset includes the diagnostic results and data sample labels of each diagnostic algorithm in the diagnostic algorithm combination scheme.

[0071] Furthermore, the training steps for evaluating the model are shown in Table 2. The diagnostic algorithm combination scheme with the highest diagnostic accuracy is selected by evaluating each combination scheme. The specific steps are as follows: First, each diagnostic algorithm combination is constructed using the diagnostic algorithms obtained in step 3. The second dataset This dataset uses a combination of diagnostic algorithms. Each diagnostic algorithm Diagnostic results With data sample labels The second dataset is then constructed using diagnostic algorithms. Based on this, train all diagnostic algorithms in the integrated scheme to obtain a new diagnostic algorithm. And obtain the diagnostic accuracy of the new diagnostic algorithm. This is used as the diagnostic accuracy result of the diagnostic algorithm combination scheme. After repeating the above steps, the diagnostic algorithm corresponding to each diagnostic algorithm combination scheme is obtained. Diagnostic accuracy Ultimately, the diagnostic scheme with the highest diagnostic accuracy is selected.

[0072] Table 2

[0073] In one embodiment, a specific device will be used for further explanation:

[0074] Step 1: Identify the Fault Modes of the Target Equipment: A comprehensive analysis and classification of the fault modes of a certain type of equipment was conducted, ultimately identifying 65 fault modes awaiting diagnosis, including severe engine wear, low fluid level in the integrated transmission system, and abnormal control of the automatic loading mechanism. For subsequent steps, we will use severe engine wear as an example; the relevant steps for the other fault modes are identical and will not be repeated.

[0075] Step 2: Develop multiple candidate diagnostic algorithm combination schemes based on the fault mode to be diagnosed. For the fault mode of severe engine wear, two candidate diagnostic algorithm combination schemes are developed. Therefore, the set of candidate diagnostic algorithm combination schemes can be represented as follows: Among them, the plan Diagnostic algorithms based on decision trees Diagnostic algorithms based on support vector machines Composition; Scheme Diagnostic algorithms based on random forests Diagnostic algorithm based on radial support vector machine and diagnostic algorithms based on logistic regression Composition. Simultaneously, a data set was constructed based on 17 types of signals, including engine lubricating oil abrasive particle information signals of ferromagnetic particles smaller than 100 micrometers, 100-300 micrometers, 300-600 micrometers, and larger than 600 micrometers, as well as engine lubricating oil viscosity, density, and contamination signals. and .

[0076] Step 3: Combination of candidate diagnostic algorithms and The relevant algorithms have been developed to meet the critical time index requirements ( The design work was carried out, and the original algorithm in the scheme was replaced with a new design algorithm. The relevant information of the new algorithm for each scheme is shown in Table 3.

[0077] Table 3

[0078] (Note: The data sampling frequency in the relevant dataset is 0.1 Hz)

[0079] Step 4: Based on the evaluation model, the diagnostic accuracy of each diagnostic algorithm combination scheme was obtained, which are respectively , Ultimately, the second candidate diagnostic algorithm combination was selected and deployed to the fault diagnosis subsystem, meeting the design target of 90% diagnostic accuracy, thus completing all related design work.

[0080] Another embodiment of the present invention discloses a fault diagnosis subsystem design device for intelligent equipment operation and maintenance decision-making, comprising:

[0081] The graphical interface module is used to obtain system design parameters and display design results;

[0082] The system design module is used to execute design methods based on design parameters to generate design results.

[0083] Furthermore, using the above method, a fault diagnosis subsystem for designing the target equipment health management system is constructed. This subsystem selects the combination of diagnostic algorithms to be deployed in the fault diagnosis subsystem and ensures that the diagnostic algorithms in the scheme meet the specified critical time indicators. It consists of two parts: a graphical interface module and a system design module. The graphical interface module acquires the design input of the fault diagnosis subsystem entered by the designer and outputs the design results; the system design module executes the above steps based on the design input information captured by the graphical interface module.

[0084] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0085] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not 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 design method for a fault diagnosis subsystem oriented towards intelligent equipment operation and maintenance decision-making, characterized in that, The specific steps are as follows: Acquire relevant data of the target equipment, and obtain multiple fault modes to be diagnosed of the target equipment based on the relevant data; Based on each of the aforementioned fault modes to be diagnosed, construct multiple candidate diagnostic algorithm combination schemes; The candidate diagnostic algorithm combination schemes are optimized based on the optimization algorithm to obtain multiple diagnostic algorithm combination schemes that meet the critical time index. The diagnostic accuracy of multiple diagnostic algorithm combinations is evaluated using an evaluation model, and the optimal diagnostic algorithm combination is obtained based on the diagnostic accuracy.

2. The design method for a fault diagnosis subsystem for intelligent equipment operation and maintenance decision-making according to claim 1, characterized in that, The steps for obtaining the fault mode to be diagnosed are as follows: Based on the relevant data, the failure mode type of the target equipment is extracted; Analyze the failure propagation chain for each failure mode type using reliability engineering methods; Multiple fault modes to be diagnosed for the target equipment are determined based on the fault propagation chain of each fault mode type.

3. The design method for a fault diagnosis subsystem for intelligent equipment operation and maintenance decision-making according to claim 1, characterized in that, The steps for obtaining the candidate diagnostic algorithm combination scheme are as follows: Based on the relevant data of the target equipment, relevant diagnostic algorithms and data sets for each of the fault modes to be diagnosed are extracted, namely the first data set; Based on the aforementioned diagnostic algorithms, a combination scheme of multiple candidate diagnostic algorithms is constructed to satisfy the system's computational resource constraints.

4. The design method of a fault diagnosis subsystem for intelligent equipment operation and maintenance decision-making according to claim 3, characterized in that, The steps for obtaining the diagnostic algorithm combination scheme are as follows: Step 31: Obtain the sampling frequency, input data length, and first data set of each candidate diagnostic algorithm in the candidate diagnostic algorithm combination scheme; Step 32: Divide the first dataset according to a preset segmentation ratio to obtain a first training dataset and a first test dataset; Step 33: Extract data samples from the first training dataset and the first test dataset based on the input data length, and construct the second training dataset and the second test dataset respectively; Step 34: Train the candidate diagnostic algorithm based on the second training dataset, and evaluate the first diagnostic accuracy of the trained candidate diagnostic algorithm at the input signal length based on the test dataset; Step 35: Update the first diagnostic accuracy to the diagnostic accuracy set, determine the updated input data length according to the diagnostic accuracy set, jump to step 33, until the iteration stop condition is reached, and obtain the optimal input data length and the optimal diagnostic accuracy estimate. Step 36: Construct an updated diagnostic algorithm based on the optimal input data length and the optimal diagnostic accuracy estimate, and replace the candidate diagnostic algorithm with the updated diagnostic algorithm to form the diagnostic algorithm combination scheme.

5. The design method for a fault diagnosis subsystem for intelligent equipment operation and maintenance decision-making according to claim 1, characterized in that, The step of obtaining the updated input data length is as follows: the maximum value in the diagnostic accuracy set is used as the updated diagnostic accuracy estimate of the diagnostic algorithm, and the position number of the updated diagnostic accuracy estimate in the diagnostic accuracy set is the updated input data length.

6. The design method for a fault diagnosis subsystem for intelligent equipment operation and maintenance decision-making according to claim 1, characterized in that, The steps for obtaining the optimal combination of diagnostic algorithms are as follows: A second dataset is constructed based on multiple combinations of the aforementioned diagnostic algorithms; The second dataset is input into the evaluation model for training and learning to obtain multiple diagnostic accuracies; The optimal combination of diagnostic algorithms is obtained based on multiple diagnostic accuracies.

7. The design method for a fault diagnosis subsystem for intelligent equipment operation and maintenance decision-making according to claim 6, characterized in that, The second dataset includes the diagnostic results and data sample labels of each diagnostic algorithm in the diagnostic algorithm combination scheme.

8. A fault diagnosis subsystem design device for intelligent equipment operation and maintenance decision-making, characterized in that, include: A graphical interface module is used to obtain the system design parameters and display the design results; The system design module is used to execute the design method based on the design parameters to generate the design results.

Images