A welding machine fault diagnosis method and system

Abstract

The present invention discloses a welding machine fault diagnosis method and system, the method comprising the following steps: collecting welding machine data under various fault states and constructing a database; preprocessing the welding machine data in the database and generating a training data set; training a hidden Markov model using the training data set to generate a fault state model corresponding to different fault states; and monitoring the welding machine state in real time based on the fault state model. The present invention can determine the welding machine fault location and fault mode by analyzing and calculating the welding machine fault state model and the real-time collected data, so as to facilitate technicians to determine a fault solution.

Description

Technical Field

[0001] The invention belongs to the field of fault diagnosis, and in particular to a welding machine fault diagnosis method and system. Background Art

[0002] In the development of modern industry, welding technology plays an indispensable role. With the promotion of welding technology, welding machines, as an important part of the welding process, have also been widely used. Welding machines have the advantages of simple operation, convenient use, and strong welds after welding. They can be used in various industrial fields such as aerospace, shipbuilding, and automobiles.

[0003] However, due to the increase in the service life of the welding machine and the impact of the environment on electronic components, many types of faults will occur. Common welding machine faults include: welding overheating, low no-load voltage, unadjustable current, air switch tripping after power supply, etc. The inside of the welding machine is composed of multiple circuits that control each other. If one of the components is damaged, the other components will also be affected. The complexity of the internal system structure of the welding machine has brought great difficulty to the operator to solve the welding machine failure. To determine the welding machine failure by traditional methods, various measuring tools are required to go through various steps to measure whether each component is damaged. Such a measurement method cannot quickly and accurately locate the fault site and determine the fault form, making it difficult to carry out subsequent maintenance work, resulting in long-term downtime and affecting production efficiency.

[0004] Based on data-driven welding machine fault diagnosis, historical data is used to establish a welding machine fault model. On this basis, through the analysis and processing of real-time data, the system intuitively displays the welding machine status. Once a welding machine fails, the fault location can be quickly and accurately determined, providing strong guarantees for maintenance personnel to formulate plans, minimize downtime as much as possible, and improve production efficiency. Summary of the invention

[0005] The object of the present invention is to provide a welding machine fault diagnosis method and system, which can monitor the welding machine status in real time and determine the welding machine fault mode.

[0006] The technical solution to achieve the purpose of the present invention is: a welding machine fault diagnosis method, the method comprising the following steps:

[0007] Step 1, collecting welding machine data under various fault conditions and building a database;

[0008] Step 2, preprocessing the welding machine data in the database to generate a training data set;

[0009] Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0010] Step 4: Based on the fault state model, real-time monitoring of the welding machine state.

[0011] Furthermore, in step 2, the welding machine data in the database is preprocessed to generate a training data set, specifically, the welding machine data is preprocessed according to the principal component analysis method PCA, and the process includes:

[0012] Step 2-1, selecting welding machine data corresponding to a certain fault state within a certain time period from the database to form a high-dimensional data matrix;

[0013] Step 2-2, use principal component analysis (PCA) to reduce the dimensionality of the high-dimensional data matrix to obtain a one-dimensional data matrix, which constitutes a training data set corresponding to the fault state.

[0014] Furthermore, in step 4, real-time monitoring of the welding machine state based on the fault state model specifically includes:

[0015] Step 4-1, real-time collection of welding machine data;

[0016] Step 4-2, preprocessing the real-time collected welding machine data using the method of step 2 to obtain a one-dimensional data matrix corresponding to the current welding machine data;

[0017] Step 4-3, calculate the maximum likelihood similarity probability between the one-dimensional data matrix and each fault state model, select the model with the largest probability value, and the fault state corresponding to the model is the current welding machine fault state.

[0018] A welding machine fault diagnosis system, the system comprising a data acquisition module, a data preprocessing module, a fault state model training module and a real-time state monitoring module;

[0019] The data acquisition module connects the upper computer and the lower computer according to the CAN bus communication protocol, collects the welding machine data under various fault states, and builds a database;

[0020] The data preprocessing module is used to preprocess the welding machine data in the database to generate a training data set;

[0021] The fault state model training module uses the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0022] The real-time status monitoring module is used to monitor the status of the welding machine in real time.

[0023] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0024] Step 1, collecting welding machine data under various fault conditions and building a database;

[0025] Step 2, preprocessing the welding machine data in the database to generate a training data set;

[0026] Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0027] Step 4: Based on the fault state model, real-time monitoring of the welding machine state.

[0028] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0029] Step 1, collecting welding machine data under various fault conditions and building a database;

[0030] Step 2, preprocessing the welding machine data in the database to generate a training data set;

[0031] Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0032] Step 4: Based on the fault state model, real-time monitoring of the welding machine state.

[0033] Compared with the prior art, the present invention has the following significant advantages: (1) a novel welding machine fault diagnosis system and a method for real-time monitoring of the welding machine status and determining the welding machine fault mode are proposed; (2) the welding machine data is analyzed and processed by a hidden Markov model (HMM), which has a strong fault mode classification capability and realizes online monitoring of the welding machine status.

[0034] The present invention is further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The figure is a flow chart of the welding machine fault diagnosis method.

[0036] Figure 2 It is a structural diagram of the specific implementation of the welding machine fault diagnosis system.

[0037] Figure 3 It is a schematic diagram of the working process of the welding machine fault diagnosis system.

[0038] Figure 4 It is a schematic diagram of welding machine fault state model training and state monitoring calculation process. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0040] In one embodiment, in combination Figure 1 , provides a welding machine fault diagnosis method, comprising the following steps:

[0041] Step 1, collecting welding machine data under various fault conditions and building a database;

[0042] Here, the welding machine data includes parameters such as current, voltage, wire feeding speed, hot wire current, etc.; the fault status includes power phase loss, internal fuse breakage, internal wire breakage, main control board damage, etc.;

[0043] Step 2, preprocessing the welding machine data in the database to generate a training data set;

[0044] Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0045] Step 4: Based on the fault state model, real-time monitoring of the welding machine state.

[0046] Further, in one embodiment, the step 2 pre-processes the welding machine data in the database to generate a training data set, specifically pre-processing the welding machine data according to the principal component analysis method PCA, and the process includes:

[0047] Step 2-1, selecting welding machine data corresponding to a certain fault state within a certain time period from the database to form a high-dimensional data matrix;

[0048] Step 2-2, use principal component analysis (PCA) to reduce the dimensionality of the high-dimensional data matrix to obtain a one-dimensional data matrix, which constitutes a training data set corresponding to the fault state.

[0049] Further, in one embodiment, the step 4 of monitoring the welding machine status in real time based on the fault status model specifically includes:

[0050] Step 4-1, real-time collection of welding machine data;

[0051] Step 4-2, preprocessing the real-time collected welding machine data using the method of step 2 to obtain a one-dimensional data matrix corresponding to the current welding machine data;

[0052] Step 4-3, calculate the maximum likelihood similarity probability between the one-dimensional data matrix and each fault state model, select the model with the largest probability value, and the fault state corresponding to the model is the current welding machine fault state.

[0053] In one embodiment, in combination Figures 2 to 4 , provides a welding machine fault diagnosis system, the system includes a data acquisition module, a data preprocessing module, a fault state model training module and a real-time state monitoring module;

[0054] The data acquisition module connects the upper computer and the lower computer according to the CAN bus communication protocol, collects the welding machine data under various fault states, and builds a database;

[0055] The data preprocessing module is used to preprocess the welding machine data in the database to generate a training data set;

[0056] The fault state model training module uses the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0057] The real-time status monitoring module is used to monitor the status of the welding machine in real time.

[0058] Furthermore, in one embodiment, the data preprocessing module includes:

[0059] A matrix construction unit is used to select welding machine data corresponding to a certain fault state within a certain time period from the database to form a high-dimensional data matrix;

[0060] The dimension reduction unit is used to perform dimension reduction processing on the high-dimensional data matrix using the principal component analysis method PCA to obtain a one-dimensional data matrix to constitute a training data set corresponding to the fault state.

[0061] Furthermore, in one embodiment, the real-time status monitoring module includes:

[0062] Real-time acquisition unit, used to collect welding machine data in real time;

[0063] A preprocessing unit is used to preprocess the real-time collected welding machine data using the preprocessing method of the data preprocessing module to obtain a one-dimensional data matrix corresponding to the current welding machine data;

[0064] The state monitoring unit is used to calculate the maximum likelihood similarity probability between the one-dimensional data matrix and each fault state model, and select the model with the largest probability value. The fault state corresponding to the model is the current welding machine fault state.

[0065] The specific definition of the welding machine fault diagnosis system can be found in the definition of the welding machine fault diagnosis method mentioned above, which will not be repeated here. Each module in the above welding machine fault diagnosis system can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0066] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0067] Step 1, collecting welding machine data under various fault conditions and building a database;

[0068] Step 2, preprocessing the welding machine data in the database to generate a training data set;

[0069] Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0070] Step 4: Based on the fault state model, real-time monitoring of the welding machine state.

[0071] For the specific limitations of each step, please refer to the limitations of the welding machine fault diagnosis method above, which will not be repeated here.

[0072] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0073] Step 1, collecting welding machine data under various fault conditions and building a database;

[0074] Step 2, preprocessing the welding machine data in the database to generate a training data set;

[0075] Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states;

[0076] Step 4: Based on the fault state model, real-time monitoring of the welding machine state.

[0077] For the specific limitations of each step, please refer to the limitations of the welding machine fault diagnosis method above, which will not be repeated here.

[0078] As a specific example, the present invention is further described.

[0079] The present invention provides a welding machine fault diagnosis method, comprising the following contents:

[0080] Step 1: Data collection: collect data of each welding machine, including welding machine current, welding machine voltage, wire feeding speed, hot wire current, and record the time node of data collection, and store it in MySQL database through SQL statements. During the collection process, the switch (CANMsg.ID >> 21) statement is used to distinguish the device model, and the switch (CANMsg.ID >> 1&0b_1111) statement is used to distinguish the welding machine data from the wire feeding machine and the hot wire machine data. The ReadArc() function and the ReadWire() function are used to read the welding machine data and the wire feeding machine and the hot wire machine data respectively.

[0081] Step 2, data preprocessing, preprocess the collected data according to the principal component analysis (PCA), and reduce the dimension of the high-dimensional data matrix composed of welding machine current, welding machine voltage, etc. by the cov(double x[][dim], int m) function to obtain a one-dimensional matrix suitable for fault state model training. The specific method is: select the data of the welding machine in a certain period of time to form a high-dimensional data matrix, first calculate the mean ave[] of each parameter, and calculate the centralization matrix cen[][] and the covariance matrix Cov[][] through the mean, obtain the eigenvalue and eigenvector of the covariance matrix, select the eigenvector corresponding to the maximum eigenvalue, and calculate the one-dimensional matrix o_pca[] suitable for fault state model training.

[0082] Step 3, fault state model training. There are many forms of welding machine failure, including power phase loss, internal fuse breakage, internal wire breakage, main control board damage, etc. Taking the internal wire breakage as an example, the data in the corresponding time period of this situation is selected from the database, and a one-dimensional data matrix is ​​obtained after data preprocessing. Taking the data matrix as the training object, the baum_welch() function based on the hidden Markov model (HMM) is used to iteratively optimize the model parameters to obtain the fault state model corresponding to the internal wire breakage. The specific model parameters include the initial probability distribution vector PI, the state transition probability matrix A, and the observation value probability matrix B.

[0083] Step 4, real-time status monitoring, taking the current welding machine status as a broken wire as an example. A segment of data containing the current time node is selected from the database, and after data preprocessing, a one-dimensional data matrix o_pca[] is obtained. The maximum likelihood similarity probability P(O|λ) between the data matrix and each fault state model is calculated through the viterbi() function based on the hidden Markov model (HMM), and the model corresponding to the maximum value is selected, that is, the current welding machine fault state is a broken wire.

[0084] The present invention can determine the fault location and fault mode of the welding machine by analyzing and calculating the fault state model of the welding machine and the real-time collected data, so as to facilitate the technicians to determine the fault solution.

[0085] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A welding machine fault diagnosis method, characterized in that: The method comprises the following steps: Step 1, collecting welding machine data under various fault conditions and building a database; Step 2: preprocess the welding machine data in the database to generate a training data set; specifically, preprocess the welding machine data according to the principal component analysis method PCA, and the process includes: Step 2-1, selecting welding machine data corresponding to a certain fault state within a certain time period from the database to form a high-dimensional data matrix; Step 2-2, using principal component analysis (PCA) to reduce the dimensionality of the high-dimensional data matrix to obtain a one-dimensional data matrix, which constitutes a training data set corresponding to the fault state; Step 3, using the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states; Step 4, based on the fault state model, real-time monitoring of the welding machine state; specifically comprising: Step 4-1, real-time collection of welding machine data; Step 4-2, preprocessing the real-time collected welding machine data using the method of step 2 to obtain a one-dimensional data matrix corresponding to the current welding machine data; Step 4-3, calculating the maximum likelihood similarity probability between the one-dimensional data matrix and each fault state model, selecting the model with the largest probability value, and the fault state corresponding to the model is the current welding machine fault state; The welding machine data in step 1 include current, voltage, wire feeding speed, and hot wire current; the fault state includes power phase loss, internal fuse breakage, internal wire breakage, and main control board damage; Step 1 specifically collects welding machine data under various fault states through the CAN bus communication protocol.

2. A welding machine fault diagnosis system based on the method of claim 1, characterized in that: The system includes a data acquisition module, a data preprocessing module, a fault state model training module and a real-time state monitoring module; The data acquisition module connects the upper computer and the lower computer according to the CAN bus communication protocol, collects the welding machine data under various fault states, and builds a database; The data preprocessing module is used to preprocess the welding machine data in the database to generate a training data set; The fault state model training module uses the training data set to train the hidden Markov model to generate fault state models corresponding to different fault states; The real-time status monitoring module is used to monitor the status of the welding machine in real time.

3. The welding machine fault diagnosis system according to claim 2, characterized in that: The data preprocessing module comprises: A matrix construction unit is used to select welding machine data corresponding to a certain fault state within a certain time period from the database to form a high-dimensional data matrix; The dimension reduction unit is used to perform dimension reduction processing on the high-dimensional data matrix using the principal component analysis method PCA to obtain a one-dimensional data matrix to constitute a training data set corresponding to the fault state.

4. The welding machine fault diagnosis system according to claim 3, characterized in that: The real-time status monitoring module comprises: Real-time acquisition unit, used to collect welding machine data in real time; A preprocessing unit is used to preprocess the real-time collected welding machine data using the preprocessing method of the data preprocessing module to obtain a one-dimensional data matrix corresponding to the current welding machine data; The state monitoring unit is used to calculate the maximum likelihood similarity probability between the one-dimensional data matrix and each fault state model, and select the model with the largest probability value. The fault state corresponding to the model is the current welding machine fault state.

5. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to claim 1 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 1 are implemented.

Images