Fan safety fault detection method based on artificial intelligence

Abstract

The invention belongs to the technical field of fault detection, and particularly relates to a fan safety fault detection method based on artificial intelligence. The method comprises the following steps: firstly collecting an original signal sample, and then carrying out preprocessing such as denoising and filtering on the original signal sample; secondly, an improved artificial bee colony algorithm ABC is adopted to optimize feature mode decomposition FMD, an intrinsic mode function IMF is obtained by screening local extreme values, an optimization problem is constructed by taking minimization of an error function and maximization of information gain as targets, and a signal mode component is obtained through a series of operations; and finally, by utilizing KDTree considering information gain, dividing dimensions according to data features to construct a tree, incorporating signal component data into the tree, and traversing leaf nodes to realize fault judgment. Various technical means are integrated, the problems that an existing fan fault detection method is low in efficiency and poor in accuracy are effectively solved, the accuracy and efficiency of fan fault detection are improved, and safe operation of the fan is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault detection, and in particular relates to a fan safety fault detection method based on artificial intelligence. Background Art

[0002] Fans are widely used as common heat dissipation devices in industrial production and daily life. However, fans may have various safety faults during long-term operation, such as motor failure, blade damage, bearing wear, etc. These faults will not only affect the normal operation of the fan and reduce the heat dissipation effect, but may even cause safety accidents, such as equipment overheating and fire. Existing fan fault detection methods mainly rely on manual detection and simple sensor monitoring. This method is inefficient, the accuracy is difficult to guarantee, and the problem can often only be discovered after the fault occurs, and it is impossible to achieve early warning and accurate prediction of faults. Summary of the invention

[0003] In view of the technical problems existing in the above-mentioned background technology, the present invention proposes a fan safety fault detection method based on artificial intelligence.

[0004] In order to achieve the above object, the technical solution adopted by the present invention comprises the following steps:

[0005] S1, first collect the original signal samples;

[0006] S2, preprocessing the original signal;

[0007] S3. Use the improved artificial bee colony algorithm ABC to optimize the characteristic mode decomposition FMD to maximize the accuracy of the extracted fault feature information and minimize the error between the extracted signal component and the real fault feature signal. Extract the signal component of the fault information from the original signal. The specific implementation steps are as follows:

[0008] S31. First, the intrinsic mode function IMF is obtained by gradually screening the local maximum and minimum values ​​of the signal: Where x(t) is the original signal, IMF i (t) is the i-th intrinsic mode function, r N (t) is the residual signal;

[0009] S32, determine the optimization target as minimizing the error function: where y i (t) is the real fault signal corresponding to the modal component;

[0010] S33. Use information gain to measure the contribution of each modal component to the fault. Information gain is expressed as: IG(f iq )=H(y i )-H(y i|f iq ), where f q is the characteristic of the modal component, H(y i ) is the entropy of the fault, H(y i |f iq ) is a known feature f iq Conditional entropy of failure under conditions;

[0011] S34. Finally, a new optimization problem is constructed to minimize the error and maximize the information gain: Where N is the number of modal components, Q is the number of features corresponding to each modal component, and the improved ABC is used to optimize FMD, and a set of signal modal components is obtained through ABC optimization;

[0012] S4. Finally, KDTree based on information gain is used to calculate information gain according to data characteristics and select partition dimensions to build a tree. Signal component data is included in the tree and fault judgment is realized based on leaf nodes.

[0013] Preferably, the original signal samples in step S1 include vibration signals, temperature signals, current signals and audio signals.

[0014] Preferably, the operations of preprocessing the original signal in step S2 include denoising, filtering, standardization and smoothing operations.

[0015] Preferably, the specific operation of improving the ABC algorithm to achieve optimization in step S34 is:

[0016] S351, first initialize and randomly generate P bee positions, representing potential solutions, P j =(x j1 ,x j2 ,...,x jm ), where P j represents the position of the jth bee, m is the dimension of the solution; set the initial search range [l min ,l max ], where l min ,l max are the minimum and maximum bounds respectively, and for each solution P j Calculate its fitness value;

[0017] S352. In each round of iteration, the search range is dynamically adjusted according to the change in fitness, and the change in fitness is evaluated: in, are the current optimal fitness and the current fitness respectively;

[0018] S353, according to the fitness change, adjust the search range, if the current fitness If it is smaller than the previous generation, the search range will be narrowed: l new = l old ·(1-α range ), where α range is the ratio of narrowing the control range, which is 0.1. If the current fitness If it is larger than the previous generation, the search range will be narrowed: l new = l old ·(1+α range ), where l old , l new They are the search ranges before and after adjustment respectively;

[0019] S354, then the bees start searching, and each hired bee is within the new search range, and the search formula is: P new =P current +Φ(P best -P current ), where Φ is a random factor between [-1,1], controlling the step size of the search;

[0020] S355, after each search, update the fitness function and evaluate the quality of the current solution, and select the solution with the best fitness as the global optimal solution;

[0021] S356. Continue iterative optimization until the predetermined maximum number of iterations of 50 is reached and the iteration is terminated.

[0022] Preferably, in step S353, in order to avoid the search range being too small, a minimum threshold is set. Prevent the search scope from being narrowed too much, which may result in ineffective search of new solution spaces.

[0023] Preferably, the specific implementation of the fault judgment in step S4 is:

[0024] S41. First, select the optimal feature f according to the information gain opt To construct the root node of the tree;

[0025] S42. Select the median of the feature based on the value of the selected feature Divide the data into two parts, the left subtree corresponds to The right subtree corresponds to

[0026] S43, recursively perform the same operation on the left subtree and the right subtree, and select the feature with the largest information gain for partitioning;

[0027] S44. Start traversing from the root node of the tree. When traversing to a leaf node, the leaf node is marked as faulty or non-faulty, that is, the fault determination result of the output signal component.

[0028] Compared with the prior art, the advantages and positive effects of the present invention are that it integrates multiple original signals to comprehensively monitor the operating status of the fan and avoid the limitations of single signal monitoring. The improved ABC algorithm is used to optimize FMD to improve the accuracy of fault feature extraction. The improved algorithm can better balance the search capability and obtain better signal modal components. KDTree considering information gain is used for fault judgment to quickly and accurately determine the fault. The tree is constructed by selecting features through information gain, recursively dividing data, and efficiently mining fault features to achieve early warning and accurate prediction of fan faults, thereby ensuring safe and stable operation of the equipment. DETAILED DESCRIPTION

[0029] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described below in conjunction with embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0030] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways than those described herein. Therefore, the present invention is not limited to the specific embodiments of the following disclosure.

[0031] Embodiment, there are many problems with traditional fan fault detection methods. For example, manual detection and simple sensor monitoring are inefficient, inaccurate, and cannot provide early warning of faults. In order to solve these problems, the present invention proposes a fan safety fault detection method based on artificial intelligence, which aims to improve the accuracy and efficiency of fan fault detection, achieve early warning, and ensure the safe and stable operation of the fan.

[0032] The collected original signal samples include vibration signals, temperature signals, current signals and audio signals. The vibration signal can reflect the balance state of the fan blades, the wear of the bearings, etc. The temperature signal can indicate whether the motor is overheated, thereby indirectly reflecting the working state of the motor; the change of the current signal is related to the load of the motor, winding failure, etc.; the audio signal can capture whether there is abnormal noise when the fan is running, such as blade scraping, motor abnormal noise, etc. By comprehensively collecting these different types of signals, the operating status of the fan can be fully monitored from multiple angles to avoid the problem of missed detection or misjudgment that may be caused by single signal monitoring. In order to improve the signal quality and facilitate subsequent accurate analysis, preprocessing operations are implemented. This includes denoising, filtering, standardization and smoothing operations on the original signal.

[0033] In order to maximize the accuracy of fault feature information extraction and minimize errors, the improved artificial bee colony algorithm ABC is used to optimize the characteristic mode decomposition FMD, accurately screen the local maximum and minimum values ​​of the signal to obtain the intrinsic mode function IMF, decompose the complex signal into multiple meaningful components, and deeply explore the internal structure. The optimization goal is clear. By minimizing the error function, the extracted signal components are close to the real fault signal, which improves the accuracy of fault feature extraction. The information gain is used to measure the contribution of the modal component to the fault and screen out the key components. The improved ABC algorithm balances global and local searches to avoid falling into local optimality, thereby more effectively obtaining the optimal signal modal components. The specific implementation is: first, by gradually screening the local maximum and minimum values ​​of the signal, the intrinsic mode function IMF is obtained: Where x(t) is the original signal, IMF i (t) is the i-th intrinsic mode function, r N (t) is the residual signal; the optimization objective is to minimize the error function: where y i (t) is the real fault signal corresponding to the modal component; the information gain is used to measure the contribution of each modal component to the fault, and the information gain is expressed as: IG(f iq )=H(y i )-H(y i |f iq ), where f q is the characteristic of the modal component, H(y i ) is the entropy of the fault, H(y i |f iq ) is a known feature f iq The conditional entropy of the fault under the condition; finally, a new optimization problem is constructed to minimize the control error and maximize the information gain: Where N is the number of modal components, Q is the number of features corresponding to each modal component, and the improved ABC is used to optimize FMD, and a set of signal modal components are obtained through ABC optimization.

[0034] The improved ABC algorithm is optimized as follows: first, initialization is performed to randomly generate P bee positions, representing potential solutions, P j =(x j1 ,x j2 ,...,x jm ), where P j represents the position of the jth bee, m is the dimension of the solution; set the initial search range [l min ,l max ], where l min ,l max are the minimum and maximum bounds respectively, and for each solution P j Calculate its fitness value; In each round of iteration, the search range is dynamically adjusted according to the change in fitness, and the change in fitness is evaluated: in, are the current optimal fitness and the current fitness respectively; according to the change of fitness, adjust the search range. If the current fitness If it is smaller than the previous generation, the search range will be narrowed: l new = l old ·(1-α range ), where α range is the ratio of narrowing the control range, which is 0.1. If the current fitness If it is larger than the previous generation, the search range will be narrowed: l new = l old ·(1+α range ), where l old , l new are the search ranges before and after the adjustment respectively; then the bees start searching, and each hired bee is in the new search range, and the search formula is: P new =P current +Φ(P best -P current ), where Φ is a random factor between [-1,1], which controls the step size of the search; after each search, the fitness function is updated and the quality of the current solution is evaluated, and the solution with the best fitness is selected as the global optimal solution; the iterative optimization is continued until the predetermined maximum number of iterations of 50 is reached and the iteration is terminated.

[0035] Finally, in order to accurately judge the fan fault, KDTree based on information gain is used. The information gain is calculated according to the data characteristics to select the partition dimension to build the tree. The most discriminative features can be prioritized as tree nodes, making data classification more efficient and accurate. When processing fan fault data, by recursively selecting the feature partition with the largest information gain, the classification can be gradually refined and the fault features can be accurately mined. Fault judgment can be achieved by traversing from the root node to the leaf node, which can quickly give the judgment result of fault or non-fault, effectively improving the efficiency and accuracy of fault diagnosis. First, the optimal feature f is selected according to the information gain. opt To build the root node of the tree; according to the value of the selected feature, select the median of the feature Divide the data into two parts, the left subtree corresponds to The right subtree corresponds to The same operation is recursively performed on the left subtree and the right subtree, and the feature with the largest information gain is selected for division; finally, the traversal starts from the root node of the tree, and when the leaf node is traversed, the leaf node is marked as faulty or non-faulty, that is, the fault judgment result of the output signal component.

[0036] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A fan safety fault detection method based on artificial intelligence, characterized in that: The following steps are involved: S1, first collect the original signal samples; S2, preprocessing the original signal; S3. Use the improved artificial bee colony algorithm ABC to optimize the characteristic mode decomposition FMD to maximize the accuracy of the extracted fault feature information and minimize the error between the extracted signal component and the real fault feature signal. Extract the signal component of the fault information from the original signal. The specific implementation steps are as follows: S31. First, the intrinsic mode function IMF is obtained by gradually screening the local maximum and minimum values ​​of the signal: Where x(t) is the original signal, IMF i (t) is the i-th intrinsic mode function, r N (t) is the residual signal; S32, determine the optimization target as minimizing the error function: where y i (t) is the real fault signal corresponding to the modal component; S33. Use information gain to measure the contribution of each modal component to the fault. Information gain is expressed as: IG(f iq )=H(y i )-H(y i |f iq ), where f q is the characteristic of the modal component, H(y i ) is the entropy of the fault, H(y i |f iq ) is a known feature f iq Conditional entropy of failure under conditions; S34. Finally, a new optimization problem is constructed to minimize the error and maximize the information gain: Where N is the number of modal components, Q is the number of features corresponding to each modal component, and the improved ABC is used to optimize FMD, and a set of signal modal components is obtained through ABC optimization; S4. Finally, KDTree based on information gain is used to calculate information gain according to data characteristics and select partition dimensions to build a tree. Signal component data is included in the tree and fault judgment is realized based on leaf nodes.

2. The method for detecting fan safety faults based on artificial intelligence according to claim 1, characterized in that: The original signal samples in step S1 include a vibration signal, a temperature signal, a current signal and an audio signal.

3. The fan safety fault detection method based on artificial intelligence according to claim 1, characterized in that: The operations of preprocessing the original signal in step S2 include denoising, filtering, standardization and smoothing operations.

4. The method for detecting fan safety faults based on artificial intelligence according to claim 1, characterized in that: The specific operation of improving the ABC algorithm to achieve optimization in step S34 is: S351, first initialize and randomly generate P bee positions, representing potential solutions, P j =(x j1 ,x j2 ,...,x jm ), where P j represents the position of the jth bee, m is the dimension of the solution; set the initial search range [l min ,l max ], where l min ,l max are the minimum and maximum bounds respectively, and for each solution P j Calculate its fitness value; S352. In each round of iteration, the search range is dynamically adjusted according to the change in fitness, and the change in fitness is evaluated: in, are the current optimal fitness and the current fitness respectively; S353, according to the fitness change, adjust the search range, if the current fitness If it is smaller than the previous generation, the search range will be narrowed: l new = l old ·(1-α range ), where α range is the ratio of narrowing the control range, which is 0.

1. If the current fitness If it is larger than the previous generation, the search range will be narrowed: l new = l old ·(1+α range ), where l old , l new They are the search ranges before and after adjustment respectively; S354, then the bees start searching, and each hired bee is within the new search range, and the search formula is: P new =P current +Φ(P best -P current ), where Φ is a random factor between [-1,1], controlling the step size of the search; S355, after each search, update the fitness function and evaluate the quality of the current solution, and select the solution with the best fitness as the global optimal solution; S356. Continue iterative optimization until the predetermined maximum number of iterations of 50 is reached and the iteration is terminated.

5. The method for detecting fan safety faults based on artificial intelligence according to claim 4, characterized in that: In step S353, in order to avoid the search range being too small, a minimum threshold is set. Prevent the search scope from being narrowed too much, which may result in ineffective search of new solution spaces.

6. The method for detecting fan safety faults based on artificial intelligence according to claim 1, characterized in that: The specific implementation of the fault judgment in step S4 is: S41. First, select the optimal feature f according to the information gain opt To construct the root node of the tree; S42. Select the median of the feature based on the value of the selected feature Divide the data into two parts, the left subtree corresponds to The right subtree corresponds to S43, recursively perform the same operation on the left subtree and the right subtree, and select the feature with the largest information gain for partitioning; S44. Start traversing from the root node of the tree. When traversing to a leaf node, the leaf node is marked as faulty or non-faulty, that is, the fault determination result of the output signal component.

Images