Equipment fault diagnosis method and system based on distributed data mining

Abstract

The invention discloses an equipment fault diagnosis method and system based on distributed data mining, and the method comprises the steps: obtaining the time sequence operation data of each piece of equipment, applying an excitation signal corresponding to each element in the operation data to the equipment, generating an excitation disturbance value of each element, and calculating the behavior response intensity of each element; the main frequency, the frequency band diffusivity and the phase response offset of each element are extracted, a response label of each element is generated, and a response label capable of reflecting possible faults of the equipment is screened out from all the response labels to serve as a fault response label; according to elements in the operation data corresponding to the fault response labels, the fault types of the devices and reasons causing related faults are determined, and fault diagnosis is carried out on each device in a distributed mode.

Description

Technical Field

[0001] This invention belongs to the field of equipment fault diagnosis technology, and more specifically, relates to a method and system for equipment fault diagnosis based on distributed data mining. Background Technology

[0002] In the current equipment fault diagnosis technology system, although sensor data-based analysis models, feature extraction algorithms, and machine learning methods are widely used, the following prominent technical problems and shortcomings still exist in multi-device, decentralized scenarios:

[0003] 1. Insufficient ability to model local dynamic behavior responses.

[0004] Existing methods often rely on statistics (such as mean, variance, kurtosis, etc.) or spectral energy distribution for feature extraction, neglecting the process of changes in the response behavior of equipment under disturbances.

[0005] The lack of identification and modeling mechanisms for key behavioral characteristics such as signal response phase and non-steady-state oscillation response makes it difficult to detect early and subtle faults.

[0006] 2. Heavily reliant on centralized data and tag-dependent structures

[0007] Mainstream models generally rely on centralized data collection and unified label training processes, such as traditional machine learning models (SVM, RF) or deep learning (CNN, LSTM) structures.

[0008] In multi-site, heterogeneous distributed deployments, data upload costs are high, labels are severely lacking, and privacy risks are significant, making it difficult to conduct effective training and deployment.

[0009] 3. Lack of a closed-loop mechanism for disturbance excitation and response.

[0010] Current diagnostic methods mostly employ passive sensing and non-disturbed guidance processes, meaning they can only observe the natural behavior of the device without stimulating potential responses, thus resulting in insufficient diagnostic sensitivity.

[0011] The lack of a closed-loop mechanism for "perturbation input - behavioral response - frequency phase characteristic evolution" makes it impossible to capture the trend of fault evolution. Summary of the Invention

[0012] To address the above technical problems, this invention proposes a device fault diagnosis method based on distributed data mining, comprising:

[0013] Acquire the timing operation data of each device, apply the excitation signal corresponding to each element in the operation data to the device, generate the excitation perturbation value of each element, calculate the behavioral response intensity of each element, and delete the elements whose behavioral response intensity is less than the preset intensity threshold.

[0014] Extract the main frequency, frequency band spread, and phase response offset of each element, and generate a response label for each element. Filter out the response labels that can reflect the possible faults in the device from all the response labels, and use them as fault response labels.

[0015] Based on the elements in the operational data corresponding to the fault response tags, the fault type of the device and the cause of the related fault are determined, wherein fault diagnosis is performed on each device in a distributed manner.

[0016] Furthermore, the excitation perturbation value for each element includes:

[0017]

[0018] Where, δ i (t) represents the excitation perturbation value of the i-th element at time t, A i Let ω be the perturbation amplitude of the i-th element. i Let be the perturbation frequency of the i-th element. Let be the perturbation phase of the i-th element.

[0019] Furthermore, calculating the behavioral response strength of each element includes:

[0020] R i (τ)=∫ t t+Δ [S i (t+τ)-S i (t)]·δ i (t)dt

[0021] Among them, R i (τ) represents the behavioral response intensity of the i-th element with a time lag window of τ, and Δ represents the change in time. i (t+τ) represents the value of the i-th element at time t+τ, S i (t) represents the value of the i-th element at time t.

[0022] Furthermore, from all response tags, those that indicate a potential device malfunction are selected, including:

[0023] Based on the main frequency, frequency band spread, and phase response offset of each element, an anomaly metric value is generated to characterize the equipment fault. The anomaly metric value is compared with a preset anomaly threshold, and the response metric value corresponding to the anomaly metric value that exceeds the preset anomaly threshold is used as the fault response metric.

[0024] Furthermore, based on the elements in the operational data corresponding to the fault response tags, the fault type of the equipment and the causes of the related faults are determined, including:

[0025] Each element corresponds to a fault type. The fault type of the device is identified by the elements in the operating data corresponding to the fault response label.

[0026] The cause of the related fault is an element in the running data corresponding to the fault response label.

[0027] Furthermore, it also includes: using the main frequency corresponding to the fault response label as the horizontal axis, the phase response offset as the vertical axis, and the frequency band diffusion as the pixel brightness to form a two-dimensional grayscale image corresponding to the fault response label.

[0028] Furthermore, before forming the two-dimensional grayscale image corresponding to the fault response label, the following steps are also included:

[0029] The main frequency is normalized according to the image width;

[0030] The phase response offset is normalized according to the image height;

[0031] The frequency band spread is normalized based on the pixel grayscale value.

[0032] Furthermore, it also includes: performing hash encoding processing on the two-dimensional grayscale image of each device to generate a hash code for the two-dimensional grayscale image.

[0033] Furthermore, it also includes: performing pairwise similarity comparisons on the hash codes of the corresponding two-dimensional grayscale images of each device, and clustering devices with similarity exceeding a preset similarity threshold into the same fault type.

[0034] This invention also proposes a device fault diagnosis system based on distributed data mining, comprising:

[0035] The excitation module is used to acquire the timing operation data of each device, apply excitation signals corresponding to each element in the operation data to the device, generate the excitation perturbation value of each element, calculate the behavioral response intensity of each element, and delete elements whose behavioral response intensity is less than a preset intensity threshold.

[0036] The fault response tag generation module is used to extract the main frequency, frequency band spread and phase response offset of each element, and generate a response tag for each element. From all the response tags, the response tags that can reflect the possible fault of the equipment are selected as the fault response tags.

[0037] The fault diagnosis module is used to determine the fault type of the device and the cause of the fault based on the elements in the operating data corresponding to the fault response label. Fault diagnosis is performed on each device in a distributed manner.

[0038] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0039] Through the above technical solutions, this invention can accurately detect equipment faults, thereby improving the operation and maintenance efficiency of maintenance personnel. Attached Figure Description

[0040] Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention;

[0041] Figure 2 This is a system structure diagram of Embodiment 2 of the present invention. Detailed Implementation

[0042] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0043] The method provided by this invention can be implemented in a terminal environment that may include one or more of the following components: a processor, a storage medium, and a display screen. The storage medium stores at least one instruction, which is loaded and executed by the processor to implement the method described in the following embodiments.

[0044] A processor may include one or more processing cores. The processor uses various interfaces and lines to connect various parts of the terminal, and performs various functions and processes data by running or executing instructions, programs, code sets or instruction sets stored in the storage medium, and by calling data stored in the storage medium.

[0045] Storage media can include random access memory (RAM) or read-only memory (ROM). Storage media can be used to store instructions, programs, code, code sets, or instructions.

[0046] The display screen is used to show the user interface of each application.

[0047] In addition, those skilled in the art will understand that the above-described structure of the terminal does not constitute a limitation on the terminal. The terminal may include more or fewer components, or combine certain components, or have different component arrangements. For example, the terminal may also include radio frequency circuits, input units, sensors, audio circuits, power supplies, and other components, which will not be described in detail here.

[0048] Example 1

[0049] like Figure 1 As shown, this embodiment proposes a device fault diagnosis method based on distributed data mining, including:

[0050] Step 101: Obtain the timing operation data of each device, apply the excitation signal corresponding to each element in the operation data to the device, generate the excitation perturbation value of each element, calculate the behavioral response intensity of each element, and delete the elements whose behavioral response intensity is less than the preset intensity threshold.

[0051] The purpose of setting the excitation signal in this embodiment is to inject an excitation signal into the device and observe how the device responds. Essentially, it is to expose the defects of the device. Suppose a bearing has a potential microcrack that is not visible during normal operation, but if it is gently tapped in the frequency domain (such as by injecting an excitation signal), the device's response (such as a vibration signal) will be nonlinearly amplified or abnormally phased in certain frequency bands.

[0052] Preferably, the timing operation data of the device in this embodiment includes multiple elements, wherein the elements can be: device temperature, device current, device load, device vibration, device speed, device pressure, etc. at each time point. This embodiment does not limit the specific device, and users can obtain the timing operation data of the corresponding device according to their own needs.

[0053] Specifically, through the equipment's PLC / frequency converter / controller, periodic micro-oscillations (excitation signals) are directly injected into the equipment's control signals. For example, adjusting the speed, voltage, and pressure setpoints generates excitation disturbance values ​​for each element, including:

[0054]

[0055] Where, δ i (t) represents the excitation perturbation value of the i-th element at time t, A i Let ω be the perturbation amplitude of the i-th element. i Let be the perturbation frequency of the i-th element. Let be the perturbation phase of the i-th element.

[0056] Specifically, calculating the behavioral response strength of each element includes:

[0057] R i (τ)=∫ t t+Δ [S i (t+τ)-S i (t)]·δ i (t)dt

[0058] Among them, R i (τ) represents the behavioral response intensity of the i-th element with a time lag window of τ, and Δ represents the change in time. i (t+τ) represents the value of the i-th element at time t+τ, S i (t) represents the value of the i-th element at time t.

[0059] Step 102: Extract the main frequency, frequency band spread and phase response offset of each element, and generate a response label for each element. Filter out the response labels that can reflect the possible faults of the device from all the response labels and use them as fault response labels.

[0060] This embodiment calculates the dominant frequency, frequency band spread, and phase response shift of each element in the following way. Specifically, the dominant frequency of each element is calculated as follows:

[0061]

[0062] Where, β i Let f be the dominant frequency of the i-th element, representing the significant frequency changes of the device under excitation, such as device temperature, device current, and device load. f is the frequency variable after performing a Fast Fourier Transform on the element. FFT is the Fast Fourier Transform.

[0063] Suppose we perform an FFT on a vibration signal sampled 1000 times per second (sampling rate 1000Hz): the resulting frequency coordinates f are an array: f = [0, 1, 2, ..., 499, 500] (unit: Hz).

[0064] The frequency band spread of each element is calculated as follows:

[0065]

[0066] Where, σ i E represents the frequency band spread of the i-th element. total E represents the total frequency domain energy. f For the frequency variable f to be higher than β i The sum of energy at all frequencies.

[0067] The phase response offset for each element is calculated as follows:

[0068] φ i =∠FFT(δ) i (t)·S i (t))

[0069] Where, φ i Let be the phase response offset of the i-th element, and ∠ be a complex number.

[0070] Specifically, the process of selecting response tags from all response tags that can indicate potential equipment failures includes: generating an anomaly metric for each response tag based on the main frequency, frequency band spread, and phase response offset of each element; comparing the anomaly metric with a preset anomaly threshold; and using response tags corresponding to anomaly metric values ​​that exceed the preset anomaly threshold as fault response tags.

[0071] Preferably, in this embodiment, the anomaly metric of the device is calculated in the following manner:

[0072] Ψ(L i )=α·D β +γ·Var(σ i )+λ·|Δφ i |

[0073] Among them, Ψ(L i The response label L generated for the i-th element i The anomaly measure, α is the deviation weight, D β Let be the deviation between the dominant frequency of the i-th element and the historical average dominant frequency, and γ be the weight of the variance, Var(σ) i ) represents the sliding window variance of the frequency band spread (e.g., within 5 seconds), λ represents the weight of the phase response offset, and Δφ i It represents the deviation between the phase response offset of the i-th element and the previous phase response offset of the i-th element.

[0074] Step 103: Determine the fault type of the device and the cause of the fault based on the elements in the operation data corresponding to the fault response label, wherein fault diagnosis is performed on each device in a distributed manner.

[0075] Specifically, based on the elements in the operational data corresponding to the fault response tags, the fault type of the equipment and the causes of the related faults are determined, including:

[0076] Each element corresponds to a fault type. The fault type of the device is identified by the elements in the operating data corresponding to the fault response label.

[0077] The cause of the related fault is an element in the running data corresponding to the fault response label.

[0078] Specifically, in this embodiment, the main frequency corresponding to the fault response label is used as the horizontal axis, the phase response offset is used as the vertical axis, and the frequency band diffusion is used as the pixel brightness to form a two-dimensional grayscale image corresponding to the fault response label.

[0079] Preferably, before forming the two-dimensional grayscale image corresponding to the fault response label, the process further includes:

[0080] The main frequency is normalized according to the image width;

[0081] The phase response offset is normalized according to the image height;

[0082] The frequency band spread is normalized based on the pixel grayscale value.

[0083] Preferably, to better describe the formation process of a two-dimensional grayscale image, this embodiment provides the following example to make the formation process of a two-dimensional grayscale image clearer, as shown below:

[0084] Normalizing the dominant frequency according to the image width specifically includes:

[0085] β i →x=round((β i -β min ) / (β max -β min )·W)

[0086] Where, β min The minimum value of the dominant frequency, β max The maximum value of the main frequency, W is the image width.

[0087] Normalizing the phase response offset according to image height specifically includes:

[0088] φ i →y=round((φ i +π) / (2π)·H)

[0089] Where H is the image height.

[0090] Normalizing the frequency band spread based on pixel grayscale values ​​specifically includes:

[0091] σ i →intensity=round(log(1+σ i ) / log(1+σ max )·255)

[0092] Where intensity is the pixel brightness (grayscale), σ max This represents the maximum value of the frequency band spread.

[0093] Specifically, in this embodiment, the two-dimensional grayscale image of each device is hash-encoded to generate a hash code for the two-dimensional grayscale image. The hash codes of the corresponding two-dimensional grayscale images of each device are compared pairwise for similarity, and devices with similarity exceeding a preset similarity threshold are clustered into the same fault type.

[0094] Example 2

[0095] like Figure 2 As shown, this embodiment proposes a device fault diagnosis system based on distributed data mining, including:

[0096] The excitation module is used to acquire the timing operation data of each device, apply excitation signals corresponding to each element in the operation data to the device, generate the excitation perturbation value of each element, calculate the behavioral response intensity of each element, and delete elements whose behavioral response intensity is less than a preset intensity threshold.

[0097] Specifically, the excitation perturbation value for each element includes:

[0098]

[0099] Where, δ i (t) represents the excitation perturbation value of the i-th element at time t, A i Let ω be the perturbation amplitude of the i-th element. i Let be the perturbation frequency of the i-th element. Let be the perturbation phase of the i-th element.

[0100] Specifically, calculating the behavioral response strength of each element includes:

[0101] R i (τ)=∫ t t+Δ [S i (t+τ)-S i (t)]·δ i (t)dt

[0102] Among them, R i (τ) represents the behavioral response intensity of the i-th element with a time lag window of τ, and Δ represents the change in time. i (t+τ) represents the value of the i-th element at time t+τ, S i (t) represents the value of the i-th element at time t.

[0103] The fault response tag generation module is used to extract the main frequency, frequency band spread and phase response offset of each element, and generate a response tag for each element. From all the response tags, the response tags that can reflect the possible fault of the equipment are selected as the fault response tags.

[0104] Specifically, the process of selecting response tags from all response tags that can indicate potential equipment failures includes: generating an anomaly metric for each response tag based on the main frequency, frequency band spread, and phase response offset of each element; comparing the anomaly metric with a preset anomaly threshold; and using response tags corresponding to anomaly metric values ​​that exceed the preset anomaly threshold as fault response tags.

[0105] The fault diagnosis module is used to determine the fault type of the device and the cause of the fault based on the elements in the operating data corresponding to the fault response label. Fault diagnosis is performed on each device in a distributed manner.

[0106] Specifically, based on the elements in the operational data corresponding to the fault response tags, the fault type of the equipment and the causes of the related faults are determined, including:

[0107] Each element corresponds to a fault type. The fault type of the device is identified by the elements in the operating data corresponding to the fault response label.

[0108] The cause of the related fault is an element in the running data corresponding to the fault response label.

[0109] Specifically, in this embodiment, the main frequency corresponding to the fault response label is used as the horizontal axis, the phase response offset is used as the vertical axis, and the frequency band diffusion is used as the pixel brightness to form a two-dimensional grayscale image corresponding to the fault response label.

[0110] Preferably, before forming the two-dimensional grayscale image corresponding to the fault response label, the process further includes:

[0111] The main frequency is normalized according to the image width;

[0112] The phase response offset is normalized according to the image height;

[0113] The frequency band spread is normalized based on the pixel grayscale value.

[0114] Specifically, in this embodiment, the two-dimensional grayscale image of each device is hash-encoded to generate a hash code for the two-dimensional grayscale image. The hash codes of the corresponding two-dimensional grayscale images of each device are compared pairwise for similarity, and devices with similarity exceeding a preset similarity threshold are clustered into the same fault type.

[0115] Example 3

[0116] This invention also proposes a storage medium storing multiple instructions for implementing the device fault diagnosis method based on distributed data mining.

[0117] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0118] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following method steps: Step 101, acquiring timing operation data of each device, applying an excitation signal corresponding to each element in the operation data to the device, generating an excitation perturbation value for each element, calculating the behavioral response intensity of each element, and deleting elements whose behavioral response intensity is less than a preset intensity threshold.

[0119] Specifically, the excitation perturbation value for each element includes:

[0120]

[0121] Where, δ i (t) represents the excitation perturbation value of the i-th element at time t, A i Let ω be the perturbation amplitude of the i-th element. i Let be the perturbation frequency of the i-th element. Let be the perturbation phase of the i-th element.

[0122] Specifically, calculating the behavioral response strength of each element includes:

[0123] R i (τ)=∫ t t+Δ [S i (t+t)-S i (t)]·δ i (t)dt

[0124] Among them, R i (τ) represents the behavioral response intensity of the i-th element with a time lag window of τ, and Δ represents the change in time. i (t+τ) represents the value of the i-th element at time t+τ, S i (t) represents the value of the i-th element at time t.

[0125] Step 102: Extract the main frequency, frequency band spread and phase response offset of each element, and generate a response label for each element. Filter out the response labels that can reflect the possible faults of the device from all the response labels and use them as fault response labels.

[0126] Specifically, the process of selecting response tags from all response tags that can indicate potential equipment failures includes: generating an anomaly metric for each response tag based on the main frequency, frequency band spread, and phase response offset of each element; comparing the anomaly metric with a preset anomaly threshold; and using response tags corresponding to anomaly metric values ​​that exceed the preset anomaly threshold as fault response tags.

[0127] Step 103: Determine the fault type of the device and the cause of the fault based on the elements in the operation data corresponding to the fault response label, wherein fault diagnosis is performed on each device in a distributed manner.

[0128] Specifically, based on the elements in the operational data corresponding to the fault response tags, the fault type of the equipment and the causes of the related faults are determined, including:

[0129] Each element corresponds to a fault type. The fault type of the device is identified by the elements in the operating data corresponding to the fault response label.

[0130] The cause of the related fault is an element in the running data corresponding to the fault response label.

[0131] Specifically, in this embodiment, the main frequency corresponding to the fault response label is used as the horizontal axis, the phase response offset is used as the vertical axis, and the frequency band diffusion is used as the pixel brightness to form a two-dimensional grayscale image corresponding to the fault response label.

[0132] Preferably, before forming the two-dimensional grayscale image corresponding to the fault response label, the process further includes:

[0133] The main frequency is normalized according to the image width;

[0134] The phase response offset is normalized according to the image height;

[0135] The frequency band spread is normalized based on the pixel grayscale value.

[0136] Specifically, in this embodiment, the two-dimensional grayscale image of each device is hash-encoded to generate a hash code for the two-dimensional grayscale image. The hash codes of the corresponding two-dimensional grayscale images of each device are compared pairwise for similarity, and devices with similarity exceeding a preset similarity threshold are clustered into the same fault type.

[0137] Example 4

[0138] This invention also proposes an electronic device, including a processor and a storage medium connected to the processor. The storage medium stores multiple instructions, which can be loaded and executed by the processor to enable the processor to perform the device fault diagnosis method based on distributed data mining.

[0139] Specifically, the electronic device in this embodiment can be a computer terminal, which may include one or more processors and a storage medium.

[0140] The storage medium can be used to store software programs and modules, such as the device fault diagnosis method based on distributed data mining in this embodiment of the invention. The corresponding program instructions / modules allow the processor to execute various functional applications and data processing by running the software programs and modules stored in the storage medium, thus realizing the aforementioned device fault diagnosis method based on distributed data mining. The storage medium may include high-speed random access storage media, and may also include non-volatile storage media, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state storage media. In some instances, the storage medium may further include storage media remotely configured relative to the processor, which can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0141] The processor can call the information and application stored in the storage medium through the transmission system to execute the following method steps: Step 101, acquire the timing operation data of each device, apply the excitation signal corresponding to each element in the operation data to the device, generate the excitation perturbation value of each element, calculate the behavior response intensity of each element, and delete the elements whose behavior response intensity is less than the preset intensity threshold.

[0142] Specifically, the excitation perturbation value for each element includes:

[0143]

[0144] Where, δ i (t) represents the excitation perturbation value of the i-th element at time t, A i Let ω be the perturbation amplitude of the i-th element. i Let be the perturbation frequency of the i-th element. Let be the perturbation phase of the i-th element.

[0145] Specifically, calculating the behavioral response strength of each element includes:

[0146] R i (τ)=∫ t t+Δ [S i (t+τ)-S i (t)]·δ i (t)dt

[0147] Among them, R i (τ) represents the behavioral response intensity of the i-th element with a time lag window of τ, and Δ represents the change in time. i (t+τ) represents the value of the i-th element at time t+τ, S i (t) represents the value of the i-th element at time t.

[0148] Step 102: Extract the main frequency, frequency band spread and phase response offset of each element, and generate a response label for each element. Filter out the response labels that can reflect the possible faults of the device from all the response labels and use them as fault response labels.

[0149] Specifically, the process of selecting response tags from all response tags that can indicate potential equipment failures includes: generating an anomaly metric for each response tag based on the main frequency, frequency band spread, and phase response offset of each element; comparing the anomaly metric with a preset anomaly threshold; and using response tags corresponding to anomaly metric values ​​that exceed the preset anomaly threshold as fault response tags.

[0150] Step 103: Determine the fault type of the device and the cause of the fault based on the elements in the operation data corresponding to the fault response label, wherein fault diagnosis is performed on each device in a distributed manner.

[0151] Specifically, based on the elements in the operational data corresponding to the fault response tags, the fault type of the equipment and the causes of the related faults are determined, including:

[0152] Each element corresponds to a fault type. The fault type of the device is identified by the elements in the operating data corresponding to the fault response label.

[0153] The cause of the related fault is an element in the running data corresponding to the fault response label.

[0154] Specifically, in this embodiment, the main frequency corresponding to the fault response label is used as the horizontal axis, the phase response offset is used as the vertical axis, and the frequency band diffusion is used as the pixel brightness to form a two-dimensional grayscale image corresponding to the fault response label.

[0155] Preferably, before forming the two-dimensional grayscale image corresponding to the fault response label, the process further includes:

[0156] The main frequency is normalized according to the image width;

[0157] The phase response offset is normalized according to the image height;

[0158] The frequency band spread is normalized based on the pixel grayscale value.

[0159] Specifically, in this embodiment, the two-dimensional grayscale image of each device is hash-encoded to generate a hash code for the two-dimensional grayscale image. The hash codes of the corresponding two-dimensional grayscale images of each device are compared pairwise for similarity, and devices with similarity exceeding a preset similarity threshold are clustered into the same fault type.

[0160] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0161] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0162] In the several embodiments provided by this invention, it should be understood that the disclosed technical content can be implemented in other ways. The system embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0163] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0164] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0165] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, optical disks, and other media capable of storing program code.

[0166] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A device failure diagnosis method based on distributed data mining, characterized by, The method comprises the following steps: obtaining time sequence operation data of each device, applying an excitation signal corresponding to each element in the operation data to the device, generating an excitation disturbance value of each element, and calculating the behavior response strength of each element, and deleting elements with a behavior response strength less than a preset strength threshold; extracting the main frequency, frequency band diffusion degree and phase response offset of each element, and generating a response label of each element, and screening a response label capable of reflecting a possible fault of the device from all response labels as a fault response label; determining the fault type of the device and the cause of the related fault according to the element in the operation data corresponding to the fault response label, wherein the fault diagnosis is performed on each device in a distributed manner.

2. The device fault diagnosis method based on distributed data mining according to claim 1, wherein, The method further comprises the following steps: wherein δ i (t) is the excitation disturbance value of the i-th element at time t, A i is the disturbance amplitude of the i-th element, ω i is the disturbance frequency of the i-th element, is the disturbance phase of the i-th element.

3. The device fault diagnosis method based on distributed data mining according to claim 2, wherein, calculating the behavior response strength of each element comprises: wherein R i S (t) is the value of the i-th element at time t. i S (t+τ) is the value of the i-th element at time t+τ. i S (t) is the value of the i-th element at time t.

4. The device fault diagnosis method based on distributed data mining according to claim 1, wherein, screening a response label capable of reflecting a possible fault of the device from all response labels comprises: generating an abnormality metric value of each response label for characterizing the abnormality of the device fault according to the main frequency, frequency band diffusion degree and phase response offset of each element, comparing the abnormality metric value with a preset abnormality threshold, and taking the response label corresponding to the abnormality metric value exceeding the preset abnormality threshold as the fault response label.

5. The device fault diagnosis method based on distributed data mining according to claim 1, wherein, determining the fault type of the device and the cause of the related fault according to the element in the operation data corresponding to the fault response label comprises: each element corresponds to a fault type, and the fault type of the device is identified by the element in the operation data corresponding to the fault response label; the cause of the related fault is the element in the operation data corresponding to the fault response label.

6. The device fault diagnosis method based on distributed data mining according to claim 1, wherein, The method further comprises the following steps: forming a two-dimensional gray scale image corresponding to the fault response label by taking the main frequency corresponding to the fault response label as the horizontal axis, the phase response offset as the vertical axis, and the frequency band diffusion degree as the pixel brightness.

7. The device fault diagnosis method based on distributed data mining according to claim 6, wherein, The method further comprises the following steps before forming the two-dimensional gray scale image corresponding to the fault response label: normalizing the main frequency according to the image width; normalizing the phase response offset according to the image height; normalizing the frequency band diffusion degree according to the pixel gray scale value.

8. The device fault diagnosis method based on distributed data mining according to claim 6, wherein, The method further comprises the following steps: hash encoding the two-dimensional gray scale image of each device to generate a hash code of the two-dimensional gray scale image.

9. The device fault diagnosis method based on distributed data mining according to claim 8, wherein, The method further comprises the following steps: comparing the hash codes of the two-dimensional gray scale images corresponding to each device in pairs according to the similarity, and clustering devices with a similarity exceeding a preset similarity threshold into the same fault type.

10. A distributed data mining-based device failure diagnosis system, characterized by, The method comprises the following steps: an excitation module for obtaining time sequence operation data of each device, applying an excitation signal corresponding to each element in the operation data to the device, generating an excitation disturbance value of each element, and calculating the behavior response strength of each element, and deleting elements with a behavior response strength less than a preset strength threshold; a fault response label generation module for extracting the main frequency, frequency band diffusion degree and phase response offset of each element, and generating a response label of each element, and screening a response label capable of reflecting a possible fault of the device from all response labels as a fault response label; A fault diagnosis module is configured to determine a fault type of the device and a cause of the related fault according to elements in the operation data corresponding to the fault response label, and the fault diagnosis is performed on each device in a distributed manner.

Images