Equipment fault rapid positioning method and system

Abstract

The invention relates to the technical field of equipment fault rapid positioning, and discloses an equipment fault rapid positioning method and system, and the method comprises the steps: 1, executing a data preprocessing strategy for a current data set, and recording the processed current data set as a first data set; 2, executing an anomaly detection strategy on the first data set, and judging whether the equipment is in a normal operation state or not; if the equipment has a fault, marking the equipment with the fault as marking equipment; the method comprises the following steps of: acquiring all equipment of which the current is monitored, numbering each equipment, and respectively recording the equipment as equipment 1, equipment 2,... equipment a; 3, executing a fault propagation analysis strategy, calculating the influence of the fault of the marked equipment on other equipment, and quantifying the influence of the marked equipment on other equipment into an influence value; and 4, executing a key fault source positioning strategy, and determining the dominant fault equipment. The fault response time is greatly shortened, and the accuracy and reliability of fault diagnosis are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of rapid equipment fault location, and in particular to a method and system for rapid equipment fault location. Background Art

[0002] Current status of related technologies in this field: With the deep integration of industrialization and informatization, equipment operation and maintenance management has become a key link in ensuring production efficiency and quality. Traditional equipment fault diagnosis usually relies on manual experience judgment or simple threshold alarms. This method often cannot locate the source of the fault in a timely and accurate manner, especially in complex equipment networks. The fault transmission effect is obvious, and the problem of a single device may cause a chain reaction, affecting the stable operation of the entire system.

[0003] Currently commonly used fault diagnosis technologies mainly include condition monitoring, predictive maintenance and expert systems. Among them, condition monitoring monitors the health status of equipment in real time by collecting equipment operating parameters. Predictive maintenance introduces data analysis based on condition monitoring to predict possible future faults. Expert systems use preset rule bases to infer and judge equipment abnormalities.

[0004] Deficiencies of existing technologies: Although existing fault diagnosis technologies have achieved certain results, they still have problems such as slow response speed, high misjudgment rate, and difficulty in tracing the root cause when faced with large-scale, highly complex equipment networks. Especially when multiple devices fail concurrently, it is difficult to quickly identify the key source of the fault, which extends the troubleshooting cycle, increases unnecessary downtime and economic losses.

[0005] Therefore, there is an urgent need for a method and system for quickly locating equipment faults to quickly and accurately locate the source of the fault, thereby improving the efficiency of fault inspection and maintenance. Summary of the invention

[0006] The present invention provides a method and system for quickly locating equipment faults, which facilitates solving the problems mentioned in the above background technology.

[0007] The present invention provides the following technical solution: a method for quickly locating equipment faults, comprising: For any device that needs to monitor current; Set monitoring intervals; The time when the device current monitoring starts is recorded as the start time. From the start time, a monitoring time is set every monitoring interval, and the device current is recorded at each monitoring time. The recorded currents are organized into a current data set in chronological order. ; By continuously monitoring the equipment current, real-time current data can be obtained, providing dynamic changes in the equipment's operating status. Data collected in chronological order can reflect the equipment's operating conditions at different time points and help capture potential fault signals.

[0008] Step 1: Execute a data preprocessing strategy on the current data set, and record the processed current data set as the first data set; Step 2: Execute an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state; If there is a fault in the equipment, the faulty equipment will be recorded as a marked equipment; Obtain all devices whose current is monitored, number each device, and record them as device 1, device 2, ... device a; Step 3: Execute the fault propagation analysis strategy, calculate the impact of the marked device fault on other devices, and quantify the impact of the marked device on other devices as an impact value; Step 4: Execute the key fault source location strategy to determine the device that causes the fault.

[0009] Preferably, executing a data preprocessing strategy on the current data set and recording the processed current data set as a first data set includes: For the current dataset ; calculate , the result is recorded as α; calculate , the result is recorded as β; For any element in the current data set I, execute the following formula; , the result is recorded as , i=1,2,…,n; Among them, The value corresponding to the execution of the data preprocessing strategy is recorded as .

[0010] Data preprocessing can effectively remove noise and ensure data accuracy. The current values ​​of different devices may have different ranges. Standardization can make the data of different devices comparable, which is helpful for subsequent analysis and comparison.

[0011] Preferably, executing an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state includes: Analysis using Long Short-Term Memory Networks:

[0012] ; ; ; , where 1≦t≦n; in, are the activation values ​​of the forget gate, input gate, and output gates, is the cell state, is the output status.

[0013] Through anomaly detection, it is possible to quickly identify whether the equipment has a fault or anomaly, especially when the equipment fails, its characteristics can be quickly captured. Automatic identification: Using long short-term memory (LSTM) to analyze the current sequence can automatically identify whether the equipment is in normal working condition without manual intervention, which improves the efficiency and accuracy of fault identification.

[0014] Preferably, executing an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state includes: Analysis using convolutional neural networks: , where W is the convolution kernel, b is the bias term, and Y is the output feature after convolution; Setting a fault threshold, which is used to determine whether the device is operating normally; Compare The relationship between Y and the fault threshold; like >fault threshold and Y>fault threshold, it is determined that the device has failed.

[0015] Preferably, the executing fault propagation analysis strategy to calculate the impact of the marked device fault on other devices includes: Constructing a fault propagation graph: Obtain the connection topology between device 1, device 2, ..., device a, and establish a device relationship matrix E;

[0016] in, Impact of a failure in device p on device q, where 1≦p≦a and 1≦q≦a.

[0017] By calculating the impact value, the propagation intensity of the fault between devices can be quantified, providing a basis for the subsequent location of the key fault source. Faulty equipment may affect the normal operation of other equipment. Understanding this impact helps to identify and locate the source of cascading faults.

[0018] By constructing a fault propagation map and device relationship matrix, we can intuitively understand the transmission path of the fault in the system. Quantifying the impact of the fault helps to assess the severity and scope of the fault, thereby providing guidance for maintenance and repair.

[0019] Preferably, the executing fault propagation analysis strategy to calculate the impact of the marked device failure on other devices includes: Get the number r of the marking device; Get the output state of the labeled device using the long short-term memory network analysis, recorded as ; For any device s other than the marked device; calculate , the result is recorded as ,in, is the fault impact value of device s affected by the fault of the marked device.

[0020] Preferably, the step of executing a key fault source location strategy to determine the device causing the main fault includes: Get the number r of the marking device; Get the output state of the labeled device using the long short-term memory network analysis, recorded as ; calculate , the result is recorded as the fault mark value; Obtain the fault mark values ​​of all devices, calculate the mean, and record the result as the fault mean; All devices whose fault mark values ​​are greater than the fault mean are obtained, and the devices are sorted from large to small according to the fault mark values. The sorting result is used as the priority of the device with the dominant current fault.

[0021] By calculating the fault tag value, the key equipment that is most likely to cause other equipment failures can be located. This helps to respond quickly and carry out targeted repairs. Optimize maintenance resources: Locating the dominant fault source can give priority to the most critical faulty equipment, thereby effectively utilizing maintenance resources and improving fault response efficiency.

[0022] A rapid equipment fault location system, comprising: Current sensing module: used to monitor the current of the device. A current sensing module is installed on each device. Data storage module: records the current data of the equipment at each monitoring moment and stores it; Data calculation module: executes a data preprocessing strategy on the current data set, and records the processed current data set as the first data set; Data judgment module: executes an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state; Fault analysis module: Executes fault propagation analysis strategy, calculates the impact of marked device failure on other devices, and quantifies the impact of marked device on other devices as impact value; executes key fault source location strategy to determine the device that dominates the fault.

[0023] The present invention has the following beneficial effects: 1. This method for quickly locating equipment faults achieves the effect of quickly and accurately locating the source of the fault because it adopts the equipment fault analysis model and fault propagation map establishment technology, greatly shortens the fault response time, and improves the accuracy and reliability of fault diagnosis.

[0024] 2. Thanks to the introduction of deep neural network technology, the accuracy of equipment status prediction has made a qualitative leap in the rapid location method of equipment faults. Especially in the face of complex dynamic environments, the model shows stronger adaptability and anti-interference performance, which helps to reduce the misdiagnosis rate and the risk of missed diagnosis;.

[0025] 3. The rapid fault location method of the equipment and the effective implementation of multimodal data fusion analysis make the fault recovery plan more targeted and effective, reduce the number of unplanned downtime caused by faults, and save a lot of maintenance resources and labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Schematic diagram of the method of the present invention.

[0027] Figure 2 Schematic diagram of the system of the present invention.

[0028] Figure 3 This is a schematic diagram of a method for determining whether a device is faulty according to the present invention.

[0029] Figure 4 Schematic diagram of the equipment method for obtaining the dominant current fault in this aspect. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] Embodiment 1, refer to Figure 1 , a method for quickly locating equipment faults, including: For any device that needs to monitor current; Set monitoring intervals; The time when the device current monitoring starts is recorded as the start time. From the start time, a monitoring time is set every monitoring interval, and the device current is recorded at each monitoring time. The recorded currents are organized into a current data set in chronological order. ; By setting monitoring intervals and recording current data regularly, the system can fully capture the operating status of the equipment. This time series data provides the basis for subsequent analysis and fault detection.

[0032] Step 1: Execute a data preprocessing strategy on the current data set, and record the processed current data set as the first data set; Step 2: Execute an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state; If there is a fault in the equipment, the faulty equipment will be recorded as a marked equipment; Obtain all devices whose current is monitored, number each device, and record them as device 1, device 2, ... device a; Step 3: Execute the fault propagation analysis strategy, calculate the impact of the marked device fault on other devices, and quantify the impact of the marked device on other devices as an impact value; Step 4: Execute the key fault source location strategy to determine the device that causes the fault.

[0033] Data preprocessing stage: clean and preprocess the real-time collected equipment operation data, eliminate noise interference, and extract effective feature parameters; Anomaly detection phase: Use machine learning algorithms to build a normal operation benchmark model for historical data, compare it with real-time data, and automatically identify abnormal indicators; Fault propagation analysis phase: Based on the interaction principle and connection topology between devices, a fault propagation map is constructed to simulate the impact range and intensity changes under different device fault states; Key fault source location stage: Combined with the fault propagation analysis results, quantitatively evaluate the impact weight of each device failure on the overall system stability, determine the leading fault devices and their ranking; eMaintenance suggestion generation stage: Based on the fault location results, targeted equipment maintenance strategies and troubleshooting guides are automatically generated to accelerate the fault repair process.

[0034] The step of executing a data preprocessing strategy on the current data set and recording the processed current data set as a first data set includes: For the current dataset ; calculate , the result is recorded as α; calculate , the result is recorded as β; For any element in the current data set I, execute the following formula; , the result is recorded as , i=1,2,…,n; Among them, The value corresponding to the execution of the data preprocessing strategy is recorded as .

[0035] Data preprocessing can remove noise, fill missing data, normalize data range, etc. Doing so can improve the accuracy and robustness of anomaly detection algorithms and reduce the probability of false positives.

[0036] The executing of the anomaly detection strategy on the first data set to determine whether the device is in a normal operating state includes: Analysis using Long Short-Term Memory Networks:

[0037] ; ; ; , where 1≦t≦n; in, are the activation values ​​of the forget gate, input gate, and output gates, is the cell state, is the output status.

[0038] LSTM can capture long-term and short-term dependencies in time series data and help identify potential abnormal patterns, especially for complex dynamic systems.

[0039] The performing of an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state includes: Analysis using convolutional neural networks: , where W is the convolution kernel, b is the bias term, and Y is the output feature after convolution; CNN is good at feature extraction and can extract useful local features from current data to effectively identify anomalies.

[0040] Setting a fault threshold, which is used to determine whether the device is operating normally; Compare The relationship between Y and the fault threshold; like >fault threshold and Y>fault threshold, it is determined that the device has failed.

[0041] In this embodiment, refer to Figure 3 , using pipelines for analysis using long short-term memory networks and analysis using convolutional neural networks.

[0042] In order to solve the challenges brought by complex and changeable fault modes to fault diagnosis, deep neural network technology is used to enhance the ability to predict equipment status. By integrating the long short-term memory (LSTM) network and the convolutional neural network (CNN), the time series characteristics of equipment operation data are fully mined to improve the accuracy of capturing abnormal behaviors. Implementation method: LSTM is used to capture the time dependency of data, and CNN is responsible for extracting spatial features. The combination of the two forms a deep hybrid neural network architecture, which significantly improves the learning ability and generalization performance of the model.

[0043] The execution of the fault propagation analysis strategy to calculate the impact of the marked device fault on other devices includes: Constructing a fault propagation graph: Obtain the connection topology between device 1, device 2, ..., device a, and establish a device relationship matrix E;

[0044] in, Impact of a failure in device p on device q, where 1≦p≦a and 1≦q≦a.

[0045] By constructing a fault propagation map, you can intuitively see the impact relationship between devices, which helps identify the fault transmission path. Systematic analysis: The device relationship matrix can help establish a system-level fault propagation model, making the fault impact analysis more systematic and structured.

[0046] The execution of the fault propagation analysis strategy to calculate the impact of the marked device fault on other devices includes: Get the number r of the marking device; Get the output state of the labeled device using the long short-term memory network analysis, recorded as ; For any device s other than the marked device; calculate , the result is recorded as ,in, is the fault impact value of device s affected by the fault of the marked device.

[0047] By calculating the fault impact value of each device, it helps to quantify the intensity of fault propagation between devices and helps to analyze the source of the fault more accurately.

[0048] The method of executing the key fault source location strategy and determining the device that causes the fault includes: Get the number r of the marking device; Get the output state of the labeled device using the long short-term memory network analysis, recorded as ; calculate , the result is recorded as the fault mark value; Obtain the fault mark values ​​of all devices, calculate the mean, and record the result as the fault mean; All devices whose fault mark values ​​are greater than the fault mean are obtained, and the devices are sorted from large to small according to the fault mark values. The sorting result is used as the priority of the device with the dominant current fault.

[0049] In this embodiment, refer to Figure 4 , showing a schematic diagram of the method for obtaining the dominant current fault.

[0050] By calculating and sorting the fault tag values, the main fault source equipment can be quickly identified and located. Such priority sorting helps maintenance personnel respond quickly and reduce fault downtime. The main fault equipment can be clearly identified, so that resources can be concentrated on repairing the problem equipment and minimizing system downtime. Decision support: This sorting mechanism can provide decision makers with a basis for locating the source of the fault and help formulate a more scientific equipment maintenance strategy.

[0051] Example 2, refer to Figure 2 , a rapid equipment fault location system, specifically: Current sensing module: used to monitor the current of the device. A current sensing module is installed on each device. The current sensing module installed on each device can monitor the device status in real time, forming a distributed monitoring network, thereby improving the precision and timeliness of data collection.

[0052] Data storage module: records the current data of the equipment at each monitoring moment and stores it; Continuous data storage enables retrospective analysis and pattern recognition, helping to better understand device behavior and performance trends.

[0053] Data calculation module: executes a data preprocessing strategy on the current data set, and records the processed current data set as the first data set; Centralize data preprocessing and computing tasks to ensure data consistency and processing efficiency.

[0054] Data judgment module: executes an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state; Provides real-time anomaly detection results to help quickly determine the operating status of the device.

[0055] Fault analysis module: Executes fault propagation analysis strategy, calculates the impact of marked device failure on other devices, and quantifies the impact of marked device on other devices as impact value; executes key fault source location strategy to determine the device that dominates the fault.

[0056] Comprehensively analyze fault propagation and location, provide strategic maintenance decision support, and optimize resource allocation.

[0057] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0058] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for quickly locating equipment faults, characterized in that: include: For any device that needs to monitor current; Set monitoring intervals; The time when the device current monitoring starts is recorded as the start time. From the start time, a monitoring time is set every monitoring interval, and the device current is recorded at each monitoring time. The recorded currents are organized into a current data set in chronological order. ; Step 1: Execute a data preprocessing strategy on the current data set, and record the processed current data set as the first data set; Step 2: Execute an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state; If there is a fault in the equipment, the faulty equipment will be recorded as a marked equipment; Obtain all devices whose current is monitored, number each device, and record them as device 1, device 2, ... device a; Step 3: Execute the fault propagation analysis strategy, calculate the impact of the marked device fault on other devices, and quantify the impact of the marked device on other devices as an impact value; Step 4: Execute the key fault source location strategy to determine the device that causes the fault.

2. The method for quickly locating equipment faults according to claim 1, characterized in that: The step of executing a data preprocessing strategy on the current data set and recording the processed current data set as a first data set includes: For the current dataset ; calculate , the result is recorded as α; calculate , the result is recorded as β; For any element in the current data set I, execute the following formula; , the result is recorded as , i=1,2,…,n; Among them, The value corresponding to the execution of the data preprocessing strategy is recorded as .

3. The method for quickly locating equipment faults according to claim 1, characterized in that: The executing of the anomaly detection strategy on the first data set to determine whether the device is in a normal operating state includes: Using LSTM network analysis: ； ； ； ； , where 1≦t≦n; in, are the activation values ​​of the forget gate, input gate, and output gates, is the cell state, is the output status.

4. The method for quickly locating equipment faults according to claim 3, characterized in that: The executing of the anomaly detection strategy on the first data set to determine whether the device is in a normal operating state includes: Analysis using convolutional neural networks: , where W is the convolution kernel, b is the bias term, and Y is the output feature after convolution; Setting a fault threshold, which is used to determine whether the device is operating normally; Compare The relationship between Y and the fault threshold; like >fault threshold and Y>fault threshold, it is determined that the device has failed.

5. The method for quickly locating equipment faults according to claim 1, characterized in that: The execution of the fault propagation analysis strategy to calculate the impact of the marked device fault on other devices includes: Constructing a fault propagation graph: Obtain the connection topology between device 1, device 2, ..., device a, and establish a device relationship matrix E; ； in, Impact of a failure in device p on device q, where 1≦p≦a and 1≦q≦a.

6. The method for quickly locating equipment faults according to claim 1, characterized in that: The execution of the fault propagation analysis strategy to calculate the impact of the marked device fault on other devices includes: Get the number r of the marking device; Get the output state of the labeled device using the long short-term memory network analysis, recorded as ; For any device s other than the marked device; calculate , the result is recorded as ,in, is the fault impact value of device s affected by the fault of the marked device.

7. The method for quickly locating equipment faults according to claim 1, characterized in that: The method of executing the key fault source location strategy and determining the device that causes the fault includes: calculate , the result is recorded as the fault mark value; Obtain the fault mark values ​​of all devices, calculate the mean, and record the result as the fault mean; All devices whose fault mark values ​​are greater than the fault mean are obtained, and the devices are sorted from large to small according to the fault mark values. The sorting result is used as the priority of the device with the dominant current fault.

8. A system for implementing the method for quickly locating equipment faults according to claim 1, characterized in that: include: Current sensing module: used to monitor the current of the device. A current sensing module is installed on each device. Data storage module: records the current data of the equipment at each monitoring moment and stores it; Data calculation module: executes a data preprocessing strategy on the current data set, and records the processed current data set as the first data set; Data judgment module: executes an anomaly detection strategy on the first data set to determine whether the device is in a normal operating state; Fault analysis module: executes fault propagation analysis strategy, calculates the impact of the marked device fault on other devices, and quantifies the impact of the marked device on other devices as an impact value; Execute key fault source location strategy to identify the equipment that is causing the fault.

Images