Method for intelligently detecting FAS debugging fault point

Abstract

The invention discloses a method for intelligently detecting an FAS debugging fault point, and the method specifically comprises the following steps: collecting real-time operation data of FAS system end equipment, and carrying out the data preprocessing; fault features are taken, and a fault diagnosis model is established; inputting the real-time data into the fault diagnosis model, performing fault detection and classification, and positioning a fault point; according to the method, the real-time data flow of the FAS end equipment and the historical fault database are collected to collect fault signals, the fault features are extracted, the fault diagnosis model is established, and when the fault diagnosis model detects that one FAS system end equipment has a fault, the fault diagnosis is performed on the FAS system end equipment. The fault diagnosis model displays the fault type of the FAS system end equipment, and displays the position of the faulted FAS system end equipment in the FAS system BIM model, so that the position of a fault point is visually presented, the fault type and position can be accurately diagnosed, and the fault diagnosis precision and efficiency are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of FAS debugging fault diagnosis, and specifically provides a method for intelligently detecting FAS debugging fault points. Background Art

[0002] As a fast and efficient public transportation means, the subway plays an important role in modern cities. It can not only meet people's travel needs but also relieve urban traffic congestion. The subway FAS, namely the fire automatic alarm system, is an important part of the urban rail transit system. Subways are generally located in indoor underground spaces with a high density of people and equipment. Once a disaster occurs, evacuation and rescue are very difficult. In addition, a large number of electromechanical devices and lines such as operating trains and environmental control systems are concentrated, which are also prone to causing fires. Therefore, the importance of the subway FAS as a safety barrier for urban rail operation is self-evident.

[0003] However, the subway integrated joint commissioning FAS system has many end devices and interfaces. Relying on experience for judgment may lead to a long diagnosis time and low diagnosis efficiency. Basic monitoring devices often have difficulty accurately diagnosing the specific type and location of faults, especially in the stage of complex faults or early faults, so that manual inspection is still required after early warning. Therefore, the fault diagnosis of the existing technology is not only inefficient but also inaccurate. Therefore, it is necessary to design a method for intelligently detecting FAS debugging fault points to improve the above problems. Summary of the Invention

[0004] To solve the above problems, the present invention provides a method for intelligently detecting FAS debugging fault points, which can accurately diagnose the fault type and location, and greatly improve the accuracy and efficiency of fault diagnosis.

[0005] The present invention provides a method for intelligently detecting FAS debugging fault points, specifically including the following steps:

[0006] S1. Collect the real-time operation data of the end devices of the FAS system and perform data preprocessing;

[0007] S2. Extract fault features and establish a fault diagnosis model;

[0008] S3. Input the real-time data into the fault diagnosis model, perform fault detection and classification, and locate the fault point;

[0009] S4. Centralize the alarm, visually display the fault information, and recommend the optimal linkage processing plan.

[0010] As a preferred embodiment of the present invention, the data preprocessing is specifically data cleaning. By means of data preprocessing, the quality of the collected real-time operation data is ensured. High-precision sensors are installed on each key terminal device of the FAS system for real-time collection of operation data such as current, voltage, and signal status. The sensors on the terminal devices are connected to the data acquisition unit of the FAS system by wired or wireless means, and the data acquisition unit is used to receive the sensor data from each terminal device.

[0011] As a preferred embodiment of the present invention, the specific steps of data cleaning are as follows: Import data from different sources into a unified data environment for preliminary exploratory analysis to understand the structure, type, distribution, and potential problems of the data. According to the characteristics of the real-time operation data of the terminal devices of the FAS system and the subsequent analysis requirements, clarify the objectives and requirements of data cleaning. Detect the missing values existing in the data set, and select appropriate processing methods according to the data characteristics and analysis requirements, such as deleting the missing values or filling the missing values. Use statistical methods to identify the outliers in the data, and decide whether to remove, correct, or retain the outliers according to the actual situation. Solve the problems of data conflicts and inconsistencies to ensure the consistency of data formats, units, naming, etc. Identify and delete duplicate data by comparing the similarity of records or unique identifiers to maintain the uniqueness of the data set, avoid redundancy and confusion, convert the data into a form suitable for analysis, eliminate the influence of feature dimensions, and improve the model convergence speed. Process problems such as missing values, outliers, and duplicate data one by one according to the above specific methods, perform data consistency checks and corrections, and perform data conversion and standardization.

[0012] As a preferred embodiment of the present invention, after each step of data cleaning, data quality inspection is carried out to ensure that the cleaning operation does not introduce new problems and improves the overall quality of the data. The cleaned data is stored in the database of the FAS system, and the data is continuously monitored.

[0013] As a preferred embodiment of the present invention, the real-time data stream of the FAS terminal device and the historical fault database are collected to collect fault signals. Time-domain analysis, frequency-domain analysis, and time-frequency domain analysis processing are performed on the fault signals. Key features are selected from the signal processing results, and the autoencoder is used to reduce the dimension of the key features to extract potential features. The importance of the features is calculated using random forest or XGBoost, and redundant features are removed for extracting fault features. The fault features are input into the AI fault diagnosis engine to establish a fault diagnosis model.

[0014] As a preferred embodiment of the present invention, the AI fault diagnosis engine includes a deep learning model training module and a fault identification module. The deep learning model training module is used to train a deep learning algorithm using a training data set, learn the relationship between fault features and fault types, and obtain a trained deep learning model. The fault identification module is used to analyze the collected fault features according to the trained deep learning model, identify the fault features of the FAS system, and determine the fault type.

[0015] As a preferred embodiment of the present invention, the data acquisition unit inputs the real-time operation data of the FAS system terminal device into the fault diagnosis model. The fault diagnosis model further includes a BIM model of the FAS system, and the FAS system terminal device is displayed in the BIM model of the FAS system. When the fault diagnosis model detects a fault in a certain FAS system terminal device, the fault diagnosis model displays the fault type of the FAS system terminal device and displays the location of the faulty FAS system terminal device in the BIM model of the FAS system, visually presenting the location of the fault point.

[0016] As a preferred embodiment of the present invention, the specific method for establishing the BIM model of the FAS system includes the following steps:

[0017] Draw a building BIM model according to the construction drawings of the railway building professional.

[0018] Based on the building BIM model, draw the BIM model of the FAS system according to the design drawings of the railway FAS professional.

[0019] Establish a data storage structure, establish data tables and databases, and digitalize the data of the FAS system terminal device according to the design drawings of the railway FAS professional.

[0020] As a preferred embodiment of the present invention, an alarm signal is generated based on the fault type and the location of the fault point. Fault handling solutions are stored in the fault diagnosis model. The corresponding fault handling solutions are sent to relevant personnel in real time through various methods such as text messages, emails, and APP push through the fault diagnosis model to ensure that the faults are handled in a timely manner.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0022] The present invention collects real-time data streams of FAS end devices and fault signals from the historical fault database, extracts fault features, establishes a fault diagnosis model, preprocesses the collected real-time operation data of the FAS system end devices to ensure data conversion and standardization, improves the overall quality of the data, then inputs the real-time data into the fault diagnosis model for fault detection and classification. When the fault diagnosis model detects a fault in a certain FAS system end device, the fault diagnosis model displays the fault type of the FAS system end device and shows the location of the faulty FAS system end device in the FAS system BIM model, intuitively presenting the location of the fault point. Furthermore, since the fault diagnosis model stores fault handling solutions, it reads the alarm signal through the fault diagnosis model and sends the corresponding fault handling solutions to relevant personnel in real time through various methods such as text messages, emails, and APP push, ensuring that the fault is processed in a timely manner, accurately diagnosing the fault type and location, and greatly improving the accuracy and efficiency of fault diagnosis. Detailed implementation mode

[0023] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other implementation manners obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope protected by the present invention.

[0024] Embodiment: This embodiment provides a method for intelligently detecting FAS commissioning fault points, which specifically includes the following steps:

[0025] Step 1: Collect the real-time operation data of the FAS system end devices and perform data preprocessing;

[0026] The FAS system end devices include fire detectors, manual alarm devices, alarm and indication devices, fire fighting linkage devices, and other auxiliary devices, etc. The fire detectors include smoke detectors, heat detectors, aspirating smoke detection systems, heat optical fiber hosts, and heat optical fibers, etc.; the manual alarm devices include manual alarm buttons, fire hydrant buttons, etc.; the alarm and indication devices include audible and visual alarms, release indicators, etc.; the fire fighting linkage devices include fire telephone hosts, fire wall-mounted telephones, and fire telephone jacks, fire door monitoring systems; the other auxiliary devices include input modules and output modules, end-of-pipe test devices, etc.; the collection includes analog signals such as vibration signals, temperature, pressure, current, voltage, etc.

[0027] Data preprocessing is specifically data cleaning. Data preprocessing ensures the quality of real-time operation data collected. High-precision sensors are installed on each key terminal device of the FAS system to collect real-time operation data such as current, voltage, and signal status. The sensors on the terminal devices are connected to the data acquisition unit of the FAS system through wired or wireless means. The data acquisition unit is used to receive sensor data from each terminal device.

[0028] The specific steps of data cleaning are: importing data from different sources into a unified data environment, conducting preliminary exploration and analysis, understanding the structure, type, distribution and potential problems of the data, clarifying the goals and requirements of data cleaning according to the real-time operation data characteristics of the terminal equipment of the FAS system and the subsequent analysis requirements, and selecting appropriate processing methods according to data characteristics and analysis requirements by detecting missing values ​​in the data set, such as deleting missing values ​​(when the number of missing values ​​is small and has little impact on the overall data) or filling missing values ​​(common methods include mean filling, median filling, mode filling and interpolation, etc.), using statistical methods (such as Z-score, IQR method, etc.) to identify missing values. The outliers in the data should be removed, corrected or retained according to the actual situation, and data conflicts and inconsistencies should be resolved to ensure consistency in data format, unit, naming, etc. Duplicate data should be identified and deleted by comparing the similarity of records or unique identifiers to maintain the uniqueness of the data set, avoid redundancy and confusion, and convert the data into a form suitable for analysis, such as categorical variable encoding (unique hot encoding, label encoding) and feature scaling (standardization, normalization), eliminate the influence of feature dimensions, and improve the convergence speed of the model. According to the above specific methods, problems such as missing values, outliers, and duplicate data should be handled one by one, and data consistency checks and corrections should be carried out, and data conversion and standardization should be performed;

[0029] After each step of data cleaning, a data quality check is performed to ensure that the cleaning operation does not introduce new problems and improves the overall quality of the data. The cleaned data is stored in the database of the FAS system and the data is continuously monitored.

[0030] Step 2: Extract fault features and establish a fault diagnosis model;

[0031] Collect the real-time data stream of FAS terminal equipment and the historical fault database to collect fault signals, perform time domain analysis, frequency domain analysis and time-frequency domain analysis on the fault signals, select key features from the signal processing results, use autoencoders to reduce the dimension of key features, extract potential features, use random forests or XGBoost to calculate feature importance, eliminate redundant features, use them to extract fault features, and input the fault features into the AI ​​fault diagnosis engine to establish a fault diagnosis model;

[0032] Time-domain analysis is used to calculate the mean value, variance, peak-to-peak value, form factor, pulse width, etc.; frequency-domain analysis is used to extract spectral features (main frequency, harmonic energy ratio) using FFT; time-frequency domain analysis is used to extract time-frequency energy distribution features using wavelet transform;

[0033] The AI fault diagnosis engine includes a deep learning model training module and a fault identification module. The deep learning model training module is used to train a deep learning algorithm using a training data set, learn the relationship between fault features and fault types, and obtain a trained deep learning model; the fault identification module is used to analyze the collected fault features according to the trained deep learning model, identify the fault features of the FAS system, and determine the fault type.

[0034] Step 3: Input the real-time data into the fault diagnosis model, perform fault detection and classification, and locate the fault point;

[0035] The data acquisition unit inputs the real-time operation data of the FAS system end devices into the fault diagnosis model. The fault diagnosis model also includes the FAS system BIM model, and the FAS system end devices are displayed in the FAS system BIM model. When the fault diagnosis model detects a fault in a certain FAS system end device, the fault diagnosis model displays the fault type of the FAS system end device and displays the location of the faulty FAS system end device in the FAS system BIM model, visually presenting the location of the fault point;

[0036] The specific method for establishing the FAS system BIM model includes the following steps:

[0037] Draw the building BIM model according to the railway building construction drawings. For the lightweight of the BIM model and to improve the loading speed, the building BIM model only needs to draw structural columns, structural beams, structural walls, structural slabs, building columns, building walls, doors and windows, and does not require detailed models such as internal reinforcement and decoration. However, with the upgrade of BIM model lightweight technology and hardware technology, the BIM model can load more models of relevant specialties;

[0038] Based on the building BIM model, draw the FAS system BIM model according to the railway FAS professional design drawings. The FAS professional equipment model needs to reach the 2-level geometric expression accuracy specified in Article 4.3.5 of the "GBT51301-2018 Building Information Model Design Delivery Standard", which meets the geometric expression accuracy requirements for rough identification such as space occupancy and main color, and the equipment coding needs to comply with the IFD standard;

[0039] Establish a data storage structure, establish data tables and databases, and digitalize the FAS system end device data according to the railway FAS professional design drawings.

[0040] Step 4: Centralize alarms, visually display fault information, and recommend the optimal linkage processing solution;

[0041] Generate an alarm signal based on the fault type and the location of the fault point. Fault handling solutions are stored in the fault diagnosis model. Read the alarm signal through the fault diagnosis model and send the corresponding fault handling solution to relevant personnel in real time through various methods such as text messages, emails, and APP push to ensure that the fault is handled in a timely manner.

[0042] The present invention collects real-time data streams and historical fault databases of FAS terminal devices to collect fault signals, extracts fault features, and establishes a fault diagnosis model. After preprocessing the real-time operation data of the FAS system terminal devices collected, it ensures data conversion and standardization, improves the overall quality of the data, and then inputs the real-time data into the fault diagnosis model for fault detection and classification. When the fault diagnosis model detects a fault in a certain FAS system terminal device, the fault diagnosis model displays the fault type of the FAS system terminal device and shows the location of the faulty FAS system terminal device in the FAS system BIM model, visually presenting the location of the fault point. Then, since the fault handling solutions are stored in the fault diagnosis model, read the alarm signal through the fault diagnosis model and send the corresponding fault handling solution to relevant personnel in real time through various methods such as text messages, emails, and APP push to ensure that the fault is handled in a timely manner. It can accurately diagnose the fault type and location, greatly improving the accuracy and efficiency of fault diagnosis.

[0043] All technical features in this embodiment can be freely combined according to actual needs.

[0044] The above embodiments are preferred implementation solutions of the present invention. In addition, the present invention can also be implemented in other ways. Any obvious replacement without departing from the concept of the technical solution of the present invention is within the protection scope of the present invention.

Claims

1. A method for intelligently detecting FAS debugging fault points, characterized in that: The specific steps include: S1. Collect the real-time operation data of the terminal equipment of the FAS system and perform data preprocessing; S2, extract fault features and establish a fault diagnosis model; S3, input the real-time data into the fault diagnosis model, perform fault detection and classification, and locate the fault point; S4: Centralized alarms, visual display of fault information, and recommendation of the best linkage processing solution.

2. A method for intelligently detecting FAS debugging fault points according to claim 1, characterized in that: Data preprocessing is specifically data cleaning. Data preprocessing ensures the quality of the collected real-time operation data. High-precision sensors are installed on each key terminal device of the FAS system to collect real-time operation data such as current, voltage, and signal status. The sensors on the terminal devices are connected to the data acquisition unit of the FAS system via wired or wireless means. The data acquisition unit is used to receive sensor data from each terminal device.

3. The method for intelligently detecting FAS debugging fault points according to claim 2, characterized in that: The specific steps of data cleaning are as follows: import data from different sources into a unified data environment, conduct preliminary exploration and analysis, understand the structure, type, distribution and potential problems of the data, clarify the goals and requirements of data cleaning according to the real-time operation data characteristics of the terminal equipment of the FAS system and the subsequent analysis requirements, detect the missing values ​​in the data set, and select appropriate processing methods according to the data characteristics and analysis requirements, such as deleting missing values ​​or filling missing values, using statistical methods to identify outliers in the data, and deciding whether to remove, correct or retain outliers based on actual conditions, resolving data conflicts and inconsistencies, ensuring consistency in data format, unit, naming, etc., identifying and deleting duplicate data by comparing the similarity of records or unique identifiers, maintaining the uniqueness of the data set, avoiding redundancy and confusion, converting data into a form suitable for analysis, eliminating the influence of characteristic dimensions, and improving the convergence speed of the model, and handling missing values, outliers, duplicate data and other issues one by one according to the above specific methods, conducting data consistency checks and corrections, and performing data conversion and standardization.

4. The method for intelligently detecting FAS debugging fault points according to claim 3, characterized in that: After each step of data cleaning, a data quality check is performed to ensure that the cleaning operation does not introduce new problems and improves the overall quality of the data. The cleaned data is stored in the database of the FAS system and the data is continuously monitored.

5. The method for intelligently detecting FAS debugging fault points according to claim 1, characterized in that: Collect the real-time data stream of FAS terminal equipment and the historical fault database to collect fault signals, perform time domain analysis, frequency domain analysis and time-frequency domain analysis on the fault signals, select key features from the signal processing results, use autoencoders to reduce the dimension of key features, extract potential features, use random forests or XGBoost to calculate feature importance, eliminate redundant features, and use them to extract fault features. The fault features are input into the AI ​​fault diagnosis engine to establish a fault diagnosis model.

6. The method for intelligently detecting FAS debugging fault points according to claim 1, characterized in that: The AI ​​fault diagnosis engine includes a deep learning model training module and a fault identification module. The deep learning model training module is used to train a deep learning algorithm using a training data set, learn the relationship between fault characteristics and fault types, and obtain a trained deep learning model; The fault identification module is used to analyze the collected fault features according to the trained deep learning model, identify the fault features of the FAS system, and determine the fault type.

7. The method for intelligently detecting FAS debugging fault points according to claim 1, characterized in that: The data acquisition unit inputs the real-time operation data of the FAS system terminal equipment into the fault diagnosis model. The fault diagnosis model also includes a FAS system BIM model. The FAS system terminal equipment is displayed in the FAS system BIM model. When the fault diagnosis model detects a fault in a FAS system terminal equipment, the fault diagnosis model displays the fault type of the FAS system terminal equipment and displays the location of the faulty FAS system terminal equipment in the FAS system BIM model, visually presenting the location of the fault point.

8. The method for intelligently detecting FAS debugging fault points according to claim 7, characterized in that: The specific method for establishing the FAS system BIM model includes the following steps: Draw the building BIM model according to the professional construction drawings of railway building construction; Based on the architectural BIM model and according to the railway FAS professional design drawings, draw the FAS system BIM model; Establish data storage structure, data tables and databases, and digitize FAS system terminal equipment according to railway FAS professional design drawings.

9. The method for intelligently detecting FAS debugging fault points according to claim 1, characterized in that: An alarm signal is generated based on the fault type and fault point location. The fault diagnosis model stores the fault handling plan. The alarm signal is read through the fault diagnosis model, and the corresponding fault handling plan is sent to relevant personnel in real time through SMS, email, APP push and other methods to ensure that the fault is handled in a timely manner.