Multi-point distributed high-precision industrial equipment operation monitoring method and system

Abstract

The invention discloses a multi-point distributed high-precision industrial equipment operation monitoring method and system, and relates to the technical field related to industrial equipment monitoring, and the method comprises the steps: independently collecting the operation parameter data of equipment through a plurality of sensor nodes; transmitting the operation parameter data to a central control unit in real time for fusion processing, and generating an equipment operation state feature set; performing fault prediction and health assessment on the equipment operation state feature set to obtain an equipment health score; and dynamically triggering an alarm signal according to the equipment health score, determining a target sensor node, and generating a target maintenance instruction. The technical problems that in the prior art, the equipment operation monitoring efficiency is low, the overall condition of the equipment is difficult to master comprehensively, a flexible monitoring and early warning mechanism is lacked, the fault position of the equipment cannot be accurately positioned and quickly responded and processed, and industrial equipment cannot efficiently and stably operate are solved. The technical effect of improving the accuracy and stability of equipment fault positioning is achieved.

Description

Technical Field

[0001] This application relates to the technical field of industrial equipment monitoring, and particularly to a multi-point distributed high-precision industrial equipment operation monitoring method and system. Background Art

[0002] In modern industrial production, the stable operation of equipment is crucial for ensuring production efficiency, product quality, and the economic benefits of enterprises. With the continuous improvement of industrial automation, the structure and functions of industrial equipment have become increasingly complex, and the monitoring and maintenance of its operating status face unprecedented challenges. Traditional equipment monitoring methods often rely on manual inspections and simple single-point monitoring devices, which are not only inefficient but also difficult to comprehensively and timely grasp the overall operating conditions of the equipment. Once a fault occurs in a critical part of the equipment, it may lead to production interruptions and cause huge losses. For example, in a large data center, the stable operation of the air conditioning system is directly related to the heat dissipation effect of the servers and the secure storage of data. If the temperature, humidity, and wind speed at the air outlets or inlets of the air conditioners are abnormal, it may cause the servers to overheat, resulting in data loss or equipment damage. Moreover, the optimal operating parameters of industrial equipment vary in different working environments and operating stages. The previous alarm mechanism with fixed thresholds cannot adapt to this dynamic change, easily resulting in false alarms or missed alarms, making it difficult to collect the operating parameters of critical parts of the equipment in real time and comprehensively, accurately evaluate the health status of the equipment, and timely detect and handle faults, thus affecting the efficient and stable operation of industrial equipment.

[0003] Therefore, in the current related technologies, there are technical problems such as low efficiency in equipment operation monitoring, difficulty in comprehensively grasping the overall condition of the equipment, lack of a flexible monitoring and warning mechanism, which further leads to the inability to accurately locate the fault position of the equipment and quickly respond to processing, resulting in the inefficient and unstable operation of industrial equipment. Summary of the Invention

[0004] By providing a multi-point distributed high-precision industrial equipment operation monitoring method and system, this application solves the technical problems in the prior art, such as low efficiency in equipment operation monitoring, difficulty in comprehensively grasping the overall condition of the equipment, lack of a flexible monitoring and warning mechanism, which further leads to the inability to accurately locate the fault position of the equipment and quickly respond to processing, resulting in the inefficient and unstable operation of industrial equipment, and achieves the technical effect of improving the accuracy and stability of equipment fault location.

[0005] The present application provides a multi-point distributed high-precision industrial equipment operation monitoring method, and the method includes: independently collecting operation parameter data of the equipment through multiple sensor nodes; transmitting the operation parameter data to a central control unit in real time for fusion processing to generate an equipment operation status feature set, wherein the equipment operation status feature set is obtained by calling operation timestamps to perform time series alignment and validity denoising on the operation parameter data, extracting effective parameter features, and performing stage comparison and fusion; using an LSTM model to perform fault prediction on the equipment operation status feature set, constructing a multi-dimensional health index weight matrix for health assessment, and obtaining an equipment health score; dynamically triggering an alarm signal according to the equipment health score, determining a target sensor node according to the alarm signal, and generating a target maintenance instruction.

[0006] The present application also provides a multi-point distributed high-precision industrial equipment operation monitoring system, including: an operation parameter data acquisition module for independently collecting operation parameter data of the equipment through multiple sensor nodes; an equipment operation status feature set generation module for transmitting the operation parameter data to a central control unit in real time for fusion processing to generate an equipment operation status feature set; an equipment health score acquisition module for performing health assessment based on the equipment operation status feature set to obtain an equipment health score; a target maintenance instruction generation module for dynamically triggering an alarm signal according to the equipment health score, determining a target sensor node according to the alarm signal, and generating a target maintenance instruction.

[0007] It is intended to solve the technical problems existing in the prior art, such as low efficiency of equipment operation monitoring, difficulty in comprehensively grasping the overall condition of the equipment, lack of a flexible monitoring and early warning mechanism, which in turn leads to the inability to accurately locate and quickly respond to the equipment fault location, resulting in the inefficient and unstable operation of industrial equipment, by means of a multi-point distributed high-precision industrial equipment operation monitoring method and system proposed in the present application. The method includes: independently collecting operation parameter data of the equipment through multiple sensor nodes; transmitting the operation parameter data to a central control unit in real time for fusion processing to generate an equipment operation status feature set, wherein the equipment operation status feature set is obtained by calling operation timestamps to perform time series alignment and validity denoising on the operation parameter data, extracting effective parameter features, and performing stage comparison and fusion; using an LSTM model to perform fault prediction on the equipment operation status feature set, constructing a multi-dimensional health index weight matrix for health assessment, and obtaining an equipment health score; dynamically triggering an alarm signal according to the equipment health score, determining a target sensor node according to the alarm signal, and generating a target maintenance instruction. The technical effect of improving the accuracy and stability of equipment fault location is achieved. Description of the Drawings

[0008] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments of the present disclosure will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the operations described above or below do not necessarily need to be performed precisely in sequence. On the contrary, according to the needs, various steps can be performed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or more steps can be removed from these processes.

[0009] Figure 1 It is a schematic flowchart of a multi-point distributed high-precision industrial equipment operation monitoring method provided by an embodiment of the present application.

[0010] Figure 2 It is a schematic structural diagram of a multi-point distributed high-precision industrial equipment operation monitoring system provided by an embodiment of the present application.

[0011] Explanation of reference numerals: Operation parameter data acquisition module 10, Equipment operation status feature set generation module 20, Equipment health score acquisition module 30, Target maintenance instruction generation module 40. Detailed implementation manners

[0012] The above description is only an overview of the technical solutions of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically gives the detailed implementation manners of the present application.

[0013] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limitations of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0014] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "first" and "second" are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application.

[0015] An embodiment of the present application provides a multi-point distributed high-precision industrial equipment operation monitoring method, as Figure 1 shown, the method includes: Step S100, independently collect the operation parameter data of the equipment through multiple sensor nodes.

[0016] Preferably, various types of sensors (such as temperature and humidity sensors, wind speed sensors, current and voltage sensors, and vibration sensors, etc.) are installed at different key parts of the equipment (such as the air outlet, return air outlet, and air inlet of the air conditioner, etc.). Each sensor serves as an independent node and respectively collects the specific operation parameters of the corresponding part of the equipment in real time. Specifically, the temperature and humidity sensor is used to measure the temperature and humidity of the environment around the key part of the equipment in real time. For example, when installed at the air outlet of the air conditioner, it can master the temperature and humidity of the outgoing air, which helps to judge the cooling or heating effect and dehumidification ability of the air conditioner; when installed at the return air outlet, it can understand the temperature and humidity state of the indoor air when it returns to the air conditioner, and assist in evaluating the overall operation efficiency of the air conditioner and its ability to adjust the environment. The wind speed sensor is installed at the air outlet and air inlet, and can measure the speed of air flow. By monitoring the wind speed, the working state of the air conditioner fan and whether the air duct is unobstructed can be judged. If the wind speed at the air outlet suddenly decreases, it may mean that the air duct is blocked or the fan fails; the current and voltage sensor is used to monitor the current and voltage values during the operation of the equipment. For air conditioning equipment, stable current and voltage are the basis for its normal operation. When the current increases abnormally or the voltage fluctuates greatly, it may indicate that there are electrical faults in the equipment, such as motor short circuit, circuit aging, etc.; the vibration sensor is used to detect the vibration situation during the operation of the equipment. Abnormal vibration may indicate that the internal components of the equipment are loose, worn, or unbalanced. For example, abnormal vibration of the air conditioner compressor may mean that there are problems with the internal parts of the compressor.

[0017] Preferably, a plurality of sensors of different types are used to monitor the entire industrial equipment based on a distributed control and response mechanism. Among them, the sensors and related control units at each monitoring point can independently complete partial tasks, such as data acquisition, preliminary processing, and simple judgment. Even if the central control system fails, each monitoring point can still continue to work, ensuring the reliability and stability of the system. When an abnormal data is detected at a certain monitoring point, it can immediately respond without waiting for the instruction of the central control system. For example, when the temperature and humidity sensor at a certain air outlet detects an abnormal increase in temperature, this monitoring point can immediately trigger the local warning device and at the same time quickly transmit the abnormal information to the central control system. And each monitoring point is independent of each other and will not be affected by the failure or abnormality of other monitoring points in terms of its own data acquisition function. For example, even if the sensor at the air inlet fails, the sensors at the air outlet and the return air outlet can still normally collect data, so as to comprehensively, real-time, and accurately obtain the operation status information of the equipment.

[0018] Further, step S100 further includes step S110 of traversing the multi-point distributed high-precision industrial equipment for operation analysis to determine multiple key operation parts; step S120 of arranging the multiple sensor nodes according to the multiple key operation parts, traversing the multiple sensor nodes for configuration to determine independent data acquisition modules; and step S130 of performing operation acquisition on the industrial equipment through the independent data acquisition modules to obtain the operation parameter data.

[0019] Preferably, for the operation analysis of the multi-point distributed high-precision industrial equipment, it may include understanding the working principle of the equipment, referring to past failure records, mining historical operation data, etc., to identify the parts in the equipment that have a key impact on its overall operation status and performance, that is, to determine the key operation parts. Then, appropriate sensor nodes are installed at these parts. For example, vibration sensors are installed at the vibration-sensitive parts of the equipment, and temperature sensors are installed where temperature needs to be monitored, etc. The type and quantity of sensors should be determined according to the characteristics of the key operation parts and the monitoring requirements to ensure that the operation parameters of the equipment can be comprehensively and accurately obtained. Then, each installed sensor node is configured in detail, including setting parameters such as the measurement range, sampling frequency, and accuracy of the sensor to ensure that the sensor can collect data as expected. For example, for a temperature sensor, an appropriate measurement range needs to be set to adapt to the temperature changes that may occur during the operation of the equipment; for a vibration sensor, an appropriate sampling frequency should be set according to the vibration characteristics of the equipment to capture key vibration information.

[0020] Preferably, each sensor node is further determined as an independent data acquisition module. Specifically, the independent data acquisition module includes a data cache unit and a communication protocol conversion unit. Among them, the data cache unit is used to temporarily store the data collected by the sensor to prevent data loss, especially in the case of communication delay or interruption; the communication protocol conversion unit is responsible for converting the data collected by the sensor according to a specific communication protocol so that it can be transmitted in a local area network or other communication networks. For example, if the sensor outputs an analog signal, the communication protocol conversion unit needs to convert it into a digital signal and encapsulate it according to the specified protocol format. Finally, the independent data acquisition module collects the real-time data of the key operating parts of the industrial equipment according to the pre-configured parameters and acquisition frequency, and finally obtains various operating parameter data of the industrial equipment, such as temperature, pressure, vibration, current, voltage, etc., for subsequent health assessment and fault diagnosis of the equipment. For example, by analyzing the temperature data, it can be judged whether the equipment has overheating problems, and by analyzing the vibration data, it can be detected whether there are loose or worn parts in the equipment.

[0021] Step S200, transmitting the operating parameter data to the central control unit in real time for fusion processing to generate a device operating state feature set. Among them, the device operating state feature set is obtained by calling the operation timestamp to perform time series alignment and validity denoising on the operating parameter data, and extracting effective parameter features for stage comparison and fusion.

[0022] Preferably, the collected operating parameter data is transmitted to the central control unit in real time through the industrial Ethernet local area network IEEE 802.3. Among them, IEEE 802.3 is the standard protocol of Ethernet. The industrial Ethernet local area network is a network technology widely used in industrial environments. Based on the Ethernet standard and optimized for the needs of industrial applications, it stipulates the communication rules of the data link layer and the physical layer (such as defining the format of data frames, transmission rate, signal coding method, etc.), ensuring that data can be transmitted stably, quickly and accurately in the industrial environment, and having characteristics such as high reliability, real-time performance and anti-interference. After each sensor node collects the operating parameter data, it will encapsulate the data into an Ethernet data frame in a specific format according to the regulations of the IEEE 802.3 protocol, which includes information such as the source address (the address of the sensor node), the destination address (the address of the central control unit) and the data content, and then transmit it in the form of an electrical signal or an optical signal through the physical medium of the industrial Ethernet local area network (such as twisted pair, optical fiber, etc.). Since the industrial Ethernet local area network has a high transmission rate (such as common 100Mbps or 1000Mbps) and low latency, it can ensure that the operating parameter data is quickly and timely transmitted to the central control unit after collection, meeting the real-time requirements.

[0023] Preferably, the central control unit receives Ethernet data frames from the industrial Ethernet local area network through its network interface. Specifically, first, the data frames are parsed according to the IEEE 802.3 protocol to extract the operating parameter data therein. Since the data formats sent by different types of sensor nodes may be different, the central control unit needs to correctly parse and identify the data of each type of sensor according to the pre-agreed protocol. For example, for the data of temperature and humidity sensors, the data is converted into actual temperature and humidity values according to a specific coding method. For the data of current and voltage sensors, corresponding decoding and range conversion are also performed to obtain accurate current and voltage values; then, the parsed operating parameter data is subjected to fusion processing, comprehensively considering the correlation between different types of data. For example, the heat exchange situation inside the device is analyzed by combining temperature and humidity data with wind speed data. By analyzing a large amount of historical data and real-time data, a model of the device operating state is established, so as to more comprehensively and accurately evaluate the current operating state of the device and generate a device operating state feature set.

[0024] Preferably, the device operating state feature set is obtained by retrieving the operation timestamp to perform time series alignment and validity denoising on the operating parameter data, extracting effective parameter features for stage comparison and fusion. Specifically, the operation timestamp records the acquisition time of the operating parameter data. Retrieving this timestamp can arrange and align the operating parameter data from different sources with inconsistent times due to network latency or acquisition frequency differences in chronological order, so that the data accurately reflects the change of the device operating state over time and has consistency and accuracy in the time dimension; the collected operating parameter data may contain noise and redundant information. Algorithms such as Kalman filtering are used to perform validity denoising on the data to remove interference factors and retain the data that truly reflects the device operating state; then, effective parameter features such as temperature and humidity change rate, instantaneous air volume fluctuation value, and current and voltage harmonic components are extracted from the denoised data to more accurately reflect the key information of the device operation; finally, according to the different data transmission bandwidths corresponding to the device operation stages (start-up period, steady state period, load fluctuation period, etc.), the effective parameter features are compared with the historical operation data, the differences and correlations between the current data and the historical data in each stage are analyzed, and then fused to form a device operating state feature set for comprehensively evaluating the device operating state.

[0025] Preferably, the device operation status feature set mainly includes the temperature and humidity change curve, the dynamic calculated value of the air volume, and the current and voltage fluctuation range. Specifically, the central control unit, based on the real-time data received from the temperature and humidity sensors, takes time as the horizontal axis and temperature and humidity as the vertical axes respectively to draw the temperature and humidity change curve, and records and updates the temperature and humidity data at each time point to form a continuous curve. By observing the trend, slope, and fluctuation of the temperature and humidity change curve, the temperature and humidity status and change trend of the device at different times can be intuitively understood. For example, the rising or falling trend of the curve can reflect whether the device is heating up or cooling down, and the fluctuation amplitude of the curve can reflect the stability of the temperature and humidity. If the temperature and humidity change curve shows abnormal fluctuations or trends beyond the normal range, it may indicate that there are faults in the device or the influence of environmental factors. Based on the wind speed data collected by the wind speed sensors at different positions of the device and the geometric parameters of the device air duct (such as air duct cross-sectional area, length, etc.), the central control unit uses the principles of fluid mechanics to dynamically calculate the air volume. For example, according to the average wind speed measured by the wind speed sensor and combined with the cross-sectional area of the air duct, the volume of air passing through the air duct per unit time, that is, the air volume, can be calculated. At the same time, considering factors such as air duct resistance and fan speed change that may exist during the operation of the device, the calculation result is corrected and optimized. By calculating the dynamic calculated value of the air volume in real time, the operation status of the device ventilation system can be timely grasped, and it can be judged whether the air duct is unobstructed and whether the fan is working properly. If the dynamic calculated value of the air volume shows obvious abnormalities, such as sudden increase or decrease, it means that there are problems with the device ventilation system.

[0026] Preferably, the central control unit conducts statistical analysis on the real-time data collected by the current and voltage sensors to determine the fluctuation range of the current and voltage. For example, the current and voltage values at each moment are recorded, and statistical parameters such as the maximum value, minimum value, and average value within a certain period of time (such as one minute or one hour) are calculated. By analyzing these statistical parameters, the fluctuation range of the current and voltage can be obtained. Under normal circumstances, the current and voltage of the device should fluctuate within a certain range. If the fluctuation range exceeds the normal threshold, it may indicate that there are electrical faults in the device, such as unstable power supply, line short circuit, or device overload, etc. By fusing and processing the device operation parameter data, a device operation status feature set is generated, which provides a data basis for the device status monitoring and fault diagnosis, helps to timely understand the operation status of the device, accurately locate device faults, and take corresponding measures to ensure the safe and stable operation of the device.

[0027] Further, step S200 further includes step S210 of transmitting the operation parameter data to the central control unit in real time to extract an operation simulation signal; step S220 of performing communication conversion on the operation simulation signal to obtain a standard digital signal; step S230 of activating the local area network to extract edge computing nodes, performing data verification on the standard digital signal according to the edge computing nodes, analyzing the operation impact according to the verification result, and setting the device operation priority; and step S240 of performing bandwidth allocation on the multiple sensor nodes according to the device operation priority to obtain multiple data transmission bandwidths.

[0028] Preferably, the operation parameter data is transmitted to the central control unit in real time through an industrial Ethernet local area network to extract an analog signal that can reflect the operation state of the device. For example, the electrical signal collected by a current-voltage sensor is processed through certain amplification, filtering, etc. to obtain an analog signal that can represent the change in the current and voltage of the device, more intuitively reflecting the change in the electrical state during device operation. Then, the analog signal is converted by a communication protocol conversion unit, that is, according to a specific communication protocol and conversion algorithm, the analog signal is converted into a standard digital signal. For example, a continuously changing analog voltage signal is converted into discrete digital values and follows certain coding rules and data formats, such as common binary coding, so that the computer can recognize and process it.

[0029] Preferably, the local area network is activated and edge computing nodes are extracted. Among them, edge computing nodes are usually computing devices close to the data source or device, with certain computing and data processing capabilities, used to share part of the computing tasks of the central control unit, improving the efficiency and real-time performance of data processing. Then, data verification is performed on the standard digital signal according to the edge computing nodes to ensure the accuracy and integrity of the data, checking whether errors or losses occur during data transmission. For example, algorithms such as cyclic redundancy check (CRC) can be used to calculate the data in the digital signal to generate a check code, and then the check code of the received data is compared with the check code generated by the sending end. If they are the same, it means the data is basically complete and correct; if they are different, it means the data may have problems. Then, according to the result of the data verification, analyze whether there are abnormalities in the device operation parameter data and the possible impacts of these abnormalities on device operation. For example, if it is found that the temperature data collected by a certain sensor exceeds the normal range and the data is verified to be correct, then it is necessary to analyze the possible impacts of this abnormal temperature on the various components of the device, such as whether it will cause a decline in device performance and whether it will accelerate component aging.

[0030] Preferably, considering the results of the operation impact analysis and the importance of the equipment, etc., set the operation priority for the equipment. For those equipment that are crucial to the entire production process and may cause serious consequences in case of failure, give a higher priority; while for some non-critical equipment, if there are some less serious abnormalities in their operation parameters, their priority can be appropriately reduced to ensure the normal operation of critical equipment first under the condition of limited resources. Finally, according to the set equipment operation priority, allocate different data transmission bandwidths to multiple sensor nodes connected to the equipment. For equipment with a high operation priority, its corresponding sensor nodes will be allocated more network bandwidth to ensure that the data collected by these sensors can be transmitted to the central control unit in a timely and accurate manner for processing, so as to better monitor and control the operation status of high-priority equipment; while for equipment with a lower priority, the bandwidth allocated to its sensor nodes is relatively less, but it can also ensure the basic data transmission requirements. For example, through devices such as network switches, different bandwidth quotas can be allocated to the sensor nodes of equipment with different priorities according to pre-set rules. For example, the sensor nodes of high-priority equipment may obtain a bandwidth of 100 Mbps, while the sensor nodes of low-priority equipment may only have a bandwidth of 10 Mbps; and use the MQTT protocol to establish a lightweight communication link within the local area network to provide different quality of service (QoS) levels. Among them, the QoS level of the data packet is set to at least one transmission guarantee, that is, in the case of unstable network, each data packet will be transmitted at least once to ensure that the data will not be lost. If the data packet is lost or the confirmation message is not received during the transmission process, the sender will re-transmit the data packet until the confirmation message from the receiver is received, so as to ensure the stable operation of industrial equipment.

[0031] Further, step S200 further includes step S250, retrieving the operation timestamp, aligning the operation parameter data in time series according to the operation timestamp to obtain aligned operation parameter data; step S260, performing effective denoising based on the aligned operation parameter data and extracting effective parameter features; step S270, comparing the effective parameter features with the historical operation data of the equipment in stages according to the multiple data transmission bandwidths, and performing fusion according to the comparison results to obtain the equipment operation status feature set.

[0032] Preferably, the running timestamp is a time mark recorded for each running parameter data, which accurately records the moment of data acquisition. By retrieving the running timestamp, the positions of each data point on the time axis are determined, and then the running parameter data is aligned in time sequence to eliminate the differences in the running parameter data caused by network latency or different sensor acquisition frequencies, that is, all the running parameter data is arranged according to its corresponding running timestamp so that they accurately correspond in the time series, forming an ordered, temporally continuous and accurate data set, namely the aligned running parameter data; the Kalman filtering algorithm is used to denoise the aligned running parameter data, that is, an algorithm that optimally estimates the system state by continuously updating the estimated value using the system state equation and the observation equation, effectively filtering out the noise in the data and retaining the true signal characteristics, and extracting the characteristic parameters from the data that can more accurately reflect the running state of the equipment, such as the temperature and humidity change rate, the instantaneous fluctuation value of the air volume, and the harmonic components of the current and voltage. Among them, the temperature and humidity change rate can reflect the dynamic changes of the environment around the equipment; the instantaneous fluctuation value of the air volume helps to understand the stability of the equipment ventilation system; the harmonic components of the current and voltage can reveal potential problems in the equipment electrical system. For example, too high harmonic content may indicate power quality problems in the equipment or failures of some components.

[0033] Preferably, according to the multiple data transmission bandwidths allocated previously, the extracted effective parameter features are compared with the historical running data of the equipment in stages, that is, the current effective parameter features are compared one by one with the historical data in different past time periods, and the differences between the current running state of the equipment and the previous normal running state or failure state are analyzed. Then, according to the comparison results, fusion is carried out, comprehensively considering the relationship between the current effective parameter features and the historical data, and the relevant information is integrated. Finally, a set of equipment running state features is generated, which includes a series of characteristic parameters that can comprehensively and accurately describe the current running state of the equipment, such as the temperature and humidity change curve, the dynamically calculated air volume value, and the current and voltage fluctuation ranges, etc., providing an important basis for the state monitoring, fault diagnosis and predictive maintenance of the equipment, helping to timely discover potential problems of the equipment and take corresponding measures to ensure the stable operation of industrial equipment.

[0034] Further, step S270 further includes step S271, retrieving the historical running data of the equipment, dividing the running stages based on the historical running data of the equipment, and determining multiple running stage information; step S272, matching the multiple running stage information according to the multiple data transmission bandwidths to generate a matching result, and performing cluster analysis on the effective parameter features and the historical running data of the equipment according to the matching result to generate a staged parameter distribution map; step S273, adding the staged parameter distribution map to the comparison result.

[0035] Preferably, various data generated during the past operation of the device are obtained from the historical data storage record of the device, including the operation parameters of the device at different times and under different working conditions, such as temperature, pressure, current, voltage, etc. Based on the historical operation data of the device, according to the characteristics and laws of the device operation, the operation process of the device is divided into different stages, mainly including the startup period (the process from the startup of the device to reaching the normal operation state), the steady state period (the stage where the device continuously operates in a stable operation state), and the load fluctuation period (the stage where the operation parameters of the device fluctuate due to load changes during the operation process), and the relevant information of each operation stage is determined, such as the time range of each stage, the typical characteristics of the device operation parameters within this stage, etc. For example, during the startup period, the current of the device may have a large peak value and the temperature gradually rises; during the steady state period, each operation parameter is relatively stable; during the load fluctuation period, some parameters (such as current, pressure) will fluctuate with the change of the load.

[0036] Preferably, according to the allocated multiple data transmission bandwidths, they are matched with the information of each operation stage. For example, for the situation where key parameter data may need to be transmitted more quickly and accurately during the startup period, a higher data transmission bandwidth is allocated to the relevant sensor nodes; while during the steady state period, the data transmission requirement is relatively low, and the bandwidth allocation can be appropriately adjusted. The generated matching result records the corresponding relationship between each operation stage and the corresponding data transmission bandwidth. According to the matching result, clustering analysis is respectively carried out on the effective parameter characteristics (such as the change rate of temperature and humidity, the instantaneous fluctuation value of air volume, and the harmonic components of current and voltage, etc.) and the historical operation data of the device in each operation stage. By calculating the similarity between the data, the similar data are grouped into one category, and then the distribution law and pattern of the device operation parameter data within each operation stage are discovered, and a stage-based parameter distribution map is generated. For example, during the steady state period, the normal fluctuation range and typical value distribution of the device operation parameters can be determined through clustering analysis; during the load fluctuation period, the change law of the parameters under different load conditions can be analyzed.

[0037] Preferably, the sliding window algorithm is used to update the baseline parameters in real time. Among them, the sliding window is a time window with a fixed size or a variable size. As time goes by, the window continuously slides on the data sequence, and the window length is dynamically adjusted according to the device operation stage. For example, during the startup period, since the device operation parameters change relatively fast, the window length may be shorter to capture the parameter changes more timely; during the steady state period, the window length can be appropriately extended. The current baseline parameters are calculated through the data within the sliding window to reflect the normal operation state of the device in the current operation stage. Finally, the generated stage-based parameter distribution map and the updated baseline parameters are added to the comparison result of the previous effective parameter characteristics and the historical operation data of the device to more accurately evaluate the device operation state, discover potential faults, and ensure the stable operation of the device.

[0038] Step S300: Use the LSTM model to perform fault prediction on the device operation status feature set, construct a multi-dimensional health index weight matrix for health assessment, and obtain the device health score.

[0039] Step S300 further includes step S310: Use the random forest algorithm to rank the feature importance of historical fault data to obtain a feature sequence; step S320: Use LSTM to perform pattern recognition on the feature sequence, and predict the potential fault trend of industrial equipment according to the recognition result; step S330: Based on the device operation status feature set and in combination with the historical fault data, perform association according to the potential fault trend to construct a multi-dimensional health index weight matrix; step S340: Evaluate the multi-dimensional health index weight matrix according to the device operation status feature set to generate the device health score.

[0040] Preferably, according to the generated device operation status feature set, comprehensively evaluate the health status of industrial equipment, and finally obtain a quantified device health score to intuitively reflect the current operation status of the device. Specifically, use the random forest algorithm to rank the feature importance of historical fault data. Among them, the random forest is an ensemble learning method composed of multiple decision trees. By synthesizing the results of multiple decision trees, it can provide more accurate and stable predictions; the historical fault data is obtained from the historical operation and maintenance records of the device. For historical fault data, the random forest algorithm can analyze the influence degree of each feature (such as temperature and humidity change rate, instantaneous air volume fluctuation value, harmonic components of current and voltage, etc.) on the occurrence of faults, calculate the importance score of each feature according to the role of the feature in the construction process of the decision tree, and rank the features from high to low according to the feature importance score to form a feature sequence, which can clearly show which features have a greater impact on device faults.

[0041] Preferably, input the feature sequence into the LSTM (Long Short-Term Memory network). The LSTM model analyzes and learns the feature sequence to identify the patterns in it. Among them, the LSTM (Long Short-Term Memory network) is a special recurrent neural network that can effectively process sequence data and can capture long-term dependencies in the data. In device fault prediction, determine the structure of the LSTM model according to the characteristics of the feature sequence and the complexity of the problem, including the number of LSTM layers, the number of neurons in each layer, the input dimension (i.e., the dimension of the feature sequence), and the output dimension (for example, the predicted fault category or fault probability), etc. Input the feature sequence in the prepared training data into the LSTM model. LSTM can learn the patterns in the historical fault data. For example, it can find the associations between specific combinations or change trends of certain features and device faults. Exemplary model configuration data is shown in Table 1:

[0042] Preferably, the device operation status feature set, historical fault data, and potential fault trends are subjected to correlation analysis. Specifically, the device operation status feature set, historical fault data, and potential fault trends are subjected to correlation analysis to reflect the importance of each indicator in evaluating the device health status. The weights are combined into a multi-dimensional matrix, that is, a multi-dimensional health index weight matrix, which can comprehensively consider the influence of multiple factors on the device health. Finally, the index values in the device operation status feature set are weighted and calculated with the multi-dimensional health index weight matrix, that is, each index value is multiplied by its corresponding weight, and then all the results are added to obtain a comprehensive evaluation value. Combining with a certain scoring standard, a device health score is generated to intuitively reflect the device health status. For example, the higher the score, the better the device health status; otherwise, it indicates that the device may have potential faults or need maintenance.

[0043] Step S400, dynamically trigger an alarm signal according to the device health score, determine a target sensor node according to the alarm signal, and generate a target maintenance instruction.

[0044] Preferably, the device health score is a quantitative evaluation result of the current operating condition of industrial equipment. Different score thresholds are usually set to correspond to different device status levels. For example, a health score of 80 - 100 indicates that the device is operating well; 60 - 80 indicates that the device has potential risks and needs attention; below 60 indicates that the device may have serious problems. When the device health score is lower than a specific threshold (such as 60), the system automatically triggers an alarm signal to indicate that the device is abnormal and requires staff to handle it. After the alarm signal is triggered, the sensor nodes that have a key impact on the device health score are determined to collect data, and the specific parts or relevant parameters that may have problems are found. That is, by analyzing the characteristic data related to the device health score and the correlation between sensors, it is further identified which sensor node data anomalies have caused the decrease in the device health score. These sensor nodes are the target sensor nodes. For example, if the decrease in the device health score is caused by too high temperature, then the sensor node responsible for monitoring the device temperature becomes the target sensor node. Finally, detailed target maintenance instructions are generated, clearly indicating which device components need to be inspected, repaired or replaced, as well as the specific maintenance operation steps and precautions, etc. For example, if the target sensor node monitors the temperature of a certain key component and the temperature of this component rises abnormally resulting in an alarm, the maintenance instruction may require checking whether the heat dissipation system of this component is normal, whether the heat dissipation fan needs to be cleaned or the heat conduction material needs to be replaced, etc. The maintenance instruction may also include information such as the recommended maintenance time and the skill requirements of the maintenance personnel to ensure that the maintenance work can be carried out efficiently and accurately, so as to ensure that the device returns to a normal and stable operating state.

[0045] Further, step S400 further includes step S410 of performing stage difference analysis based on the multiple operation stage information to determine multiple difference data; step S420 of constructing a dynamic adaptive threshold according to the multiple difference data, matching the device health score with the dynamic adaptive threshold, and updating the dynamic adaptive threshold in real time to generate a threshold matching result; step S430 of performing hierarchical alarm according to the threshold matching result and triggering the alarm signal.

[0046] Preferably, the device has different operating characteristics and parameter performances in different operating stages (start-up period, steady-state period, load fluctuation period, etc.). By comparing and analyzing the historical data and real-time operating parameter data of the device in each operating stage, the differences between different stages are found. For example, in the start-up period, parameters such as the current and temperature of the device may have large fluctuations and changes, while in the steady-state period, these parameters are relatively stable. Analyze these differences to determine the key parameter data that can reflect the differences between stages, that is, the difference data. For example, the difference values in aspects such as the maximum current and temperature rise rate between the start-up period and the steady-state period; then construct dynamically changing thresholds for the operating parameters of the device based on multiple difference data, which can be adaptively adjusted according to the operating stage and real-time data of the device. For example, considering that the device operation is unstable and various parameters fluctuate greatly in the start-up period, a wider threshold range is set for the parameters in the start-up period to avoid frequent alarms due to normal fluctuations; as the device enters the steady-state period and the operation gradually stabilizes, the threshold range is gradually narrowed to enable the system to more accurately monitor the operating state of the device; in the load fluctuation period, since the parameters will fluctuate with the change of the load, a sliding window algorithm is introduced to dynamically update the upper and lower limits of the threshold according to the real-time data within the sliding window to adapt to the parameter fluctuations caused by the load change.

[0047] Preferably, compare the device health score with the constructed dynamic adaptive threshold. During the operation of the device, continuously update the dynamic adaptive threshold according to the real-time collected data, so that the threshold can always reflect the normal range of the current operating stage of the device. Specifically, by comparing the device health score with the threshold, judge whether the health status of the device exceeds the normal range to obtain the threshold matching result. For example, if the device health score is lower than the lower limit of the threshold set in the steady-state period, it indicates that there may be a problem with the device; if the score is within the threshold range, it means that the device is operating normally. Then, according to the matching situation between the device health score and the threshold, and the severity of the over-limit parameters, the alarms are divided into first-level alarms and second-level alarms. Among them, when the device health score seriously exceeds the threshold range, that is, when the severity of the over-limit parameters is relatively high, trigger a first-level alarm, automatically generate a device shutdown instruction, and push the instruction to the operation and maintenance terminal to stop the device operation in time to avoid more serious damage or safety accidents; when the device health score only slightly exceeds the threshold range and the severity of the over-limit parameters is relatively low, trigger a second-level alarm, generate a warning log, record the abnormal situation of the device, and adjust the monitoring frequency to the intensive mode to collect the device operation data more frequently, monitor the state change of the device in real time, and discover potential problems in time and take corresponding measures. Thus, the dynamic monitoring and hierarchical alarm of the device operating state are realized, which can more effectively ensure the safe and stable operation of the device, discover and handle device failures in time, and reduce the device downtime and maintenance costs.

[0048] Further, step S400 further includes step S440 of mapping the physical positions of the multiple sensor nodes to the industrial equipment to construct a mapping relation table; step S450 of traversing the multiple sensor nodes according to the alarm signal for positioning to determine the target sensor node; step S460 of retrieving the mapping relation table with the target sensor node as an index to determine the fault location information; step S470 of performing maintenance analysis based on the fault location information in combination with the equipment topology structure to determine the fault maintenance path, constructing a visual positioning map based on the fault maintenance path, and generating the target maintenance instruction according to the visual positioning map.

[0049] Preferably, the physical positions of multiple sensors are matched and mapped to the equipment, that is, the specific installation positions of each sensor node on the equipment are determined, and this correspondence is sorted into a table form, that is, a mapping relation table is constructed. For example, information such as sensor node A is installed at the motor part of the equipment and sensor node B is installed at the air duct inlet of the equipment will be recorded in the mapping relation table; when an alarm signal is triggered during the operation of the equipment, it indicates that the equipment may have an abnormal situation. According to the operation parameter data associated with the alarm signal, all sensor nodes are checked and analyzed one by one, that is, by comparing the data collected by the sensor nodes with the data range during normal operation, to find out which sensor nodes have abnormal data. The sensor nodes with abnormal data are the target sensor nodes. For example, if the alarm signal is caused by too high equipment temperature, then check all the sensor nodes related to temperature monitoring to determine which nodes' temperature data exceeds the normal range, which are the target sensor nodes.

[0050] Preferably, the target sensor node is used as an index to search in the constructed mapping relation table. Specifically, by searching the mapping relation table, the physical position information of the industrial equipment corresponding to the target sensor node can be obtained, which is the fault location information. For example, if the target sensor node is the aforementioned sensor node A, through retrieving in the mapping relation table, it is known that it is installed at the motor part of the equipment, then the fault location can be determined to be at the motor of the equipment; then, in combination with the equipment topology structure (i.e., the connection relationship and layout between each component of the equipment), the fault is deeply analyzed. Specifically, in combination with the structural characteristics of the equipment and the feasibility of maintenance, the best path for reaching the fault location for maintenance operation is determined starting from the overall layout of the equipment. For example, some components of the equipment may be blocked by other components and it is necessary to disassemble some external components first to reach the fault location, then a reasonable disassembly and installation sequence needs to be planned to determine the fault maintenance path.

[0051] Preferably, according to the determined fault maintenance path, a graphical tool is used to visually display the topological structure of the device and the fault maintenance path, and a visual positioning map is constructed to clearly see the specific information of each component, fault location, and maintenance path of the device; the operation and maintenance personnel can quickly understand the fault location and specific operation steps of maintenance by viewing the visual positioning map, and generate detailed target maintenance instructions according to the information displayed on the visual positioning map, which may include specific maintenance operation steps, precautions, etc., so that the maintenance personnel can accurately and efficiently carry out fault repair work. For example, the maintenance instructions may state to first turn off the power of the device, then use specific tools to disassemble a certain component, and then check and repair the faulty component, etc., so as to quickly and accurately locate the fault location when the device fails, and formulate a reasonable maintenance plan to improve the maintenance efficiency and operation stability of the device.

[0052] Further, step S470 further includes step S471, mapping the fault maintenance path to the device topology and binding coordinates with the multiple sensor nodes to generate multiple binding coordinates; step S472, performing marking according to the multiple binding coordinates to determine multiple fault regions; step S473, performing troubleshooting and splicing on the multiple fault regions through a human-computer interface to construct the visual positioning map.

[0053] Preferably, the fault maintenance path is mapped onto the device topology, that is, the route that the maintenance personnel need to pass through is clearly marked in the layout diagram of the device, and this path is bound with the multiple sensor nodes. Specifically, since each sensor node has its specific physical position on the device, it can be represented by coordinates. By establishing the corresponding relationship between the fault maintenance path and the sensor node coordinates, specific coordinate values are determined for the key positions on the path to form multiple binding coordinates. For example, in the three-dimensional spatial layout of the device, sensor node A is located at the position of (X1, Y1, Z1), and the fault maintenance path passes near this sensor node, then this coordinate is bound to the corresponding position on the path to generate a binding coordinate; then, based on these coordinates, markings are made on the device topology to more clearly define the fault regions. Specifically, through the analysis and processing of the binding coordinates, the area covered by the coordinates is divided into different parts to obtain the fault regions. For example, along the fault maintenance path, the regions where the motor is located, the connecting pipes are located, etc., which may be related to the fault, can be determined according to the binding coordinates.

[0054] Preferably, the human-machine interaction interface is a window for the operator to interact with the system. Through the human-machine interaction page, multiple determined fault areas are investigated and pieced together. Specifically, each fault area is inspected and analyzed in detail to determine whether there are faults or clues related to faults in this area. Then, the individual fault areas are combined and integrated on the interface to form a complete image, i.e., a visualization positioning map, which graphically shows the fault location of the equipment and the areas related to the fault. The operator can perform operations on the interface, such as zooming in, zooming out, rotating, etc., to view the details of each area more clearly, thereby making more effective fault diagnosis and maintenance decisions, helping to quickly determine the root cause of the fault and formulate a maintenance plan to ensure the stable operation of the equipment.

[0055] In the above text, reference is made to Figure 1 a method for monitoring the operation of a multi-point distributed high-precision industrial equipment according to an embodiment of the present invention is described in detail. Next, reference will be made to Figure 2 describe a multi-point distributed high-precision industrial equipment operation monitoring system according to an embodiment of the present invention.

[0056] A multi-point distributed high-precision industrial equipment operation monitoring system according to an embodiment of the present invention is used to solve the technical problems existing in the prior art, such as low efficiency of equipment operation monitoring, difficulty in comprehensively grasping the overall condition of the equipment, and lack of a flexible monitoring and warning mechanism, which in turn leads to the inability to accurately locate the fault position of the equipment and respond quickly for processing, resulting in the inefficiency and instability of the industrial equipment operation. It achieves the technical effect of improving the accuracy and stability of equipment fault positioning. As Figure 2 shown, a multi-point distributed high-precision industrial equipment operation monitoring system includes: an operating parameter data acquisition module 10, an equipment operating state feature set generation module 20, an equipment health score acquisition module 30, and a target maintenance instruction generation module 40.

[0057] The operating parameter data acquisition module 10 is used to independently acquire the operating parameter data of the equipment through multiple sensor nodes; the equipment operating state feature set generation module 20 is used to transmit the operating parameter data to the central control unit in real time for fusion processing to generate an equipment operating state feature set. Among them, the equipment operating state feature set is obtained by calling the operation timestamp to perform time series alignment and effective denoising on the operating parameter data, and extracting effective parameter features for stage comparison and fusion; the equipment health score acquisition module 30 is used to use the LSTM model to perform fault prediction on the equipment operating state feature set, construct a multi-dimensional health index weight matrix for health assessment, and obtain an equipment health score; the target maintenance instruction generation module 40 is used to dynamically trigger an alarm signal according to the equipment health score, determine the target sensor node according to the alarm signal, and generate a target maintenance instruction.

[0058] Next, the specific configuration of the operating parameter data acquisition module 10 will be described in detail. The operating parameter data acquisition module 10 further includes: traversing multi-point distributed high-precision industrial equipment for operation analysis to determine multiple key operating parts; arranging the multiple sensor nodes according to the multiple key operating parts, traversing the multiple sensor nodes for configuration to determine independent data acquisition modules; and performing operation acquisition on the industrial equipment through the independent data acquisition modules to obtain the operating parameter data.

[0059] Next, the specific configuration of the device operating state feature set generation module 20 will be described in detail. The device operating state feature set generation module 20 further includes: transmitting the operating parameter data to the central control unit in real time to extract an operation simulation signal; performing communication conversion on the operation simulation signal to obtain a standard digital signal; activating a local area network to extract edge computing nodes, performing data verification on the standard digital signal according to the edge computing nodes, performing operation impact analysis based on the verification result, and setting the device operation priority; and allocating bandwidths to the multiple sensor nodes according to the device operation priority to obtain multiple data transmission bandwidths.

[0060] Next, the specific configuration of the device operating state feature set generation module 20 will be further described in detail. The device operating state feature set generation module 20 further includes: retrieving an operation timestamp, aligning the operating parameter data in time sequence according to the operation timestamp to obtain operation parameter aligned data; performing validity denoising based on the operation parameter aligned data to extract effective parameter features; comparing the effective parameter features with the device historical operation data in stages according to the multiple data transmission bandwidths, and performing fusion according to the comparison result to obtain the device operating state feature set.

[0061] Next, the specific configuration of the device operating state feature set generation module 20 will be further described in detail. The device operating state feature set generation module 20 further includes: retrieving the device historical operation data, dividing operation stages based on the device historical operation data to determine multiple operation stage information; matching the multiple operation stage information according to the multiple data transmission bandwidths to generate a matching result, performing clustering analysis on the effective parameter features and the device historical operation data according to the matching result to generate a staged parameter distribution map; and adding the staged parameter distribution map to the comparison result.

[0062] Next, the specific configuration of the device health score obtaining module 30 will be described in detail. The device health score obtaining module 30 further includes: using the random forest algorithm to rank the feature importance of historical fault data to obtain a feature sequence; performing pattern recognition on the feature sequence through LSTM, and predicting the potential fault trend of industrial equipment according to the recognition result; based on the device operation state feature set and in combination with the historical fault data, associating according to the potential fault trend to construct a multi-dimensional health index weight matrix; evaluating the multi-dimensional health index weight matrix according to the device operation state feature set to generate the device health score.

[0063] Next, the specific configuration of the target maintenance instruction generation module 40 will be described in detail. The target maintenance instruction generation module 40 further includes: performing phase difference analysis based on the multiple operation phase information to determine a plurality of difference data; constructing a dynamic adaptive threshold according to the plurality of difference data, matching the device health score with the dynamic adaptive threshold, and updating the dynamic adaptive threshold in real time to generate a threshold matching result; performing hierarchical alarm according to the threshold matching result to trigger the alarm signal.

[0064] Next, the specific configuration of the target maintenance instruction generation module 40 will be further described in detail. The target maintenance instruction generation module 40 further includes: mapping the plurality of sensor nodes to the physical positions of the industrial equipment to construct a mapping relationship table; traversing the plurality of sensor nodes according to the alarm signal for positioning to determine the target sensor node; retrieving the mapping relationship table using the target sensor node as an index to determine the fault location information; performing maintenance analysis according to the fault location information in combination with the device topology structure to determine the fault maintenance path, constructing a visual positioning map based on the fault maintenance path, and generating the target maintenance instruction according to the visual positioning map.

[0065] Next, the specific configuration of the target maintenance instruction generation module 40 will be further described in detail. The target maintenance instruction generation module 40 further includes: mapping the fault maintenance path to the device topology structure and binding coordinates with the plurality of sensor nodes to generate a plurality of bound coordinates; performing labeling according to the plurality of bound coordinates to determine a plurality of fault regions; performing troubleshooting and splicing on the plurality of fault regions through a human-computer interaction interface to construct the visual positioning map.

[0066] The multi-point distributed high-precision industrial equipment operation monitoring system provided by the embodiment of the present invention can execute the multi-point distributed high-precision industrial equipment operation monitoring method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0067] Although this application makes various references to certain modules in the system according to the embodiments of this application, however, any number of different modules can be used and run on the user terminal and / or the server. The various units and modules included are only divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of mutual distinction and are not used to limit the protection scope of the present invention.

[0068] The above specific embodiments do not constitute a limitation to the protection scope of this application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of this application shall be included within the protection scope of this application.

Claims

1. A multi-point distributed high-precision industrial equipment operation monitoring method, characterized in that: The method comprises: Independently collect equipment operating parameter data through multiple sensor nodes; The operation parameter data is transmitted to the central control unit in real time for fusion processing to generate a device operation status feature set, wherein the device operation status feature set is obtained by calling the operation timestamp to perform time sequence alignment and validity denoising on the operation parameter data, extracting valid parameter features for stage comparison and fusion; The LSTM model is used to predict faults of the equipment operation status feature set, and a multi-dimensional health indicator weight matrix is ​​constructed to perform health assessment to obtain the equipment health score; An alarm signal is dynamically triggered according to the equipment health score, a target sensor node is determined according to the alarm signal, and a target maintenance instruction is generated.

2. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 1, characterized in that: The operating parameter data of the equipment is collected independently by multiple sensor nodes, and the method includes: Traverse multi-point distributed high-precision industrial equipment for operation analysis and identify multiple key operating parts; Arrange the plurality of sensor nodes according to the plurality of key operation parts, traverse the plurality of sensor nodes for configuration, and determine an independent data acquisition module; The operation parameter data of the industrial equipment is obtained by performing operation collection on the industrial equipment through the independent data collection module.

3. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 1, characterized in that: The operating parameter data is transmitted to a central control unit in real time, the method comprising: Transmitting the operating parameter data to the central control unit in real time to extract an operating simulation signal; Performing communication conversion on the operation analog signal to obtain a standard digital signal; Activate the local area network to extract the edge computing node, perform data verification on the standard digital signal according to the edge computing node, perform operation impact analysis based on the verification result, and set the equipment operation priority; Bandwidths are allocated to the multiple sensor nodes according to the device operation priorities to obtain multiple data transmission bandwidths.

4. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 3, characterized in that: The operating parameter data is transmitted to a central control unit in real time for fusion processing to generate a device operating status feature set, the method comprising: Retrieving the operation timestamp, performing time alignment on the operation parameter data according to the operation timestamp, and obtaining the operation parameter alignment data; Performing effectiveness denoising based on the operation parameter alignment data to extract effective parameter features; The effective parameter characteristics are compared with the historical operation data of the equipment in stages according to the multiple data transmission bandwidths, and are integrated according to the comparison results to obtain the equipment operation status feature set.

5. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 4, characterized in that: The effective parameter characteristics are compared with the historical operation data of the equipment in stages according to the multiple data transmission bandwidths, and the method includes: Retrieving historical operation data of the device, dividing the operation stages based on the historical operation data of the device, and determining multiple operation stage information; Matching multiple operation stage information according to the multiple data transmission bandwidths to generate a matching result, performing cluster analysis on the effective parameter characteristics and the historical operation data of the equipment according to the matching result to generate a phased parameter distribution map; The phased parameter distribution map is added to the alignment result.

6. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 1, characterized in that: The LSTM model is used to predict the faults of the equipment operation status feature set, a multi-dimensional health indicator weight matrix is ​​constructed for health assessment, and the equipment health score is obtained. The method includes: The random forest algorithm is used to sort the feature importance of historical fault data and obtain the feature sequence; Performing pattern recognition on the feature sequence through LSTM, and predicting the potential failure trend of the industrial equipment according to the recognition result; Based on the equipment operation status feature set combined with the historical fault data, the potential fault trend is associated to construct a multi-dimensional health indicator weight matrix; The multi-dimensional health indicator weight matrix is ​​evaluated according to the equipment operation status feature set to generate the equipment health score.

7. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 5, characterized in that: Dynamically triggering an alarm signal according to the device health score, the method comprising: Performing a phase difference analysis based on the plurality of operation phase information to determine a plurality of difference data; constructing a dynamic adaptive threshold according to the plurality of difference data, matching the device health score with the dynamic adaptive threshold, updating the dynamic adaptive threshold in real time, and generating a threshold matching result; A graded alarm is performed according to the threshold matching result to trigger the alarm signal.

8. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 1, characterized in that: Determining a target sensor node according to the alarm signal and generating a target maintenance instruction, the method includes: Mapping the multiple sensor nodes to the physical locations of the industrial equipment to construct a mapping relationship table; Traversing the plurality of sensor nodes for positioning according to the alarm signal to determine a target sensor node; Using the target sensor node as an index to search the mapping relationship table to determine fault location information; A maintenance analysis is performed based on the fault location information in combination with the equipment topology structure to determine a fault maintenance path, a visual location map is constructed based on the fault maintenance path, and the target maintenance instruction is generated based on the visual location map.

9. A multi-point distributed high-precision industrial equipment operation monitoring method as claimed in claim 8, characterized in that: Building a visual location map based on the fault maintenance path, the method includes: Mapping the fault maintenance path to the device topology structure and performing coordinate binding with the multiple sensor nodes to generate multiple binding coordinates; Marking according to the multiple binding coordinates to determine multiple fault areas; The multiple fault areas are checked and spliced ​​through a human-computer interaction interface to construct the visual positioning map.

10. A multi-point distributed high-precision industrial equipment operation monitoring system, characterized in that: The system is used to implement a multi-point distributed high-precision industrial equipment operation monitoring method according to any one of claims 1 to 9, and the system comprises: An operating parameter data collection module is used to independently collect operating parameter data of the device through multiple sensor nodes; The equipment operation status feature set generation module is used to transmit the operation parameter data to the central control unit in real time for fusion processing to generate the equipment operation status feature set, wherein the equipment operation status feature set is obtained by calling the operation timestamp to perform time alignment and validity denoising on the operation parameter data, extracting valid parameter features for stage comparison and fusion; The equipment health score acquisition module is used to use the LSTM model to predict faults on the equipment operation status feature set, construct a multi-dimensional health indicator weight matrix for health assessment, and obtain the equipment health score; The target maintenance instruction generation module is used to dynamically trigger an alarm signal according to the equipment health score, determine the target sensor node according to the alarm signal, and generate a target maintenance instruction.

Images