Equipment health management visualization method and system based on Qt framework

Abstract

The invention relates to the technical field of equipment maintenance and management, in particular to an equipment health management visualization method and system based on a Qt framework, a characteristic spectrum system forming module obtains equipment real-time data of multi-source sensing, signal decoupling analysis is carried out, and a decoupling characteristic spectrum is obtained; performing frequency domain conversion and modulation analysis on the decoupling characteristic spectrum to form a multi-dimensional characteristic spectrum system; the fault feature mapping network construction module analyzes modal correlation of the multi-dimensional feature pedigree based on a Qt framework to obtain a fault feature mapping network; the health trend graph generation module constructs a mixed time sequence prediction model based on the fault feature mapping network, performs equipment residual life prediction and degradation trend evaluation, and forms a health trend graph of the equipment; and the reliability assessment display module performs reliability assessment according to the baseline threshold of the equipment health state assessment in combination with the health trend map, obtains the health state of the equipment, and performs visual display. According to the invention, the accuracy of equipment fault judgment can be improved, and equipment health management is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment maintenance management, in particular to an equipment health management visualization method and system based on a Qt framework. BACKGROUND

[0002] Equipment health management mainly establishes the bionic function of "self-monitoring, self-maintenance, self-repair, self-strengthening, self-compensation, and self-adaptation" for equipment, effectively solving the long-term self-care of the "health state" and the self-monitoring and rehabilitation of the "sub-health" state. The existing equipment health management mostly adopts a PHM (Prognostics and Health Management) system, which is a comprehensive management platform integrating data acquisition, state monitoring, fault prediction, and decision support, and can realize active maintenance and health management of equipment or projects. Through real-time monitoring, data analysis and prediction, the PHM system reduces the equipment failure risk through predictive maintenance and improves the operation efficiency.

[0003] However, there are various types of equipment and the data sensors used are also characterized by a large number of types and quantities, so the PHM system needs to integrate data (such as vibration, temperature, current, etc.) from different sensors and different formats. The traditional method has problems such as difficulty in data synchronization, inaccurate feature extraction, and low signal-to-noise ratio, which affects the accuracy of equipment fault judgment. SUMMARY

[0004] To solve the above technical problems, the present application provides an equipment health management visualization method based on a Qt framework, comprising: Obtaining real-time data of equipment from multiple sources of sensors, performing signal decoupling analysis to obtain a decoupling characteristic spectrum; performing frequency domain conversion and modulation analysis on the decoupling characteristic spectrum to form a multi-dimensional characteristic spectrum system; Analyzing the modal correlation of the multi-dimensional characteristic spectrum system based on the Qt framework to obtain a fault feature mapping network; Constructing a hybrid time series prediction model based on the fault feature mapping network to perform equipment residual life prediction and degradation trend evaluation, and forming a health trend atlas of the equipment; According to the baseline threshold of the equipment health state evaluation, combining the health trend atlas to perform reliability evaluation, obtaining the health state of the equipment, and visualizing the health state.

[0005] As a further optimization scheme, the equipment health management visualization method based on the Qt framework further comprises: generating a maintenance decision scheme based on the health state of the equipment, determining a maintenance method according to the maintenance decision scheme when the equipment is abnormal, forming a maintenance suggestion, and pushing the equipment abnormality and the maintenance suggestion to a pre-associated terminal device.

[0006] As a further optimization scheme, the real-time data of the equipment is obtained by the following way: In the equipment side, a high-performance edge computing gateway with multiple industrial protocol analysis modules is deployed, multiple source sensors are connected, and the original data of the equipment are synchronously collected through a multi-source sensor synchronous acquisition card, and a precise clock protocol is used for multi-sensor data tagging synchronization. The original data are preliminarily preprocessed, including removing abnormal noise points by using a hybrid filtering algorithm based on median filtering and wavelet threshold denoising, and compressing the denoised original data; The compressed original data are transmitted to a data lake, a unified data model mapping mechanism is established in the data lake, and data format unified conversion and quality checking are performed to form equipment real-time data for classified storage.

[0007] As a further optimization scheme, a hybrid time series prediction model is constructed based on a fault feature mapping network to predict the remaining life and assess the degradation trend of the equipment, and a health trend atlas of the equipment is formed, which specifically includes: A state equation of the equipment state parameters changing with time is constructed as an equipment degradation model; The most relevant key feature values related to degradation are extracted from the fault feature mapping network, an observation equation of the state parameters and observation features is established to form a feature mapping of the equipment; A particle filter (PF) algorithm is run in a work thread in the Qt framework; based on the prior knowledge of the initial health state of the device, the initial values of each state are represented by particles, and combined with the initial weight, N particles are generated; Importance sampling, weight updating, weight normalization and resampling processing are implemented; Based on the state equation, the remaining useful life (RUL) samples corresponding to the particles are predicted by iterating forward with the time step as the iteration number, and the RUL samples of all particles constitute the probability distribution of the RUL; the health trend atlas is drawn in the graphical view framework of Qt.

[0008] As a further optimization scheme, the baseline threshold for equipment health state assessment is optimized, specifically including: A digital twin equipment model is constructed based on the Qt framework, which integrates the physical mechanism of the equipment; Real-time equipment data are obtained, and the model parameters of the digital twin equipment model are dynamically updated using the real-time equipment data, so that the digital twin equipment model is synchronized with the physical entity; Based on the updated digital twin equipment model, expected health state data of the equipment under various working conditions are simulated and generated; According to the expected health state data and the real-time equipment data, the baseline threshold for initial equipment health state assessment is dynamically adjusted and optimized.

[0009] As a further optimization scheme, a digital twin equipment model integrating equipment physical mechanism is built based on Qt framework, specifically including: A three-dimensional structure model of the equipment is obtained, imported into multi-physics simulation software, and based on the equipment physical mechanism, a multi-physics coupling model including fluid dynamics, thermodynamics and structural mechanics is built based on the three-dimensional structure model; The high-dimensional multi-physics coupling model is reduced to a low-dimensional state space model by model reduction technology, and the state space model is derived; The derived state space model is embedded into the digital twin running platform based on Qt framework, that is, the digital twin equipment model integrating equipment physical mechanism is formed.

[0010] As a further optimization scheme, the digital twin equipment model is kept synchronized with the physical entity, specifically including: The digital twin equipment model is expressed in a nonlinear state space form, and the time-varying parameters therein are clearly defined; For time-varying parameters, the corresponding real-time parameter values are determined and obtained from real-time equipment data, and machine learning algorithms are used for learning and compensation; Through learning and compensation, the model parameter update is realized, and the digital twin equipment model is kept synchronized with the physical entity.

[0011] As a further optimization scheme, the expected health state data of the equipment under various working conditions is simulated and generated, specifically including: Using high-performance computing technology and historical operation data, a data set covering sufficient working conditions is generated, including input working condition parameters and output health state indicators of the digital twin equipment model; A deep neural network structure is constructed to learn the complex mapping from input to output of the data set, and a digital twin proxy model based on deep neural network is obtained; Based on the updated model parameters of the digital twin equipment model, the digital twin proxy model is used for rapid deduction, and the expected health state data of the equipment under various working conditions is output.

[0012] As a further optimization scheme, the baseline threshold for initial equipment health state assessment is dynamically adjusted and optimized, specifically including: One or more key performance indicators representing the health state of the equipment are selected, and a control chart based on exponential weighted moving average is used to set a dynamic threshold; The distribution characteristics of the key performance indicators are continuously monitored, and if a significant change in the distribution is detected, a threshold relearning process is triggered, that is, the baseline threshold is recalculated using the equipment real-time data in the recent period of time; According to the recalculated reference threshold, the baseline threshold of the initial equipment health state evaluation is dynamically adjusted and optimized.

[0013] The application also provides an equipment health management visualization system based on a Qt framework, comprising: A feature spectrum forming module is configured to acquire real-time equipment data sensed by multiple sources, perform signal decoupling analysis to obtain a decoupled feature spectrum, and perform frequency domain conversion and modulation analysis on the decoupled feature spectrum to form a multi-dimensional feature spectrum. A fault feature mapping network construction module is configured to analyze modal correlation of the multi-dimensional feature spectrum based on the Qt framework to obtain a fault feature mapping network. A health trend atlas generation module is configured to construct a hybrid time series prediction model based on the fault feature mapping network, perform equipment residual life prediction and degradation trend evaluation, and form a health trend atlas of the equipment. A reliability evaluation display module is configured to perform reliability evaluation according to the baseline threshold of the equipment health state evaluation in combination with the health trend atlas to obtain the health state of the equipment and perform visual display.

[0014] The equipment health management visualization method and system based on the Qt framework provided by the application can detect multi-source sensing data (i.e., real-time equipment data) of the equipment through multiple sensors, perform signal decoupling analysis on signal representation of the real-time equipment data to obtain a decoupled feature spectrum, perform frequency domain conversion and modulation analysis to form a multi-dimensional feature spectrum, analyze modal correlation of the multi-dimensional feature spectrum of the equipment based on the Qt framework, express the modal correlation of the multi-dimensional feature spectrum in a mapping network to obtain a fault feature mapping network, perform equipment residual life prediction and degradation trend evaluation according to the fault feature mapping network, express the prediction and evaluation process and results in a graphical form to form a health trend atlas of the equipment, use a baseline threshold (which can be pre-set or imported later) as a reference standard for equipment health state evaluation, perform reliability evaluation in combination with the health trend atlas to obtain the health state of the equipment, and perform visual display of the health trend atlas and the health state of the equipment. The application integrates data from different sensors and in different formats to solve problems such as difficulty in data synchronization, inaccurate feature extraction, and low signal-to-noise ratio in traditional methods, finally improves the accuracy of equipment fault judgment, and achieves the purpose of effective equipment health management.

[0015] Additional features and advantages of the application are set forth in the following description, and in part will be apparent to those skilled in the art from the description, or can be learned by practice of the application. The purposes and other advantages of the present application can be realized and attained by the structure particularly pointed out in the written description and claims of this patent application.

[0016] The technical solutions of the present application are described in further detail below with reference to the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application together with the embodiments thereof, and explain the present application, but are not intended to limit the present application. In the drawings: Figure 1 A flowchart of a Qt framework-based equipment health management visualization method in an embodiment of the present application; Figure 2 A flowchart of an equipment real-time data acquisition method used in an embodiment of the Qt framework-based equipment health management visualization method of the present application; Figure 3 A flowchart of a health trend map generation process used in an embodiment of the Qt framework-based equipment health management visualization method of the present application; Figure 4 A flowchart of a baseline threshold optimization method used in an embodiment of the Qt framework-based equipment health management visualization method of the present application; Figure 5 A schematic diagram of a Qt framework-based equipment health management visualization system in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not intended to limit the present application.

[0019] As shown in Figure 1 An embodiment of the present application provides a Qt framework-based equipment health management visualization method, which comprises: S100: acquiring equipment real-time data of multi-source sensing, performing signal decoupling analysis to obtain a decoupled feature spectrum; performing frequency domain conversion and modulation analysis on the decoupled feature spectrum to form a multi-dimensional feature spectrum system; S200: analyzing the modal correlation of the multi-dimensional feature spectrum system based on the Qt framework to obtain a fault feature mapping network; S300: constructing a hybrid time series prediction model based on the fault feature mapping network, performing equipment residual life prediction and degradation trend evaluation, and forming a health trend map of the equipment; S400: performing reliability evaluation based on the baseline threshold of the equipment health state evaluation, combining the health trend map to obtain the health state of the equipment, and performing visualization display.

[0020] The working principle of the technical solution is as follows: the scheme detects the multi-source sensing data (i.e., equipment real-time data) of the equipment through a multi-source sensor, performs signal decoupling analysis on the signal representation of the equipment real-time data, and obtains a decoupling characteristic spectrum; then, frequency domain conversion and modulation analysis are performed to form a multi-dimensional characteristic spectrum series; based on the Qt framework, the multi-dimensional characteristic spectrum series modal correlation of the equipment is analyzed, the multi-dimensional characteristic spectrum series modal correlation is expressed by using a mapping network to obtain a fault feature mapping network, the residual life prediction and degradation trend evaluation of the equipment are implemented according to the fault feature mapping network, the prediction and evaluation process and results are expressed by using a graph to form a health trend graph of the equipment; and a baseline threshold (which can be pre-set or imported later) is used as a reference standard for the health state evaluation of the equipment, and the reliability evaluation is performed in combination with the health trend graph, so that the health state of the equipment is obtained, and the health trend graph and the health state of the equipment can be visually displayed.

[0021] The technical solution has the following beneficial effects: the scheme is suitable for various and rapidly changing equipment, considers that the data sensors used by the equipment also have the characteristics of a large number of types and quantities, integrates data (such as vibration, temperature, and current) from different sensors and in different formats, solves the problems of difficulty in data synchronization, inaccurate feature extraction, and low signal-to-noise ratio in traditional methods, improves the accuracy of equipment fault judgment, and achieves the purpose of effective equipment health management.

[0022] In one embodiment, the method further includes: generating a maintenance decision scheme based on the health state of the equipment, determining a maintenance method according to the maintenance decision scheme when the equipment is abnormal, forming a maintenance suggestion, and pushing the equipment abnormality and the maintenance suggestion to a terminal device pre-associated with the equipment.

[0023] The working principle and beneficial effects of the technical solution are as follows: based on the health state of the equipment obtained in the foregoing, a maintenance decision scheme of the equipment is generated, if the equipment real-time data of a certain equipment shows that the equipment is abnormal, a maintenance method can be determined according to the maintenance decision scheme, a maintenance suggestion is formed, and the detected equipment abnormality and the recommended maintenance suggestion can be pushed to a terminal device pre-associated with the equipment in real time through an edge computing node, so that the maintenance personnel can timely understand the situation and take measures.

[0024] In one embodiment, as shown in Figure 2 The equipment real-time data in the S100 step is obtained by the following method: S110: deploying a high-performance edge computing gateway with a built-in multi-industry protocol analysis module on the equipment side, connecting a multi-source sensor, synchronously collecting raw data of the equipment through a multi-source sensor synchronous acquisition card, and performing multi-sensor data tag synchronization by using a precise clock protocol; S120: preliminary preprocessing of the original data, including removing abnormal noise points by using a hybrid filtering algorithm based on median filtering and wavelet threshold denoising, and compressing the denoised original data; S130: transmitting the compressed original data to the data lake, establishing a unified data model mapping mechanism in the data lake, and performing data format unified conversion and quality checking to form equipment real-time data for classified storage.

[0025] The working principle of the above technical solution is: a high-performance edge computing gateway (such as Atlas500) is deployed on the equipment side, which is built-in with multiple industrial protocol analysis modules (supporting ModbusTCP / IP, OPCUA, Profinet, EtherCAT, etc.); The gateway synchronously collects original data, such as data including vibration acceleration sensors, temperature sensors, and current sensors, through a synchronous acquisition card (such as a sampling rate of 51.2kHz), and uses a precise clock protocol (such as IEEE1588) for multi-sensor data tagging synchronization to ensure the timing consistency of the data. The collected original data is preliminarily preprocessed in the edge gateway, including removing abnormal noise points by using a hybrid filtering algorithm based on median filtering and wavelet threshold denoising, and compressing the data (such as compressing to 20% of the original size by using the piecewise aggregate approximation (PAA) algorithm) and temporarily storing it in the local time series database; The compressed original data can be transmitted to the data lake through the MQTT protocol, which can be a cloud data lake (such as Aliyun DataLakeFormation). A unified data model mapping mechanism is established in the data lake, which can use ApacheAvro format to define standard data patterns (Schema), and convert different sources and formats of original data into a unified Parquet columnar storage format; Quality checking of the data into the lake can include outlier detection based on the 3σ criterion, data integrity check (ensuring that the continuous data stream is interrupted for no more than 5ms), and data drift detection based on machine learning (using KS test to compare the distribution of real-time data and historical data). The checked data is stored in categories, such as: high-frequency vibration original data is stored in object storage (OSS) for deep analysis, feature data is stored in time series database (TSDB), and metadata and event data are stored in relational database (RDS), thereby forming a complete data asset directory. Thus, second-level collection and management of massive data points can be realized, and the data availability rate reaches 99.95%.

[0026] The beneficial effects of the above technical solution are as follows: In the initial preprocessing of the original data, this solution uses a hybrid filtering algorithm based on median filtering and wavelet threshold denoising to remove abnormal noise points in the original data, thereby improving the data quality and avoiding the impact of data quality issues on data usability; in addition, the original data is compressed and then transmitted to the data lake; by establishing a unified data model mapping mechanism, the original data is subjected to unified format conversion and quality verification to form real-time equipment data, and the real-time equipment data is classified and stored, realizing unified data format and classified management, which facilitates retrieval and use when needed later.

[0027] In one embodiment, such as Figure 3 As shown, step S300 specifically includes: S310: Construct state equations for the changes in equipment state parameters over time as a model for equipment degradation; for example: in, Indicates the first The state vector of the state parameters at time step. Represents the state function. Indicates the first The state vector at time t, Represents model parameters (e.g., material constants). Indicates the first Time-matter noise; S320: Extract the key feature values ​​most relevant to degradation from the fault feature mapping network, establish observation equations bridging state parameters and observation features, and form the feature mapping of the equipment; for example: in, Indicates the first State observation data at any given time, Represents the observation function, Indicates the first Observational noise at any given moment; S330: In the Qt framework, a worker thread is created to run the Particle Filter (PF) algorithm; based on the prior knowledge of the initial health state of the device, the initial values ​​of each state are represented by particles, and N particles are generated by combining the initial weights. S340: Implement importance sampling, weight update, weight normalization, and resampling; specifically: Whenever new state observation data is available, perform the following: Importance sampling: For each particle, predict its state at the next time step based on the physical degradation model; Weight update: calculate the importance weight of each particle, and implement weight update based on the observation likelihood function, i.e. the probability of the current state observation data occurring under the particle state; this means that particles whose predicted state is more consistent with the actual observation data will obtain higher weights; Weight normalization and resampling: to avoid the particle degeneracy problem (a few particles have extremely high weights, and the rest have almost zero weights), calculate the effective particle number, and perform resampling when it is lower than the threshold; resampling means copying high-weight particles and eliminating low-weight particles according to the weight distribution, so as to concentrate resources in the high-probability area; after resampling, the particle weights are reset to 1 / N; S350: forward iteration prediction based on the state equation, the iteration number multiplied by the time step as the remaining useful life (RUL) sample corresponding to the particle, and the remaining useful life samples of all particles constitute the probability distribution of the remaining useful life; draw the health trend map in the graphical view framework of Qt; specifically: Implementing remaining useful life (RUL) calculation: for each particle, starting from its current state, iteratively predict forward according to the state equation until its state first exceeds the preset failure threshold, and this iteration number multiplied by the time step is the remaining useful life sample corresponding to the particle, and the remaining useful life samples of all particles constitute the probability distribution of the remaining useful life; Trend map generation: draw the following contents in the graphical view framework of Qt (such as QtCharts or QCustomPlot) to form the health trend map: Degradation trajectory: draw the curve of the posterior mean of state estimation changing over time; Uncertainty boundary: use a translucent color band to display the confidence interval (such as 95% confidence interval) of the state estimation; Remaining useful life distribution: display the distribution of the remaining useful life in the form of a probability density function (PDF) or a cumulative distribution function (CDF) graph, and label key statistics such as median and mean; Failure threshold line: clearly mark the failure threshold in the degradation trajectory graph.

[0028] The working principle and beneficial effects of the above technical solution are: by introducing particle filtering and Bayesian inference technology, the uncertainty management and probability prediction of the remaining life of the equipment are realized, and more abundant basis is provided for the maintenance decision of the equipment; among them, the state observation data is compared with the baseline threshold of the same working condition, so as to judge whether the equipment state is normal; the probability of the normal state of the equipment can be used as its reliability.

[0029] In one embodiment, as shown in Figure 4 the baseline threshold for equipment health state evaluation in S400 is optimized, specifically including: S410: build a digital twin equipment model integrating equipment physical mechanism based on a Qt framework (a cross-platform C++ graphical user interface application development framework); S420: acquire equipment real-time data, dynamically update model parameters of the digital twin equipment model using the equipment real-time data, and keep the digital twin equipment model synchronized with the physical entity; S430: simulate and generate expected health state data of the equipment under various working conditions based on the updated digital twin equipment model; S440: dynamically adjust and optimize the baseline threshold for the initial set equipment health state evaluation according to the expected health state data and the equipment real-time data.

[0030] The working principle of the above technical solution is as follows: The baseline threshold for the equipment health state evaluation is considered. Although the initially set baseline threshold also reflects the working condition fluctuation, the baseline threshold generally needs to be determined after a large amount of historical data of the equipment is accumulated and studied, so as to ensure its accuracy and reliability. However, it is very difficult to construct an accurate baseline threshold for new equipment or equipment lacking a large amount of historical data. In addition, the baseline threshold may also drift with the equipment wear and environmental changes, which affects the accuracy of the equipment health evaluation. In order to implement visual equipment health management for new equipment or equipment lacking a large amount of historical data, the present scheme builds a digital twin equipment model integrating equipment physical mechanism, synchronizes the digital twin equipment model with the physical entity following the equipment real-time data, and on this basis, simulates the expected health state data of the equipment under various working conditions, and dynamically adjusts and optimizes the initially set baseline threshold according to the expected health state data.

[0031] The beneficial effects of the above technical solution are as follows: The present scheme dynamically adjusts and optimizes the initially set baseline threshold, improves the equipment and working condition matching of the baseline threshold as a reference standard, avoids the problem that it is difficult to ensure the accuracy of the baseline threshold for new equipment or equipment lacking a large amount of historical data, thereby making the health management of the equipment more scientific, and the health management mode more in line with the actual situation, which is conducive to popularization and application.

[0032] In one embodiment, the step S410 specifically comprises: acquire a three-dimensional structure model diagram of the equipment, import a multi-physical field simulation software, build a multi-physical field coupling model including fluid dynamics, thermodynamics and structural mechanics based on the equipment physical mechanism and the three-dimensional structure model diagram; use model reduction (MOR) technology (such as proper orthogonal decomposition (POD) or balanced truncation) to reduce the high-dimensional multi-physical field coupling model to a low-dimensional state space model and export it; Embed the derived state space model into a digital twin running platform based on the Qt framework, that is, form a digital twin equipment model that integrates the physical mechanism of the equipment.

[0033] The working principle and beneficial effects of the technical solution are: based on the three-dimensional structural model of the equipment, a multi-physics field coupling model including fluid dynamics, thermodynamics and structural mechanics is constructed through multi-physics field simulation software, so that the model seems to have a skeleton (structural mechanics) and blood vessels (fluid dynamics, thermodynamics), laying a foundation for subsequent synchronization with the physical entity; the model is reduced to a low-dimensional state space model by model reduction (MOR) technology, realizing moderate simplification of the model and its operation, avoiding too complex and time-consuming operation, and improving the operation processing efficiency; then embedded into a digital twin running platform based on the Qt framework, thereby forming a digital twin equipment model that integrates the physical mechanism of the equipment.

[0034] In one embodiment, the S420 step includes: Expressing the digital twin equipment model in a nonlinear state space form, and explicitly the time-varying parameters therein, that is, the model parameters that need to be updated; For time-varying parameters, determine and obtain the corresponding real-time parameter values from the equipment real-time data, and use machine learning algorithms for learning and compensation; Through learning and compensation, the model parameter update is realized, and the digital twin equipment model is kept synchronized with the physical entity.

[0035] The working principle and beneficial effects of the technical solution are: by using the nonlinear state space form expression method, for the time-varying parameters therein, the corresponding real-time parameter values are determined and obtained from the equipment real-time data, and machine learning algorithms are used for learning and compensation, thereby realizing model parameter update, so that the digital twin equipment model can be kept synchronized with the physical entity; through this scheme, the latest collected equipment real-time data can be included in the consideration of the optimization baseline threshold, thereby continuously enriching the data basis for the optimization of the baseline threshold, further improving the use fitness of the baseline threshold obtained therefrom, and thereby ensuring the effectiveness of equipment health management.

[0036] In one embodiment, the S430 step includes: Using high-performance computing (HPC) technology and historical operation data, a data set covering sufficient working conditions (all working conditions that the real equipment can encounter) is generated, and the data set includes input working condition parameters and output health status indicators of the digital twin equipment model; Construct a deep neural network structure to learn the complex mapping from input to output, and obtain a digital twin agent model based on a deep neural network; Based on the model parameters of the updated digital twin equipment model, a digital twin agent model is used to implement rapid deduction, and expected health state data of the equipment under various working conditions is output.

[0037] The working principle and beneficial effects of the technical solution are as follows: the scheme constructs a deep neural network structure to learn the complex mapping from input to output of a data set containing sufficient working condition data, forms a digital twin agent model, and implements rapid deduction to output expected health state data of the equipment under various working conditions; the scheme uses high-performance computing technology to provide an operation basis for updating and adjusting the baseline threshold.

[0038] In one embodiment, the S440 step includes: One or more key performance indicators (KPIs) representing the health state of the equipment are selected, and an exponentially weighted moving average (EWMA) based control chart is used to set a dynamic threshold; The distribution characteristics of the key performance indicators are continuously monitored, and if a significant change in the distribution is detected, a threshold relearning process is triggered, that is, the baseline threshold is recalculated using the equipment real-time data of the last period of time (for example, the past week); According to the recalculated baseline threshold, the baseline threshold initially set for the equipment health state evaluation is dynamically adjusted and optimized.

[0039] The working principle and beneficial effects of the technical solution are as follows: the scheme continuously monitors the distribution characteristics of the key performance indicators, and when the distribution changes significantly, the threshold is relearned, the baseline threshold can be recalculated using the equipment real-time data of the past week (the reference standard data corresponding to a certain point of the baseline threshold), and the deviation value of the calculated baseline threshold from the reference standard data corresponding to the point of the baseline threshold can be used to dynamically adjust the corresponding point of the initially set baseline threshold one by one to optimize the baseline threshold.

[0040] As shown in Figure 5 The embodiment of the present application provides an equipment health management visualization system based on a Qt framework, which comprises: A feature spectrum forming module 10 is configured to acquire equipment real-time data sensed by multiple sources, perform signal decoupling analysis, and obtain decoupled feature spectra; and perform frequency domain conversion and modulation analysis on the decoupled feature spectra to form a multi-dimensional feature spectrum. A fault feature mapping network construction module 20 is configured to analyze the modal correlation of the multi-dimensional feature spectrum based on the Qt framework to obtain a fault feature mapping network. A health trend graph generation module 30 is configured to construct a hybrid time series prediction model based on the fault feature mapping network, perform equipment residual life prediction and degradation trend evaluation, and form a health trend graph of the equipment. The reliability evaluation display module 40 is used for reliability evaluation according to a baseline threshold of the equipment health state evaluation, in combination with the health trend graph, obtaining the health state of the equipment, and visual display.

[0041] The working principle of the technical solution is as follows: in the scheme, the multi-source sensor detects the multi-source sensing data (i.e. equipment real-time data) of the equipment, the feature spectrum formation module performs signal decoupling analysis on the signal representation of the equipment real-time data to obtain decoupling features, then performs frequency domain conversion and modulation analysis to form a multi-dimensional feature spectrum, the fault feature mapping network construction module analyzes the multi-dimensional feature spectrum modal correlation of the equipment based on the Qt framework, expresses the multi-dimensional feature spectrum modal correlation in a mapping network mode to obtain a fault feature mapping network, the health trend graph generation module implements equipment residual life prediction and degradation trend evaluation according to the fault feature mapping network, and forms an equipment health trend graph by graphically expressing the prediction and evaluation process and results, and the reliability evaluation display module uses a baseline threshold (which can be pre-set or imported later) as a reference standard for equipment health state evaluation, in combination with the health trend graph, to perform reliability evaluation, thereby obtaining the health state of the equipment, and the equipment health trend graph and health state can be visually displayed.

[0042] The beneficial effects of the technical solution are as follows: the scheme is aimed at various and rapidly changing equipment, taking into account the characteristics of the data sensors used by the equipment, which have many types and quantities, integrating data from different sensors and different formats (such as vibration, temperature, current, etc.), solving the problems of difficult data synchronization, inaccurate feature extraction, and low signal-to-noise ratio in traditional methods, improving the accuracy of equipment fault judgment, and achieving the purpose of effective equipment health management.

[0043] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A visualization method for equipment health management based on the Qt framework, characterized in that, include: Acquire real-time equipment data from multiple sources of sensors, perform signal decoupling analysis, and obtain the decoupling characteristic spectrum; Frequency domain transformation and modulation analysis are performed on the decoupled characteristic spectrum to form a multidimensional characteristic spectrum system; Based on the Qt framework, the modal correlation of the multidimensional feature spectrum is analyzed to obtain the fault feature mapping network; A hybrid time-series prediction model is constructed based on a fault feature mapping network to predict the remaining service life of equipment and assess its degradation trend, thereby generating a health trend map of the equipment. Based on the baseline threshold of equipment health status assessment, reliability assessment is conducted in conjunction with health trend maps to obtain the equipment health status and display it visually.

2. The equipment health management visualization method based on the Qt framework according to claim 1, characterized in that, Also includes: Maintenance decision plans are generated based on the health status of the equipment. When the equipment malfunctions, maintenance methods are determined according to the maintenance decision plan, maintenance recommendations are generated, and the equipment malfunction and maintenance recommendations are pushed to pre-associated terminal devices.

3. The equipment health management visualization method based on the Qt framework according to claim 1, characterized in that, Real-time equipment data is obtained through the following methods: A high-performance edge computing gateway with built-in parsing modules for multiple industrial protocols is deployed on the equipment side to connect to multiple source sensors. The raw data of the equipment is collected synchronously through a multi-source sensor synchronous acquisition card, and a precision clock protocol is used to mark and synchronize the multi-sensor data. The raw data undergoes preliminary preprocessing, including the removal of abnormal noise points using a hybrid filtering algorithm based on median filtering and wavelet threshold denoising, and the denoised raw data is then compressed. The compressed raw data is transmitted to the data lake, where a unified data model mapping mechanism is established. Data format conversion and quality verification are performed to generate real-time equipment data, which is then categorized and stored.

4. The equipment health management visualization method based on the Qt framework according to claim 1, characterized in that, A hybrid time-series prediction model is constructed based on a fault feature mapping network to predict the remaining service life of equipment and assess its degradation trend, thereby generating a health trend map of the equipment, specifically including: Construct state equations for the changes in equipment state parameters over time as a model for equipment degradation; Extract the key feature values ​​most relevant to degradation from the fault feature mapping network, establish the observation equation that bridges the state parameters and observation features, and form the feature mapping of the equipment; In the Qt framework, a worker thread is created to run the particle filter algorithm; based on the prior knowledge of the initial health state of the device, the initial values ​​of each state are represented by particles, and N particles are generated by combining the initial weights. Implement importance sampling, weight update, weight normalization, and resampling processes; Based on forward iterative prediction using the state equation, the remaining lifespan sample corresponding to the particle is obtained by multiplying the number of iterations by the time step. The remaining lifespan samples of all particles constitute the probability distribution of the remaining lifespan. A health trend map is then drawn in the Qt graphical view framework.

5. The equipment health management visualization method based on the Qt framework according to claim 1, characterized in that, Optimizations were implemented for the baseline thresholds of equipment health status assessment, specifically including: A digital twin equipment model integrating the physical mechanisms of equipment was constructed based on the Qt framework; Acquire real-time equipment data and use the real-time equipment data to dynamically update the model parameters of the digital twin equipment model, so that the digital twin equipment model keeps synchronized with the physical entity. Based on the updated digital twin equipment model, simulate and generate expected health status data of the equipment under various working conditions; Based on the expected health status data and real-time equipment data, the baseline threshold for the initial equipment health status assessment is dynamically adjusted and optimized.

6. The equipment health management visualization method based on the Qt framework according to claim 5, characterized in that, Based on the Qt framework, a digital twin equipment model integrating the physical mechanisms of the equipment is constructed, specifically including: Obtain the three-dimensional structural model of the equipment, import it into multiphysics simulation software, and construct a multiphysics coupled model including fluid dynamics, thermodynamics and structural mechanics based on the physical mechanism of the equipment on the basis of the three-dimensional structural model. Using model reduction techniques, high-dimensional multiphysics coupled models are reduced to low-dimensional state-space models and derived. The exported state-space model is embedded into a Qt-based digital twin runtime platform, thus forming a digital twin equipment model that integrates the physical mechanism of the equipment.

7. The equipment health management visualization method based on the Qt framework according to claim 5, characterized in that, To keep the digital twin equipment model synchronized with the physical entity, specifically including: The digital twin equipment model is expressed as a nonlinear state-space form, and the time-varying parameters are clearly defined. For parameters that change over time, the corresponding real-time parameter values ​​are determined and obtained from the equipment's real-time data, and machine learning algorithms are used for learning and compensation. By learning and compensation, model parameters are updated, enabling the digital twin equipment model to remain synchronized with the physical entity.

8. The equipment health management visualization method based on the Qt framework according to claim 5, characterized in that, Simulate and generate expected health status data for equipment under various operating conditions, specifically including: Using high-performance computing technology and historical operating data, a dataset covering all operating conditions is generated. The dataset includes input operating condition parameters and output health status indicators of the digital twin equipment model. By constructing a deep neural network structure to learn the complex mapping from input to output of the dataset, a digital twin agent model based on deep neural networks is obtained. Based on the updated digital twin equipment model parameters, a digital twin proxy model is used to perform rapid simulations and output the expected health status data of the equipment under various working conditions.

9. The equipment health management visualization method based on the Qt framework according to claim 5, characterized in that, The baseline thresholds for the initial equipment health status assessment are dynamically adjusted and optimized, specifically including: Select one or more key performance indicators that can characterize the health status of the equipment, and use a control chart based on the exponential weighted moving average to set dynamic thresholds. Continuously monitor the distribution characteristics of key performance indicators. If a significant change in the distribution is detected, trigger the threshold relearning process, which involves recalculating the baseline threshold using real-time equipment data from the most recent period. Based on the recalculated baseline threshold, the baseline threshold for the initial equipment health status assessment is dynamically adjusted and optimized.

10. A visualization system for equipment health management based on the Qt framework, characterized in that, include: The feature spectrum formation module is used to acquire real-time equipment data from multi-source sensors, perform signal decoupling analysis, and obtain decoupled feature spectra. Frequency domain transformation and modulation analysis are performed on the decoupled characteristic spectrum to form a multidimensional characteristic spectrum system; The fault feature mapping network construction module is used to analyze the modal correlation of multidimensional feature spectrum based on the Qt framework and obtain the fault feature mapping network. The health trend map generation module is used to construct a hybrid time series prediction model based on the fault feature mapping network, perform equipment remaining life prediction and degradation trend assessment, and generate a health trend map of the equipment. The reliability assessment and display module is used to conduct reliability assessments based on the baseline threshold of equipment health status assessment and in conjunction with health trend maps, to obtain the health status of the equipment and to display it visually.

Images