Remote monitoring and fault diagnosis method and system for environmental simulation system

Abstract

The present invention relates to the field of simulated environment monitoring, and in particular to a remote monitoring and fault diagnosis method and system for an environmental simulation system. The method comprises the following steps: identifying all distributed multidimensional sensor nodes, performing spatial topological distribution analysis, and constructing multiple node area models; performing all-round environmental simulation operation status monitoring on an environmental simulation specimen box, and performing nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features; performing dynamic state rendering on multiple node area models based on the nonlinear operation status correlation features to construct multiple environmental simulation data twin models; performing operation simulation monitoring processing on multiple environmental simulation data twin models, and performing parameter difference analysis between models to generate parameter difference features of different models. The present invention realizes efficient and accurate environmental simulation system fault analysis and diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of simulated environment monitoring, and in particular to a remote monitoring and fault diagnosis method and system for an environmental simulation system. Background Art

[0002] With the continuous development of industrial automation and intelligent technology, the application of environmental simulation systems in multiple fields has gradually increased, especially in manufacturing, energy management, aerospace and other industries, and has become an indispensable key technology. Environmental simulation systems are mainly used to simulate and simulate various physical environmental conditions, such as temperature, humidity, pressure, vibration, electromagnetic interference and other factors that affect equipment operation. By simulating real or extreme environments, the performance and stability of the equipment can be predicted and verified, thereby providing a scientific basis for equipment design, optimization and fault prediction.

[0003] However, with the increasing complexity and automation level of industrial equipment, traditional environmental simulation systems often have problems such as single monitoring means, poor real-time performance, and frequent manual intervention. Existing environmental simulation systems often rely on manual inspections and regular checks, and are unable to remotely monitor the status of equipment in a timely and effective manner. Due to the particularity of the working environment, they often face the situation of incomplete manual monitoring and delayed data processing. In addition, the large amount of data generated during the environmental simulation process often needs to rely on expert experience for judgment, and there is a lack of systematic and automated fault diagnosis methods, resulting in inaccurate equipment failure predictions, high maintenance costs, and even affecting production safety.

[0004] In order to adapt to the development needs of intelligent production and industrial Internet, how to achieve real-time remote monitoring and fault diagnosis of environmental simulation systems has become an urgent problem to be solved. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention proposes a remote monitoring and fault diagnosis method and system for an environmental simulation system to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a remote monitoring and fault diagnosis method for an environmental simulation system, comprising the following steps:

[0007] Step S1: Identify all distributed multi-dimensional sensor nodes, perform spatial topological distribution analysis, and construct multiple node area models;

[0008] Step S2: Performing all-round environmental simulation operation status monitoring on the environmental simulation sample box, and performing nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features;

[0009] Step S3: Dynamically render multiple node area models based on nonlinear operating state association characteristics to construct multiple environmental simulation data twin models;

[0010] Step S4: Performing operation simulation monitoring on multiple environmental simulation data twin models, and performing parameter difference analysis between models to generate parameter difference characteristics of different models;

[0011] Step S5: performing a local model abnormality feature analysis based on the parameter difference characteristics of different models, and performing a potential environmental simulation fault analysis to obtain potential environmental simulation fault data of the abnormal area;

[0012] Step S6: Inferring environmental fault factors from the potential environmental simulated fault data in the abnormal area, making dynamic adaptive fault diagnosis decisions, and building an adaptive fault diagnosis strategy to perform remote monitoring and fault diagnosis operations.

[0013] The present invention helps to fully understand the structure of the environmental monitoring network by identifying and analyzing the spatial topological distribution of distributed multidimensional sensor nodes. Constructing multiple node area models can provide a detailed description of the status of each environmental area, providing a basis for subsequent monitoring and analysis. Performing all-round environmental simulation operation status monitoring helps to understand the state changes of the simulated environment in real time. Obtaining nonlinear operation state correlation characteristics can reveal the potential correlation between environmental states and provide features for subsequent analysis. Dynamic state rendering of multiple node area models based on nonlinear correlation characteristics can more intuitively present the state changes of the models. Constructing an environmental simulation data twin model helps to simulate the operation conditions under different environmental conditions and provide a basis for comparative analysis. Performing operation simulation monitoring on the environmental simulation data twin model can compare the operation conditions between different models. Generating parameter difference characteristics of different models can help to discover differences between models and provide a basis for anomaly detection. Performing local model anomaly feature analysis based on the parameter difference characteristics of different models can locate abnormal areas. Obtaining potential environmental simulation fault data of abnormal areas helps to identify potential fault points in the environmental simulation system. Inferring the fault factors of the environmental simulation system helps to quickly locate the root cause of the fault. Constructing an adaptive fault diagnosis strategy can improve the accuracy and efficiency of fault diagnosis, ensuring the smooth progress of remote monitoring and fault diagnosis operations.

[0014] Preferably, step S1 includes the following steps:

[0015] Step S11: Identify the distributed multi-dimensional sensors in the environmental simulation sample box and mark all the distributed multi-dimensional sensor nodes;

[0016] Step S12: performing node position calculation on all distributed multi-dimensional sensor nodes to generate position information of all sensor nodes;

[0017] Step S13: Perform spatial topological distribution analysis on the location information of all sensor nodes to generate node spatial topological distribution data;

[0018] Step S14: performing spatial topology distribution modeling on the node spatial topology distribution data to construct a node spatial topology distribution map;

[0019] Step S15: dividing the node space topology distribution map into regions to obtain multiple node topology regions;

[0020] Step S16: Perform three-dimensional point cloud modeling on multiple node topology areas to construct multiple node area models.

[0021] The present invention helps to accurately monitor environmental parameters by identifying distributed multidimensional sensors in an environmental simulation sample box and marking nodes. Marking sensor nodes can provide accurate identification for subsequent position calculation and data analysis. Calculating the position information of sensor nodes can provide the specific position of each sensor in the sample box. Generating sensor node position information helps to establish spatial relationships and provides a basis for subsequent analysis. Analyzing the spatial topological distribution of sensor nodes can reveal the positional relationship between sensors. Generating node spatial topological distribution data helps to understand the structure and layout of the sensor network. Establishing a node spatial topological distribution map can intuitively display the connection relationship between sensor nodes. Constructing a node spatial topological distribution map helps to form an overall sensor network model. Dividing the node spatial topological distribution map into regions can group and manage sensor nodes. Obtaining multiple node topological regions helps to independently analyze and monitor different regions. Performing three-dimensional point cloud modeling can more vividly present the spatial structure of the node region. Constructing multiple node region models helps to deeply understand the sensor layout of each region and provide visual support for remote monitoring and fault diagnosis.

[0022] Preferably, step S2 includes the following steps:

[0023] Step S21: performing all-round environmental simulation operation status monitoring on the environmental simulation sample box based on the distributed multi-dimensional sensor, and extracting multi-dimensional environmental simulation operation monitoring parameters, wherein the multi-dimensional environmental simulation operation monitoring parameters include environmental temperature parameters, environmental humidity parameters, and environmental wind pressure parameters;

[0024] Step S22: performing temperature fluctuation analysis on the ambient temperature parameters to generate ambient temperature fluctuation characteristics;

[0025] Step S23: calculating the fluctuation amplitude of the ambient temperature fluctuation characteristics to generate the ambient temperature fluctuation amplitude;

[0026] Step S24: performing temperature fluctuation amplitude fitting based on the ambient temperature fluctuation amplitude to construct an ambient temperature fluctuation amplitude curve;

[0027] Step S25: Performing environmental state nonlinear correlation mining on the environmental temperature fluctuation amplitude curve, the environmental humidity parameter, and the environmental wind pressure parameter to obtain nonlinear operating state correlation features.

[0028] The present invention uses distributed multi-dimensional sensors to conduct all-round monitoring of the environmental simulation sample box, and can obtain environmental parameter data in real time. Extracting multi-dimensional parameters such as ambient temperature, humidity and wind pressure helps to fully understand the internal environmental status of the sample box, and analyzing ambient temperature fluctuations can reveal the laws and trends of ambient temperature changes. Generating ambient temperature fluctuation characteristics helps to identify the pattern and periodicity of temperature fluctuations. Calculating the ambient temperature fluctuation amplitude can quantify the size and intensity of temperature fluctuations. Generating fluctuation amplitude data helps to further analyze the changes in temperature fluctuations. By fitting the temperature fluctuation amplitude, a fluctuation amplitude model can be established. Constructing an ambient temperature fluctuation amplitude curve helps to visualize the changing trends and laws of temperature fluctuations. Analyzing the correlation between the ambient temperature fluctuation amplitude curve and humidity and wind pressure parameters can discover the nonlinear relationship between different parameters. Extracting nonlinear operating state correlation characteristics helps to identify the complex interactions between environmental parameters, providing a deeper understanding for environmental state analysis and fault diagnosis.

[0029] Preferably, the specific steps of step S25 are:

[0030] Perform spatial humidity distribution analysis on environmental humidity parameters to obtain spatial humidity distribution data of the sample box;

[0031] Conduct long-term periodic change analysis on the humidity distribution data of the sample box space to generate periodic change data of humidity distribution;

[0032] The humidity distribution periodic change data is subjected to humidity change trend evolution, thereby generating humidity change trend evolution characteristics;

[0033] Perform multi-time-frequency decomposition on the environmental wind pressure parameters to obtain wind pressure spectra at different frequencies;

[0034] Perform main frequency component analysis on each spectrum of wind pressure spectra at different frequencies and extract the main frequency component of each spectrum;

[0035] Calculate the power spectrum density of the main frequency component of each spectrum graph to generate the power spectrum density of each spectrum graph;

[0036] Perform power spectrum density distribution identification on the power spectrum density of each spectrum graph to generate wind pressure density distribution data;

[0037] The nonlinear correlation mining of environmental states is performed on the evolution characteristics of humidity change trends, ambient temperature fluctuation amplitude curves and wind pressure density distribution data to obtain the nonlinear operating state correlation characteristics.

[0038] By analyzing the spatial distribution of environmental humidity parameters within a sample box, the present invention can identify humidity non-uniformity, facilitating adjustments to humidity control equipment to ensure uniform humidity distribution. Obtaining spatial humidity distribution data within the sample box can be used to detect potential areas of abnormal humidity and identify humidity issues in advance. By analyzing long-term cyclical changes in humidity distribution, it is possible to identify trends in humidity over time, helping to predict the cyclical patterns of humidity variation. Extracting cyclical humidity variation characteristics helps understand the cyclical patterns of humidity variation, providing a more accurate reference for humidity control and regulation. Extracting the evolutionary characteristics of humidity variation trends provides a deeper understanding of historical trends in humidity variation, helping to predict future humidity changes. Analyzing the evolutionary characteristics of humidity variation trends helps identify underlying patterns in humidity variation, providing a reference for timely adjustments to the sample box's environmental conditions. Processing environmental wind pressure parameters through multi-time-frequency decomposition can identify patterns in wind pressure variation at different frequencies, helping to understand the frequency domain characteristics of wind pressure. Obtaining a wind pressure spectrum can help analyze the frequency distribution of wind pressure, providing a basis for assessing environmental conditions under wind influence. Analyzing the dominant frequency components of the wind pressure spectrum at different frequencies can identify key wind pressure frequency components, helping to understand the primary characteristics of wind pressure at different frequencies. Extracting wind pressure density distribution data can describe the density distribution of wind pressure at different frequencies, supporting quantitative analysis of wind pressure characteristics. Combining humidity trend evolution characteristics, temperature fluctuation amplitude curves, and wind pressure density distribution data for nonlinear correlation mining helps to gain a deeper understanding of the complex relationships between different environmental parameters and provide a comprehensive environmental status assessment. Extracting nonlinear operating status correlation features can help monitor environmental state changes, promptly identify potential problems, and provide more accurate information support for fault diagnosis and prediction.

[0039] Preferably, the specific steps of step S3 are:

[0040] Step S31: performing environmental simulation feature evolution based on nonlinear operating state correlation features, thereby generating all-round environmental operating state evolution data;

[0041] Step S32: performing regional node state matching on the omnidirectional environment operation state evolution data according to the position information of all sensor nodes, thereby obtaining corresponding state evolution data of the node area;

[0042] Step S33: Use the corresponding state evolution data of the node area to dynamically render the state of multiple node area models and construct multiple environmental simulation data twin models.

[0043] The present invention can capture the complex nonlinear relationship between environmental parameters through the nonlinear operating state association characteristics, provide a more accurate description of the evolution of environmental simulation characteristics, and generate all-round environmental operating state evolution data to help monitor subtle changes in environmental parameters and provide early warning of environmental anomalies, thereby ensuring the stability and reliability of system operation. Regional node state matching is performed on the all-round environmental operating state evolution data according to the sensor node position information, which helps to correspond environmental parameters with specific locations, and realize monitoring and diagnosis of environmental states in different regions. Obtaining the corresponding state evolution data of the node area can provide detailed evolution information of the environmental state of each region, providing an accurate basis for subsequent fault diagnosis and problem location. Regional node state matching is performed on the all-round environmental operating state evolution data according to the sensor node position information, which helps to correspond environmental parameters with specific locations, and realize monitoring and diagnosis of environmental states in different regions. Obtaining the corresponding state evolution data of the node area can provide detailed evolution information of the environmental state of each region, providing an accurate basis for subsequent fault diagnosis and problem location.

[0044] Preferably, the specific steps of step S4 are:

[0045] Step S41: performing operation simulation monitoring processing on multiple environmental simulation data twin models, and extracting simulation monitoring data of each twin model;

[0046] Step S42: performing a time series state change trend analysis on the simulated monitoring data of each twin model to generate a time series state change trend of each twin model;

[0047] Step S43: performing situation evolution prediction on the time series state change trend of each twin model, thereby obtaining situation evolution prediction data of each twin model;

[0048] Step S44: Perform inter-model parameter difference analysis on each twin model situation evolution prediction data to generate parameter difference characteristics of different models.

[0049] The present invention can monitor the operating status of multiple environmental simulation data twin models in real time by performing operation simulation monitoring processing on them, detect abnormal situations in time, and ensure the normal operation of the environmental simulation system. Extracting the simulation monitoring data of each twin model helps to understand the performance of each model and provide data support for subsequent analysis and evaluation. Performing time-series state change trend analysis on the simulation monitoring data of each twin model can reveal the change pattern of the model state over time and help identify potential state evolution trends. Generating the time-series state change trend of each twin model helps to understand the evolution process of the model state and provide a basis for subsequent situation evolution prediction. By performing situation evolution prediction on the time-series state change trend of each twin model, the future development trend of the model state can be predicted, helping to take measures to deal with problems in advance. Obtaining the situation evolution prediction data of each twin model can provide early warning information for the remote monitoring system and strengthen the real-time monitoring and management of the environmental simulation system. Performing inter-model parameter difference analysis on the situation evolution prediction data of each twin model helps to compare the state change characteristics between different models and provide a reference for model performance evaluation and optimization. Generating parameter difference characteristics of different models can help identify the differences between models and provide key information support for system fault diagnosis and performance optimization.

[0050] Preferably, the specific steps of step S5 are:

[0051] Step S51: performing local model abnormality feature analysis based on parameter difference features of different models, and extracting local model abnormality feature data;

[0052] Step S52: locating abnormal features of the abnormal feature data of the local model and extracting the abnormal node area model;

[0053] Step S53: Preview the environmental parameters at multiple time points on the abnormal node area model to obtain the environmental preview parameters at multiple time points;

[0054] Step S54: performing potential environmental simulation fault analysis on the environmental rehearsal parameters at multiple time points to obtain potential environmental simulation fault data of abnormal areas.

[0055] The present invention uses local model anomaly feature analysis based on parameter differences between different models to help identify anomalies within the model and identify potential problem points in advance. Extracting local model anomaly feature data provides key data support for subsequent anomaly detection and diagnosis, helping to improve system stability and reliability. Localizing the anomaly feature data within the local model anomaly feature data helps accurately locate the abnormal node area model, quickly pinpointing the problem, and shortening fault diagnosis and repair time. Extracting the abnormal node area model allows for in-depth analysis of environmental parameter changes in the area, providing foundational data for subsequent fault analysis and prediction. Pre-running the abnormal node area model at multiple time points simulates the development trends of anomalies at different time points, helping to understand the problem's evolution. Obtaining environmental preview parameters at multiple time points provides detailed data support for subsequent fault analysis and prediction, facilitating accurate assessment of potential risks. Performing potential environmental simulation fault analysis on the environmental preview parameters at multiple time points helps identify potential fault points in the abnormal area and predict the potential failure mode. Obtaining potential environmental simulation fault data for the abnormal area provides clues and direction for fault diagnosis and problem resolution, improving the system's fault response capabilities and efficiency.

[0056] Preferably, the specific steps of step S6 are:

[0057] Step S61: inferring environmental fault factors based on potential environmental simulated fault data in the abnormal area to obtain environmental fault factor data;

[0058] Step S62: generating a fault warning signal based on the environmental fault factor data;

[0059] Step S63: making dynamic adaptive fault diagnosis decisions based on the fault warning signal and constructing an adaptive fault diagnosis strategy;

[0060] Step S64: Upload the adaptive fault diagnosis strategy to the cloud server to perform remote monitoring and fault diagnosis operations.

[0061] The present invention can help identify the root cause of the fault by inferring the environmental fault factors of the potential environmental simulation fault data in the abnormal area, deeply analyze the occurrence mechanism of the problem, and obtain environmental fault factor data, which can provide an important reference for subsequent fault warning and diagnosis, and help to accurately predict the fault situation. The generation of fault warning signals based on environmental fault factor data can detect potential fault risks in advance and warn system operation and maintenance personnel, which helps to avoid the serious impact of faults on the system. The generation of fault warning signals can improve the real-time monitoring capability of the system, take timely measures to prevent the occurrence of faults, and ensure the stable operation of the system. Dynamic adaptive fault diagnosis decisions made according to fault warning signals can flexibly adjust the diagnosis strategy according to real-time conditions, improve the accuracy and efficiency of fault diagnosis, and construct an adaptive fault diagnosis strategy to help take corresponding measures according to different situations, quickly respond to fault events, and reduce losses caused by faults. Uploading the adaptive fault diagnosis strategy to the cloud server can realize remote monitoring and fault diagnosis operations, improve the response speed and flexibility of the system, and execute the fault diagnosis strategy through the cloud server to achieve remote collaboration and management, ensuring that the system can obtain timely fault diagnosis support at any time and any place.

[0062] In this specification, a remote monitoring and fault diagnosis system for an environmental simulation system is provided, which is used to execute the remote monitoring and fault diagnosis method for the environmental simulation system as described above, including:

[0063] The spatial topology distribution module is used to identify all distributed multi-dimensional sensor nodes, perform spatial topology distribution analysis, and build multiple node area models;

[0064] The nonlinear mining module is used to conduct all-round environmental simulation operation status monitoring of the environmental simulation sample box and conduct nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features;

[0065] Dynamic state rendering module, used to dynamically render multiple node area models based on nonlinear operating state association characteristics and build multiple environmental simulation data twin models;

[0066] The parameter difference module is used to perform operation simulation monitoring on multiple environmental simulation data twin models and perform parameter difference analysis between models to generate parameter difference characteristics of different models;

[0067] The potential fault analysis module is used to analyze the abnormal characteristics of local models based on the parameter difference characteristics of different models, and to analyze the potential environmental simulation faults to obtain the potential environmental simulation fault data of the abnormal area;

[0068] The fault diagnosis decision module is used to infer environmental fault factors based on potential environmental simulated fault data in abnormal areas, make dynamic adaptive fault diagnosis decisions, and build adaptive fault diagnosis strategies to perform remote monitoring and fault diagnosis operations.

[0069] The present invention helps to establish the cognition of the overall structure of the system by identifying distributed multi-dimensional sensor nodes and performing spatial topological distribution analysis, improves the visualization and management efficiency of the system, and constructs multiple node area models to help understand the correlation between nodes, providing a basis for subsequent environmental simulation and monitoring. Nonlinear correlation mining of environmental states can reveal the complex nonlinear relationship between environmental parameters, help understand the deep-level characteristics of the system operation state, and obtain nonlinear operation state correlation characteristics, which can provide important clues for subsequent model establishment and fault diagnosis. Dynamic state rendering based on nonlinear operation state correlation characteristics can more intuitively show the operation state changes of multiple node area models, help identify abnormal situations, and construct environmental simulation data twin models to help simulate the system operation state under different conditions, provide a reference for fault prediction and analysis, and improve the environmental simulation data. Parameter difference analysis based on twin models can help compare the operating conditions of different models, identify potential abnormal characteristics, and provide support for fault diagnosis. Generating parameter difference characteristics of different models helps to understand the performance differences of the system under different conditions and guide subsequent fault analysis and repair. Local model abnormal characteristic analysis based on parameter difference characteristics can effectively identify local abnormal conditions and help discover potential fault points in advance. Potential environmental simulation fault analysis can help predict fault conditions in abnormal areas and provide guidance for fault troubleshooting and repair. Environmental fault factor inference and dynamic adaptive fault diagnosis decision-making can help quickly and accurately diagnose the cause of the fault and reduce the impact of system failures. Building an adaptive fault diagnosis strategy and performing remote monitoring and fault diagnosis operations can improve the stability and reliability of the system and reduce the interference of faults on the normal operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 A schematic flow chart of the steps of a remote monitoring and fault diagnosis method for an environmental simulation system according to the present invention;

[0071] Figure 2 Detailed implementation flow chart of step S1;

[0072] Figure 3 Detailed implementation flow chart of step S2;

[0073] Figure 4 Schematic diagram of the detailed implementation steps of step S3. DETAILED DESCRIPTION

[0074] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0075] This application provides a remote monitoring and fault diagnosis method and system for an environmental simulation system. The execution entities of the remote monitoring and fault diagnosis method and system include, but are not limited to, the following: mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc., which can be considered as general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of an audio and image management system, an information management system, and a cloud data management system.

[0076] See also Figures 1 to 4 The present invention provides a remote monitoring and fault diagnosis method for an environmental simulation system, the remote monitoring and fault diagnosis method for an environmental simulation system comprising the following steps:

[0077] Step S1: Identify all distributed multi-dimensional sensor nodes, perform spatial topological distribution analysis, and construct multiple node area models;

[0078] Step S2: Performing all-round environmental simulation operation status monitoring on the environmental simulation sample box, and performing nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features;

[0079] Step S3: Dynamically render multiple node area models based on nonlinear operating state association characteristics to construct multiple environmental simulation data twin models;

[0080] Step S4: Performing operation simulation monitoring on multiple environmental simulation data twin models, and performing parameter difference analysis between models to generate parameter difference characteristics of different models;

[0081] Step S5: performing a local model abnormality feature analysis based on the parameter difference characteristics of different models, and performing a potential environmental simulation fault analysis to obtain potential environmental simulation fault data of the abnormal area;

[0082] Step S6: Inferring environmental fault factors from the potential environmental simulated fault data in the abnormal area, making dynamic adaptive fault diagnosis decisions, and building an adaptive fault diagnosis strategy to perform remote monitoring and fault diagnosis operations.

[0083] The present invention helps to fully understand the structure of the environmental monitoring network by identifying and analyzing the spatial topological distribution of distributed multidimensional sensor nodes. Constructing multiple node area models can provide a detailed description of the status of each environmental area, providing a basis for subsequent monitoring and analysis. Performing all-round environmental simulation operation status monitoring helps to understand the state changes of the simulated environment in real time. Obtaining nonlinear operation state correlation characteristics can reveal the potential correlation between environmental states and provide features for subsequent analysis. Dynamic state rendering of multiple node area models based on nonlinear correlation characteristics can more intuitively present the state changes of the models. Constructing an environmental simulation data twin model helps to simulate the operation conditions under different environmental conditions and provide a basis for comparative analysis. Performing operation simulation monitoring on the environmental simulation data twin model can compare the operation conditions between different models. Generating parameter difference characteristics of different models can help to discover differences between models and provide a basis for anomaly detection. Performing local model anomaly feature analysis based on the parameter difference characteristics of different models can locate abnormal areas. Obtaining potential environmental simulation fault data of abnormal areas helps to identify potential fault points in the environmental simulation system. Inferring the fault factors of the environmental simulation system helps to quickly locate the root cause of the fault. Constructing an adaptive fault diagnosis strategy can improve the accuracy and efficiency of fault diagnosis, ensuring the smooth progress of remote monitoring and fault diagnosis operations.

[0084] In the embodiment of the present invention, see Figure 1 , is a flowchart of the steps of a remote monitoring and fault diagnosis method for an environmental simulation system of the present invention. In this example, the steps of the remote monitoring and fault diagnosis method for an environmental simulation system include:

[0085] Step S1: Identify all distributed multi-dimensional sensor nodes, perform spatial topological distribution analysis, and construct multiple node area models;

[0086] In this embodiment, sensor nodes in the environmental simulation system are typically distributed, with each sensor responsible for collecting specific environmental data, such as temperature, humidity, pressure, gas composition, flow rate, and vibration. Each node contains one or more sensor modules that collect the corresponding environmental parameters and transmit the data to a central control platform via a network. Each sensor node has different functions and measurement accuracy, so it is necessary to classify each node by type. For example, nodes can be categorized as temperature and humidity sensor nodes, pressure sensor nodes, and flow sensor nodes. A node identification algorithm is used to identify the specific type and deployment location of each node and obtain the multi-dimensional data it collects. Sensor nodes typically transmit data using standard communication protocols (such as LoRa, Zigbee, Wi-Fi, Bluetooth, or Modbus). During this process, it is necessary to identify the communication protocol used by each node to ensure accurate decoding and processing of the data during subsequent data collection and analysis. The spatial location of each sensor node is accurately identified. Node location can be determined using the Global Positioning System (GPS) or indoor positioning technologies (such as Wi-Fi positioning, Bluetooth positioning, and ultra-wideband (UWB) positioning). This information provides the necessary coordinate data for subsequent topological analysis and spatial modeling. After identifying and locating all sensor nodes, spatial topological analysis is performed based on their physical locations. Spatial topological relationships generally refer to the connectivity between nodes and their relative positional layout. To perform topological analysis, the geometric positions of the nodes (i.e., X, Y, and Z coordinates) must be analyzed and a spatial graph model constructed based on this information. Proximity relationships between nodes can be identified by calculating the Euclidean distance between them. Clustering algorithms such as K-means and DBSCAN (density-based clustering) can be used for this step to group similar nodes into the same cluster, forming several regions or sub-regions. Node density is a key metric in spatial topological analysis. Counting the number of sensor nodes in each region helps identify areas with dense data collection and those with relatively sparse data collection. Based on this, spatial partitioning can be performed to divide the environment into multiple zones with similar monitoring densities and functions. For example, one zone may be characterized by frequent temperature and humidity fluctuations, while another may be characterized by more drastic changes in air pressure. Optimize spatial topology to ensure effective communication and collaboration between sensor nodes within each area. Graph theory methods such as the shortest path algorithm and the minimum spanning tree algorithm can be used to optimize node connectivity and ensure efficient and robust information transfer. Furthermore, network analysis must consider factors such as signal attenuation and communication delay between sensor nodes.To visualize the spatial topological distribution results, a Geographical Information System (GIS) or 3D modeling software (such as Blender or Unity) can be used to generate a spatial layout diagram of the nodes within the region. This helps understand the spatial relationships between nodes and their distribution. Based on the results of the spatial topological distribution analysis, the sensor nodes within the region are divided into different regional models based on factors such as their physical location, measurement data type, and distribution density. Each regional model includes the spatial distribution of sensor nodes within the region, measurement data, and the interactions between nodes. During model construction, finite element models (FEM), physics-based modeling methods (such as CFD (computational fluid dynamics) models), or data-driven modeling methods can be used to establish simulation models for each region based on the environmental characteristics of each region. For example, for areas with dense temperature and humidity sensors, a model simulating heat conduction and airflow can be constructed; for areas with significant pressure fluctuations, a fluid dynamics model can be constructed. When constructing multiple regional models, it is important to ensure that the data from each node within the region can be coordinated. Data fusion techniques (such as Kalman filtering and weighted averaging) can be used to integrate the monitoring data from each sensor node within the region to form a more accurate regional status assessment. Finally, multiple regional models need to be integrated into the overall environmental monitoring system. This can be achieved through a hierarchical modeling approach, where each regional model is treated as an independent module. The overall model simulates the dynamic changes of the environment by connecting the interrelationships between different regions. This integration process can also be optimized using model integration methods in systems engineering, such as system dynamics models.

[0087] Step S2: Performing all-round environmental simulation operation status monitoring on the environmental simulation sample box, and performing nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features;

[0088] In this example, the environmental parameters to be monitored are determined, such as temperature, humidity, air pressure, light intensity, and gas concentration. Appropriate sensors are selected for data collection based on the experimental objectives. Sensors are strategically arranged within the environmental simulation chamber to ensure full coverage of the monitoring area. Sensors should be evenly distributed to capture representative environmental data. A data acquisition system is set up to ensure real-time collection and storage of sensor output data. Data is typically collected using a microcontroller (such as an Arduino or Raspberry Pi) and transmitted to a central database via a communication interface (such as Wi-Fi or Bluetooth). The monitoring system is activated to begin comprehensive environmental status monitoring. Ensure stable system operation and continuous data recording. The monitoring system should record the environmental parameters of each sensor at different time points in real time. A scheduled sampling mechanism, such as sampling every minute, is established to ensure data timeliness. The collected data is stored in a database, typically a relational database (such as MySQL) or a time series database (such as InfluxDB) to accommodate time series data storage requirements. The collected environmental data is cleaned, addressing missing values ​​and outliers to ensure data quality. The data is standardized or normalized to facilitate subsequent analysis. Select appropriate nonlinear association mining methods, such as: Neural Networks: Use deep learning models (such as multilayer perceptrons and LSTMs) to capture nonlinear relationships in the data. Support Vector Regression (SVR): Capture nonlinear features in the data. Random Forest: Analyze nonlinear relationships between features through ensemble learning methods. Use the organized dataset to train the selected model, taking environmental parameters as input, and the model outputs nonlinear association features. Usually, it is divided into training sets and validation sets for cross-validation. Extract nonlinear operating state association features from the trained model, such as the degree of mutual influence between parameters and the importance of key features. Use data visualization tools (such as Matplotlib and Seaborn) to display the extracted nonlinear association features and display the relationship between different environmental parameters in the form of graphs. Record the analysis results of the nonlinear operating state association features in a database for subsequent analysis and application. At the same time, feedback the results to the monitoring system to optimize sensor layout or monitoring parameters.

[0089] Step S3: Dynamically render multiple node area models based on nonlinear operating state association characteristics to construct multiple environmental simulation data twin models;

[0090] In this embodiment, the nonlinear operating state association features extracted in the previous step are checked to ensure the integrity and accuracy of the data, including the characteristic information of different node areas. According to the experimental objectives, the environmental parameters that need to be dynamically rendered (such as temperature, humidity, air pressure, etc.) are selected, and the rendering effects (such as color changes, dynamic graphics, etc.) are designed. A suitable rendering engine is selected, such as: Unity: suitable for complex three-dimensional environment simulation and dynamic interaction, Three.js: for three-dimensional visualization on web pages, Blender: for high-quality visual effects and animation production, and a rendering tool is used to build a three-dimensional model of each node area. According to the position and functional characteristics of the node, the corresponding environmental scene is designed, and the nonlinear operating state association features are mapped to the three-dimensional model. The specific methods include: color coding: using color gradients to represent different states according to parameters such as temperature or humidity (for example, red Indicates high temperature, blue indicates low temperature), shape change: dynamically adjust the shape or size of the model according to the change of parameters to enhance the visualization effect, and write scripts so that the environmental status can be updated in real time. For example, set a timer to regularly obtain the latest environmental parameters from the data source and update the appearance of the model in real time to achieve dynamic effects. For example, use animation to show the airflow and temperature changes in the node area to enhance the user's perception of environmental changes. Determine the structure of the twin model, including input (environmental parameters), output (rendering effect) and internal logic (data processing and rendering algorithm), integrate the environmental simulation data of multiple node areas into the twin model, ensure that the model can fully represent the overall environmental status, and optimize the parameters of the twin model by comparing the simulation results with the actual environmental data to improve its accuracy and reliability. Conduct verification tests to ensure that the model can accurately reflect changes under different environmental conditions.

[0091] Step S4: Performing operation simulation monitoring on multiple environmental simulation data twin models, and performing parameter difference analysis between models to generate parameter difference characteristics of different models;

[0092] In this embodiment, the input parameters of each environmental simulation data twin model are collected, including environmental conditions, sensor data and other related variables, so as to perform a unified simulation operation, determine the key parameters that need to be monitored, such as temperature, humidity, air flow velocity, etc., to ensure that these parameters can reflect the operating status and performance of the model, and select appropriate software or tools for operation simulation monitoring. Commonly used tools include: MATLAB / Simulink: for system simulation and dynamic model analysis, Python: combined with NumPy, Pandas and Matplotlib for data processing and visualization, start the environmental simulation data twin model, and perform operation simulation. Each model should be run under the same initial conditions to ensure the effectiveness of the comparison. During the simulation operation, the monitoring data output by each model is collected in real time, and the time series data of key parameters are recorded for subsequent analysis. Evaluate the operating status of each model, monitor its stability, response speed and parameter changes, ensure that the model operates as expected, organize the collected output parameters of each model to form a structured data set for subsequent analysis, and the data set should include the parameter names, timestamps and corresponding values ​​of different models. Select appropriate statistical analysis methods for parameter difference analysis. Commonly used methods include: analysis of variance (ANOVA): compare parameter differences between multiple models, t-test: compare two groups of model parameters, cluster analysis: identify similarities and differences between models, principal component analysis (PCA): dimensionality reduction analysis to identify the main components of parameter differences, and based on the results of parameter difference analysis, put forward model optimization suggestions to support subsequent environmental monitoring and management decisions, collect user feedback on the analysis results to evaluate the effectiveness and operability of the results, and make further model adjustments and optimizations based on the feedback.

[0093] Step S5: performing a local model abnormality feature analysis based on the parameter difference characteristics of different models, and performing a potential environmental simulation fault analysis to obtain potential environmental simulation fault data of the abnormal area;

[0094] In this embodiment, based on historical operating data and industry standards, the threshold of parameter abnormality characteristics is defined. For example, the standard deviation, Z-score, or quantile based on data distribution is used to set the standard for abnormal values. A suitable abnormality detection algorithm is selected, such as: Isolation Forest: used for outlier detection of high-dimensional data, Local Outlier Factor (LOF): suitable for detecting local anomalies, Support Vector Machine (SVM): boundary-based anomaly detection. The integrated parameter difference feature data is input into the selected anomaly detection model to identify abnormal features. The abnormality detection results of each local model are recorded, including the abnormality type, severity, and impact range. The detected abnormal features are organized into a data set containing node ID, abnormality type, detection time, and related parameters to facilitate subsequent analysis and determine the appropriate potential fault analysis technology. For example: Fault Tree Analysis (FTA): used to identify the causes and consequences of failures; Event Tree Analysis (ETA): used to analyze the consequences of failures and their paths; Data-based fault prediction models: use historical data for predictions, apply selected fault analysis methods to associate abnormal characteristics with failure modes, and identify potential causes of failures. For example, use fault tree analysis to build a fault model, identify the root cause of the abnormality, organize the analysis results, generate a potential environmental simulation fault data report on the abnormal area, record the fault type, potential cause, affected area and response measures, and propose improvement suggestions and preventive measures based on the results of potential fault analysis to support system maintenance and management decisions, collect user feedback on fault analysis results, evaluate its effectiveness, and continuously optimize the analysis process and model parameters based on feedback.

[0095] Step S6: Inferring environmental fault factors from the potential environmental simulated fault data in the abnormal area, making dynamic adaptive fault diagnosis decisions, and building an adaptive fault diagnosis strategy to perform remote monitoring and fault diagnosis operations.

[0096] In this embodiment, the potential environmental simulation fault data of the abnormal area is integrated, including the environmental parameters when the fault occurs (such as temperature, humidity, pressure, airflow, etc.), the time point of the fault occurrence and the corresponding parameter value are recorded to ensure the integrity of the data, and the key features related to the fault are identified through preliminary data analysis (such as correlation analysis and principal component analysis) for subsequent inference. The selected model is trained using the integrated data to infer the environmental factors related to the fault, and the model performance is evaluated to ensure accuracy (cross-validation and other methods can be used). The inference results of each fault factor, including the degree of impact and probability of occurrence, are recorded, and a fault diagnosis model that can be updated in real time is constructed. Appropriate algorithms are selected, such as: fuzzy logic system: the advantages of handling uncertainty and fuzzy information, reinforcement learning: continuously optimizing decision-making strategies based on feedback, and real-time access to environmental monitoring data. Use it as dynamic input to ensure that the model is updated according to the latest data, use dynamic input data and fault factor inference results to generate fault diagnosis decisions, identify the current fault status, provide processing suggestions for different faults, support rapid response, and design adaptive fault diagnosis strategies based on the fault diagnosis results. The strategies include: Fault identification: clarify the characteristics and processing methods of different faults, response strategy: formulate corresponding response measures for different fault conditions (such as issuing alarms, automatically adjusting parameters, notifying maintenance personnel, etc.), evaluate the effectiveness of the diagnosis strategy through real-time monitoring and historical data, adjust and optimize the strategy based on feedback information, upload the constructed adaptive fault diagnosis strategy to the cloud server, ensure remote access and operation, integrate the cloud monitoring system, receive the status information of each environmental simulation model in real time, and use adaptive strategies for fault diagnosis.

[0097] In this embodiment, refer to Figure 2 , is a flowchart of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0098] Step S11: Identify the distributed multi-dimensional sensors in the environmental simulation sample box and mark all the distributed multi-dimensional sensor nodes;

[0099] Step S12: performing node position calculation on all distributed multi-dimensional sensor nodes to generate position information of all sensor nodes;

[0100] Step S13: Perform spatial topological distribution analysis on the location information of all sensor nodes to generate node spatial topological distribution data;

[0101] Step S14: performing spatial topology distribution modeling on the node spatial topology distribution data to construct a node spatial topology distribution map;

[0102] Step S15: dividing the node space topology distribution map into regions to obtain multiple node topology regions;

[0103] Step S16: Perform three-dimensional point cloud modeling on multiple node topology areas to construct multiple node area models.

[0104] In this embodiment, image processing technology or manual inspection is used to identify all distributed multidimensional sensors in the environmental simulation sample box, and the type, function and preliminary position of the sensor are recorded. A unique identifier (ID) is assigned to each identified sensor node to facilitate subsequent data processing and analysis. A node list is created, containing basic information of each sensor (such as type, location, status, etc.). A laser rangefinder, GPS, or 3D scanner is used to accurately measure the spatial position (x, y, z coordinates) of each sensor node. The measured position data is recorded in a database or spreadsheet to ensure the accuracy and completeness of the data. The accuracy of the position calculation of each node is ensured by cross-validating the measurement data. Re-measurement is performed when necessary. An appropriate topological analysis method is selected, such as: an adjacency matrix: used to represent the connection relationship between nodes; a graph theory algorithm: such as the Dijkstra algorithm, used to calculate the shortest path between nodes; based on the node position data, the distance and connection relationship between nodes are calculated to generate topological distribution data, and the analysis results are stored in a database to form a spatial topological distribution data set that can be used for subsequent analysis. Appropriate modeling tools and software (such as Use MATLAB, Python's NetworkX library, or Gephi) to perform spatial topology modeling. Input the spatial topology distribution data into the selected modeling tool, create node and edge models, define node attributes (such as location and type) and edge weights (such as distance and connection strength), generate a node spatial topology distribution map, and intuitively display the connection relationship and spatial layout between nodes. Select an appropriate region partitioning algorithm, such as: K-means clustering: divide the nodes into multiple regions, DBSCAN: a density-based clustering algorithm used to discover regions of arbitrary shapes, apply the selected algorithm to partition the node spatial topology distribution map into regions, and generate multiple node topology regions. Select an appropriate 3D modeling tool (such as PCL (PointCloudLibrary), Blender, or MATLAB) for point cloud modeling. Convert the node position data after region division into point cloud data format, ensuring that the location, attributes, and other information of each point are complete. Use the selected tool to generate a 3D point cloud model, visualize the spatial position and attributes of the nodes, and optimize the generated point cloud model to improve the model's visualization and rendering performance.

[0105] In this embodiment, refer to Figure 3 , is a flowchart of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0106] Step S21: performing all-round environmental simulation operation status monitoring on the environmental simulation sample box based on the distributed multi-dimensional sensor, and extracting multi-dimensional environmental simulation operation monitoring parameters, wherein the multi-dimensional environmental simulation operation monitoring parameters include environmental temperature parameters, environmental humidity parameters, and environmental wind pressure parameters;

[0107] Step S22: performing temperature fluctuation analysis on the ambient temperature parameters to generate ambient temperature fluctuation characteristics;

[0108] Step S23: calculating the fluctuation amplitude of the ambient temperature fluctuation characteristics to generate the ambient temperature fluctuation amplitude;

[0109] Step S24: performing temperature fluctuation amplitude fitting based on the ambient temperature fluctuation amplitude to construct an ambient temperature fluctuation amplitude curve;

[0110] Step S25: Performing environmental state nonlinear correlation mining on the environmental temperature fluctuation amplitude curve, the environmental humidity parameter, and the environmental wind pressure parameter to obtain nonlinear operating state correlation features.

[0111] In this embodiment, distributed multi-dimensional sensors, including temperature sensors, humidity sensors, and wind pressure sensors, are arranged in an environmental simulation specimen box. The sensors are ensured to be evenly distributed to comprehensively monitor the environmental status. A suitable data acquisition system (such as Arduino, Raspberry Pi, LabVIEW, etc.) is selected, the connection between the sensors and the data acquisition device is set up, and a program is written to collect data at a regular interval, for example, reading sensor data once every minute to ensure that the collection frequency can capture environmental changes. The collected environmental monitoring parameters (temperature, humidity, wind pressure) are stored in a database (such as MySQL, MongoDB) in real time. Each data record should include information such as timestamp, sensor ID, parameter value, etc., and the implementation Data backup mechanism to prevent data loss, clean the ambient temperature data, remove outliers caused by sensor failure or environmental interference, you can use methods such as: Z-score method: calculate the Z-score of each data point, set a threshold (such as ±3), and treat data outside this range as outliers, apply smoothing technology (such as moving average) to reduce noise, ensure the smoothness of temperature data, calculate the statistical characteristics of temperature data, such as: mean: reflects the overall temperature level, standard deviation: reflects the degree of temperature fluctuation, maximum and minimum values: identify extreme temperature values, record the time series characteristics of temperature fluctuations, extract the fluctuation pattern within the time window (for example, using sliding window technology), extract the temperature fluctuation features (such as mean, standard deviation) and calculate the temperature fluctuation characteristics (such as mean, standard deviation) The data (standard deviation, fluctuation range, etc.) are stored in the database to form a data set for subsequent analysis. The calculation formula for the fluctuation amplitude is usually defined as: amplitude = maximum value - minimum value, or the standard deviation is used as a measure of the fluctuation amplitude. Apply the above formula to calculate the temperature fluctuation characteristics and generate the temperature fluctuation amplitude for each time period. Select a suitable fitting model: According to the characteristics of the fluctuation amplitude data, select a suitable fitting method: Linear regression model: suitable for simple linear relationships, polynomial regression: suitable for data with nonlinear relationships, better capture fluctuation trends, smoothing spline: used to process complex nonlinear relationships in the data, use statistical software (such as Python's SciPy or R) to fit the temperature fluctuation amplitude. Fit the degree, use a visualization library (such as Matplotlib) to draw the temperature fluctuation amplitude curve, collect data sets of ambient temperature fluctuation amplitude, humidity parameters, and wind pressure parameters, ensure that these data are aligned in time to form a complete feature set, and select suitable nonlinear association mining methods. Commonly used methods are: Support Vector Machine Regression (SVR): suitable for capturing complex nonlinear relationships, Neural Network: Use Multi-Layer Perceptron (MLP) to model nonlinear relationships, Random Forest: Evaluate nonlinear relationships between features, mine important features, use the prepared data set to train the selected model, use cross-validation to evaluate model performance, ensure the robustness and accuracy of the model, and extract nonlinear operating state association features from the trained model.Identify the relationship between ambient temperature, humidity, and wind pressure, and record key indicators such as importance scores.

[0112] In this embodiment, the specific steps of step S25 are:

[0113] Perform spatial humidity distribution analysis on environmental humidity parameters to obtain spatial humidity distribution data of the sample box;

[0114] Conduct long-term periodic change analysis on the humidity distribution data of the sample box space to generate periodic change data of humidity distribution;

[0115] The humidity distribution periodic change data is subjected to humidity change trend evolution, thereby generating humidity change trend evolution characteristics;

[0116] Perform multi-time-frequency decomposition on the environmental wind pressure parameters to obtain wind pressure spectra at different frequencies;

[0117] Perform main frequency component analysis on each spectrum of wind pressure spectra at different frequencies and extract the main frequency component of each spectrum;

[0118] Calculate the power spectrum density of the main frequency component of each spectrum graph to generate the power spectrum density of each spectrum graph;

[0119] Perform power spectrum density distribution identification on the power spectrum density of each spectrum graph to generate wind pressure density distribution data;

[0120] The nonlinear correlation mining of environmental states is performed on the evolution characteristics of humidity change trends, ambient temperature fluctuation amplitude curves and wind pressure density distribution data to obtain the nonlinear operating state correlation characteristics.

[0121] In this embodiment, environmental humidity parameters are collected from distributed multidimensional sensors, ensuring that the data includes timestamps and sensor locations, and spatial interpolation methods (such as Kriging interpolation or inverse distance weighting) are used to perform spatial analysis on the humidity data to generate a spatial distribution map of humidity, record the humidity values ​​of each area, and form spatial humidity distribution data of the sample box. Visualization tools (such as Matplotlib or Plotly) are used to display the humidity spatial distribution map to intuitively display the changes in humidity in the sample box. The humidity distribution data are organized into time series to ensure the continuity and integrity of the data. Fourier transform or wavelet transform is used to perform frequency domain analysis on the humidity data to identify periodic changes and generate humidity distribution periodic change data. The humidity changes in different time periods are recorded and the generated humidity periodic change data are stored in a database for subsequent analysis. Regression analysis methods (such as linear regression or polynomial regression) are used to analyze the humidity distribution periodic change data to identify the change trend, extract important humidity change trend evolution characteristics from the regression model, record the direction and amplitude of the trend, denoise the wind pressure parameter data to ensure data quality, and use short-time Fourier transform to analyze the humidity data. Perform multi-time-frequency decomposition of wind pressure data using STFT or wavelet transform to generate wind pressure spectra at different frequencies. Record the frequency components and amplitudes of each spectra in preparation for main frequency component analysis. Use the Welch method or adaptive spectrum estimation method to calculate the power spectral density (PSD) of each spectra to obtain the power spectral density of each spectra. Store the calculated power spectral density data in a database for subsequent analysis. Perform distribution identification on the power spectral density of each spectra. Use statistical analysis methods (such as histograms or kernel density estimation) to generate wind pressure density distribution data. Collect the evolution characteristics of humidity change trends, ambient temperature fluctuation amplitude curves, and wind pressure density distribution data to ensure data consistency. Select appropriate nonlinear association mining methods, such as random forest regression: used to capture complex nonlinear relationships between features. Neural network: analyze nonlinear relationships between features through multi-layer perceptron (MLP). Use the prepared dataset to train the selected model. Evaluate model performance through cross-validation. Extract nonlinear operating state association features from the trained model. Record key indicators related to environmental status.

[0122] In this embodiment, refer to Figure 4 , is a flowchart of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0123] Step S31: performing environmental simulation feature evolution based on nonlinear operating state correlation features, thereby generating all-round environmental operating state evolution data;

[0124] Step S32: performing regional node state matching on the omnidirectional environment operation state evolution data according to the position information of all sensor nodes, thereby obtaining corresponding state evolution data of the node area;

[0125] Step S33: Use the corresponding state evolution data of the node area to dynamically render the state of multiple node area models and construct multiple environmental simulation data twin models.

[0126] In this embodiment, a suitable mathematical model (such as a state space model or a Markov model) is selected to describe the evolution of the environmental state. The model should be able to capture nonlinear relationships and dynamic changes. For example, a recurrent neural network (RNN) or a long short-term memory network (LSTM) is used to process time series data. A data set based on nonlinear operating state association features is collected and prepared, including parameters such as temperature, humidity, and wind force. The model is trained using a deep learning framework (such as TensorFlow or PyTorch). The input is historical state data and the output is the predicted future state. The model parameters are optimized through a backpropagation algorithm to minimize the prediction error. The trained evolution model is used to predict the future environmental state and generate all-round environmental operating state evolution data, including temperature, humidity, and wind pressure at different time points in the time series. The location information of all sensor nodes is extracted from the previous step to ensure that the coordinates of each node correspond to its function. A spatial interpolation method (such as Kriging interpolation or inverse distance weighting) is used to integrate the all-round environmental operating state evolution data with the sensor nodes. The generated environmental operation state evolution data is matched with the geographic location of the node, and the corresponding state evolution data of each regional node is determined. A database containing the state evolution of the node area is generated, including the state information of each node at each time point. A 3D visualization tool (such as Unity, Blender, or Three.js) is used to create a dynamic state rendering model. The state evolution data corresponding to the node area is imported into the selected rendering tool, ensuring that the data format meets the requirements of the model. Different visualization effects are set according to the node state evolution data. For example, temperature is represented by color changes, red represents high temperature, blue represents low temperature, and humidity and wind pressure are visualized through different graphic effects (such as flow animation). Run the rendering model to observe whether the dynamic effect accurately reflects the state evolution process. Adjust the rendering parameters as needed to optimize the visual effect to make it more intuitive and easy to understand. Display the rendering results to relevant personnel to facilitate environmental monitoring and decision support. The user can view the environmental status at different time points through an interactive interface.

[0127] In this embodiment, step S4 includes the following steps:

[0128] Step S41: performing operation simulation monitoring processing on multiple environmental simulation data twin models, and extracting simulation monitoring data of each twin model;

[0129] Step S42: performing a time series state change trend analysis on the simulated monitoring data of each twin model to generate a time series state change trend of each twin model;

[0130] Step S43: performing situation evolution prediction on the time series state change trend of each twin model, thereby obtaining situation evolution prediction data of each twin model;

[0131] Step S44: Perform inter-model parameter difference analysis on each twin model situation evolution prediction data to generate parameter difference characteristics of different models.

[0132] In this embodiment, it is ensured that all environmental simulation data twin models have been configured and have all the input parameters and initial conditions required for operation. According to the characteristics of different models, the time step, simulation cycle and other relevant parameters of the simulation operation are set, and the simulation operation of each twin model is started. The monitoring process is usually implemented through programming languages ​​​​(such as Python, MATLAB) or dedicated software (such as Simulink). During the simulation operation, the output monitoring data of each twin model is recorded in real time, including key parameters such as temperature, humidity, and wind speed. The data is stored in the form of a time series to ensure that the state data of each time step is complete. The extracted monitoring data is cleaned, missing values ​​​​and outliers are processed to ensure data quality. The monitoring data of each twin model is organized into a time series format for subsequent analysis. Statistical methods (such as moving average and exponential smoothing) are used to analyze the state change trend of each twin model, and the change rate and growth rate of each time period are calculated to identify trends. Data visualization tools (such as Matplotlib and Seaborn) are used to draw the state change trend chart of each twin model to show the changes in key parameters. Ensure that the chart is clear and easy to understand, and mark important change points and trends. According to the characteristics of the data, select a suitable time series prediction model, such as :Autoregressive integrated moving average (ARIMA) model: suitable for linear time series data, long short-term memory network (LSTM): suitable for processing complex nonlinear relationships and long-term dependent data, divide the state change trend data into training set and test set, usually use 70% of the data for training and 30% of the data for verification, use the selected prediction algorithm to train the training data, optimize the model parameters, ensure that the model can accurately reflect the change trend of the data, use the trained model to predict the future state, generate the situation evolution prediction data of each twin model, and use visualization tools to draw the situation evolution prediction of each twin model A graph is produced to show the comparison between the prediction results and the actual data. The situation evolution prediction data of all twin models are organized into a comprehensive data set, including the key parameters of each model. Statistical analysis methods (such as t-test, analysis of variance ANOVA) or machine learning methods (such as cluster analysis) are used to compare the parameter differences between different models. The parameters of each model are compared, and the features of significant differences are identified. The parameter difference features are recorded and a difference analysis report is generated, including the parameter mean, standard deviation and significance of each model. Visualization tools are used to draw charts of the difference analysis results (such as box plots and bar charts) to intuitively display the parameter differences between different models.

[0133] In this embodiment, step S5 includes the following steps:

[0134] Step S51: performing local model abnormality feature analysis based on parameter difference features of different models, and extracting local model abnormality feature data;

[0135] Step S52: locating abnormal features of the abnormal feature data of the local model and extracting the abnormal node area model;

[0136] Step S53: Preview the environmental parameters at multiple time points on the abnormal node area model to obtain the environmental preview parameters at multiple time points;

[0137] Step S54: performing potential environmental simulation fault analysis on the environmental rehearsal parameters at multiple time points to obtain potential environmental simulation fault data of abnormal areas.

[0138] In this embodiment, parameter difference feature data of different models are integrated into a data set, including key parameters of each model and their differences. The threshold of abnormal features, such as standard deviation multiples, Z-score, etc., is determined based on historical data to define under what circumstances the parameters are considered abnormal. A machine learning algorithm (such as isolation forest, local outlier factor (LOF) or support vector machine (SVM)) is used to detect abnormal features of the local model. The integrated data is input into the selected anomaly detection algorithm to identify abnormal features and extract abnormal feature data of the local model. The type and severity of the abnormal features are recorded for subsequent analysis. Based on the severity and type of the abnormal features, positioning criteria are set, such as the threshold of the outlier value, the range of influence, etc. The specific abnormal node area is identified based on the extracted abnormal feature data. The abnormal features are associated with the node area using geographic information system (GIS) tools or spatial analysis methods to determine the affected area and generate a model data set containing the abnormal node area. According to the data, the environmental parameters and status information of these areas are recorded to facilitate subsequent processing. According to the required accuracy and complexity, a suitable pre-rehearsal model is selected, such as a physics-based simulation model or a data-driven prediction model. The necessary initial conditions and parameters, including the environmental status data within the time period, are collected to input into the pre-rehearsal model. Multi-point environmental parameter pre-rehearsals are performed on the selected abnormal node area model, usually in units of hours, days or weeks. The changes in environmental parameters at each time point, such as temperature, humidity, wind speed, etc., are monitored and recorded. Fault tree analysis (FTA), event tree analysis (ETA) or data-driven methods, such as anomaly detection and prediction models, are used to analyze the potential factors leading to failures based on the multi-point environmental parameter pre-rehearsal data, determine which parameter changes will trigger environmental simulation failures, and record the characteristics of these potential failures. The analysis results are sorted out to generate a data report on the potential environmental simulation failures in the abnormal area, including the failure type, impact range and severity assessment.

[0139] In this embodiment, step S6 includes the following steps:

[0140] Step S61: inferring environmental fault factors based on potential environmental simulated fault data in the abnormal area to obtain environmental fault factor data;

[0141] Step S62: generating a fault warning signal based on the environmental fault factor data;

[0142] Step S63: making dynamic adaptive fault diagnosis decisions based on the fault warning signal and constructing an adaptive fault diagnosis strategy;

[0143] Step S64: Upload the adaptive fault diagnosis strategy to the cloud server to perform remote monitoring and fault diagnosis operations.

[0144] In this embodiment, the potential environmental simulation fault data extracted from the previous step is integrated to ensure data integrity and consistency, including the time of fault occurrence, environmental parameter changes and abnormal characteristics. A suitable fault factor inference model is selected, such as a causal inference model, a Bayesian network or a decision tree, to analyze the relationship between potential faults and environmental factors. The integrated potential fault data is input into the selected inference model to analyze how each environmental factor affects the occurrence of the fault, identify key factors, evaluate the degree of influence of each environmental factor on the occurrence of the fault, generate environmental fault factor data, record the factors with greater influence and their weights, and organize the inference results into a structured data set, including the name of each fault factor, the degree of influence and the corresponding parameter value. Based on historical data and industry standards, an early warning threshold is set for each environmental fault factor, and the circumstances under which the early warning is triggered are defined. The data inferred from the fault factor is integrated with the early warning standard to prepare for generating an early warning signal. A monitoring system is developed to monitor the environmental fault factor data in real time. When a factor exceeds the set threshold, a fault early warning signal is generated. The generated early warning signal is formatted into an easy-to-understand format, including the signal timestamp, fault type and severity. The system collects generated fault warning signals and combines them with the system's historical fault data. A dynamic adaptive fault diagnosis strategy framework is designed based on machine learning algorithms (such as decision trees, random forests, or neural networks). Historical fault data and warning signals are used to train the adaptive diagnosis model, and model parameters are optimized to improve diagnostic accuracy. Upon receiving a new fault warning signal, the diagnostic model is automatically triggered for real-time analysis, generating corresponding fault diagnosis decisions (such as recommended repair measures and emergency response plans). Each generated fault diagnosis decision is recorded in the system database for subsequent tracking and auditing. The system ensures that the cloud server is configured with the necessary software environment and can process uploaded diagnostic strategies and data. The adaptive fault diagnosis strategy is organized into a format suitable for cloud storage, ensuring the clarity and completeness of the data structure. The adaptive fault diagnosis strategy is uploaded to the cloud server using an API or data transmission protocol (such as an HTTP POST request). The uploaded data is confirmed to be correct on the cloud server, ensuring that the strategy can be accessed by remote monitoring and fault diagnosis systems. A monitoring mechanism is set up in the cloud system to track the implementation of fault warning signals and diagnostic decisions in real time to ensure efficient system operation.

[0145] In this embodiment, a remote monitoring and fault diagnosis system for an environmental simulation system is provided, which is used to execute the remote monitoring and fault diagnosis method for the environmental simulation system described above, including:

[0146] The spatial topology distribution module is used to identify all distributed multi-dimensional sensor nodes, perform spatial topology distribution analysis, and build multiple node area models;

[0147] The nonlinear mining module is used to conduct all-round environmental simulation operation status monitoring of the environmental simulation sample box and conduct nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features;

[0148] Dynamic state rendering module, used to dynamically render multiple node area models based on nonlinear operating state association characteristics and build multiple environmental simulation data twin models;

[0149] The parameter difference module is used to perform operation simulation monitoring on multiple environmental simulation data twin models and perform parameter difference analysis between models to generate parameter difference characteristics of different models;

[0150] The potential fault analysis module is used to analyze the abnormal characteristics of local models based on the parameter difference characteristics of different models, and to analyze the potential environmental simulation faults to obtain the potential environmental simulation fault data of the abnormal area;

[0151] The fault diagnosis decision module is used to infer environmental fault factors based on potential environmental simulated fault data in abnormal areas, make dynamic adaptive fault diagnosis decisions, and build adaptive fault diagnosis strategies to perform remote monitoring and fault diagnosis operations.

[0152] The present invention helps to establish the cognition of the overall structure of the system by identifying distributed multi-dimensional sensor nodes and performing spatial topological distribution analysis, improves the visualization and management efficiency of the system, and constructs multiple node area models to help understand the correlation between nodes, providing a basis for subsequent environmental simulation and monitoring. Nonlinear correlation mining of environmental states can reveal the complex nonlinear relationship between environmental parameters, help understand the deep-level characteristics of the system operation state, and obtain nonlinear operation state correlation characteristics, which can provide important clues for subsequent model establishment and fault diagnosis. Dynamic state rendering based on nonlinear operation state correlation characteristics can more intuitively show the operation state changes of multiple node area models, help identify abnormal situations, and construct environmental simulation data twin models to help simulate the system operation state under different conditions, provide a reference for fault prediction and analysis, and improve the environmental simulation data. Parameter difference analysis based on twin models can help compare the operating conditions of different models, identify potential abnormal characteristics, and provide support for fault diagnosis. Generating parameter difference characteristics of different models helps to understand the performance differences of the system under different conditions and guide subsequent fault analysis and repair. Local model abnormal characteristic analysis based on parameter difference characteristics can effectively identify local abnormal conditions and help discover potential fault points in advance. Potential environmental simulation fault analysis can help predict fault conditions in abnormal areas and provide guidance for fault troubleshooting and repair. Environmental fault factor inference and dynamic adaptive fault diagnosis decision-making can help quickly and accurately diagnose the cause of the fault and reduce the impact of system failures. Building an adaptive fault diagnosis strategy and performing remote monitoring and fault diagnosis operations can improve the stability and reliability of the system and reduce the interference of faults on the normal operation of the system.

[0153] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0154] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A remote monitoring and fault diagnosis method for an environmental simulation system, characterized in that: The following steps are involved: Step S1: Identify all distributed multi-dimensional sensor nodes, perform spatial topological distribution analysis, and construct multiple node area models; Step S2: Performing all-round environmental simulation operation status monitoring on the environmental simulation sample box, and performing nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features; Step S3: Dynamically render multiple node area models based on nonlinear operating state association characteristics to construct multiple environmental simulation data twin models; Step S4: Performing operation simulation monitoring on multiple environmental simulation data twin models, and performing parameter difference analysis between models to generate parameter difference characteristics of different models; Step S5: performing a local model abnormality feature analysis based on the parameter difference characteristics of different models, and performing a potential environmental simulation fault analysis to obtain potential environmental simulation fault data of the abnormal area; Step S6: Inferring environmental fault factors based on potential environmental simulated fault data in the abnormal area, making dynamic adaptive fault diagnosis decisions, and building an adaptive fault diagnosis strategy to perform remote monitoring and fault diagnosis operations; Among them, the specific steps of step S1 are: Step S11: Identify the distributed multi-dimensional sensors in the environmental simulation sample box and mark all the distributed multi-dimensional sensor nodes; Step S12: performing node position calculation on all distributed multi-dimensional sensor nodes to generate position information of all sensor nodes; Step S13: Perform spatial topological distribution analysis on the location information of all sensor nodes to generate node spatial topological distribution data; Step S14: performing spatial topology distribution modeling on the node spatial topology distribution data to construct a node spatial topology distribution map; Step S15: dividing the node space topology distribution map into regions to obtain multiple node topology regions; Step S16: performing three-dimensional point cloud modeling on multiple node topology areas to construct multiple node area models; Among them, the specific steps of step S2 are: Step S21: performing all-round environmental simulation operation status monitoring on the environmental simulation sample box based on the distributed multi-dimensional sensor, and extracting multi-dimensional environmental simulation operation monitoring parameters, wherein the multi-dimensional environmental simulation operation monitoring parameters include environmental temperature parameters, environmental humidity parameters, and environmental wind pressure parameters; Step S22: performing temperature fluctuation analysis on the ambient temperature parameters to generate ambient temperature fluctuation characteristics; Step S23: calculating the fluctuation amplitude of the ambient temperature fluctuation characteristics to generate the ambient temperature fluctuation amplitude; Step S24: performing temperature fluctuation amplitude fitting based on the ambient temperature fluctuation amplitude to construct an ambient temperature fluctuation amplitude curve; Step S25: Performing environmental state nonlinear correlation mining on the ambient temperature fluctuation amplitude, the curve ambient humidity parameter, and the ambient wind pressure parameter to obtain nonlinear operating state correlation features.

2. The method according to claim 1, characterized in that The specific steps of step S25 are: Perform spatial humidity distribution analysis on environmental humidity parameters to obtain spatial humidity distribution data of the sample box; Conduct long-term periodic change analysis on the humidity distribution data of the sample box space to generate periodic change data of humidity distribution; The humidity distribution periodic change data is subjected to humidity change trend evolution, thereby generating humidity change trend evolution characteristics; Perform multi-time-frequency decomposition on the environmental wind pressure parameters to obtain wind pressure spectra at different frequencies; Perform main frequency component analysis on each spectrum of wind pressure spectra at different frequencies and extract the main frequency component of each spectrum; Calculate the power spectrum density of the main frequency component of each spectrum graph to generate the power spectrum density of each spectrum graph; Perform power spectrum density distribution identification on the power spectrum density of each spectrum graph to generate wind pressure density distribution data; The nonlinear correlation mining of environmental states is performed on the evolution characteristics of humidity change trends, ambient temperature fluctuation amplitude curves and wind pressure density distribution data to obtain the nonlinear operating state correlation characteristics.

3. The method according to claim 1, characterized in that The specific steps of step S3 are: Step S31: performing environmental simulation feature evolution based on nonlinear operating state correlation features, thereby generating all-round environmental operating state evolution data; Step S32: performing regional node state matching on the omnidirectional environment operation state evolution data according to the position information of all sensor nodes, thereby obtaining corresponding state evolution data of the node area; Step S33: Use the corresponding state evolution data of the node area to dynamically render the state of multiple node area models and construct multiple environmental simulation data twin models.

4. The remote monitoring and fault diagnosis method of an environmental simulation system according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: performing operation simulation monitoring processing on multiple environmental simulation data twin models, and extracting simulation monitoring data of each twin model; Step S42: performing a time series state change trend analysis on the simulated monitoring data of each twin model to generate a time series state change trend of each twin model; Step S43: performing situation evolution prediction on the time series state change trend of each twin model, thereby obtaining situation evolution prediction data of each twin model; Step S44: Perform inter-model parameter difference analysis on each twin model situation evolution prediction data to generate parameter difference characteristics of different models.

5. The remote monitoring and fault diagnosis method of an environmental simulation system according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing local model abnormality feature analysis based on parameter difference features of different models, and extracting local model abnormality feature data; Step S52: locating abnormal features of the abnormal feature data of the local model and extracting the abnormal node area model; Step S53: Preview the environmental parameters at multiple time points on the abnormal node area model to obtain the environmental preview parameters at multiple time points; Step S54: performing potential environmental simulation fault analysis on the environmental rehearsal parameters at multiple time points to obtain potential environmental simulation fault data of abnormal areas.

6. The remote monitoring and fault diagnosis method of an environmental simulation system according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: inferring environmental fault factors based on potential environmental simulated fault data in the abnormal area to obtain environmental fault factor data; Step S62: generating a fault warning signal based on the environmental fault factor data; Step S63: making dynamic adaptive fault diagnosis decisions based on the fault warning signal and constructing an adaptive fault diagnosis strategy; Step S64: Upload the adaptive fault diagnosis strategy to the cloud server to perform remote monitoring and fault diagnosis operations.

7. A remote monitoring and fault diagnosis system for an environmental simulation system, characterized in that: A remote monitoring and fault diagnosis method for an environmental simulation system according to claim 1, comprising: The spatial topology distribution module is used to identify all distributed multi-dimensional sensor nodes, perform spatial topology distribution analysis, and build multiple node area models; The nonlinear mining module is used to conduct all-round environmental simulation operation status monitoring of the environmental simulation sample box and conduct nonlinear correlation mining of the environmental status to obtain nonlinear operation status correlation features; Dynamic state rendering module, used to dynamically render multiple node area models based on nonlinear operating state association characteristics and build multiple environmental simulation data twin models; The parameter difference module is used to perform operation simulation monitoring on multiple environmental simulation data twin models and perform parameter difference analysis between models to generate parameter difference characteristics of different models; The potential fault analysis module is used to analyze the abnormal characteristics of local models based on the parameter difference characteristics of different models, and to analyze the potential environmental simulation faults to obtain the potential environmental simulation fault data of the abnormal area; The fault diagnosis decision module is used to infer environmental fault factors based on potential environmental simulated fault data in abnormal areas, make dynamic adaptive fault diagnosis decisions, and build adaptive fault diagnosis strategies to perform remote monitoring and fault diagnosis operations.

Images