Ventilator intelligent operation and maintenance analysis method and system based on big data
Through the intelligent operation and maintenance analysis method of big data, edge computing and cloud processing, combined with machine learning and graph parallel computing, the shortcomings of fault diagnosis and performance optimization in the operation and maintenance analysis of the ventilation fan are solved, and more accurate fault identification and equipment health are achieved. Predicted, reducing downtime and maintenance costs.
Patent Information
- Application Number
- CN202411473761.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-22
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing ventilation fan operation and maintenance analysis technology has shortcomings in fault diagnosis and performance optimization, especially in complex operating conditions, the diagnosis accuracy is low and the system optimization model and algorithm support is lacking.
Using intelligent operation and maintenance analysis methods based on big data, the real-time data acquisition of data by deploying monitoring sensors, edge computing and cloud processing are used to establish a distributed operation and maintenance information database, and the graph model is constructed using machine learning and graph parallel computing to evaluate the ventilation fan equipment. Running status, and display the fault propagation path and running status through visual tools.
It improves the real-time and accuracy of operation and maintenance decisions, can identify potential faults and influencing factors, predict equipment health status in advance, reduce downtime and maintenance costs, and improve operation and maintenance efficiency and equipment reliability.
Smart Images

Figure CN118982348A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of ventilator operation and maintenance, and in particular to a ventilator intelligent operation and maintenance analysis method and system based on big data. Background Art
[0002] A ventilator is a device that achieves air flow transmission and distribution through mechanical means. It is widely used in industrial, commercial, agricultural and residential environments to meet the needs of ventilation, exhaust and air conditioning. Its technical fields cover mechanical engineering, thermodynamics, fluid mechanics and control system engineering. According to its structure and functional principle, ventilators can be divided into centrifugal, axial flow, mixed flow and cross flow types. Its basic functional principle is to use the rotating impeller to apply mechanical energy to the air to make the air flow.
[0003] Specifically, the fan drives the impeller to rotate through the motor, and the blades of the impeller exert centrifugal force or axial force on the air, so that the air is sucked in from the air inlet and discharged from the air outlet. In a centrifugal fan, the air flows radially, and the centrifugal force generated by the rotation of the impeller throws the air to the surroundings, thereby generating high-pressure airflow. In an axial flow fan, the air flows axially, and the axial force generated by the rotation of the impeller pushes the air forward, which is suitable for occasions that require large air volume and low air pressure. Mixed flow fans combine the characteristics of centrifugal and axial flow fans, which can generate higher wind pressure and provide larger air volume. Cross flow fans generate wind pressure through the lateral flow of the impeller, and are often used in air conditioners and air purifiers.
[0004] The application scenarios of ventilators are very wide. In the industrial field, ventilators are used in ventilation and exhaust systems to improve the air quality in the workshop, remove harmful gases and dust, and protect the health and safety of workers. In commercial and public buildings, ventilators are an important part of the heating, ventilation and air conditioning system (HVAC), used to adjust indoor temperature and humidity and provide a comfortable environment. In addition, ventilators are used in agriculture to ventilate livestock houses, help regulate temperature and humidity, and promote the healthy growth of animals. In residential environments, ventilators are used for exhaust and indoor ventilation in kitchens and bathrooms to improve the air quality of the living environment.
[0005] The operation and maintenance analysis of ventilators mainly involves equipment monitoring, fault diagnosis, maintenance and performance optimization. The operation and maintenance analysis of existing technologies usually relies on sensors and data analysis technology to monitor the operating status of ventilators in real time, detect and diagnose faults in a timely manner, and reduce downtime and maintenance costs. However, the existing technologies still have some shortcomings and areas for improvement in operation and maintenance analysis.
[0006] For example, fault diagnosis is an important part of operation and maintenance analysis. Existing technologies mostly use rule- and model-based fault diagnosis methods. The diagnostic accuracy of these methods under complex working conditions needs to be improved. Especially when multiple faults coexist and the fault types are complex, traditional diagnostic methods often have difficulty locating problems quickly and accurately. Moreover, the performance optimization of ventilators involves multiple aspects such as air flow, wind pressure, and energy consumption. Existing optimization technologies mainly rely on experience and trial and error, and lack systematic optimization models and algorithm support. With the development of artificial intelligence and big data technologies, data-driven optimization methods will become the future development direction. Summary of the invention
[0007] Based on this, it is necessary to provide a big data-based intelligent operation and maintenance analysis method and system for ventilators to address the above technical issues.
[0008] In a first aspect, the present invention provides a method for intelligent operation and maintenance analysis of a ventilator based on big data, the method comprising the following steps: S1. Deploy monitoring sensors to collect real-time operating data of ventilator equipment; S2. Use edge computing devices to preliminarily process the operation data, and transmit the operation data that meets the preset standards to the cloud through the communication network; S3. Establish a distributed architecture operation and maintenance information database to store real-time uploaded operation data and historical operation data, and use machine learning to revise and update the operation and maintenance information database; S4, extracting the state characteristics of the operation data, defining nodes and edges, building a graph model, and evaluating the operation state of the ventilator equipment through graph parallel computing and analysis; S5. Use visualization tools to display graph data and evaluation results, provide fault propagation path and operation status monitoring, and regulate and maintain the control parameters and body status of the ventilator equipment.
[0009] Further, using the edge computing device to preliminarily process the operation data and transmitting the operation data that meets the preset standards to the cloud through the communication network includes the following steps: S21, reading the real-time collected operation data, and removing non-numeric values contained in the operation data; S22. Monitor the operating conditions of the ventilator equipment in real time, align the operating data from different sources according to the timestamp information, and exclude the operating data during the equipment shutdown period; S23, using a sliding window method to remove non-steady-state values and noise values in the operating data; S24, performing non-dimensionalization and normalization processing on the operation data; S25, defining the screening process of steps S21 to S24 as a preset standard, and then transmitting the operating data that meets the preset standard to the cloud through the communication network.
[0010] Furthermore, the state characteristics of the operating data are extracted, and nodes and edges are defined to construct a graph model. Through graph parallel computing and analysis, the operating state of the ventilator equipment is evaluated, including the following steps: S41, establishing a feature extraction parallel channel to synchronize file data in the operation and maintenance information library; S42. Establish a feature extraction model based on an artificial intelligence algorithm, analyze the types and values of operating parameters in the operating data, and extract state characteristics of the ventilator equipment; wherein the state characteristics include basic parameter characteristics, parameter level characteristics, operating trend characteristics, and parameter fluctuation characteristics; S43, converting the operation data and state characteristics into graph data, defining nodes and edges and their attributes, and constructing a graph model; S44. Select a graph computing framework, evaluate the operating status and health index of the ventilator equipment based on the graph computing results, and identify key nodes and potential faults.
[0011] Furthermore, establishing a parallel channel for feature extraction and synchronizing file data in the operation and maintenance information library includes the following steps: S411, establishing multiple feature extraction parallel channels in the operation and maintenance database, each feature extraction parallel channel is randomly divided into a number of segments, and each segment is distributed and stores multiple data nodes; S412, using the operation and maintenance information library to store the operation data and historical operation data in segments, and each segment of data has timestamp information and is randomly distributed in the data nodes of each segment; S413, according to the running data with the same timestamp information, multiple feature extraction parallel channels are used to synchronously extract, a multi-channel parallel computing task is established, and the file data in each channel is synchronously processed.
[0012] Furthermore, a feature extraction model is established based on an artificial intelligence algorithm to analyze the types and values of operating parameters in the operating data, and the state characteristics of the ventilator equipment are extracted, including the following steps: S421. Extracting basic parameter features based on the types of operating parameters included in the operating data; S422, setting a classification threshold, comparing the threshold level corresponding to the value of each type of operating parameter at the current moment in turn, and determining the parameter level characteristics of the operating parameter at the current moment; S423, set the sampling period, arbitrarily select an operating parameter, and select N operating parameters before and after the operating parameter as windows for data aggregation calculation and feature generation, calculate the arithmetic mean respectively, and then perform deviation calculation on the 2N operating parameters to obtain the deviation actual value, and obtain the trend data by subtracting the two deviation actual values before and after, and finally obtain the operating trend feature through cyclic calculation; S424. Collect differential values of operating parameters in the operating data, use a time series model to perform fluctuation analysis on the operating parameters, and obtain parameter fluctuation characteristics.
[0013] Furthermore, the calculation formula for trend data includes: In the formula, Represents trend data; Before N The deviation of an operating parameter from the actual value; After indicating N The deviation of an operating parameter from the actual value; Represents an arbitrarily selected operating parameter; Indicates the front of an arbitrarily selected operating parameter N operating parameters; Indicates the first operating parameter of an arbitrarily selected operating parameter; Indicates the value of an arbitrarily selected operating parameter. N operating parameters; Represents the arithmetic mean.
[0014] Furthermore, the running data and state features are converted into graph data, and nodes, edges and their attributes are defined. Building a graph model includes the following steps: S431, defining node types and node attributes thereof, wherein the node types include sensor nodes, device component nodes, and operation status nodes; S432, defining edge types and edge attributes, wherein edge types include sensor-component edges, component-component edges, and component-state edges; S433, converting the operation data and status characteristics into a graph data format, and creating a graph database in the operation and maintenance information library to store and manage the graph data; S434. Use the application programming interface in the graph database to create nodes and add corresponding node attributes; then define the edges between nodes and add edge attributes; create nodes and edges through batch import to build a complete graph model.
[0015] Furthermore, the graph computing framework is selected, and based on the graph computing results, the operating status and health index of the ventilator equipment are evaluated, and key nodes and potential faults are identified, including the following steps: S441. Import the graph model into the graph computing framework using an import tool in the graph database; S442, using a graph traversal algorithm to traverse the graph model, extract relevant features of nodes and edges, construct a feature vector, and use a community detection algorithm to detect communities in the graph model, identify node clusters with similar operating status, and discover potential fault areas; S443, using a graph isomorphism matching algorithm to match subgraph patterns, identify potential fault points, and use a single-source shortest path algorithm to calculate the shortest path between key nodes and analyze the fault propagation path; S444. Based on the extracted feature vector and its weight, the health index of the ventilator equipment is calculated, and a health assessment report is generated to evaluate the current operating status of the ventilator equipment.
[0016] Furthermore, the calculation formula of the health index is: In the formula, HI It represents the health index; Indicates i The weight of the feature vector; Indicates i The nonlinear transformation function of the eigenvectors; Indicates i The time decay factor of the eigenvectors; Indicates i The health threshold function of the feature vector; represents the weight factor of the interaction term; represents the mutual terms between eigenvectors.
[0017] In a second aspect, the present invention further provides a ventilator intelligent operation and maintenance analysis system based on big data, the system comprising: A monitoring module is used to collect the operation data of the ventilator equipment in real time through monitoring sensors; A processing module, used to preliminarily process the operation data using the edge computing device, and transmit the operation data that meets the preset standards to the cloud through the communication network; The storage module is used to establish a distributed architecture operation and maintenance information library, store real-time uploaded operation data and historical operation data, and use machine learning to correct and update the operation and maintenance information library; The analysis module is used to extract the status characteristics of the operating data, define nodes and edges, build a graph model, and evaluate the operating status of the ventilator equipment through graph parallel computing analysis; The display module is used to display graph data and evaluation results using visualization tools, provide fault propagation path and operation status monitoring, and regulate and maintain the control parameters and body status of the ventilator equipment.
[0018] The beneficial effects of the present invention are: 1. Through edge computing and distributed architecture, the efficiency of data processing and storage is greatly improved, the system response time is reduced, and the operation and maintenance decisions are made more real-time and accurate. With the application of machine learning, the operation and maintenance information database can learn from big data and continuously optimize, making the analysis results more reliable and durable. In addition, through graph models and parallel computing technology, the analysis of equipment status is more comprehensive and in-depth, and potential failure points and influencing factors can be identified, and the health status of equipment can be predicted in advance, thereby reducing equipment downtime and maintenance costs. It can help operation and maintenance personnel make decisions quickly and take necessary maintenance measures, further improving the overall operation and maintenance efficiency and equipment reliability.
[0019] 2. By displaying graph data and evaluation results through visualization tools, the equipment operating status and fault propagation path are clear at a glance, which is convenient for operation and maintenance personnel to make decisions and operations; at the same time, through real-time monitoring and status evaluation, the operating status of the equipment can be fully understood, problems can be discovered and solved in a timely manner, equipment downtime can be reduced, and maintenance costs can be reduced; the decision-making ability and operating efficiency of operation and maintenance personnel are improved, and it has significant technical advantages and application value. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 is a flow chart of a method for intelligent operation and maintenance analysis of a ventilator based on big data according to an embodiment of the present invention; Figure 2 It is a system principle block diagram of a ventilator intelligent operation and maintenance analysis system based on big data according to an embodiment of the present invention.
[0021] Figure numbers: 1. Monitoring module; 2. Processing module; 3. Storage module; 4. Analysis module; 5. Display module. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. 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.
[0023] See also Figure 1 , provides a ventilator intelligent operation and maintenance analysis method based on big data, the method comprising the following steps: S1. Deploy monitoring sensors to collect the operating data of the ventilator equipment in real time.
[0024] Specifically, the operating data of the ventilator equipment refers to the parameter data collected by the sensor, which includes various types of operating parameters. The types of operating parameters include the following aspects: 1. Air volume: Air volume is the volume of air delivered by the fan, usually expressed in cubic meters per hour (m³ / h) or cubic meters per second (m³ / s). The size of the air volume directly affects the ventilation effect. The air volume can be controlled by adjusting the angle of the impeller, changing the speed or adjusting the size of the air inlet.
[0025] 2. Wind pressure: Wind pressure is the pressure generated by the fan to overcome the system resistance, usually expressed in Pascal (Pa) or millimeter water column (mmH2O). The control of wind pressure can be achieved by changing the design of the impeller or adjusting the resistance of the ventilation duct.
[0026] 3. Speed: The speed is the rotation speed of the fan impeller, usually expressed in revolutions per minute (RPM). The speed can be precisely controlled by adjusting the frequency of the motor through the frequency converter.
[0027] 4. Power: Power is the electrical energy consumed by the fan when it is running, usually expressed in watts (W) or kilowatts (kW). The size of the power directly affects the energy efficiency of the fan. Energy consumption can be reduced by optimizing the design and control.
[0028] 5. Efficiency: Efficiency is the ratio of the fan output power to the input power, indicating the effectiveness of the fan energy conversion. Efficiency can be improved by improving the impeller design and material selection.
[0029] 6. Noise: Noise is the sound produced when the fan is running, usually expressed in decibels (dB). The noise level can be reduced by adopting noise reduction materials and optimizing structural design.
[0030] 7. Temperature: Temperature is the heat of the ventilator and its surroundings. Excessive temperature may cause the equipment to overheat and be damaged. By installing temperature sensors and cooling systems, the temperature can be monitored and controlled.
[0031] S2. Use edge computing devices to preliminarily process operating data, and transmit operating data that meets preset standards to the cloud through the communication network.
[0032] Among them, edge computing devices (such as industrial PCs and embedded devices) are deployed near the ventilators to preliminarily process sensor data. Edge computing devices can perform preliminary processing such as data filtering, denoising, and anomaly detection to ensure data quality; at the same time, they can cache data temporarily to prevent data loss when communication is interrupted.
[0033] In the description of the present invention, using edge computing devices to preliminarily process operating data and transmitting operating data that meets preset standards to the cloud through a communication network includes the following steps: S21. Read the real-time collected operation data, and remove non-numeric values (NaN value, Not aNumber) contained in the operation data.
[0034] S22. Monitor the operating conditions of the ventilator equipment in real time, align the operating data from different sources according to the timestamp information, and eliminate the operating data during the equipment shutdown period.
[0035] S23. Use the sliding window method to eliminate non-steady-state values and noise values in the operating data.
[0036] Specifically, since there are a lot of monitored operating data and historical operating data, it is inevitable that there are non-steady-state values and noise values in the data. These data need to be processed, and the more commonly used sliding window method is used to eliminate the non-steady-state values. The 3r principle of the sliding window method is adopted. The principle of this method is that the window moves along the time axis of the data, and judges one by one whether the data in the window is in a steady state 01.
[0037] S24. Perform dimensionless and normalized processing on the operating data.
[0038] S25, defining the screening process of steps S21 to S24 as a preset standard, and then transmitting the operating data that meets the preset standard to the cloud through the communication network.
[0039] S3. Establish an operation and maintenance information database with a distributed architecture to store real-time uploaded operation data and historical operation data, and use machine learning methods to correct and update the operation and maintenance information database.
[0040] Specifically, according to the type of data that needs to be stored, including real-time operation data, historical operation data, and graph data, the database architecture is designed, and appropriate distributed database technologies such as Apache Cassandra, Amazon DynamoDB, and MongoDB are selected. Data models and table structures are established, and real-time data tables, historical data tables, and related indexes are defined. Efficient data writing and querying are achieved, batch writing and asynchronous processing mechanisms are used to improve data processing efficiency, and data compression and archiving strategies are used to manage historical data and optimize storage space.
[0041] In addition, machine learning algorithms, such as time series analysis, regression models, and classification models, are used to train prediction and anomaly detection models. The models are trained using historical data, and cross-validation and holdout methods are used to evaluate the models to ensure the accuracy and generalization of the models. The online learning mechanism is added to update the model parameters in real time to adapt to the dynamic changes of data, and the model is retrained regularly and updated with new data to prevent the model from becoming outdated. The trained model is deployed to monitor and analyze new data in real time, detect and mark abnormal data, and correct and update the data based on the model analysis results to maintain the data quality of the operation and maintenance information database.
[0042] S4. Extract the status features of the operating data, define nodes and edges, build a graph model, and evaluate the operating status of the ventilator equipment through graph parallel computing analysis.
[0043] Graph computing is a technology used to process and analyze graph data, which consists of nodes (vertices) and edges (edges) to represent the relationship between entities. Graph computing technology has significant advantages in large-scale data analysis, especially when dealing with complex relationship networks.
[0044] In the intelligent operation and maintenance analysis method of ventilators, graph computing technology can be used in the following aspects: 1. Fault correlation analysis: Analyze the relationship between the various components of the fan through graph calculation, identify the fault propagation path and correlation, and improve the accuracy and efficiency of fault diagnosis.
[0045] 2. Predictive maintenance optimization: Use graph computing technology to build equipment health maps, predict component degradation trends, optimize maintenance strategies, and reduce unexpected downtime.
[0046] 3. Operation status monitoring: Through graph calculation, the correlation between different sensor data is analyzed to monitor the operation status of the ventilator in real time and detect abnormal situations.
[0047] 4. Resource scheduling optimization: In large ventilation systems, graph computing is used to optimize resource allocation and scheduling to improve the overall efficiency of the system.
[0048] In the description of the present invention, extracting the state characteristics of the operating data, defining nodes and edges, constructing a graph model, and evaluating the operating state of the ventilator equipment through graph parallel computing analysis include the following steps: S41. Establish a parallel channel for feature extraction and synchronize file data in the operation and maintenance information library.
[0049] In the description of the present invention, establishing a feature extraction parallel channel and synchronizing file data in the operation and maintenance information library includes the following steps: S411. Establish multiple feature extraction parallel channels in the operation and maintenance database. Each feature extraction parallel channel is randomly divided into several segments. Each segment is distributed and stores multiple data nodes.
[0050] S412. Use the operation and maintenance information library to store the operation data and historical operation data in segments, and each segment of data has timestamp information and is randomly distributed in the data nodes of each segment.
[0051] S413, according to the running data with the same timestamp information, multiple feature extraction parallel channels are used to synchronously extract, a multi-channel parallel computing task is established, and the file data in each channel is synchronously processed.
[0052] S42. Establish a feature extraction model based on artificial intelligence algorithm, analyze the type and value of the operating parameters in the operating data, and extract the state characteristics of the ventilator equipment. The state characteristics include basic parameter characteristics, parameter level characteristics, operating trend characteristics, and parameter fluctuation characteristics.
[0053] In the description of the present invention, a feature extraction model is established based on an artificial intelligence algorithm, the types and values of operating parameters in the operating data are analyzed, and the state characteristics of the ventilator equipment are extracted, including the following steps: S421. Extract basic parameter features based on the types of operating parameters included in the operating data.
[0054] Among them, it is necessary to preliminarily classify various operating parameters in the operating data, including basic operating parameters such as temperature, pressure, speed, current, voltage, etc., and extract the basic characteristics of each type of parameter, such as mean, maximum value, minimum value, standard deviation, etc.
[0055] S422, setting a classification threshold, comparing the threshold level corresponding to the value of each type of operating parameter at the current moment in turn, and determining the parameter level characteristics of the operating parameter at the current moment.
[0056] S423. Set the sampling period, select an operating parameter arbitrarily, and select N operating parameters before and after the operating parameter as windows for data aggregation calculation and feature generation, calculate the arithmetic mean respectively, and then perform deviation calculation on the 2N operating parameters to obtain the deviation from the actual value, and obtain the trend data by subtracting the two deviations from the actual values, and finally obtain the operating trend characteristics through cyclic calculation.
[0057] In the description of the present invention, the calculation formula of trend data includes: In the formula, Represents trend data; Before N The deviation of an operating parameter from the actual value; After indicating N The deviation of an operating parameter from the actual value; Represents an arbitrarily selected operating parameter; Indicates the front of an arbitrarily selected operating parameter N operating parameters; Indicates the first operating parameter of an arbitrarily selected operating parameter; Indicates the value of an arbitrarily selected operating parameter. N operating parameters; Represents the arithmetic mean.
[0058] S424. Collect differential values of operating parameters in the operating data, use a time series model to perform fluctuation analysis on the operating parameters, and obtain parameter fluctuation characteristics.
[0059] Specifically, a differential calculation is performed on each operating parameter to obtain the difference of each time point relative to the previous time point, and the differential values of all time points are combined into a differential sequence as input data for the time series model.
[0060] The autoregressive integrated moving average model (ARIMA) is selected to model the difference series, fit the model to the difference series, and estimate the model parameters; then the difference series data is used to fit the selected time series model to obtain the model parameters and fitting results. According to the fitted time series model, the fluctuation characteristics of the operating parameters are analyzed, the volatility and variance indicators are calculated, and the fluctuation characteristics of the parameters are extracted. Finally, by analyzing the output results of the model, the fluctuation characteristics of the operating parameters of the ventilator equipment are obtained.
[0061] S43. Convert the operation data and state characteristics into graph data, define nodes and edges and their attributes, and build a graph model.
[0062] In the description of the present invention, the operation data and state characteristics are converted into graph data, and nodes, edges and their attributes are defined. Building a graph model includes the following steps: S431. Define node types and their node attributes, where the node types include sensor nodes, device component nodes, and operating status nodes.
[0063] S432. Define edge types and edge attributes, where edge types include sensor-component edges, component-component edges, and component-state edges.
[0064] S433. Convert the operation data and status characteristics into a graph data format, and create a graph database in the operation and maintenance information library to store and manage the graph data.
[0065] S434. Use the application programming interface in the graph database to create nodes and add corresponding node attributes. Then define the edges between nodes and add edge attributes. Create nodes and edges through batch import to build a complete graph model.
[0066] S44. Select a graph computing framework, evaluate the operating status and health index of the ventilator equipment based on the graph computing results, and identify key nodes and potential faults.
[0067] In the description of the present invention, selecting a graph computing framework, evaluating the operating status and health index of the ventilator equipment based on the graph computing results, and identifying key nodes and potential faults include the following steps: S441. Use the import tool in the graph database to import the graph model into the graph computing framework.
[0068] Specifically, graph databases and graph computing frameworks provide tools for importing and processing graph data. Graph databases (such as Neo4j and JanusGraph) can store and manage large-scale graph data, while graph computing frameworks (such as Apache Giraph, GraphX, and Neo4j Graph Data Science Library) provide efficient graph algorithms and computing capabilities.
[0069] S442. Use a graph traversal algorithm to traverse the graph model, extract relevant features of nodes and edges, construct a feature vector, and use a community detection algorithm to detect communities in the graph model, identify node clusters with similar operating status, and discover potential fault areas.
[0070] Specifically, graph traversal algorithms (such as depth-first search DFS and breadth-first search BFS) are used to traverse the graph model and extract the features of nodes and edges. Community detection algorithms (such as Louvain algorithm and Girvan-Newman algorithm) are used to identify communities or clusters in the graph, which usually represent sets of nodes with similar characteristics in the graph.
[0071] Use the graph traversal algorithm to traverse the graph model and extract the features of each node and edge. By calculating the degree of the node, the weight of the edge, etc., the extracted node and edge features are combined into a feature vector to represent the state of each node.
[0072] The community detection algorithm is applied to divide the graph model into communities, identify node clusters with similar operating status, and discover potential fault areas by analyzing the node characteristics in the community.
[0073] S443. Use the graph isomorphism matching algorithm to match subgraph patterns, identify potential fault points, and use the single-source shortest path algorithm to calculate the shortest path before the key nodes and analyze the fault propagation path.
[0074] Specifically, the graph isomorphism matching algorithm (VF2 algorithm) is used to match specific subgraph patterns in a graph to identify potential failure points. The VF2 algorithm is an algorithm for graph isomorphism detection that can match a small subgraph in a large graph. It uses recursive search and backtracking techniques to find matches while maintaining the graph structure and node relationships. Its implementation steps include the following aspects: 1) Prepare data Large image: An operating status diagram of a ventilator device, where nodes represent device components or sensors and edges represent the relationships between them.
[0075] Subgraph pattern: represents a known failure mode graph, containing a specific configuration of several nodes and edges.
[0076] 2) Define matching conditions Node matching conditions: Node attributes (such as type, status value) need to meet specific conditions.
[0077] Edge matching conditions: Edge attributes (such as relationship type and weight) need to meet specific conditions.
[0078] 3) Application of VF2 algorithm Initialization: Define an empty node map.
[0079] Recursive search: gradually map the nodes of the subgraph to the nodes of the large graph, and check whether the mapping satisfies the matching conditions of nodes and edges.
[0080] Backtracking: During the recursive process, if the current mapping cannot meet the conditions, backtrack to the previous step and try other mapping possibilities.
[0081] Termination: When all subgraph nodes are successfully mapped to the large graph nodes, a match is found; otherwise, the search continues until all possibilities are traversed.
[0082] 4) Identify the fault point All locations matching the subgraph pattern are found in the larger graph, and the nodes at these locations are marked as potential failure points.
[0083] The single-source shortest path algorithm (Dijkstra algorithm) is used to calculate the shortest path from a key node to other nodes to analyze the fault propagation path. The Dijkstra algorithm calculates the shortest path from a single source node to all other nodes in a weighted graph, maintains a priority queue, selects the node with the shortest current path at each step, and updates the shortest path of its neighboring nodes. Its implementation steps include the following aspects: 1) Initialization Define source nodes: key nodes (failure points).
[0084] Initialization distance: Set the distance from the source node to itself to 0, and the initial distance of other nodes to infinity.
[0085] Initialize the priority queue: add all nodes to the priority queue, and the initial priority is the distance value of the node.
[0086] 2) Algorithm execution Select the node with the smallest distance: extract the node with the smallest current distance from the priority queue.
[0087] Update neighbor node distances: For each neighbor node of the node, calculate the path distance from the source node through the current node to the neighbor node. If the distance is less than the shortest distance currently recorded by the neighbor node, update the shortest distance of the neighbor node and re-add the neighbor node to the priority queue.
[0088] Continue selection: Repeat the above process until the priority queue is empty.
[0089] 3) Path reconstruction The predecessor nodes of each node are recorded so that the shortest path can be reconstructed while calculating the shortest distance.
[0090] Trace back the predecessor nodes from the target node and reconstruct the shortest path from the source node to the target node.
[0091] 4) Fault propagation path analysis The shortest path calculation results are used to analyze the propagation path from the fault point (source node) to other related nodes.
[0092] Evaluate the possible fault impact scope and propagation path to provide a basis for subsequent fault handling and maintenance decisions.
[0093] S444. Based on the extracted feature vector and its weight, the health index of the ventilator equipment is calculated, and a health assessment report is generated to evaluate the current operating status of the ventilator equipment.
[0094] In the description of the present invention, the calculation formula of the health index is: In the formula, HI It represents the health index; Indicates i The weight of the feature vector; Indicates i The nonlinear transformation function of the eigenvectors; Indicates i The time decay factor of the eigenvectors; Indicates i The health threshold function of the feature vector; represents the weight factor of the interaction term; represents the mutual terms between eigenvectors.
[0095] Among them, the eigenvector needs to be transformed nonlinearly to better reflect its impact on health status. The present invention uses Sigmoid function to transform and calculate .
[0096] The influence of the eigenvector may decay over time, so a time decay factor is introduced , represents the weight of the eigenvalue at different time points, and its calculation formula is: , where represents the decay rate; e represents a natural constant.
[0097] Health Threshold Function It is used to evaluate whether the characteristic value is within the normal range. Characteristics beyond the normal range have a greater impact on the health index. The present invention uses logical judgment or conditional function to define it, that is, the characteristic vector is 1 if it is within the normal range, and c if it exceeds the normal range, where c is the set health index.
[0098] The interaction between features also plays an important role in health status, which can be expressed by the interaction term To indicate that the interaction term needs to be calculated using the product between features.
[0099] S5. Use visualization tools to display graph data and evaluation results, provide fault propagation path and operation status monitoring, and regulate and maintain the control parameters and body status of the ventilator equipment.
[0100] See also Figure 2 , and also provides a ventilator intelligent operation and maintenance analysis system based on big data, the system includes: The monitoring module 1 is used to collect the operating data of the ventilator equipment in real time through monitoring sensors.
[0101] Processing module 2 is used to use edge computing equipment to preliminarily process the operating data and transmit the operating data that meets the preset standards to the cloud through the communication network.
[0102] The storage module 3 is used to establish an operation and maintenance information base of a distributed architecture, store real-time uploaded operation data and historical operation data, and use machine learning methods to correct and update the operation and maintenance information base.
[0103] The analysis module 4 is used to extract the state characteristics of the operation data, define nodes and edges, build a graph model, and evaluate the operation state of the ventilator equipment through graph parallel computing and analysis.
[0104] Display module 5 is used to display graph data and evaluation results using visualization tools, provide fault propagation path and operation status monitoring, and regulate and maintain control parameters and body status of the ventilator equipment.
[0105] In summary, with the help of the above technical solutions of the present invention, through edge computing and distributed architecture, the data processing and storage efficiency is greatly improved, the system response time is reduced, and the operation and maintenance decision is made more real-time and accurate; with the help of machine learning, the operation and maintenance information library can learn from big data and continuously optimize, making the analysis results more reliable and durable; in addition, through the graph model and parallel computing technology, the analysis of equipment status is more comprehensive and in-depth, and potential fault points and influencing factors can be identified, and the health status of the equipment can be predicted in advance, thereby reducing equipment downtime and maintenance costs; it can help operation and maintenance personnel make decisions quickly and take necessary maintenance measures, further improving the overall operation and maintenance efficiency and equipment reliability. By displaying graph data and evaluation results through visualization tools, the equipment operation status and fault propagation path are clear at a glance, which is convenient for operation and maintenance personnel to make decisions and operations; at the same time, through real-time monitoring and status evaluation, the operation status of the equipment can be fully grasped, problems can be discovered and solved in time, equipment downtime can be reduced, and maintenance costs can be reduced; the decision-making ability and operation efficiency of operation and maintenance personnel are improved, and it has significant technical advantages and application value.
[0106] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.
Claims
1. A ventilator intelligent operation and maintenance analysis method based on big data, characterized in that: The method comprises the following steps: S1. Deploy monitoring sensors to collect real-time operating data of ventilator equipment; S2. Use edge computing devices to preliminarily process the operation data, and transmit the operation data that meets the preset standards to the cloud through the communication network; S3. Establish a distributed architecture operation and maintenance information database to store real-time uploaded operation data and historical operation data, and use machine learning to revise and update the operation and maintenance information database; S4, extracting the state characteristics of the operation data, defining nodes and edges, building a graph model, and evaluating the operation state of the ventilator equipment through graph parallel computing and analysis; S5. Use visualization tools to display graph data and evaluation results, provide fault propagation path and operation status monitoring, and regulate and maintain the control parameters and body status of the ventilator equipment; The extracting of the state characteristics of the operating data, defining nodes and edges, constructing a graph model, and evaluating the operating state of the ventilator equipment through graph parallel computing and analysis include the following steps: S41, establishing a feature extraction parallel channel to synchronize file data in the operation and maintenance information library; S42, establishing a feature extraction model based on an artificial intelligence algorithm, analyzing the types and values of the operating parameters in the operating data, and extracting the state characteristics of the ventilator equipment; wherein the state characteristics include basic parameter characteristics, parameter level characteristics, operating trend characteristics, and parameter fluctuation characteristics; S43, converting the operation data and the state characteristics into graph data, defining nodes and edges and their attributes, and constructing a graph model; S44. Select a graph computing framework, evaluate the operating status and health index of the ventilator equipment based on the graph computing results, and identify key nodes and potential faults.
2. The intelligent operation and maintenance analysis method of a ventilator based on big data according to claim 1 is characterized in that: The method of using the edge computing device to preliminarily process the operation data and transmit the operation data that meets the preset standards to the cloud through the communication network includes the following steps: S21, reading the operating data collected in real time, and removing non-numeric values contained in the operating data; S22, monitoring the operating condition of the ventilator equipment in real time, aligning the operating data from different sources according to the timestamp information, and eliminating the operating data during the equipment shutdown period; S23, using a sliding window method to remove non-steady-state values and noise values in the operating data; S24, performing non-dimensionalization and normalization processing on the operation data; S25, defining the screening process of steps S21 to S24 as a preset standard, and then transmitting the operating data that meets the preset standard to the cloud through the communication network.
3. The intelligent operation and maintenance analysis method of a ventilator based on big data according to claim 1 is characterized in that: The establishment of a parallel channel for feature extraction and synchronization of file data in the operation and maintenance information library comprises the following steps: S411, establishing multiple feature extraction parallel channels in the operation and maintenance database, each of the feature extraction parallel channels is randomly divided into a number of segments, and each segment is distributed and stores multiple data nodes; S412, using the operation and maintenance information library to store the operation data and historical operation data in segments, and each segment of data has timestamp information and is randomly distributed in the data nodes of each segment; S413, according to the running data with the same timestamp information, multiple feature extraction parallel channels are used to synchronously extract, a multi-channel parallel computing task is established, and the file data in each channel is synchronously processed.
4. The intelligent operation and maintenance analysis method for ventilators based on big data according to claim 3 is characterized in that: The feature extraction model is established based on the artificial intelligence algorithm, the type and value of the operating parameters in the operating data are analyzed, and the state characteristics of the ventilator equipment are extracted, which includes the following steps: S421. Extracting basic parameter features based on the types of operating parameters included in the operating data; S422, setting a classification threshold, comparing the threshold level corresponding to the value of each type of operating parameter at the current moment in turn, and determining the parameter level characteristics of the operating parameter at the current moment; S423, set the sampling period, arbitrarily select an operating parameter, and select N operating parameters before and after the operating parameter as windows for data aggregation calculation and feature generation, calculate the arithmetic mean respectively, and then perform deviation calculation on the 2N operating parameters to obtain the deviation actual value, and obtain the trend data by subtracting the two deviation actual values before and after, and finally obtain the operating trend feature through cyclic calculation; S424: Collect differential values of operating parameters in the operating data, use a time series model to perform fluctuation analysis on the operating parameters, and obtain parameter fluctuation characteristics.
5. The intelligent operation and maintenance analysis method for ventilators based on big data according to claim 4 is characterized in that: The calculation formula of the trend data includes: In the formula, Represents trend data; Before N The deviation of an operating parameter from the actual value; After indicating N The deviation of an operating parameter from the actual value; Represents an arbitrarily selected operating parameter; Indicates the front of an arbitrarily selected operating parameter N operating parameters; Indicates the first operating parameter of an arbitrarily selected operating parameter; Indicates the value of an arbitrarily selected operating parameter. N operating parameters; Represents the arithmetic mean.
6. The intelligent operation and maintenance analysis method for ventilators based on big data according to claim 4 is characterized in that: The step of converting the operation data and the state characteristics into graph data, defining nodes and edges and their attributes, and constructing a graph model comprises the following steps: S431, define node type and its node attributes; S432, define edge type and edge attributes; S433, converting the operation data and state characteristics into a graph data format, and creating a graph database in the operation and maintenance information library to store and manage the graph data; S434. Use the application programming interface in the graph database to create nodes and add corresponding node attributes; then define the edges between nodes and add edge attributes; create nodes and edges through batch import to build a complete graph model.
7. The intelligent operation and maintenance analysis method for ventilators based on big data according to claim 6 is characterized in that: The node types include sensor nodes, device component nodes and operating status nodes; the edge types include sensor-component edges, component-component edges and component-status edges.
8. The intelligent operation and maintenance analysis method for ventilators based on big data according to claim 7 is characterized in that: The method of selecting a graph computing framework, evaluating the operating status and health index of the ventilator equipment based on the graph computing results, and identifying key nodes and potential faults includes the following steps: S441. Import the graph model into the graph computing framework using an import tool in the graph database; S442, using a graph traversal algorithm to traverse the graph model, extract relevant features of nodes and edges, construct a feature vector, and use a community detection algorithm to detect communities in the graph model, identify node clusters with similar operating status, and discover potential fault areas; S443, using a graph isomorphism matching algorithm to match subgraph patterns, identify potential fault points, and use a single-source shortest path algorithm to calculate the shortest path between key nodes and analyze the fault propagation path; S444. Based on the extracted feature vector and its weight, the health index of the ventilator equipment is calculated, and a health assessment report is generated to evaluate the current operating status of the ventilator equipment.
9. The intelligent operation and maintenance analysis method for ventilators based on big data according to claim 8 is characterized in that: The calculation formula of the health index is: In the formula, HI It represents the health index; Indicates i The weight of the feature vector; Indicates i The nonlinear transformation function of the eigenvectors; Indicates i The time decay factor of the eigenvectors; Indicates i The health threshold function of the feature vector; represents the weight factor of the interaction term; represents the mutual terms between eigenvectors.
10. A ventilator intelligent operation and maintenance analysis system based on big data, used to implement the ventilator intelligent operation and maintenance analysis method based on big data as described in any one of claims 1 to 9, characterized in that: The system includes: A monitoring module is used to collect the operation data of the ventilator equipment in real time through monitoring sensors; A processing module, used to preliminarily process the operation data using the edge computing device, and transmit the operation data that meets the preset standards to the cloud through the communication network; The storage module is used to establish a distributed architecture operation and maintenance information library, store real-time uploaded operation data and historical operation data, and use machine learning to correct and update the operation and maintenance information library; The analysis module is used to extract the status characteristics of the operating data, define nodes and edges, build a graph model, and evaluate the operating status of the ventilator equipment through graph parallel computing analysis; The display module is used to display graph data and evaluation results using visualization tools, provide fault propagation path and operation status monitoring, and regulate and maintain the control parameters and body status of the ventilator equipment.
Citation Information
Patent Citations
Power equipment operation and maintenance evaluation method based on power data analysis
CN116992399A
Dynamic adjustment method for performance degradation and life extension of wind turbine generator
CN117028169A
Distributed household photovoltaic power station operation and maintenance monitoring method and system
CN117236554A
Intelligent operation and maintenance cloud platform for industrial equipment
CN118154174A
Distributed photovoltaic power station intelligent monitoring platform and method
CN118174453A
Cited By
Trackless rubber-tyred vehicle multi-terminal data fusion system
CN119249359A
Remote monitoring system for magnetic suspension air feeder based on Internet
CN120990907A
