Remote monitoring method and system for running state of medium-high pressure valve

Abstract

The invention discloses a remote monitoring method and system for the running state of a medium-high pressure valve, and the method comprises the steps: collecting the multi-dimensional running parameters of the valve in real time through a multi-sensor array disposed on the medium-high pressure valve; according to the multi-dimensional operation parameters, an operation stability index of the operation stability of the medium-high pressure valve in the operation process is calculated; performing feature extraction on the multi-dimensional operation parameters by using a deep learning model; and based on the extracted features, comprehensively evaluating the operation state of the valve by using a fuzzy logic algorithm, generating a state evaluation report of the health state of the valve, and sending the multi-dimensional operation parameters, the operation stability index and the state evaluation report to a remote control center through a wireless network. According to the embodiment of the invention, the valve can be monitored in real time and accurately evaluated, so that the operation safety and reliability of the valve are improved, and intelligent state monitoring and management are realized.

Description

Technical Field

[0001] This invention belongs to the field of medium and high pressure valve technology, and in particular, it relates to a method and system for remote monitoring of the operating status of medium and high pressure valves. Background Technology

[0002] Medium and high pressure valves are widely used in industries such as petroleum, chemical, and power, serving as crucial equipment for fluid control and safety management. With the increasing automation and intelligence of industrial production, the demand for valve operation monitoring and maintenance is becoming increasingly urgent. Traditional valve monitoring methods mainly rely on manual inspections and periodic maintenance. These methods are not only inefficient but also fail to provide real-time monitoring of valve operating status, easily leading to accidents. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for remote monitoring of the operating status of medium and high pressure valves, in order to overcome the shortcomings of the prior art, enable real-time monitoring and accurate evaluation of valves, improve their operational safety and reliability, and realize intelligent status monitoring and management.

[0004] One embodiment of this application provides a method for remotely monitoring the operating status of medium and high pressure valves, the method comprising: A multi-sensor array installed on medium and high pressure valves is used to collect multi-dimensional operating parameters of the valves in real time. These multi-dimensional operating parameters include at least: vibration, temperature, pressure, and flow rate. Based on the multi-dimensional operating parameters, the operating stability index of the medium and high pressure valve during operation is calculated. The multi-dimensional operating parameters are feature extracted using a deep learning model, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state. Based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generate a valve health status assessment report, and send the multi-dimensional operating parameters, the operating stability index, and the status assessment report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

[0005] Optionally, the formula for calculating the operational stability index is:

[0006] Among them, the To ensure operational stability, the aforementioned The above The above The above The normalized values ​​for vibration, temperature, pressure, and flow rate are... The above Let be the instantaneous first derivative of the vibration and temperature at the i-th time point. The average value of the instantaneous first derivative of the vibration, where n is the number of time points. The above The above The above The above The above These are the corresponding weighting coefficients.

[0007] Optionally, the step of using a deep learning model to extract features from the multi-dimensional operating parameters, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state, includes: The multi-dimensional operating parameters are constructed into a multi-channel input matrix, with each channel corresponding to one operating parameter; Based on the multi-channel input matrix, multiple sets of parallel convolutional layers are used, each set of convolutional layers having convolutional kernels of different sizes, to extract the spatial features of the multi-dimensional operating parameters at different feature scales. Among them, small-scale convolutional kernels are used to extract local detail features, while large-scale convolutional kernels are used to capture global structural features. The extracted spatial features are organized into time series data, wherein the spatial features are mapped into time series form to form a new set of input features for capturing patterns that evolve over time. By utilizing a bidirectional LSTM layer, the forward and backward dependencies in the time series data are captured simultaneously to capture the temporal dependencies and obtain the temporal features of the multi-dimensional operating parameters. The spatial features and the temporal features are merged into a high-dimensional spatiotemporal feature vector through feature concatenation operations.

[0008] Optionally, based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status and generate a valve health status assessment report, including: Based on the spatiotemporal feature vector, multiple fuzzy sets are obtained, each fuzzy set corresponding to a specific valve operating state feature; A fuzzy rule base is constructed, in which fuzzy rules evaluate the health status of valves based on the combination of multi-dimensional features. Each rule consists of a premise and a conclusion. The premise part uses fuzzy logic operators to combine fuzzy sets of multiple features, and the conclusion part gives the health status evaluation of the valve. A fuzzy inference engine is constructed, wherein the fuzzy inference engine adopts the Mamdani or Sugeno fuzzy inference algorithm. The fuzzy inference engine receives the input feature vector, converts it into the membership degree of the fuzzy set through the fuzzification process, performs inference using the fuzzy rule base, calculates the activation strength of each rule, and generates fuzzy output based on the conclusion part of the rule. The fuzzy outputs of all activation rules are aggregated, using either the max-min synthesis method or the weighted average method. The aggregated result is a comprehensive fuzzy output that represents the membership of the valve's health status in different state categories. The aggregated fuzzy output is defuzzified to obtain a clear health status assessment value. The defuzzification methods include the centroid method, the maximum membership method, and the average maximum method. The result of the defuzzification is a specific numerical value or level, representing the health status of the valve. Based on the deblurring results, a valve health status assessment report is generated, which includes a numerical assessment of the current health status, analysis of possible causes of failure, recommended maintenance measures, and predictions of future status.

[0009] Another embodiment of this application provides a remote monitoring system for the operating status of medium and high pressure valves, the system comprising: The acquisition module is used to acquire multi-dimensional operating parameters of the valve in real time through a multi-sensor array installed on the medium and high pressure valve. The multi-dimensional operating parameters include at least: vibration, temperature, pressure and flow rate. The calculation module is used to calculate the operational stability index of the medium and high pressure valve during operation based on the multi-dimensional operating parameters. The extraction module is used to extract features from the multi-dimensional operating parameters using a deep learning model, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state. The evaluation module is used to comprehensively evaluate the operating status of the valve based on the extracted features and using a fuzzy logic algorithm, generate a valve health status evaluation report, and send the multi-dimensional operating parameters, the operating stability index and the status evaluation report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

[0010] Another embodiment of this application provides a storage medium storing a computer program, wherein the computer program is configured to execute the method described in any of the preceding claims when running.

[0011] Another embodiment of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the method described in any of the preceding claims.

[0012] Compared with existing technologies, the present invention provides a remote monitoring method for the operating status of medium and high pressure valves. This method utilizes a multi-sensor array installed on the valve to collect multi-dimensional operating parameters in real time. Based on these parameters, an operating stability index is calculated to assess the valve's operational stability. A deep learning model is used to extract features from the multi-dimensional operating parameters. Based on these extracted features, a fuzzy logic algorithm is employed to comprehensively evaluate the valve's operating status, generating a valve health status assessment report. The multi-dimensional operating parameters, the operating stability index, and the assessment report are then transmitted to a remote control center via a wireless network. This enables real-time monitoring and accurate evaluation of the valve, improving its operational safety and reliability, and achieving intelligent status monitoring and management. Attached Figure Description

[0013] Figure 1 A hardware structure block diagram of a computer terminal for a remote monitoring method for the operating status of medium and high pressure valves provided in an embodiment of the present invention; Figure 2 A flowchart illustrating a method for remotely monitoring the operating status of medium- and high-pressure valves according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a remote monitoring system for the operating status of medium and high pressure valves provided in an embodiment of the present invention. Detailed Implementation

[0014] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0015] This invention first provides a method for remote monitoring of the operating status of medium and high pressure valves. This method can be applied to electronic devices, such as computer terminals, specifically ordinary computers.

[0016] The following detailed explanation uses a computer terminal as an example. Figure 1 This is a hardware structure block diagram of a computer terminal for a remote monitoring method of the operating status of medium and high pressure valves provided in an embodiment of the present invention. (See diagram below.) Figure 1 As shown, the computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.

[0017] The non-volatile storage medium can store an operating system and a computer program. This computer program includes program instructions that, when executed, enable the processor to perform any method for remotely monitoring the operating status of medium- and high-pressure valves.

[0018] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0019] The internal memory provides an environment for the execution of computer programs in non-volatile storage media. When the computer program is executed by the processor, it enables the processor to perform any method for remotely monitoring the operating status of medium and high pressure valves.

[0020] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0021] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0022] See Figure 2 The present invention provides a method for remote monitoring of the operating status of medium and high pressure valves, which may include the following steps: S201, through a multi-sensor array installed on the medium and high pressure valve, collects multi-dimensional operating parameters of the valve in real time, wherein the multi-dimensional operating parameters include at least: vibration, temperature, pressure and flow rate; By installing multi-sensor arrays on medium- and high-pressure valves, comprehensive monitoring of valve operation can be achieved. These sensors collect multi-dimensional parameters related to valve operation in real time, such as vibration, temperature, pressure, and flow rate. These parameters reflect the actual working conditions of the valve; vibration data reveals the mechanical state within the valve, temperature changes indicate fluid characteristics and potential overheating problems, pressure detection reflects the system's fluid dynamics, and flow rate data is a crucial indicator of valve performance. Through the integration of multi-dimensional data, potential faults and anomalies can be detected promptly, ensuring stable valve operation.

[0023] This real-time monitoring method has significant practical implications. First, by systematically monitoring multi-dimensional valve operating parameters, early warnings of valve health can be provided, reducing equipment failure rates and preventing accidents caused by valve malfunctions, thereby improving safety and reliability. Second, the data provided by real-time monitoring offers a foundation for subsequent data analysis and optimization, promoting intelligent management and maintenance strategies, improving the overall efficiency of the production system, and reducing maintenance costs. Finally, this method also provides decision support for operators, enabling them to make reasonable adjustments and optimizations based on real-time data, further ensuring the stable operation of the system.

[0024] In practical implementation, by tightly integrating sensor arrays into key components of medium- and high-pressure valves, such as the valve body, stem, and inlet / outlet areas, comprehensive monitoring of various important parameters can be achieved. Sensors can utilize MEMS (Micro-Electro-Mechanical Systems) technology to achieve small size and high sensitivity. Simultaneously, the sensor placement should consider the direction and velocity of fluid flow to ensure accurate flow and pressure data acquisition. For data transmission, a wireless communication module can be incorporated, enabling real-time transmission of collected data to a centralized management system, or preliminary analysis via edge computing before uploading.

[0025] For example, vibration sensors mounted on top of valves can monitor the valve's vibration frequency and amplitude in real time to detect bearing wear or failure. Simultaneously, temperature sensors can be placed both externally and internally to monitor the thermal load on the valve materials, thereby inferring fluid temperature changes. Pressure sensors can be installed at the valve's inlet and outlet to assess the smoothness of fluid flow by monitoring the pressure difference. If abnormally increased vibration or excessive temperature is detected, the system can issue an early warning signal, prompting timely maintenance. Furthermore, after data acquisition, deep learning models are used to extract features from these multi-dimensional parameters. This feature information is combined with an operational stability index to ultimately generate a health status assessment report. This approach not only improves the intelligence level of data processing but also continuously optimizes valve operating performance, ensuring efficient and accurate remote monitoring.

[0026] S202, Based on the multi-dimensional operating parameters, calculate the operating stability index of the medium and high pressure valve during operation; Calculating the Operating Stability Index (RSI) of medium- and high-pressure valves based on the aforementioned multi-dimensional operating parameters is a crucial step. This calculation relies primarily on multiple real-time acquired operating parameters, such as vibration, temperature, pressure, and flow rate. Normalizing these parameters eliminates the influence of different dimensions on the calculation results, making the weighted sums of different parameters comparable. Simultaneously, by introducing statistical quantities such as the instantaneous first derivative, the dynamic changes in operating status can be effectively captured. The calculation formula integrates the contributions and trends of various parameters, forming a comprehensive stability index that provides a quantitative assessment of the health status of medium- and high-pressure valves during operation.

[0027] The calculation of the Reliability Stability Index (RSI) provides an important early warning mechanism for the safe operation of valves. Through real-time monitoring and calculation, operation and maintenance personnel can promptly identify potential failure risks, predict the operating status of valves, and thus take corresponding maintenance and upkeep measures. This data-driven decision-making enhances valve reliability, reduces the threat of sudden failures to production safety, and ultimately improves overall operating efficiency and economy.

[0028] Specifically, a formula for calculating an operational stability index can be:

[0029] Among them, the To ensure operational stability, the aforementioned The above The above The above The normalized values ​​for vibration, temperature, pressure, and flow rate are... The above Let be the instantaneous first derivative of the vibration and temperature at the i-th time point. The average value of the instantaneous first derivative of the vibration, where n is the number of time points. The above The above The above The above The above These are the corresponding weighting coefficients.

[0030] The calculation formula for the Operating Stability Index (RSI) integrates the influence of multi-dimensional operating parameters, effectively combining the importance of different characteristics in valve operation through weighted summation. Normalization of each parameter ensures dimensional consistency and allows for comparison of different parameters under the same standard. Multiple weighting coefficients in the formula quantify the contribution of each parameter to overall stability, reflecting their priority and relative importance in actual operation. Furthermore, the formula incorporates sensitivity to fluctuations and trends through statistical analysis methods (such as instantaneous first derivatives and average values), ensuring that the RSI reflects not only the current state but also its dynamic characteristics. This design guarantees the comprehensiveness and accuracy of the RSI as an indicator of operational health, thereby more effectively predicting valve operational stability.

[0031] S203, use a deep learning model to extract features from the multi-dimensional operating parameters, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state; In the monitoring of the operational status of medium and high-pressure valves, feature extraction using deep learning models is crucial for achieving efficient and accurate monitoring. This deep learning model combines a convolutional neural network (CNN) and a long short-term memory network (LSTM). The former focuses on extracting spatial features between operating parameters, while the latter emphasizes capturing the dynamic features of parameters changing over time. This method generates a spatiotemporal feature vector that accurately reflects various states and potential problems during valve operation. The multi-layered convolutional structure of the CNN effectively identifies local features, while the LSTM network ensures effective learning from long-sequence data. This allows the model to not only capture instantaneous features but also understand the influence of historical information, thus comprehensively analyzing the valve's health status.

[0032] The deep learning model combining convolutional neural networks and long short-term memory networks has significantly improved the intelligence level of valve operation status monitoring. By efficiently extracting the spatiotemporal features of operating parameters, the system can identify potential fault risks earlier and more accurately, enabling early warning. This not only helps prevent safety accidents caused by valve failures but also improves the targeting and efficiency of maintenance work, thereby reducing downtime and maintenance costs, and ultimately improving overall production efficiency and safety levels.

[0033] Specifically, the multi-dimensional operating parameters can be constructed into a multi-channel input matrix, with each channel corresponding to one operating parameter; In this step, the first step is to organize the multi-dimensional operating parameters (such as vibration, temperature, pressure, and flow rate) into a structured input matrix. Each operating parameter will be treated as an independent channel, and the values ​​corresponding to these channels at the same point in time will be grouped together. This method of constructing a multi-channel input matrix ensures that the model can simultaneously consider the interrelationships between different parameters during analysis, rather than relying solely on the changes of a single parameter.

[0034] By constructing a multi-channel input matrix, the operating status of the valve can be comprehensively reflected, enhancing the model's ability to understand complex data. The advantage of this approach is that the model can simultaneously utilize information from various parameters, thereby improving the effectiveness and accuracy of feature extraction and providing richer input data for subsequent analysis.

[0035] In practice, real-time data must first be acquired from various sensors installed on the valves. For example, a vibration sensor might collect vibration data every 10 seconds, and temperature, pressure, and flow sensors do the same. This acquired data is aligned using time points as indices to ensure consistent sampling times for different parameters. Subsequently, a three-dimensional matrix is ​​constructed in the format (number of time points, parameter type), where rows represent time points and columns represent different operating parameters. Assuming 100 time points and 4 parameters, the final multi-channel input matrix will have a shape of (100, 4), a structure that facilitates subsequent convolution operations.

[0036] Based on the multi-channel input matrix, multiple sets of parallel convolutional layers are used, each set of convolutional layers having convolutional kernels of different sizes, to extract the spatial features of the multi-dimensional operating parameters at different feature scales. Among them, small-scale convolutional kernels are used to extract local detail features, while large-scale convolutional kernels are used to capture global structural features. In this step, the characteristics of convolutional neural networks are utilized, and multiple sets of parallel convolutional layers are used to process the constructed multi-channel input matrix. Convolutional kernels of different sizes can extract features at multiple levels, thereby capturing the local details and global structure of the operating parameters. Small-scale convolutional kernels (such as 3x3) mainly focus on local changes, while large-scale convolutional kernels (such as 7x7) can capture a wider range of patterns and trends.

[0037] This multi-convolutional layer design enables the model to recognize features at different levels, thereby improving its ability to describe valve states. Even in the presence of noise in the parameter data, this processing method can effectively filter out unnecessary information, ensuring that the extracted features are more accurate, and thus improving the overall effectiveness and robustness of monitoring.

[0038] In practice, for each channel's input matrix, three different convolutional kernel sizes are selected: 3x3, 5x5, and 7x7. Independent convolution computations are performed for each kernel, generating different feature maps. After convolution, these feature maps are processed using an activation function (e.g., ReLU) to enhance the model's response to key features. Through parallel processing, all convolutional outputs are integrated into a single feature set, which simultaneously reflects local details and global information, providing a sufficient information foundation for subsequent time-series processing.

[0039] The extracted spatial features are organized into time series data, wherein the spatial features are mapped into time series form to form a new set of input features for capturing patterns that evolve over time. In this step, the spatial features extracted from the previous convolutional layers are reorganized to meet the requirements of time series analysis. By reorganizing the convolutional features, new time series data is formed, which not only preserves the spatial information of the features but also increases the ability to capture temporal change patterns.

[0040] Converting spatial features into a time-series data format allows subsequent models to more effectively analyze the dynamic changes of valves. For example, the state of a valve at different points in time may exhibit periodic changes. Through this time-series approach, the model can capture the patterns of these changes, thereby enabling a more accurate assessment of the valve's operating status.

[0041] In implementation, the output features of the convolutional layer must first be arranged in chronological order. For example, the feature maps at each time point can be arranged column-wise to form a new feature matrix. Then, a sliding window technique is used to extract time-series samples from the feature matrix. Assuming each sample contains feature information from 10 time points, multiple time-series samples can be generated for subsequent LSTM processing. This approach not only organizes the feature data but also provides rich time-series samples for model training, ensuring the model can effectively learn temporal dependencies.

[0042] By utilizing a bidirectional LSTM layer, the forward and backward dependencies in the time series data are captured simultaneously to capture the temporal dependencies and obtain the temporal features of the multi-dimensional operating parameters. In this step, a bidirectional long short-term memory network (Bi-LSTM) is used to process the processed time series data. Bidirectional LSTM can simultaneously focus on both forward and backward information in the time series, combining the two types of information to improve the ability to capture temporal dependencies.

[0043] By introducing a bidirectional LSTM, the model can more comprehensively understand the contextual information of time series data. This is especially important when dealing with valve operating states, where historical information and future trends are crucial for determining the current state. Extracting bidirectional features helps enhance the model's ability to identify complex dynamic changes, further improving the accuracy of state assessment.

[0044] In the implementation, the prepared time-series data is first input into a bidirectional LSTM network. Each layer of the network maintains two hidden states: one for forward propagation (from past to future) and one for backward propagation (from future to past). Simultaneously, by setting appropriate time steps and hidden layer dimensions, the LSTM layers can effectively learn long-term dependencies in the input sequence. Finally, the outputs of the bidirectional LSTM are converged to form a more representative global temporal feature vector. This vector contains dependency information for each time step in the time series, providing a crucial temporal data foundation for valve state assessment.

[0045] The spatial features and the temporal features are merged into a high-dimensional spatiotemporal feature vector through feature concatenation operations.

[0046] In this step, the model concatenates the extracted spatial and temporal features to form a high-dimensional spatiotemporal feature vector. This integration effectively combines multiple feature dimensions for subsequent analysis and processing.

[0047] The feature stitching process helps to comprehensively consider different types of data, thereby forming a comprehensive understanding of the valve's operating status. The introduction of spatiotemporal feature vectors enables the model to consider both the local features of the instantaneous state and the dynamic changes over time during comprehensive evaluation, improving the accuracy of fault detection and condition assessment.

[0048] In implementation, the spatial features from the convolutional layer are first aligned with the temporal features extracted by the bidirectional LSTM to ensure consistency of the feature vectors in the temporal dimension. Next, a feature concatenation operation is used to combine the spatial feature vector (e.g., with a shape of (batch size, feature dimension)) and the temporal feature vector (possibly of the same shape) horizontally, generating a new high-dimensional feature vector. This new feature vector contains both spatial and temporal information, providing comprehensive input data for subsequent fuzzy logic evaluation, ensuring more accurate judgment and analysis of the valve's operating status.

[0049] S204. Based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generate a valve health status assessment report, and send the multi-dimensional operating parameters, the operating stability index, and the status assessment report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

[0050] In remote monitoring methods for medium and high-pressure valves, fuzzy logic algorithms are used to comprehensively assess the valve's operating status and generate a valve health status assessment report. This is to accurately determine the valve's operational safety and reliability under specific conditions. By extracting multi-dimensional operating parameters, including vibration, temperature, pressure, and flow rate, and combining them with a calculated operational stability index, the assessment results not only rely on a single parameter but also reflect the valve's overall performance in actual operation. This comprehensive assessment provides valve maintenance personnel with detailed operational status information, ensuring that the valve operates within normal ranges, avoiding potential failures and accidents, and thus improving the overall system's safety and stability.

[0051] The significance of this comprehensive assessment method lies in its ability to promptly detect potential operational anomalies, provide corresponding health status reports, and transmit multi-dimensional operating parameters, operational stability indices, and status assessment reports to the remote control center in real time via wireless network. This real-time monitoring and feedback mechanism enables efficient management of valve status, reduces the cost and risk of manual inspections, improves the intelligence level of the monitoring system, and thus promotes the scientific decision-making for the safe operation and maintenance of medium and high-pressure valves.

[0052] Specifically, multiple fuzzy sets can be obtained based on the spatiotemporal feature vector, and each fuzzy set corresponds to a specific valve operating state feature; In this step, spatiotemporal feature vectors are used as input to generate multiple fuzzy sets. These fuzzy sets are constructed by analyzing historical data and expert knowledge, aiming to capture the characteristics of different valve operating states, such as normal, warning, and fault categories. Each fuzzy set represents the degree of membership of a certain operating state in the feature space. This representation can better handle uncertainty and fuzziness, ensuring that the evaluation process takes into account multiple possible operating states.

[0053] The purpose of this step is to lay the foundation for subsequent fuzzy logic reasoning. By clearly defining the fuzzy set for each operating state, the sensitivity and accuracy of the assessment can be improved, thus providing effective support for the comprehensive assessment of health status. This approach ensures that state determination not only relies on definite numerical values ​​but also considers potential fuzziness, enhancing the scientific rigor of the assessment.

[0054] This step can be implemented through the following process. First, using historical fault data and normal operation data, a fuzzy set containing different operating states (such as normal, minor fault, severe fault, etc.) is established. For example, the fuzzy set for the normal state can be defined as "low vibration, moderate temperature, stable pressure, normal flow." Then, a membership function is constructed for each state, such as using a triangular or trapezoidal function to describe the degree of membership of each feature. Next, the data from the spatiotemporal feature vector is input into these functions to calculate the membership degree of each feature in each fuzzy set. In this way, the system can generate nested fuzzy sets for each valve operating state, allowing subsequent reasoning processes to leverage the characteristics of different fuzzy sets for more accurate evaluation.

[0055] A fuzzy rule base is constructed, in which fuzzy rules evaluate the health status of valves based on the combination of multi-dimensional features. Each rule consists of a premise and a conclusion. The premise part uses fuzzy logic operators to combine fuzzy sets of multiple features, and the conclusion part gives the health status evaluation of the valve. In this step, a fuzzy rule base is constructed by analyzing the relationships between multiple fuzzy sets. These rules consist of premises and conclusions. The premises combine multiple operational state features using fuzzy logic operators (such as "AND" and "OR"), while the conclusions describe the health status of the valve. For example, a rule could be constructed: "If the vibration is normal and the temperature is moderate, then the valve is healthy." This rule reflects the health assessment result under a specific combination of features.

[0056] The significance of this step lies in providing an operational framework that allows complex, multi-dimensional features to be comprehensively evaluated through simple logical relationships. This rule-based approach is not only easy to understand and implement, but also makes the evaluation process highly scalable and modifiable, adaptable to new features or standards that may emerge in the future.

[0057] In practical implementation, a fuzzy rule base can be constructed through a combination of expert knowledge and data-driven methods. Experts can formulate initial fuzzy rules based on industry experience, while data analysis techniques can discover potential feature combinations and patterns through the mining and analysis of historical data. For example, cluster analysis can identify characteristic patterns under different health states, and combined with the initial rules formulated by experts, a fuzzy rule set covering multiple health states can be formed. Subsequently, the system can map multi-dimensional operating parameters collected in real time to a fuzzy set through fuzzy computation, and then apply fuzzy inference to derive a health status assessment.

[0058] A fuzzy inference engine is constructed, wherein the fuzzy inference engine adopts the Mamdani or Sugeno fuzzy inference algorithm. The fuzzy inference engine receives the input feature vector, converts it into the membership degree of the fuzzy set through the fuzzification process, performs inference using the fuzzy rule base, calculates the activation strength of each rule, and generates fuzzy output based on the conclusion part of the rule. In this step, the construction of the fuzzy inference engine is crucial. The Mamdani or Sugeno fuzzy inference algorithm is used to process the input feature vector. First, the input feature vector is converted into membership degrees of various fuzzy sets through a fuzzification process. Then, using the constructed fuzzy rule base, the activation strength of each rule is calculated, and the final fuzzy output is generated.

[0059] The role of the fuzzy inference engine is to integrate complex operating parameters and fuzzy rules into an actionable health status assessment output. By effectively capturing the relationship between input features and fuzzy rules, the fuzzy inference engine can provide reliable information support for the decision-making level. This mechanism not only improves the flexibility of the assessment but also provides real-time monitoring systems with the ability to process dynamic data.

[0060] In its implementation, the inference engine first receives spatiotemporal feature vectors of multi-dimensional operating parameters. Then, through fuzzification, it transforms these vectors into membership degrees of fuzzy sets. For example, if the input features are "normal vibration" and "moderate temperature," the engine calculates their membership degrees in the corresponding fuzzy sets. Subsequently, combining this with the previously built fuzzy rule base, the inference engine evaluates the preconditions of each fuzzy rule and calculates the activation strength. If all preconditions of a rule are met, the activation strength is greatly enhanced; otherwise, it is weakened. Finally, by calculating the fuzzy outputs of all activation rules, a comprehensive fuzzy output is generated, representing the membership degree of each health state. In this way, the system can dynamically update its judgment of the valve's health state based on real-time data.

[0061] The fuzzy outputs of all activation rules are aggregated, using either the max-min synthesis method or the weighted average method. The aggregated result is a comprehensive fuzzy output that represents the membership of the valve's health status in different state categories. In this step, the outputs of all activated fuzzy rules are aggregated into a single comprehensive fuzzy output to clearly represent the valve's health status across various state categories. The aggregation method can be either max-min synthesis or weighted average, depending on the activation intensity of each rule, to obtain a more accurate health status assessment.

[0062] The significance of this process lies in providing a quantitative approach to comprehensive assessment, unifying the results of multiple fuzzy rules into a standardized health status assessment. By aggregating different fuzzy outputs, the system can more comprehensively reflect the health status of valves, supporting maintenance personnel in considering more factors when making decisions.

[0063] Aggregation can be achieved by calculating a weighted average of the output and activation strength of each rule. For example, the weight of each rule can be set based on its accuracy during training. During aggregation, the fuzzy outputs of each rule can be multiplied by their corresponding weights and then summed. The resulting comprehensive output can more comprehensively represent the overall health status of the valve. For example, if the output of a certain rule is significantly higher than others, the system will assign it a larger weight, thereby improving the reliability of the comprehensive evaluation. The output can be a floating numerical value or a membership degree, indicating the current health status of the valve.

[0064] The aggregated fuzzy output is defuzzified to obtain a clear health status assessment value. The defuzzification methods include the centroid method, the maximum membership method, and the average maximum method. The result of the defuzzification is a specific numerical value or level, representing the health status of the valve. The aggregated fuzzy output is then transformed into a specific health status assessment value through a defuzzification process. This process can employ various defuzzification methods, such as the centroid method, the maximum membership method, and the mean maximum method, to ensure that the obtained results are intuitively interpretable.

[0065] The significance of this process lies in its clarification of complex fuzzy reasoning, resulting in a health status assessment that provides maintenance personnel with easily understandable and practical information. This not only improves the system's user-friendliness but also facilitates the effective application of real-time monitoring results, thereby enhancing the monitoring and maintenance efficiency of medium- and high-pressure valves.

[0066] In practical implementation, the centroid method can be chosen as the defuzzification approach. This involves calculating the centroid of the output based on the outputs and membership degrees of all fuzzy sets. When applying this method, the output value and corresponding membership degree of each fuzzy set are first weighted, and the weighted results are summed. These results are then used to calculate the centroids (coordinates), forming a clear numerical value or level representing the valve's health status. For example, if multiple fuzzy outputs correspond to membership degrees of 0.6, 0.3, and 0.1, then through weighted calculation, a specific value representing the valve's current health status is obtained. This value can further indicate the valve's operating condition and provide maintenance personnel with a basis for decision-making, helping them formulate appropriate maintenance measures and predict future conditions.

[0067] Based on the deblurring results, a valve health status assessment report is generated, which includes a numerical assessment of the current health status, analysis of possible causes of failure, recommended maintenance measures, and predictions of future status.

[0068] In this stage, based on the defuzzified health status assessment values, the system will integrate previous monitoring and analysis results to generate a detailed valve health status assessment report. The report will first clearly list the valve's current health status assessment values ​​and, combined with multi-dimensional operating parameters and fuzzy logic reasoning results, quantitatively describe the health status. Simultaneously, the system will analyze possible causes of failure, such as abnormal vibration or excessive temperature, and further propose corresponding maintenance suggestions and improvement measures to ensure stable valve operation. Furthermore, it will combine historical data and trend analysis to predict the valve's future operating status, which will help to perform maintenance or replacement in advance, thereby reducing the risk of failure.

[0069] The significance of generating a health status assessment report lies in its ability to enable maintenance personnel and management to quickly obtain information about the valve's operating status and potential problems through structured information delivery. This report not only provides a current condition assessment but also helps users develop scientific maintenance strategies, optimize resource allocation, and proactively address potential issues through fault cause analysis and maintenance recommendations, effectively improving the valve's operational safety and reliability.

[0070] In practice, the system generates a health status assessment report using the defuzzification results as follows: First, based on the defuzzification output, a numerical assessment of the current health status is generated. For example, if the assessment result is 85, the system includes this value in the corresponding section of the report. Then, combining operational data and the results of fuzzy rule inference, the system analyzes possible causes of the current health status, such as abnormal vibration levels and temperature increases. This can be achieved by comparing historical fault data with current monitoring parameters. Next, based on the analysis results, the system automatically generates maintenance recommendations, such as "It is recommended to check the condition of the sealing rings and replace worn parts promptly." Finally, using trend analysis and historical data, the system predicts future status, such as "It is expected that the valve's health status will remain within a good range in the next cycle, but attention should be paid to the trend of rising temperature." Ultimately, this information is integrated into a structured assessment report and sent to a remote control center via a wireless network, allowing relevant personnel to access and review it at any time, thereby ensuring efficient valve health monitoring and management.

[0071] As can be seen, by installing a multi-sensor array on the medium- and high-pressure valve, multi-dimensional operating parameters of the valve are collected in real time; based on the multi-dimensional operating parameters, an operating stability index is calculated to measure the operating stability of the medium- and high-pressure valve during operation; features are extracted from the multi-dimensional operating parameters using a deep learning model; based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generating a valve health status assessment report, and the multi-dimensional operating parameters, the operating stability index, and the status assessment report are sent to a remote control center via a wireless network. This enables real-time monitoring and accurate evaluation of the valve, improving its operational safety and reliability, and achieving intelligent status monitoring and management.

[0072] Another embodiment of the present invention provides a remote monitoring system for the operating status of medium and high pressure valves, see [link to relevant documentation]. Figure 3 The system may include: The acquisition module 301 is used to acquire multi-dimensional operating parameters of the valve in real time through a multi-sensor array installed on the medium and high pressure valve. The multi-dimensional operating parameters include at least: vibration, temperature, pressure and flow rate. The calculation module 302 is used to calculate the operation stability index of the medium and high pressure valve during operation based on the multi-dimensional operation parameters. The extraction module 303 is used to extract features from the multi-dimensional operating parameters using a deep learning model, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state. The evaluation module 304 is used to comprehensively evaluate the operating status of the valve based on the extracted features using a fuzzy logic algorithm, generate a valve health status evaluation report, and send the multi-dimensional operating parameters, the operating stability index, and the status evaluation report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

[0073] As can be seen, by installing a multi-sensor array on the medium- and high-pressure valve, multi-dimensional operating parameters of the valve are collected in real time; based on the multi-dimensional operating parameters, an operating stability index is calculated to measure the operating stability of the medium- and high-pressure valve during operation; features are extracted from the multi-dimensional operating parameters using a deep learning model; based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generating a valve health status assessment report, and the multi-dimensional operating parameters, the operating stability index, and the status assessment report are sent to a remote control center via a wireless network. This enables real-time monitoring and accurate evaluation of the valve, improving its operational safety and reliability, and achieving intelligent status monitoring and management.

[0074] This invention also provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0075] Specifically, in this embodiment, the storage medium can be configured to store a computer program for performing the following steps: S201, through a multi-sensor array installed on the medium and high pressure valve, collects multi-dimensional operating parameters of the valve in real time, wherein the multi-dimensional operating parameters include at least: vibration, temperature, pressure and flow rate; S202, Based on the multi-dimensional operating parameters, calculate the operating stability index of the medium and high pressure valve during operation; S203, use a deep learning model to extract features from the multi-dimensional operating parameters, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state; S204. Based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generate a valve health status assessment report, and send the multi-dimensional operating parameters, the operating stability index, and the status assessment report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

[0076] As can be seen, by installing a multi-sensor array on the medium- and high-pressure valve, multi-dimensional operating parameters of the valve are collected in real time; based on the multi-dimensional operating parameters, an operating stability index is calculated to measure the operating stability of the medium- and high-pressure valve during operation; features are extracted from the multi-dimensional operating parameters using a deep learning model; based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generating a valve health status assessment report, and the multi-dimensional operating parameters, the operating stability index, and the status assessment report are sent to a remote control center via a wireless network. This enables real-time monitoring and accurate evaluation of the valve, improving its operational safety and reliability, and achieving intelligent status monitoring and management.

[0077] This invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0078] Specifically, the aforementioned electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the aforementioned processor, and the input / output device is connected to the aforementioned processor.

[0079] Specifically, in this embodiment, the processor can be configured to perform the following steps via a computer program: S201, through a multi-sensor array installed on the medium and high pressure valve, collects multi-dimensional operating parameters of the valve in real time, wherein the multi-dimensional operating parameters include at least: vibration, temperature, pressure and flow rate; S202, Based on the multi-dimensional operating parameters, calculate the operating stability index of the medium and high pressure valve during operation; S203, use a deep learning model to extract features from the multi-dimensional operating parameters, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state; S204. Based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generate a valve health status assessment report, and send the multi-dimensional operating parameters, the operating stability index, and the status assessment report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

[0080] As can be seen, by installing a multi-sensor array on the medium- and high-pressure valve, multi-dimensional operating parameters of the valve are collected in real time; based on the multi-dimensional operating parameters, an operating stability index is calculated to measure the operating stability of the medium- and high-pressure valve during operation; features are extracted from the multi-dimensional operating parameters using a deep learning model; based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generating a valve health status assessment report, and the multi-dimensional operating parameters, the operating stability index, and the status assessment report are sent to a remote control center via a wireless network. This enables real-time monitoring and accurate evaluation of the valve, improving its operational safety and reliability, and achieving intelligent status monitoring and management.

[0081] The above description, based on the embodiments shown in the figures, details the structure, features, and effects of the present invention. The above description is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or equivalent embodiments modified to have equivalent changes, that do not exceed the spirit covered by the specification and figures, should be within the protection scope of the present invention.

Claims

1. A method for remotely monitoring the operating status of medium and high pressure valves, characterized in that, The method includes: A multi-sensor array installed on medium and high pressure valves is used to collect multi-dimensional operating parameters of the valves in real time. These multi-dimensional operating parameters include at least: vibration, temperature, pressure, and flow rate. Based on the multi-dimensional operating parameters, the operating stability index of the medium and high pressure valve during operation is calculated. The multi-dimensional operating parameters are feature extracted using a deep learning model, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state. Based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generate a valve health status assessment report, and send the multi-dimensional operating parameters, the operating stability index, and the status assessment report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

2. The method according to claim 1, characterized in that, The formula for calculating the operational stability index is as follows: ； Among them, the To ensure operational stability, the aforementioned The above The above The above The normalized values ​​for vibration, temperature, pressure, and flow rate are... The above Let be the instantaneous first derivative of the vibration and temperature at the i-th time point. The average value of the instantaneous first derivative of the vibration, where n is the number of time points. The above The above The above The above The above These are the corresponding weighting coefficients.

3. The method according to claim 2, characterized in that, The process of extracting features from the multi-dimensional operating parameters using a deep learning model, wherein the deep learning model combines convolutional neural networks and long short-term memory networks to capture the spatiotemporal features of the valve's operating state, includes: The multi-dimensional operating parameters are constructed into a multi-channel input matrix, with each channel corresponding to one operating parameter; Based on the multi-channel input matrix, multiple sets of parallel convolutional layers are used, each set of convolutional layers having convolutional kernels of different sizes, to extract the spatial features of the multi-dimensional operating parameters at different feature scales. Among them, small-scale convolutional kernels are used to extract local detail features, while large-scale convolutional kernels are used to capture global structural features. The extracted spatial features are organized into time series data, wherein the spatial features are mapped into time series form to form a new set of input features for capturing patterns that evolve over time. By utilizing a bidirectional LSTM layer, the forward and backward dependencies in the time series data are captured simultaneously to capture the temporal dependencies and obtain the temporal features of the multi-dimensional operating parameters. The spatial features and the temporal features are merged into a high-dimensional spatiotemporal feature vector through feature concatenation operations.

4. The method according to claim 3, characterized in that, Based on the extracted features, a fuzzy logic algorithm is used to comprehensively evaluate the valve's operating status, generating a valve health status assessment report, including: Based on the spatiotemporal feature vector, multiple fuzzy sets are obtained, each fuzzy set corresponding to a specific valve operating state feature; A fuzzy rule base is constructed, in which fuzzy rules evaluate the health status of valves based on the combination of multi-dimensional features. Each rule consists of a premise and a conclusion. The premise part uses fuzzy logic operators to combine fuzzy sets of multiple features, and the conclusion part gives the health status evaluation of the valve. A fuzzy inference engine is constructed, wherein the fuzzy inference engine adopts the Mamdani or Sugeno fuzzy inference algorithm. The fuzzy inference engine receives the input feature vector, converts it into the membership degree of the fuzzy set through the fuzzification process, performs inference using the fuzzy rule base, calculates the activation strength of each rule, and generates fuzzy output based on the conclusion part of the rule. The fuzzy outputs of all activation rules are aggregated, using either the max-min synthesis method or the weighted average method. The aggregated result is a comprehensive fuzzy output that represents the membership of the valve's health status in different state categories. The aggregated fuzzy output is defuzzified to obtain a clear health status assessment value. The defuzzification methods include the centroid method, the maximum membership method, and the average maximum method. The result of the defuzzification is a specific numerical value or level, representing the health status of the valve. Based on the deblurring results, a valve health status assessment report is generated, which includes a numerical assessment of the current health status, analysis of possible causes of failure, recommended maintenance measures, and predictions of future status.

5. A remote monitoring system for the operating status of medium and high pressure valves, characterized in that, The system includes: The acquisition module is used to acquire multi-dimensional operating parameters of the valve in real time through a multi-sensor array installed on the medium and high pressure valve. The multi-dimensional operating parameters include at least: vibration, temperature, pressure and flow rate. The calculation module is used to calculate the operational stability index of the medium and high pressure valve during operation based on the multi-dimensional operating parameters. The extraction module is used to extract features from the multi-dimensional operating parameters using a deep learning model, wherein the deep learning model combines a convolutional neural network and a long short-term memory network to capture the spatiotemporal features of the valve's operating state. The evaluation module is used to comprehensively evaluate the operating status of the valve based on the extracted features and using a fuzzy logic algorithm, generate a valve health status evaluation report, and send the multi-dimensional operating parameters, the operating stability index and the status evaluation report to a remote control center via a wireless network for remote monitoring of the operating status of medium and high pressure valves.

6. The system according to claim 5, characterized in that, The formula for calculating the operational stability index is as follows: ； Among them, the To ensure operational stability, the aforementioned The above The above The above The normalized values ​​for vibration, temperature, pressure, and flow rate are... The above Let be the instantaneous first derivative of the vibration and temperature at the i-th time point. The average value of the instantaneous first derivative of the vibration, where n is the number of time points. The above The above The above The above The above These are the corresponding weighting coefficients.

7. The system according to claim 6, characterized in that, The extraction module is specifically used for: The multi-dimensional operating parameters are constructed into a multi-channel input matrix, with each channel corresponding to one operating parameter; Based on the multi-channel input matrix, multiple sets of parallel convolutional layers are used, each set of convolutional layers having convolutional kernels of different sizes, to extract the spatial features of the multi-dimensional operating parameters at different feature scales. Among them, small-scale convolutional kernels are used to extract local detail features, while large-scale convolutional kernels are used to capture global structural features. The extracted spatial features are organized into time series data, wherein the spatial features are mapped into time series form to form a new set of input features for capturing patterns that evolve over time. By utilizing a bidirectional LSTM layer, the forward and backward dependencies in the time series data are captured simultaneously to capture the temporal dependencies and obtain the temporal features of the multi-dimensional operating parameters. The spatial features and the temporal features are merged into a high-dimensional spatiotemporal feature vector through feature concatenation operations.

8. The system according to claim 7, characterized in that, The evaluation module is specifically used for: Based on the spatiotemporal feature vector, multiple fuzzy sets are obtained, each fuzzy set corresponding to a specific valve operating state feature; A fuzzy rule base is constructed, in which fuzzy rules evaluate the health status of valves based on the combination of multi-dimensional features. Each rule consists of a premise and a conclusion. The premise part uses fuzzy logic operators to combine fuzzy sets of multiple features, and the conclusion part gives the health status evaluation of the valve. A fuzzy inference engine is constructed, wherein the fuzzy inference engine adopts the Mamdani or Sugeno fuzzy inference algorithm. The fuzzy inference engine receives the input feature vector, converts it into the membership degree of the fuzzy set through the fuzzification process, performs inference using the fuzzy rule base, calculates the activation strength of each rule, and generates fuzzy output based on the conclusion part of the rule. The fuzzy outputs of all activation rules are aggregated, using either the max-min synthesis method or the weighted average method. The aggregated result is a comprehensive fuzzy output that represents the membership of the valve's health status in different state categories. The aggregated fuzzy output is defuzzified to obtain a clear health status assessment value. The defuzzification methods include the centroid method, the maximum membership method, and the average maximum method. The result of the defuzzification is a specific numerical value or level, representing the health status of the valve. Based on the deblurring results, a valve health status assessment report is generated, which includes a numerical assessment of the current health status, analysis of possible causes of failure, recommended maintenance measures, and predictions of future status.

9. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method of any one of claims 1-4 when it is run.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method of any one of claims 1-4.

Images