Vacuum pump station management method, system and equipment and storage medium

Abstract

The invention provides a vacuum pump station management method, system and equipment and a storage medium, and relates to the field of vacuum pump management. The method comprises the steps of obtaining key operation parameters of a target vacuum pump station, and performing fusion processing based on the key operation parameters to obtain a multi-dimensional operation feature vector; inputting the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result; based on the fault judgment result and a preset fault processing strategy, a target fault processing strategy is determined, and the vacuum pump station is maintained based on the target fault processing strategy; and inputting the multi-dimensional operation feature vector into a preset operation parameter optimization model, outputting a target optimization parameter, and regulating and controlling the target vacuum pump station based on the target optimization parameter. Through the method, the management efficiency and reliability of the vacuum pump station are improved.

Description

Technical Field

[0001] The present application relates to the field of vacuum pump management, and particularly to a method, a system, a device and a storage medium for managing a vacuum pump station. Background Art

[0002] Vacuum equipment is a key equipment indispensable in the manufacturing of many high-tech industries such as semiconductors, photovoltaics, and new energy. Among them, as the core unit for obtaining and maintaining vacuum, the operating state of the vacuum pump station directly affects the stability of the vacuum process and the consistency of product quality. With the continuous improvement of process requirements and the continuous increase in equipment complexity, the operation and maintenance management of the vacuum pump station faces huge challenges.

[0003] Currently, the operation and maintenance management of the vacuum pump station mainly relies on regular manual inspections and experience judgments. Engineers evaluate the health status of the equipment based on the operating parameters of the pump station such as current, speed, vibration, etc., combined with their own experience, and formulate corresponding maintenance plans. However, relying on manual labor to manage the vacuum pump station is inefficient and unreliable. Summary of the Invention

[0004] The present application provides a method, a system, a device and a storage medium for managing a vacuum pump station, which can improve the efficiency and reliability of vacuum pump station management.

[0005] In a first aspect, the present application provides a method for managing a vacuum pump station, the method comprising: Obtaining key operating parameters of a target vacuum pump station based on a sensor matrix arranged in the target vacuum pump station, and performing fusion processing on the key operating parameters to obtain a multi-dimensional operating feature vector; Inputting the multi-dimensional operating feature vector into a preset fault judgment model to obtain a fault judgment result; Determining a target fault handling strategy based on the fault judgment result and a preset fault handling strategy, and maintaining the vacuum pump station based on the target fault handling strategy; Inputting the multi-dimensional operating feature vector into a preset operating parameter optimization model, outputting target optimization parameters, and regulating the target vacuum pump station based on the target optimization parameters.

[0006] By adopting the above technical scheme, the vacuum pump station intelligent management method of the present invention obtains the key operating parameters of the pump station and performs multi-dimensional fusion processing on them to form a comprehensive operating feature vector, which provides a comprehensive and accurate data basis for subsequent fault diagnosis, maintenance decision-making and operation optimization. Since the operating status of the vacuum pump station is affected by the interaction of multiple parameters, it is difficult for a single parameter to fully reflect the health level of the equipment. The present invention adopts multiple feature extraction methods such as time domain, frequency domain, time-frequency domain, etc. to characterize the equipment status from different dimensions, and performs feature fusion to form an information-rich and comprehensive operating feature vector. This "wide fusion" data processing method effectively improves the completeness and accuracy of equipment status perception, and lays a solid foundation for intelligent management.

[0007] On this basis, the present invention inputs the multi-dimensional operation feature vector into the preset fault judgment model to achieve rapid and accurate diagnosis of equipment faults. Due to the integration of equipment mechanism knowledge and operation big data, and the use of extended fault sample strategy to train and optimize the model, the fault judgment model has strong generalization ability and robustness, can cope with complex and changeable actual working conditions, and reduce the rate of missed detection and false positives. By timely discovering and locating potential faults of equipment, it is helpful to take targeted measures before the performance of the equipment deteriorates or is damaged, thereby minimizing unplanned downtime and improving equipment reliability and process continuity.

[0008] According to the fault judgment result, the present invention also proposes an intelligent maintenance decision-making method. By analyzing the fault type and severity and matching it with the preset fault handling strategy library, the optimal maintenance plan is automatically recommended. At the same time, a feedback verification mechanism for diagnostic confidence is introduced to re-judge the diagnostic results with low confidence, thereby improving the reliability of maintenance decisions. This knowledge-based intelligent decision-making method, on the one hand, utilizes the historical operation and maintenance experience of the equipment to improve the pertinence and effectiveness of maintenance; on the other hand, it formalizes and standardizes the decision-making process, reduces the arbitrariness of manual experience judgment, and realizes the specialization and refinement of maintenance management.

[0009] On the basis of diagnosis and decision-making, the present invention further realizes the intelligent optimization of the operating parameters of the pump station. By analyzing the performance parameters of the pump station to determine the optimization constraints, and integrating the historical optimization operation data to build an optimization model, the optimal control parameters are adaptively output, and the pump station is controlled online. Compared with the empirical parameter adjustment, this method can continuously tap the performance potential of the equipment while ensuring the safe and stable operation of the equipment, so that the pump station always works in the best working conditions, effectively improving the vacuum process level and production efficiency. At the same time, the optimization control forms a closed loop with fault diagnosis and maintenance decision-making, realizing "maintaining the best with the best", and promoting the simultaneous improvement of equipment reliability and management level.

[0010] In the second aspect of the present application, a vacuum pump station management method system is provided, including: A data acquisition module, configured to acquire key operation parameters of a target vacuum pump station based on a sensor matrix arranged in the target vacuum pump station, and perform fusion processing on the key operation parameters to obtain a multi-dimensional operation feature vector; A fault judgment module, configured to input the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result; A fault maintenance module, configured to determine a target fault handling strategy based on the fault judgment result and a preset fault handling strategy, and perform maintenance on the vacuum pump station based on the target fault handling strategy; An operation parameter optimization module, configured to input the multi-dimensional operation feature vector into a preset operation parameter optimization model, output target optimization parameters, and perform regulation on the target vacuum pump station based on the target optimization parameters.

[0011] In the third aspect of the present application, a computer storage medium is provided. The computer storage medium stores multiple instructions, and the instructions are suitable for being loaded and executed by a processor to perform the above method steps.

[0012] In the fourth aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs the above method.

[0013] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By acquiring the key operation parameters of the pump station and performing multi-dimensional fusion processing on them to form a comprehensive operation feature vector, the present application provides a comprehensive and accurate data basis for subsequent fault diagnosis, maintenance decision-making, and operation optimization. Since the operation state of the vacuum pump station is affected by the interaction of multiple parameters, a single parameter is difficult to comprehensively reflect the health level of the equipment. The present invention uses various feature extraction methods such as time domain, frequency domain, and time-frequency domain to describe the equipment state from different dimensions and perform feature fusion to form an operation feature vector with rich information and comprehensive dimensions. This "wide fusion" data processing method effectively improves the completeness and accuracy of equipment state perception and lays a solid foundation for intelligent management.

[0014] 2. Based on the fault judgment results, the present invention also proposes an intelligent maintenance decision-making method. By analyzing the fault type and severity and matching them with a preset fault handling strategy library, the optimal maintenance plan is automatically recommended. At the same time, a feedback verification mechanism for diagnostic confidence is introduced to re-judge the diagnostic results with low confidence, improving the reliability of maintenance decisions. This knowledge-based intelligent decision-making method, on the one hand, utilizes the historical operation and maintenance experience of the equipment to improve the pertinence and effectiveness of maintenance; on the other hand, formalizes and standardizes the decision-making process, reduces the arbitrariness of manual experience judgment, and realizes the specialization and refinement of maintenance management.

[0015] 3. The present application determines the optimization constraints by analyzing the performance parameters of the pumping station, constructs an optimization model by integrating historical optimal operation data, and adaptively outputs the optimal control parameters to implement online regulation of the pumping station. Compared with empirical parameter tuning, this method can continuously explore the performance potential of the equipment on the premise of ensuring the safe and stable operation of the equipment, enabling the pumping station to always operate under the best working conditions, effectively improving the vacuum process level and production efficiency. At the same time, the optimization control, fault diagnosis, and maintenance decision-making form a closed loop, realizing "maintaining excellence with excellence", and promoting the synchronous improvement of the reliability of the equipment and the management level. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic flow chart of a vacuum pumping station management method provided by an embodiment of the present application; Figure 2 It is an architecture diagram of a vacuum pumping station management system provided by an embodiment of the present application; Figure 3 It is a schematic structural diagram of an electronic device provided by the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0018] In the description of the embodiments of the present application, words such as "for example" or "for illustration" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "for example" or "for illustration" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, using words such as "for example" or "for illustration" is intended to present relevant concepts in a specific manner.

[0019] In the description of the embodiments of the present application, the term "plurality" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The terms "include", "comprise", "have" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0020] To facilitate the understanding of the method and system provided by the embodiments of the present application, before introducing the embodiments of the present application, the background of the embodiments of the present application will be introduced first.

[0021] Vacuum equipment is a key equipment indispensable for the manufacturing of many high-tech industries and plays a crucial role in fields such as semiconductors, photovoltaics, and new energy. As the core unit for vacuum acquisition and maintenance, the operating state of the vacuum pump station is directly related to the stability of the vacuum process and the consistency of product quality. However, with the increasing process requirements and the continuous increase in equipment complexity, the operation and maintenance management of the vacuum pump station is facing unprecedented challenges.

[0022] Currently, the operation and maintenance management of the vacuum pump station mainly relies on regular manual inspections and experience-based judgments. Engineers need to make a comprehensive assessment of the health status of the equipment based on various operating parameters of the pump station, such as current, rotational speed, vibration, etc., and combined with the experience accumulated over the years, and then formulate corresponding maintenance plans. Although this traditional management mode ensures the operation of the equipment to a certain extent, the problems of its low efficiency and unreliability are becoming increasingly prominent.

[0023] First of all, it is difficult for manual inspections to achieve all-weather and real-time equipment monitoring, and there are supervision blind spots and time lags. Secondly, experience-based judgments are easily affected by personnel capabilities and subjective factors, and it is difficult to ensure the accuracy and consistency of diagnosis and decision-making. Moreover, in the face of ever-changing equipment and processes, it is difficult for personnel's professional skills to be quickly iterated and adapted. More critically, manual management is difficult to process massive and high-dimensional equipment operation data, unable to discover the hidden associations and laws therein, and difficult to accurately depict and predict the equipment state.

[0024] Therefore, in today's high-speed development of advanced manufacturing, the traditional manual management mode is already difficult to meet the requirements of efficient and reliable operation and maintenance of vacuum pump stations. More and more enterprises have begun to realize that promoting the upgrade of vacuum pump station management towards automation and intelligence has become the only way to improve the comprehensive operation level of equipment, ensure safe and efficient production, and build the core competitiveness of the industry. This urgently requires industry colleagues to go deep into the front line of the process, analyze the characteristics of equipment, and explore the implementation path of intelligent operation and maintenance of vacuum pump stations from a forward-looking technical perspective and innovative management thinking, so as to provide strong support for the transformation and upgrading of China's high-end manufacturing industry.

[0025] After the background introduction of the above content, those skilled in the art can understand the problems existing in the prior art. Next, the technical solutions in the embodiments of the present application will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments.

[0026] Based on the above background technology, further, please refer to Figure 1 , Figure 1 which is a schematic flow chart of a vacuum pump station management method provided by an embodiment of the present application. This system can be implemented depending on a computer program or run as an independent tool application. Specifically, in the embodiment of the present application, this method can be applied to a server, but can also be applied to electronic devices such as a server. A vacuum pump station management method includes the following steps: S101, obtain the key operation parameters of the target vacuum pump station, and perform fusion processing based on the key operation parameters to obtain a multi-dimensional operation feature vector; Specifically, in order to comprehensively and accurately perceive the operation state of the vacuum pump station, it is first necessary to obtain the key operation parameters of the pump station. By deeply analyzing the process flow and equipment composition of the vacuum pump station, the key equipment that has the greatest impact on the operation state of the pump station is identified, such as vacuum pumps, cold traps, vacuum gauges, etc., and the corresponding key process parameters, such as pump current, ultimate vacuum, cold trap temperature, etc. After determining the monitoring objects and monitoring indicators, data collection is realized by arranging sensors at key positions, and various sensing signals are synchronously sampled, filtered and amplified, and A / D converted by using hardware facilities such as bus technology and data acquisition cards, and finally a unified digital data stream is formed to provide data input for subsequent fusion processing.

[0027] Based on the obtained key operating parameters, the multi-source heterogeneous data fusion technology is further adopted to comprehensively process the parameter data. For different types of parameters, feature extraction methods such as time domain, frequency domain, and time-frequency domain are respectively used to extract sensitive features that can reflect the operating state of the equipment. For continuous variables such as pump current, its time-domain statistics such as mean, variance, and peak-to-peak value are mainly extracted; for discrete variables such as pump speed, the number of jumps and the duration are focused on; for periodic signals such as pump vibration, its frequency-domain features are obtained through spectrum analysis. On the basis of feature extraction, data dimensionality reduction methods such as principal component analysis and independent component analysis are used to fuse various features, remove redundancy and noise between features, and finally form a multi-dimensional operating feature vector with lower dimensions and more concentrated information. This fusion processing can explore the internal relationship of parameter data from two dimensions of time and space, reveal the evolution law of the equipment state, and thus realize the dynamic and global characterization of the operating state of the pump station.

[0028] On the basis of the above embodiments, as an alternative embodiment, before obtaining the key operating parameters of the target vacuum pump station based on the sensor matrix arranged in the target vacuum pump station, it further includes: S201 Obtain the process flow and equipment characteristics of the target vacuum pump, and identify key equipment based on the process flow and the equipment characteristics to obtain key equipment information; Specifically, it is first necessary to obtain the process flow diagram and equipment layout diagram of the target vacuum pump station, understand the basic principles and paths of vacuum acquisition and gas transportation therein, and master the structural characteristics, working principles, and technical parameters of each equipment. On this basis, the analytic hierarchy process, expert evaluation method, etc. are used to comprehensively evaluate the importance of each equipment to the performance of the pump station from the perspectives of equipment performance, system matching, fault impact, etc., and form a quantitative importance weight index. Sort and screen the evaluation results according to the importance degree, and the list of key equipment with the greatest impact on the operating state of the pump station can be obtained, which is used as the key object for subsequent monitoring, diagnosis, and optimization control. Further, it is also necessary to obtain detailed information such as the nameplate parameters, design drawings, and factory inspection reports of these key equipment, which are used as important references for evaluating their health status and performance level.

[0029] By carrying out the above-mentioned key equipment identification work, we can more focusedly monitor and analyze those "bull nose" equipment that has the most significant impact on the performance of the pump station, which not only reduces the workload of data collection and processing, but also helps to grasp the main contradictions between the evolution of equipment performance and the development of faults, and avoid interference and loss of focus during the analysis process. At the same time, a comprehensive grasp of the technical parameters and performance requirements of key equipment also provides important support for the subsequent construction of accurate health assessment models and optimized control strategies. Managers can focus limited resources on the most needed and valuable equipment, and formulate more targeted and differentiated management strategies, thereby effectively improving the performance of the pump station while minimizing management costs and improving management efficiency.

[0030] S202: Determine sensor types and measurement point locations based on the key equipment information, and arrange the sensor matrix in the target vacuum pump station based on the sensor types and the measurement point locations.

[0031] Specifically, we first need to clarify the monitoring needs and purposes of each key equipment, and determine the most critical physical parameters for equipment status assessment, such as temperature, pressure, flow, vibration, noise, etc., based on their failure modes and degradation mechanisms. At the same time, we comprehensively consider the technical parameters, process requirements, environmental conditions and other factors of the equipment, and put forward clear requirements for the range, accuracy, frequency, power consumption, etc. of the monitoring parameters. On this basis, we comprehensively investigate the performance indicators, interface methods, supplier qualifications, etc. of various sensors, and through quantitative evaluation and scheme comparison, find out the sensor model that meets the monitoring needs and has the best cost-effectiveness, and make necessary customized improvements based on the on-site installation conditions, and finally determine the collection plan and sensor configuration list for each key parameter.

[0032] After determining the sensor type, it is further necessary to select the best measuring point location for the specific equipment. The ideal measuring point location should be able to sensitively reflect the key performance parameters of the equipment and have good stability, reliability and maintainability. By analyzing the internal structure, heat and mass transfer characteristics, stress distribution law, etc. of the equipment, the key components and weak points that play a decisive role in the performance of the equipment are identified, and sensor measuring points are arranged at locations such as fault-prone points, stress concentration points, and points where the temperature field changes drastically. At the same time, the actual working conditions such as the installation method, vibration level, corrosion environment, and electromagnetic interference of the equipment are comprehensively considered to optimize the installation method and wiring path of the sensor, such as using wireless transmission and anti-interference cables, to improve the stability and reliability of data acquisition. If necessary, simulation analysis and experimental verification of sensor point layout are also required to evaluate the sensitivity and representativeness of different measuring point layout schemes and select the best point layout combination scheme.

[0033] By optimizing the sensor type and measuring point location, the quality and reliability of monitoring data can be significantly improved, and information redundancy and false alarms can be reduced. A reasonably matched sensor type can fully consider the equipment characteristics and monitoring requirements, reduce the acquisition cost while meeting the accuracy requirements, and relieve the pressure of data analysis and storage. The optimized measuring point layout can focus on the key information of the equipment state, achieve comprehensive and in-depth condition monitoring with the fewest measuring points, maximize the representativeness of the collected data without affecting the equipment performance, and reduce information loss. High-quality monitoring data is the prerequisite for intelligent diagnosis and optimal control, which can better describe the health state of the equipment, reveal its degradation law, provide a reliable basis for anomaly detection, fault diagnosis, life prediction, etc., and ultimately realize the refinement and intelligence of equipment management and ensure the long-term stable operation of the vacuum pump station.

[0034] Based on the above embodiments, as an optional embodiment, fusion processing is performed on the key operating parameters to obtain a multi-dimensional operating feature vector, including: S301, preprocess the key operating parameters to obtain the preprocessed key operating parameters, and perform time-domain analysis on the preprocessed key operating parameters to obtain time-domain characteristic parameters; S302, perform time-frequency analysis on the preprocessed key operating parameters to obtain time-frequency characteristic parameters; S303, perform frequency-domain analysis on the preprocessed key operating parameters to obtain frequency-domain characteristic parameters; S304, perform feature fusion on the time-domain characteristic parameters, time-frequency characteristic parameters, and frequency-domain characteristic parameters to obtain a multi-dimensional operating feature vector.

[0035] Specifically, it is first necessary to perform necessary preprocessing on the collected key operating parameters, including outlier removal, data smoothing, noise filtering, data normalization, etc., to improve the data quality and eliminate the influence of parameters with different dimensions. On this basis, time-domain analysis methods are used to calculate the time-domain statistical characteristics such as the mean, variance, root mean square, and kurtosis of the preprocessed parameters, so as to describe the state characteristics such as the vibration level, temperature field distribution, and energy fluctuation of the equipment. Then, time-frequency analysis tools such as short-time Fourier transform and wavelet transform are used to extract the spectral components and energy distribution characteristics of the parameters at different time scales, and to reveal the dynamic evolution law of the equipment operating state over time. Further, Fourier transform is performed on the preprocessed parameter signal to obtain frequency-domain characteristic parameters such as the amplitude spectrum, phase spectrum, and multiple-frequency characteristics in the frequency domain, analyze the natural frequency of the equipment, the rotational frequency components of key components, etc., and realize the accurate description of the equipment structure characteristics and early faults.

[0036] After obtaining the characteristic parameters in various fields, feature dimension reduction methods such as principal component analysis and independent component analysis are used to remove the information redundancy between different characteristic parameters and extract the core feature vectors that can best represent the equipment state. At the same time, combined with expert experience and mechanism knowledge, the importance of the characteristics in each field is weighted and evaluated to form a multi-dimensional comprehensive feature vector that integrates the key information in the time domain, frequency domain, and time-frequency domain. This feature vector comprehensively depicts the operating state of the equipment from three core dimensions of time, frequency, and energy, with the advantages of comprehensive information, significant features, and optimal dimensions, and can more finely and dynamically perceive and evaluate the health level of the equipment.

[0037] Through the above multi-dimensional feature fusion method based on key operating parameters, the internal connections and regularities of equipment state data can be maximally explored, the interactions and causal relationships between different physical quantities can be revealed, and the leap from data to information and from information to knowledge can be achieved. The high-quality multi-dimensional feature vector provides a more reliable and comprehensive data basis for intelligent fault diagnosis, life prediction, etc., and greatly improves the accuracy and predictability of equipment health management. At the same time, in the process of feature extraction, the combination of expert knowledge and machine learning methods is also an important manifestation of the transformation from experience-driven to data-driven, promoting the complementary integration of mechanism modeling and data mining, and forming a more scientific and efficient new equipment management model. With the help of multi-dimensional state features, managers can comprehensively and intuitively master the health status of the equipment, timely warn of abnormal risks, optimize maintenance strategies, and finally achieve the refinement, prevention, and intelligence of equipment management, ensuring the safe and economic operation of the vacuum pump station.

[0038] S102, input the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result; Specifically, after obtaining the multi-dimensional operation feature vector of the vacuum pump station, in order to realize the intelligent optimization of the equipment operation state, it is necessary to input this feature vector into a preset operation parameter optimization model. The operation parameter optimization model is an intelligent optimizer based on reinforcement learning. By analyzing historical optimization cases and operation data, it autonomously learns the optimal control strategy to make the target performance index reach the optimal under the premise of meeting the constraints of equipment safety and stable operation.

[0039] The key to building an operating parameter optimization model lies in the determination of the optimization objective and the design of the reinforcement learning algorithm. First, according to the technological characteristics and management requirements of the vacuum pump station, select performance indicators that can comprehensively reflect the operation quality and efficiency of the equipment as the optimization objective, such as the comprehensive equipment efficiency, energy consumption index, failure rate, etc. On the basis of clarifying the optimization direction, combined with the mechanism model and constraint conditions of the pump station, use reinforcement learning algorithms such as Q-learning and policy gradient to build the optimization model framework. By designing a reasonable state space, action space, and reward function, the intelligent agent can seek a balance between the objective and the constraints, continuously try and improve the control strategy, and finally learn the optimal parameter combination.

[0040] When the newly collected multi-dimensional operation feature vector is input into the trained operation parameter optimization model, the model can dynamically adjust the key control parameters of the pump station, such as pump speed, cold trap temperature, gas volume, etc., based on the current equipment state and optimization objective, and output a set of real-time optimized target parameter values. Based on the target optimization parameters output by the model, the control system can accurately adjust the operation conditions of the vacuum pump station, continuously optimize and improve the equipment performance. Through the coordinated cooperation of online optimization control, fault diagnosis, and maintenance decision-making, the closed-loop self-optimization of equipment management is truly realized, and the vacuum pump station is always maintained in the best health state and production state.

[0041] On the basis of the above embodiments, as an alternative embodiment, before obtaining the fault judgment result by inputting the multi-dimensional operation feature vector into the preset fault judgment model, it further includes: S401, obtaining the vacuum pump station fault instance sample data, and performing extended extraction on the vacuum pump station fault instance sample data to obtain a multi-dimensional fault sample feature vector; S402, building a preliminary fault judgment model, and training the preliminary fault judgment model based on the multi-dimensional fault sample feature vector. When the fault judgment result reaches the preset accuracy threshold, the preset fault judgment model is obtained.

[0042] Specifically, first of all, it is necessary to widely collect the failure cases of vacuum pumping stations under various working conditions, including common failure types such as sudden change of vacuum degree, reduced pumping rate, abnormal vibration and noise, and fluctuation of power consumption, so as to obtain a typical sample set reflecting different failure modes and different degradation stages. On this basis, referring to the fusion processing method for constructing the operation feature vector mentioned above, the time series parameter data of each fault sample is preprocessed and feature extracted, and the characteristic parameters of the fault sample are calculated from different dimensions by using time domain analysis, frequency domain analysis, time-frequency analysis and other methods, and finally a multi-dimensional fault sample feature vector isomorphic to the operation feature vector is formed through feature selection and fusion strategies. In the extraction process, it is necessary to make full use of the latest technological achievements in the fields of signal processing and anomaly detection, design targeted feature engineering solutions, maximize the development of effective information of fault samples, and make up for the problem of insufficient sample quantity. At the same time, it is also necessary to comprehensively use simulation experiments, data interpolation, rotating machinery knowledge, etc. to expand and enhance the original fault samples, and to mine more feature dimensions and value combinations from the stock data without distortion, so as to comprehensively cover various fault conditions that may occur in vacuum pumping stations.

[0043] After obtaining high-quality multi-dimensional fault sample feature vectors, it is necessary to build a preliminary fault judgment model as the starting point for subsequent training optimization. According to the complexity of the vacuum pump station equipment structure and fault mechanism, different types of machine learning models such as support vector machines, decision trees, and artificial neural networks can be selected, and the model's network structure, loss function, learning rate and other hyperparameters are preliminarily set with reference to the prior knowledge of the diagnosis object to build a basic model with initial diagnostic capabilities. Then, the extended extracted fault sample feature vectors are divided into training sets and test sets, and the preliminary model is repeatedly optimized through closed-loop training-testing-feedback-correction. By continuously feeding a large number of fault feature samples to the model, it can fully learn the correspondence between fault signs and fault types, and at the same time introduce cross-validation, regularization and other methods to improve the generalization performance and anti-interference ability of the model. When the model's fault judgment accuracy, precision and other indicators on the test set continuously reach the preset performance threshold, the training process can be terminated, and finally a preset fault judgment model with excellent performance and strong robustness is obtained.

[0044] Through the above-mentioned preset diagnostic model training method based on extended fault samples, the value-added development of fault data is fully utilized, multi-dimensional fine-grained features of fault signals are maximally extracted from limited samples, the information dimension and sample capacity for model learning are expanded, and the feasibility and diagnostic performance of data-driven intelligent diagnosis are significantly improved. High-quality sample features provide a comprehensive and balanced reference standard for model training, enabling the model to more meticulously and comprehensively depict the internal mechanism and evolution law of vacuum pump station faults, enhancing its adaptability to complex working conditions, and achieving a leap from shallow pattern recognition to deep knowledge mining. At the same time, mechanism knowledge and expert experience are incorporated into the training process to achieve the collaborative enhancement of data and knowledge, further ensuring the interpretability and credibility of diagnostic results. The generalization ability of the model is also fully improved, and it can be applied to the fault diagnosis of vacuum pump stations of different types and in different application scenarios, reducing the costs of model reconstruction and secondary development, and clearing the obstacles for the engineering application of intelligent diagnosis systems.

[0045] S103. Based on the fault judgment result and the preset fault handling strategy, determine the target fault handling strategy, and maintain the vacuum pump station based on the target fault handling strategy; Specifically, after obtaining the fault judgment result of the vacuum pump station, in order to achieve the rapid handling of faults and the timely repair of equipment, it is necessary to determine the optimal target fault handling strategy based on this diagnosis result and the preset fault handling strategy. The fault handling strategy is a set of standardized fault response plans pre-developed according to factors such as fault type, severity, and occurrence stage, covering the entire process from fault confirmation, cause analysis to repair decision-making and effect evaluation. By applying standardized handling strategies throughout the entire life cycle of the vacuum pump station, the fault downtime can be minimized to the greatest extent, and the repair efficiency and success rate can be improved.

[0046] The process of determining the target fault handling strategy is essentially a rule-based decision-making matching process. First, according to the structural characteristics and fault mechanism of the pump station equipment, summarize and induce the handling experience of various faults to form a complete fault handling knowledge base. Each handling strategy in the knowledge base contains key elements such as fault characteristics, judgment conditions, handling steps, and precautions, so that it can be quickly retrieved and matched during specific applications. When a new fault judgment result is generated, the system can search for the handling strategy that meets the judgment conditions in the knowledge base according to the fault type and relevant parameters, and combine the current equipment status and maintenance resource situation to select the optimal solution from the alternative strategies as the target strategy to guide subsequent maintenance operations.

[0047] After determining the target fault handling strategy, the maintenance personnel can carry out the equipment maintenance work in an orderly manner according to the specific guidance provided by the strategy. Generally speaking, the fault handling strategy will specify in detail the steps, methods and requirements for fault handling, such as the inspection items for fault confirmation, the diagnostic ideas for fault causes, the basic process of repair operations, the assessment criteria for effect evaluation, etc., so that the maintenance process is controllable and traceable. The maintenance personnel only need to strictly execute the instructions of the strategy to quickly lock the fault location, accurately judge the fault cause, and take targeted repair measures, ultimately achieving the purpose of troubleshooting and resuming production. Through the guidance and standardization of the preset strategy, the maintenance efficiency and standardization of front-line personnel can be significantly improved, and the risk of secondary damage caused by human factors such as misoperation and missed judgment can be reduced.

[0048] Based on the above embodiments, as an alternative embodiment, a target fault handling strategy is determined based on the fault judgment result and the preset fault handling strategy, including: Determine the fault type and fault severity according to the fault judgment result, and match the fault type and fault severity with the preset fault handling strategy, and select the fault handling strategy with the highest matching degree as the target fault handling strategy.

[0049] Specifically, it is first necessary to comprehensively analyze the diagnostic results given by the fault judgment model, and extract decision key elements such as fault type and fault severity. Generally, information extraction technologies based on rules or machine learning are used to design a hierarchical classification system and keyword dictionary for fault types and severity levels, and then the semantic understanding and quantitative evaluation of the diagnostic result text are carried out, and they are transformed into structured and computable fault attribute parameters. For example, the fault type is divided into several major categories such as bearing fault, rotor fault, lubrication fault, etc., and the fault severity is quantified as a risk index from 0 to 10. Through methods such as template matching and regular expressions, the diagnostic results are automatically classified and scored, and finally a standardized fault description vector is formed.

[0050] After clarifying the detailed attributes of the fault judgment result, it is necessary to match and compare it with the preset fault handling strategy library to find the most suitable disposal plan for the current fault situation. The preset strategy library adopts various forms such as rules, cases, decision trees, etc., and formulates standardized processing procedures and key control points for different types and degrees of faults, covering the whole process from fault detection, risk assessment to maintenance decision-making and resource scheduling. When matching, an inference method based on similarity calculation is adopted, comprehensively considering multiple feature dimensions such as fault type, severity, and pump station working conditions, calculating the similarity between the fault to be processed and each fault case in the knowledge base, and screening out a batch of historical cases highly similar to the current fault by setting a matching degree threshold. Then, the adaptability evaluation and optimization of the processing strategies of these alternative cases are carried out. Combining the real-time operation parameters, environmental conditions, resource status, etc. of the pump station, the expected effects, implementation difficulties, and economic costs of each strategy are quantitatively analyzed, and weighted ranking is carried out. Finally, the strategy with the highest comprehensive matching degree is selected as the optimal target fault handling strategy. This strategy optimization mechanism makes full use of a large amount of historical knowledge and real-time dynamic information, can give the best disposal plan in a very short time after the fault occurs, and greatly improves the timeliness and reliability of fault handling.

[0051] S104. Input the multi-dimensional operation feature vector into the preset operation parameter optimization model, output the target optimization parameters, and regulate the target vacuum pump station based on the target optimization parameters.

[0052] Specifically, after obtaining the multi-dimensional operation feature vector of the vacuum pump station, in order to realize the intelligent optimization control of the equipment operation state, it is necessary to input this feature vector into the preset operation parameter optimization model, output the optimized target control parameters, and based on these parameters, carry out real-time regulation of the operation conditions of the vacuum pump station. The operation parameter optimization model is a data-driven intelligent optimizer. Through in-depth learning of a large amount of historical operation data and process mechanism knowledge, it has mastered the best control strategies under various working conditions of the equipment, and can dynamically calculate the optimal combination of control parameters according to the actual state and performance requirements of the equipment, guiding the vacuum pump station to continuously operate near the best working point.

[0053] The key to building an optimized model for operating parameters lies in the selection of optimization objectives and the design of algorithm models. First, considering the process requirements and management objectives of the vacuum pump station comprehensively, one or more quantifiable and assessable key performance indicators are selected as the optimization objective function, such as system energy consumption, gas pumping speed, ultimate vacuum degree, etc. On the basis of clarifying the optimization direction, a mathematical model for multi-objective optimization is built by comprehensively using vacuum physical models and machine learning algorithms. By learning from historical optimization cases, the internal relationship between the pump station state and the optimal control parameters is extracted. Through parameter optimization and iterative calculation, on the premise of meeting the equipment safety and production constraints, the best balance point of each performance indicator is found, forming a mapping set of equipment state - optimal parameters. When a new feature vector is input into the model, the mapping relationship most similar to the current state can be quickly searched, and the corresponding target optimization parameters can be output.

[0054] Based on the target optimization parameters output by the model, the vacuum control system will timely adjust the operating settings of the pump station and implement set-point tracking control for key control variables to ensure that the operating parameters of equipment such as vacuum pumps, cold traps, and pipelines meet the requirements of the optimization objectives. Specifically, the optimized control instructions are sent to the on-site programmable logic controller (PLC) in the form of digital or analog signals, and then the PLC drives each actuator such as frequency converters and regulating valves to realize the dynamic adjustment of parameters such as the rotational speed of the pump motor, the power of the cold trap, and the opening degree of the gas extraction valve, so that the actual operating state of the pump station converges quickly and smoothly to the optimization objective and continuously remains within the optimal region. At the same time, the real-time state data of the optimization control process is also fed back to the optimization model for evaluating the optimization effect and improving the control strategy, forming a closed-loop control mechanism of equipment state perception - optimization decision - execution feedback.

[0055] Through the organic combination of the above operating parameter optimization model and the intelligent control system, the continuous optimization and adaptive adjustment of the performance of the vacuum pump station can be achieved. Compared with the traditional method of determining parameters based on experience, this intelligent optimization control strategy has the advantages of fast response, significant effect, and global optimality. Limited by personnel experience and cognition, the traditional method often has difficulty accurately grasping the dynamic characteristics of equipment and the fluctuations of external working conditions, and it is difficult to ensure the timeliness and effectiveness of regulation. The intelligent optimization model can refine the rules from a large amount of historical data, learn the internal mechanism and control strategy of the vacuum system, accurately perceive the subtle changes in the equipment state, quickly infer the optimal control decision, and continuously correct the optimization direction through closed-loop feedback, making the control process more adaptive and robust.

[0056] On the basis of the above embodiments, as an alternative embodiment, before inputting the multi-dimensional operation feature vector into the preset operation parameter optimization model, outputting the target optimization parameters, and regulating the target vacuum pump station based on the target optimization parameters, it further includes: S501. Obtain the performance parameters of the target vacuum pump station and determine the constraint conditions based on the performance parameters; Specifically, before inputting the extracted multi-dimensional operation feature vector into the preset operation parameter optimization model to obtain the optimal target parameter settings and implementing parameter regulation on the vacuum pump station accordingly, in order to ensure the feasibility and safety of the optimization plan, it is also necessary to fully consider the actual performance constraints of the pump station equipment, obtain the key performance parameters of the pump station, and determine the hard constraint conditions that must be followed during the optimization control process based on this. The vacuum pump station is a complex system composed of multiple vacuum pumps, drive motors, pipeline valves, etc. Its operation regulation must be strictly limited within the range permitted by the equipment's own performance, such as the rated power of the pump, the safe speed of the motor, and the pressure-bearing capacity of the pipeline. Once the actual operation parameters exceed the performance limit of the equipment, it will not only damage the equipment itself but also endanger the safe and stable operation of the entire vacuum system. Therefore, it is necessary to have a full understanding and mastery of the performance parameters of the pump station equipment, clarify the constraint boundaries that must be strictly observed when formulating the optimization control plan, maximize the equipment performance to achieve energy conservation and efficiency improvement while always being vigilant about the safety red line and avoiding blind advancement. Transforming the equipment performance parameters into computable and quantifiable mathematical constraint conditions and embedding them in the objective function and solution algorithm of the optimization model can make the optimization control process more standardized and reliable, and truly achieve the dynamic balance and system optimization of multiple objectives such as equipment performance, energy consumption level, and safe operation.

[0057] During specific implementation, first, it is necessary to comprehensively obtain the factory nameplate parameters of each piece of equipment in the vacuum pump station, including the pumping speed, ultimate vacuum degree, and rated power of the vacuum pump, the rated power, speed, and efficiency of the motor, and the nominal diameter and pressure grade of the pipeline valve. These parameters specify the maximum performance limit of the equipment under the design conditions. During the actual operation process, the performance parameters of the equipment will also be affected by many factors such as production batches, manufacturing processes, service life, corrosion, and wear. It is necessary to refer to the regular inspection records and status evaluation reports of the pump station to obtain the dynamic change data of the equipment performance parameters at different life cycle stages, conduct trend extrapolation and safety margin correction on them, and finally determine the actual performance upper limit of the equipment that can be utilized during the optimization control process.

[0058] Based on the obtained performance parameters, they need to be further transformed into quantitative constraint conditions that can be used to optimize model calculations. According to the structural topology of the pump station system and the physical properties of equipment parameters, the constraint conditions can be divided into different types such as pipeline constraints, pump constraints, and motor constraints. Taking pipeline constraints as an example, it mainly includes that the pipeline pressure shall not exceed the nominal pressure, the pipeline flow rate shall not exceed the critical flow rate, etc.; pump constraints mainly include that the gas extraction rate of the pump shall not exceed the rated pumping speed, the outlet pressure shall not be lower than the ultimate vacuum degree, etc.; motor constraints mainly include that the motor power shall not exceed the rated power, the rotational speed shall not exceed the maximum safe rotational speed, etc. When quantitatively describing various constraints, it is necessary to strictly follow the laws of physics and relevant engineering specifications, extract the characteristic parameters corresponding to the constraint terms, establish a functional mapping relationship between the constraint conditions and the optimization variables, form a structured and computable inequality or equation form, and specify the strength order and logical relationship of each constraint. At the same time, considering the satisfiability and convergence of the constraint conditions, reasonably set the threshold range of the constraints to avoid being too strict or too loose.

[0059] After forming the performance constraint conditions, it is also necessary to embed them into the objective function and solution algorithm of the optimization model. First, different weights can be assigned to the energy consumption term, production efficiency term, etc. in the optimization objective function, then the constraint conditions are equalized and multiplied by the Lagrange multiplier, and finally a composite objective function containing hard constraints is formed. When solving the model, methods such as the penalty function method and the gradient projection method can be used to couple the constraint conditions with the extreme value problem of the objective function to ensure that the solution obtained by each round of iterative optimization strictly satisfies the performance constraints, and finally search for the control parameter combination with the optimal energy consumption and the largest output under the constraint conditions. This optimization control method with embedded constraints fully considers the actual performance level and physical limits of the equipment, makes the optimization scheme closer to the engineering reality, has stronger operability and robustness, and can greatly improve the success rate and benefit level of the optimization control.

[0060] S502, obtain the historical operation sample data of the target vacuum pump station, construct a preliminary operation parameter optimization model based on the constraint conditions, and train the operation parameter optimization model based on the historical operation sample data to obtain the preset operation parameter optimization model.

[0061] Specifically, it is first necessary to comprehensively obtain the historical operation archive data of the vacuum pump station and conduct systematic screening and preprocessing on it. Through data collection from the pump station production management system, equipment monitoring system, energy management system, etc., multi-source heterogeneous data such as the operation condition parameters, equipment condition parameters, energy consumption parameters, and output parameters of the pump station system over a period of time are obtained. Then, referring to the data collection time, sampling frequency, data quality, etc., high-quality sample data with good continuity and high integrity is selected, and incomplete, discontinuous, and obviously abnormal dirty data is excluded. And the physical dimensions, numerical ranges, etc. of the sample data are uniformly normalized to improve the regularity and comparability of the data. Next, the time series characteristics of the sample data are extracted. Through data smoothing, anomaly detection, etc., random perturbations and abnormal interferences are removed to highlight the change trends and correlations of the data. Finally, data dimensionality reduction methods such as principal component analysis and correlation analysis are used to extract the key characteristic parameters that have a significant impact on the system energy efficiency, and a time series dependence relationship matrix between the parameters is constructed to guide the subsequent model structure design.

[0062] After obtaining a high-quality historical operation sample library, it is necessary to build a preliminary operation parameter optimization model framework based on the performance constraint conditions. According to the physical structure and control logic of the pump station system, a hybrid modeling paradigm that combines mechanism modeling and data-driven is adopted, and various machine learning models such as neural networks, support vector machines, and deep reinforcement learning are selected according to local conditions to model the core components of the pump station system and describe the independent characteristics and dynamic coupling relationships of each component. At the same time, the aforementioned performance constraint conditions are embedded in the objective function, constraint equation, and state equation of the optimization model to form a hybrid optimization model prototype that can not only fully reflect the internal laws of the pump station but also strictly follow the equipment performance boundaries. To further improve the modeling efficiency, a modular and parameterized modeling paradigm can be adopted to achieve flexible combination and parameter tuning of the component models and quickly adapt to different pump station operating conditions.

[0063] After the initial optimization model is constructed, it is necessary to use historical operation sample data to repeatedly train, learn, and dynamically correct the model. The sample data set is divided into a training set and a validation set according to a certain ratio. Through methods such as cross-validation, the model fully learns historical experience on the training set and evaluates the learning effect on the validation set. Multiple evaluation indicators such as mean absolute error, root mean square error, and coefficient of determination are used to objectively measure the fitting degree between the model output and the actual energy consumption. For problems such as unqualified model accuracy and poor generalization, methods such as adding regularization terms, dynamic learning rates, and early stopping mechanisms are used to continuously optimize the model structure and parameters, improve the model performance until the predetermined evaluation criteria are met. During this period, attention should also be paid to the degree to which the model meets the performance constraint conditions. The penalty function method and other methods are used to consider the constraint terms and the objective function together to ensure that the model output always meets the equipment safety boundary. As the model continuously learns and strengthens on a large number of historical samples, its internal control laws are gradually discovered and improved, and its optimization and solution capabilities will also continue to increase. Eventually, a customized exclusive optimization model that best fits the actual working conditions of the pumping station will be formed.

[0064] First, it is necessary to comprehensively obtain the historical operation archive data of the vacuum pumping station and conduct systematic screening and preprocessing on it. Through data collection from the pumping station production management system, equipment monitoring system, energy management system, etc., multi-source heterogeneous data such as the operation condition parameters, equipment condition parameters, energy consumption parameters, and output parameters of the pumping station system over a period of time are obtained. Then, referring to the data collection time, sampling frequency, data quality, etc., high-quality sample data with good continuity and high integrity is selected, and incomplete, discontinuous, and obviously abnormal dirty data is removed. And the physical dimensions, numerical ranges, etc. of the sample data are uniformly normalized to improve the regularity and comparability of the data. Then, the time series characteristics of the sample data are extracted. Through data smoothing, anomaly detection, etc., random disturbances and abnormal interferences are removed to highlight the change trend and correlation of the data. Finally, data dimensionality reduction methods such as principal component analysis and correlation analysis are used to extract the key characteristic parameters that have a significant impact on the system energy efficiency, and a time series dependence relationship matrix between the parameters is constructed to guide the subsequent model structure design.

[0065] After obtaining a high-quality historical operation sample library, it is necessary to build a preliminary operating parameter optimization model framework based on performance constraints. According to the physical structure and control logic of the pump station system, a hybrid modeling paradigm combining mechanism modeling and data-driven is adopted. A variety of machine learning models such as neural networks, support vector machines, and deep reinforcement learning are selected according to local conditions to model the core components of the pump station system and describe the independent characteristics and dynamic coupling relationships of each component. At the same time, the aforementioned performance constraints are embedded in the objective function, constraint equation, and state equation of the optimization model to form a prototype of a hybrid optimization model that can fully reflect the internal laws of the pump station and strictly follow the performance boundaries of the equipment. In order to further improve the modeling efficiency, a modular and parameterized modeling paradigm can be adopted to achieve flexible combination and parameter setting of component models, and quickly adapt to different pump station operating conditions.

[0066] After the initial optimization model is built, the historical operation sample data needs to be used to repeatedly train, learn and dynamically correct the model. The sample data set is divided into a training set and a validation set according to a certain ratio. Through cross-validation and other methods, the model can fully learn historical experience on the training set and evaluate the learning effect on the validation set. Multiple evaluation indicators such as mean absolute error, root mean square error, and determination coefficient are used to objectively measure the degree of fit between the model output and the actual energy consumption. In view of the problems of substandard model accuracy and poor generalization, the model structure and parameters are continuously optimized by adding regularization terms, dynamic learning rate, early stopping mechanism and other methods to improve the model performance until the predetermined evaluation criteria are met. During this period, attention should also be paid to the degree to which the model satisfies the performance constraints. The penalty function method and other methods are used to unify the constraints and the objective function to ensure that the model output always meets the equipment safety boundary. With the continuous learning and strengthening of the model on massive historical samples, its internal control laws are gradually discovered and improved, and the optimization and solution capabilities will continue to increase, and a customized and exclusive optimization model that best fits the actual working conditions of the pump station will eventually be formed.

[0067] See also Figure 2 , Figure 2 A vacuum pump station management system architecture diagram provided in an embodiment of the present application, the vacuum pump station management system may include: A data acquisition module 1 is used to acquire key operating parameters of a target vacuum pump station based on a sensor matrix arranged in the target vacuum pump station, and to perform fusion processing based on the key operating parameters to obtain a multi-dimensional operating feature vector; Fault judgment module 2, used for inputting the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result; The fault maintenance module 3 is used to determine a target fault handling strategy based on the fault judgment result and a preset fault handling strategy, and to perform maintenance on the vacuum pump station based on the target fault handling strategy; An operation parameter optimization module 4 is configured to input a multi-dimensional operation feature vector into a preset operation parameter optimization model, output target optimization parameters, and control a target vacuum pump station based on the target optimization parameters.

[0068] It should be noted that when the system provided in the above embodiments realizes its functions, only the division of the above function modules is used for illustration. In actual applications, the above functions can be allocated to different function modules according to needs, that is, the internal structure of the device is divided into different function modules to complete all or part of the functions described above. In addition, the systems and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiments and will not be repeated here.

[0069] Please refer to Figure 3 This application also discloses an electronic device. Figure 3 It is a schematic structural diagram of an electronic device disclosed in an embodiment of this application. The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, and at least one communication bus 302 or end-to-end wireless communication.

[0070] Among them, the communication bus 302 is used to realize the connection and communication between these components.

[0071] Among them, the user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may further include a standard wired interface and a wireless interface.

[0072] Among them, the network interface 304 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).

[0073] Among them, the processor 301 may include one or more processing cores. The processor 301 connects various parts within the entire server through various interfaces and lines, and executes various functions of the server and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and by calling the data stored in the memory 305. Optionally, the processor 301 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 301 may integrate a combination of one or several of a central processing unit (CPU), a graphics processing unit (GPU), and a modem, etc. Among them, the CPU mainly processes the operating system, user interface, application programs, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 301 and may be implemented separately by a single chip.

[0074] Among them, the memory 305 may include random access memory (RAM) and may also include read-only memory. Optionally, the memory 305 includes a non-transitory computer-readable medium. The memory 305 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area. Among them, the program storage area may store instructions for implementing the operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store the data involved in the above-mentioned various method embodiments. The memory 305 is optionally further a storage system located at least one away from the aforementioned processor 301. Refer to Figure 3 , the memory 305, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for a vacuum pump station management method.

[0075] In Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an interface for the user to input and obtain the data input by the user; while the processor 301 can be used to call the application program stored in the memory 305 for the vacuum pump station management method. When executed by one or more processors 301, the electronic device 300 is caused to execute the method of one or more of the above embodiments. It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be adopted in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0076] In several implementation manners provided by this application, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some service interfaces. The indirect couplings or communication connections of the system or modules can be in electrical or other forms.

[0077] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, that is, they can be located in one place, or they can be distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0078] This application embodiment also provides a computer storage medium. The computer storage medium can store multiple instructions, and the instructions are suitable for being loaded and executed by the processor to perform the vacuum pump station management method of the above Figure 1 shown embodiment. The specific execution process can refer to the specific description of the Figure 1 shown embodiment and will not be elaborated here.

[0079] In addition, in each embodiment of this application, the functional modules can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0080] When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present application. And the aforementioned memory includes: various media such as USB flash drives, mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0081] The above are only exemplary embodiments of the present disclosure, and the scope of the present disclosure cannot be limited thereby. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure still fall within the scope covered by the present disclosure. Those skilled in the art will readily think of other implementation manners of the present disclosure after considering the specification and the practice of the present disclosure.

[0082] The present application aims to cover any variations, uses, or adaptive changes of the present disclosure, which follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the technical field not recorded in the present disclosure. The specification and the embodiments are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A vacuum pump station management method, characterized in that: The method comprises: Acquire key operating parameters of the target vacuum pump station based on a sensor matrix arranged in the target vacuum pump station, and perform fusion processing based on the key operating parameters to obtain a multi-dimensional operating feature vector; Inputting the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result; Based on the fault judgment result and the preset fault handling strategy, determine a target fault handling strategy, and maintain the vacuum pump station based on the target fault handling strategy; The multi-dimensional operation characteristic vector is input into a preset operation parameter optimization model, target optimization parameters are output, and the target vacuum pump station is regulated based on the target optimization parameters.

2. The method according to claim 1, characterized in that Before acquiring the key operating parameters of the target vacuum pump station based on the sensor matrix arranged in the target vacuum pump station, the method further includes: Acquire the process flow and equipment characteristics of the target vacuum pump, and identify key equipment based on the process flow and equipment characteristics to obtain key equipment information; The sensor type and the measuring point position are determined based on the key equipment information, and the sensor matrix is ​​arranged in the target vacuum pump station based on the sensor type and the measuring point position.

3. The method according to claim 1, characterized in that The fusion processing is performed based on the key operation parameters to obtain a multi-dimensional operation feature vector, including: Preprocessing the key operating parameters to obtain preprocessed key operating parameters, and performing time domain analysis on the preprocessed key operating parameters to obtain time domain characteristic parameters; Performing time-frequency analysis on the preprocessed key operating parameters to obtain time-frequency characteristic parameters; Performing frequency domain analysis on the preprocessed key operating parameters to obtain frequency domain characteristic parameters; The time domain feature parameters, the time-frequency feature parameters and the frequency domain feature parameters are subjected to feature fusion to obtain the multi-dimensional operation feature vector.

4. The method according to claim 1, characterized in that: Before inputting the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result, the method further includes: Acquire vacuum pump station fault instance sample data, and perform expansion extraction on the vacuum pump station fault instance sample data to obtain a multi-dimensional fault sample feature vector; A preliminary fault judgment model is constructed, and the preliminary fault judgment model is trained based on the multi-dimensional fault sample feature vector. When the fault judgment result reaches a preset accuracy threshold, the preset fault judgment model is obtained.

5. The method according to claim 1, characterized in that The determining of a target fault handling strategy based on the fault judgment result and a preset fault handling strategy includes: The fault type and fault severity are determined according to the fault judgment result, and the fault type and fault severity are matched with the preset fault handling strategy, and the fault handling strategy with the highest matching degree is selected as the target fault handling strategy.

6. The method according to claim 1, characterized in that Before determining a target fault handling strategy based on the fault judgment result and the preset fault handling strategy, the method further includes: Analyzing the confidence of the fault judgment result, and when the confidence is less than a preset confidence threshold, re-inputting the multi-dimensional operation feature vector into the preset fault judgment model to obtain a new fault judgment result; The determining of a target fault handling strategy based on the fault judgment result and a preset fault handling strategy includes: Based on the new fault judgment result and the preset fault handling strategy, a target fault handling strategy is determined.

7. The method according to claim 1, characterized in that Before inputting the multi-dimensional operation characteristic vector into a preset operation parameter optimization model, outputting target optimization parameters, and regulating the target vacuum pump station based on the target optimization parameters, the method further includes: Acquiring performance parameters of the target vacuum pump station, and determining constraint conditions based on the performance parameters; The historical operation sample data of the target vacuum pump station is obtained, and a preliminary operation parameter optimization model is constructed based on the constraint conditions, and the operation parameter optimization model is trained based on the historical operation sample data to obtain the preset operation parameter optimization model.

8. A vacuum pump station management system, characterized in that: The system comprises: A data acquisition module, used for acquiring key operating parameters of a target vacuum pump station based on a sensor matrix arranged in the target vacuum pump station, and performing fusion processing based on the key operating parameters to obtain a multi-dimensional operating feature vector; A fault judgment module, used for inputting the multi-dimensional operation feature vector into a preset fault judgment model to obtain a fault judgment result; A fault maintenance module, used to determine a target fault handling strategy based on the fault judgment result and a preset fault handling strategy, and to maintain the vacuum pump station based on the target fault handling strategy; The operation parameter optimization module is used to input the multi-dimensional operation characteristic vector into a preset operation parameter optimization model, output target optimization parameters, and regulate the target vacuum pump station based on the target optimization parameters.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It includes a processor, a memory and a transceiver, the memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method according to any one of claims 1 to 7.

Images