A sensing and control system fault prediction method and device
By dividing the infection control system equipment into working stages and acquiring parameter information in real time, a fault prediction model is constructed using a health assessment model, kernel principal component analysis algorithm, and random forest algorithm. This solves the problem of inaccurate fault prediction in the infection control system and improves the reliability of the system and the stability of the disinfection work chain.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- LAOKEN MEDICAL TECH
- Filing Date
- 2022-12-07
- Publication Date
- 2026-05-12
AI Technical Summary
Existing fault prediction methods for sensor control systems are inaccurate, resulting in low reliability of these systems.
The devices in the sensing and control system are divided into multiple device sets according to their working stages. Device parameter information is acquired in real time, and a health assessment model is used to evaluate and determine fault characteristics. Finally, a fault prediction model is constructed using the kernel principal component analysis algorithm and the random forest algorithm to predict faults.
It improves the accuracy of fault prediction, enhances the reliability of the infection control system, and ensures the stability of the disinfection work chain.
Smart Images
Figure CN115840919B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of sensing and control technology, and more specifically, to a method and apparatus for predicting faults in sensing and control systems. Background Technology
[0002] In the field of infection control, equipment malfunctions or potential malfunctions can lead to work failures, delays, and other adverse effects on cleaning, drying, disinfection, and sterilization processes. This, in turn, can affect the stable and timely supply of sterilized medical devices and even cause medical accidents such as surgical interruptions. In fact, this instability and unreliability is the most prominent problem in the field of infection control.
[0003] To address the issues of instability and unreliability, in addition to conventional methods of improving the reliability of individual devices, real-time monitoring of the status of all aspects of the entire sensing and control field and prediction and diagnosis of faults has become a promising and effective approach. This allows for timely and proactive measures such as inspection, maintenance, and repair.
[0004] Currently, fault prediction and diagnosis technology has been successfully applied in the industrial field. With the rapid development of the Industrial Internet of Things and intelligent sensing technology, more and more industrial production equipment is beginning to adopt factory equipment fault prediction and diagnosis systems in order to timely and accurately monitor the production process or equipment status in real time and predict faults. This fault prediction method is also applicable in the field of sensing and control. However, existing fault prediction methods have inaccuracy problems, resulting in low reliability of sensing and control systems. Summary of the Invention
[0005] The purpose of this invention is to provide a method and apparatus for predicting faults in a sensing and control system, so as to improve the problem that the fault prediction in the prior art is inaccurate, resulting in low reliability of the sensing and control system.
[0006] In a first aspect, embodiments of this application provide a method for predicting faults in a sensing and control system, comprising the following steps:
[0007] The devices in the sensing and control system are divided according to their working stages to obtain multiple device sets, each device set including one or more devices;
[0008] Real-time acquisition of parameter information of each of the devices in the device cluster;
[0009] Based on the parameter information of the devices in each of the aforementioned device sets, an assessment is conducted using the corresponding health assessment model for each device set to obtain the assessment results for each working stage.
[0010] Based on the evaluation results of each working stage, fault characteristics are determined from the parameter information of each of the equipment in the equipment cluster;
[0011] Based on the fault characteristics, a fault prediction model constructed using the kernel principal component analysis algorithm and the random forest algorithm is used to predict the faults of the sensing and control system, and the system fault prediction results are obtained.
[0012] Based on the first aspect, in some embodiments of the present invention, the construction of the health assessment model includes the following five steps:
[0013] Historical data samples of each device set are obtained respectively. The historical data samples of each device set include historical device parameter information and the health status of the corresponding working stage.
[0014] Based on historical data samples from each device set, an evaluation model based on the XGBoost algorithm is trained to obtain a health evaluation model for each device set.
[0015] 0. Based on the first aspect, in some embodiments of the present invention, based on the evaluation results of the various working stages, from...
[0016] The fault characteristics are determined from the parameter information of each of the aforementioned devices, including the following steps:
[0017] The evaluation results of each work stage are compared with the corresponding stage thresholds to obtain multiple comparison results;
[0018] Based on the multiple comparison results, the abnormal working phase is determined;
[0019] 5. Based on the abnormal working stage, extract the parameter information of the corresponding centralized equipment;
[0020] Based on the parameter information of the corresponding equipment in the equipment set and the corresponding comparison results, the fault characteristics are determined.
[0021] Based on the first aspect, in some embodiments of the present invention, the fault characteristics are determined according to the parameter information of the corresponding equipment in the centralized equipment and the corresponding comparison results, including the following steps:
[0022] The comparison results corresponding to the abnormal working phase are used as weights.
[0023] 0. The parameter information of the corresponding devices in the device set is used as the feature parameter;
[0024] The fault characteristics are obtained by multiplying the feature parameters by the weights.
[0025] Based on the first aspect, in some embodiments of the present invention, the following steps are also included:
[0026] Obtain the fault training set and fault test set;
[0027] Based on the fault training set, the kernel principal component analysis algorithm and the random forest algorithm are used to construct the model, resulting in 5 initial prediction models.
[0028] The initial prediction model is optimized using the fault test set to obtain a fault prediction model.
[0029] Based on the first aspect, in some embodiments of the present invention, an initial prediction model is obtained by using the kernel principal component analysis algorithm and the random forest algorithm to construct a model based on the fault training set, including the following steps:
[0030] The feature samples in the fault training set are normalized to obtain preprocessed feature samples; random noise is added to the preprocessed feature samples to obtain new feature samples.
[0031] The kernel principal component analysis algorithm is used to extract kernel principal components from the new feature samples to obtain a new feature sample matrix.
[0032] Based on the new feature sample matrix and the fault training set, the pre-set random forest model is trained to obtain the initial prediction model.
[0033] Based on the first aspect, in some embodiments of the present invention, the following steps are also included:
[0034] Based on the system fault prediction results, the faulty equipment is identified, and fault alert information is generated based on the faulty equipment.
[0035] Secondly, embodiments of this application provide a fault prediction device for a sensing and control system, comprising:
[0036] The device partitioning module is used to partition the devices in the sensing and control system according to the working stage to obtain multiple device sets, each device set including one or more devices;
[0037] The parameter information acquisition module is used to acquire parameter information of each device in the device set in real time.
[0038] The evaluation module is used to evaluate each device set based on its parameter information and the corresponding health evaluation model, and to obtain the evaluation results for each working stage.
[0039] The fault feature determination module is used to determine fault features from the parameter information of each of the equipment in the equipment set based on the evaluation results of each working stage.
[0040] The fault prediction module is used to predict the faults of the sensing and control system based on the fault characteristics and a fault prediction model constructed using the kernel principal component analysis algorithm and the random forest algorithm, so as to obtain the system fault prediction result.
[0041] Thirdly, embodiments of this application provide an electronic device including a memory for storing one or more programs; and a processor. When the one or more programs are executed by the processor, the method as described in any one of the first aspects above is implemented.
[0042] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of the first aspects above.
[0043] The embodiments of the present invention have at least the following advantages or beneficial effects:
[0044] This invention provides a method and apparatus for predicting faults in a sensor control system. The method involves dividing the devices in the sensor control system according to their operating stages, resulting in multiple device sets, each set including one or more devices. Parameter information of the devices in each set is then acquired in real time. Based on the parameter information of each device set, a health assessment model corresponding to that set is used for evaluation, yielding evaluation results for each operating stage. Fault characteristics are determined from the parameter information of the devices in each set based on the evaluation results of each operating stage. By applying the health assessment model to different operating stages, the health status of each stage can be determined, and the health status can be used to identify fault characteristics that may affect the sensor control system, indicating that these are valid fault characteristics and ensuring their effectiveness. Finally, based on the fault characteristics, a fault prediction model constructed using kernel principal component analysis (KPI) and random forest algorithms is used to predict the faults in the sensor control system, yielding system fault prediction results. Since KPI can effectively select feature samples, and random forest has good noise tolerance and high evaluation accuracy, the prediction accuracy of the obtained fault prediction model is greatly improved, resulting in more accurate system fault prediction results. Furthermore, the fault characteristics obtained are all valid fault characteristics, which further improves the accuracy of system fault prediction and thus improves the reliability of the entire infection control system's disinfection work chain. Attached Figure Description
[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0046] Figure 1 A flowchart of a fault prediction method for a sensing and control system provided in an embodiment of the present invention;
[0047] Figure 2 A structural block diagram of a fault prediction device for a sensing and control system provided in an embodiment of the present invention;
[0048] Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of the present invention.
[0049] Icons: 110 - Device partitioning module; 120 - Parameter information acquisition module; 130 - Evaluation module; 140 - Fault characteristic determination module; 150 - Fault prediction module; 101 - Memory; 102 - Processor; 103 - Communication interface. Detailed Implementation
[0050] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.
[0051] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0052] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0053] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0054] In the description of this application, it should be noted that the terms "upper", "lower", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship that the product of this application is usually placed in. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.
[0055] Example
[0056] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the various embodiments and features described below can be combined with each other.
[0057] Please refer to Figure 1 , Figure 1 This invention provides a flowchart of a method for predicting faults in an infection control system. The method can be applied to infection control systems in hospitals. The method includes the following steps:
[0058] Step S110: Divide the devices in the sensing and control system according to their working stages to obtain multiple device sets, each device set including one or more devices. In this embodiment, the sensing and control system may include three working stages: cleaning, drying, and disinfection, depending on the different working steps. The devices involved in each working stage are different and are selected and configured according to the actual application scenario. For example, the disinfection stage includes ultraviolet lamp sterilizers and air sterilizers, while the drying stage includes hot air blowers and exhaust fans. It should be noted that the sensing and control system can also divide its working stages in other ways; this embodiment does not limit this.
[0059] Step S120: Real-time acquisition of parameter information from each of the centralized devices; In this embodiment, the aforementioned parameter information refers to the real-time parameter information of the devices during operation, including internal technical parameter information and real-time collected information, such as data collected by common sensors in the devices, the speed of the motors in the devices, the speed of the fans, etc. Specifically, different parameter information can be set according to the different devices actually used. By acquiring the device parameter information in real time, the operating status of the devices can be understood in a timely manner, which facilitates subsequent device fault diagnosis and analysis.
[0060] Step S130: Based on the parameter information of the devices in each of the device sets, the health assessment model corresponding to each device set is used to perform the assessment to obtain the assessment results of each working stage. In this embodiment, a corresponding health assessment model can be pre-constructed for different working stages to assess the health status of the devices in the current working stage.
[0061] The health assessment models can be constructed using the same or different methods. To save computational resources and facilitate rapid construction of health assessment models, the same construction method can be used, which may include the following steps:
[0062] First, historical data samples are obtained for each device set. The historical data samples for each device set include historical device parameter information and the health status of the corresponding working stage. In this embodiment, the historical data samples include multiple sets of historical data, wherein the historical data includes device parameter information and the health status of the corresponding working stage.
[0063] Then, based on the historical data samples of each device set, an evaluation model based on the XGBoost algorithm is trained to obtain the health evaluation model corresponding to each device set.
[0064] In this embodiment, XGBoost typically transforms multi-class classification tasks into multiple binary classification tasks, such as "Is this work phase normal?". Therefore, the model consists of three sub-models, each outputting a score to determine whether an input sample belongs to that category. The health assessment model outputs the status of the work phase, with three options: normal, low efficiency, and high energy consumption. The health assessment model is essentially a classification model based on the XGBoost algorithm. Specifically, the assessment model optimizes by minimizing the loss function and uses backpropagation for gradient boosting.
[0065] KMeans clustering can be used to establish state intervals. The KMeans clustering algorithm automatically groups similar samples into a single category. Multiple historical health assessment data points can be acquired and used as samples. Two clusters are defined: normal and abnormal. Initial centroids are determined for each cluster. KMeans clustering is then applied to each historical health assessment data point to obtain two categories. The state category of the historical health assessment data within each category is extracted, resulting in a state interval for each category. Establishing state intervals through KMeans clustering makes the output scores more discriminative. A health assessment model based on the XGBoost algorithm is built to score work stages between normal and abnormal states, presenting the work stage health assessment results through state intervals. The assessment model uses equipment parameter information as input data. Equipment parameter information includes equipment technical parameters and real-time data collected by the equipment. The health assessment model extracts the intermediate layer probabilities of the XGBoost model to score the health status of the work stage, presenting the work stage health assessment results through state intervals.
[0066] In this embodiment, the health assessment model outputs a score, and can also automatically generate explicit state intervals based on the output score.
[0067] In practice, a health assessment model can be built based on the XGBoost algorithm. First, historical data samples, including normal and abnormal samples, are acquired. Then, using the XGBoost algorithm, parameters are set, and an assessment model is built to determine whether the work phase is abnormal. Specifically, the model's accuracy and generalization are checked to see if they meet the requirements. If they do, a scoring layer is extracted, and a health assessment is performed on all historical data samples. If not, the XGBoost algorithm is used again, parameters are set, and an assessment model is built to determine whether the work phase is abnormal. After the health assessment, the scoring is checked for discriminative power. If it does, health status intervals are formed through clustering, completing the health assessment model construction. If not, the XGBoost algorithm is used again, parameters are set, and an assessment model is built to determine whether the work phase is abnormal.
[0068] In this embodiment, each working stage can be pre-built with a corresponding health assessment model according to the above process. Then, based on the parameter information of each device, the health status of each working stage can be assessed, thereby obtaining the health assessment results of each working stage. This allows for the prediction of faults in the entire infection control system based on the assessment results in the later stages.
[0069] Step S140: Based on the evaluation results of each working stage, determine the fault characteristics from the parameter information of each of the equipment in the equipment cluster; after obtaining the evaluation results, it can be determined which working stages are abnormal, thereby determining the characteristic parameters that may affect the sensing and control system. Specifically, this can be achieved through the following steps:
[0070] First, the evaluation results of each work stage are compared with the corresponding stage thresholds to obtain multiple comparison results. In this embodiment, the evaluation results are scores. Stage thresholds for different work stages can be preset. The comparison process can be to subtract the evaluation results from the stage thresholds to obtain the difference, which is used as the comparison result. It should be noted that the stage thresholds can be set separately according to different work stages or according to actual needs.
[0071] Then, based on the multiple comparison results, the abnormal working stage is determined. In this embodiment, if the comparison result is small, it indicates that it is close to the stage threshold; conversely, if the comparison result is large, it indicates that it is far from the stage threshold. Larger values in the comparison results can be filtered out to find the corresponding working stage, which is then designated as the abnormal working stage. For example, if the comparison result for the cleaning stage is 0.1, the comparison result for the drying stage is 0.1, and the comparison result for the disinfection stage is 0.8, then the abnormal working stage is the disinfection stage. It should be noted that the aforementioned abnormal working stage can be one or more working stages.
[0072] Then, based on the abnormal working stage, the parameter information of the corresponding equipment in the equipment center is extracted;
[0073] Finally, based on the parameter information of the corresponding devices in the device set and the corresponding comparison results, the fault characteristics are determined. In this embodiment, since a larger comparison result indicates a larger difference from the stage threshold, the probability of a fault is higher. Therefore, when determining fault characteristics, the comparison result can be considered as a weight to reflect the probability of a fault occurring. Specifically, this includes the following steps:
[0074] The first step is to use the comparison results corresponding to the abnormal working phase as weights;
[0075] The second step is to use the parameter information of the corresponding equipment in the device set as feature parameters;
[0076] The third step is to multiply the feature parameters by the weights to obtain the fault features.
[0077] By considering weights, fault characteristics can better reflect the health status of the working stage, thereby helping to improve the accuracy of fault prediction in the later stages.
[0078] Step S150: Based on the fault characteristics, a fault prediction model constructed using the kernel principal component analysis algorithm and the random forest algorithm is used to predict the fault of the sensing and control system, and the system fault prediction result is obtained.
[0079] In this embodiment, the above-mentioned fault prediction model can be constructed by combining the kernel principal component analysis algorithm and the random forest algorithm, specifically including the following steps:
[0080] First, obtain the fault training set and fault test set. These sets can be historical data. The training set is used to train the model, and the test set is used to validate and optimize the model. The sample data in both sets represents fault data for the entire sensor control system, including parameter information for each device and the system fault categories.
[0081] Then, based on the fault training set, the kernel principal component analysis (KPCA) algorithm and the random forest algorithm are used to construct the model and obtain the initial prediction model. KPCA has a strong feature selection capability in high-dimensional space and random forest has an excellent fault identification capability. The original feature samples can be mapped to the high-dimensional feature space to extract principal components to construct new feature samples. Then, the random forest model is used to diagnose the faults in the sensing and control system.
[0082] The initial prediction model described above can be constructed through the following steps:
[0083] The first step is to normalize the feature samples in the fault training set to obtain preprocessed feature samples. The normalization process can be normalized to the range [0, 1] to eliminate the influence of dimensions and help speed up model training.
[0084] The second step is to add random noise to the preprocessed feature samples to obtain new feature samples. Since some noise may have been introduced into the preprocessed feature samples, random noise can be added to the preprocessed feature samples in order to test the model's anti-interference ability.
[0085] Specifically: Let D1 be the preprocessed feature sample matrix before adding noise, and let D2 be the new feature sample matrix after adding noise.
[0086] D2(i,j)=D1(i,j)×[1+a×rands(1)]
[0087] Where a is the noise control coefficient, which can take values of 0.2, 0.5, and 0.8; rands(1) is a random function used to generate random numbers from -1 to 1, i is a row in the feature sample matrix, and j is a column in the feature sample matrix.
[0088] The third step involves using kernel principal component analysis (KPCA) to extract kernel principal components from the new feature samples, resulting in a new feature sample matrix. KPCA is a nonlinear extension of linear principal component analysis (PCA). It extracts principal components using a nonlinear method; specifically, KPCA maps the new feature samples to a high-dimensional space using a mapping function, performing PCA analysis in that high-dimensional space. KPCA is not only suitable for solving nonlinear feature extraction problems, but it also provides a greater number of features and higher feature quality than PCA. The kernel principal component extraction using KPCA described above is existing technology and will not be elaborated further here.
[0089] The fourth step involves training the pre-set random forest model using the new feature sample matrix and the fault training set to obtain the initial prediction model. Appropriate random forest model parameters are selected, including the number of pre-selected variables for tree nodes and the number of decision trees in the random forest. Then, the random forest model is trained using the fault training set with the new feature sample matrix to obtain the initial prediction model. Random forests utilize a bootstrap resampling technique to repeatedly and randomly sample k samples (k is generally the same as N) with replacement from the original training sample set N to generate a new training sample set. Then, n classification trees are generated based on the bootstrap sample set to form a random forest. Essentially, it is an improvement on the decision tree algorithm, merging multiple decision trees together, with each tree's construction depending on an independently sampled set. The above-described random forest model is existing technology and will not be elaborated further here.
[0090] Finally, the initial prediction model is optimized using the fault test set to obtain a fault prediction model. This optimization involves verifying the initial prediction model using the fault test set and adjusting its parameters to obtain a more optimized fault prediction model.
[0091] By using fault features as input to the fault prediction model, system fault prediction results can be obtained. Since kernel principal component analysis can effectively select feature samples, and random forests have good tolerance to noise and high evaluation accuracy, the prediction accuracy of the obtained fault prediction model is greatly improved, thus making the obtained system fault prediction results more accurate.
[0092] In the above implementation process, the devices in the sensing and control system are divided according to their working stages to obtain multiple device sets, each set including one or more devices. Then, the parameter information of each device in each set is acquired in real time. Based on the parameter information of each device set, a health assessment model corresponding to that set is used for evaluation to obtain the evaluation results for each working stage. Based on the evaluation results of each working stage, fault characteristics are determined from the parameter information of each device in each set. By applying the health assessment model to different working stages, the health status of each working stage can be determined. Furthermore, the health status is used to identify fault characteristics that may affect the sensing and control system, indicating that these are valid fault characteristics, thus ensuring the validity of the fault characteristics. Finally, based on the fault characteristics, a fault prediction model constructed using kernel principal component analysis (KPI) and random forest algorithms is used to predict the faults in the sensing and control system, obtaining the system fault prediction results. Since KPI can effectively select feature samples, and random forest has good tolerance to noise and high evaluation accuracy, the prediction accuracy of the obtained fault prediction model is greatly improved, thus making the obtained system fault prediction results more accurate. Furthermore, the fault characteristics obtained are all valid fault characteristics, which further improves the accuracy of system fault prediction and thus improves the reliability of the entire infection control system's disinfection work chain.
[0093] After obtaining the system fault prediction results, the process also includes the following steps:
[0094] Step S160: Based on the system fault prediction results, identify the faulty equipment and generate a fault alert message based on the faulty equipment. The system fault prediction results determine the fault type and thus the faulty equipment, generating a fault alert message to remind staff to perform maintenance, thereby preventing faults and improving the reliability of equipment operation.
[0095] Correspondingly, control information can be generated based on the system fault prediction results to control the operation of faulty equipment and ensure the normal operation of the sensing and control system.
[0096] Based on the same inventive concept, this invention also proposes a fault prediction device for a sensing and control system, please refer to... Figure 2 , Figure 2 A structural block diagram of a fault prediction device for a sensing and control system provided in an embodiment of the present invention. The fault prediction device for the sensing and control system includes:
[0097] The device segmentation module 110 is used to segment the devices in the sensing and control system according to the working stage to obtain multiple device sets, wherein each device set includes one or more devices;
[0098] The parameter information acquisition module 120 is used to acquire parameter information of each device in the device set in real time.
[0099] The evaluation module 130 is used to evaluate each device set according to the parameter information of each device set and the corresponding health evaluation model of each device set to obtain the evaluation results of each working stage.
[0100] The fault feature determination module 140 is used to determine fault features from the parameter information of each of the equipment in the equipment set based on the evaluation results of each working stage.
[0101] The fault prediction module 150 is used to predict the fault of the sensing and control system based on the fault characteristics and a fault prediction model constructed using the kernel principal component analysis algorithm and the random forest algorithm, so as to obtain the system fault prediction result.
[0102] In the above implementation process, the device partitioning module 110 divides the devices in the sensing and control system according to the working stages, resulting in multiple device sets, each device set including one or more devices; the parameter information acquisition module 120 acquires the parameter information of the devices in each device set in real time; the evaluation module 130 evaluates each device set using the corresponding health assessment model based on the parameter information of the devices in each device set, obtaining the evaluation results for each working stage; the fault feature determination module 140 determines the fault features from the parameter information of the devices in each device set based on the evaluation results of each working stage; by applying the health assessment model to perform health assessments for different working stages, the health status of each working stage can be determined, and the fault features that may affect the sensing and control system to malfunction can be determined based on the health status, indicating that they are valid fault features, thus ensuring the validity of the fault features. The fault prediction module 150 uses a fault prediction model constructed based on the kernel principal component analysis algorithm and the random forest algorithm to predict the faults in the sensing and control system based on the fault features, obtaining the system fault prediction results. Because kernel principal component analysis can effectively select feature samples, and random forests have good tolerance to noise and high evaluation accuracy, the prediction accuracy of the obtained fault prediction model is greatly improved, thus making the system fault prediction results more accurate. Furthermore, the obtained fault features are all valid fault features, further improving the accuracy of system fault prediction and consequently enhancing the reliability of the entire elimination work chain.
[0103] Please see Figure 3 , Figure 3This is a schematic structural block diagram of an electronic device provided in an embodiment of this application. The electronic device includes a memory 101, a processor 102, and a communication interface 103. The memory 101, processor 102, and communication interface 103 are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The memory 101 can be used to store software programs and modules, such as the program instructions / modules corresponding to the fault prediction device 100 of the sensing and control system provided in this embodiment of the application. The processor 102 executes various functional applications and data processing by executing the software programs and modules stored in the memory 101. The communication interface 103 can be used to communicate with other node devices for signaling or data.
[0104] The memory 101 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.
[0105] The processor 102 can be an integrated circuit chip with signal processing capabilities. The processor 102 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0106] Understandable. Figure 3 The structure shown is for illustrative purposes only; the electronic device may also include components that are more advanced than those shown. Figure 3 The more or fewer components shown, or having the same Figure 3 The different configurations shown. Figure 3 The components shown can be implemented using hardware, software, or a combination thereof.
[0107] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.
[0108] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.
[0109] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0110] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
[0111] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within this application. No reference numerals in the claims should be construed as limiting the scope of the claims.
Claims
1. A method for predicting faults in a sensing and control system, characterized in that, Includes the following steps: The devices in the sensing and control system are divided according to their working stages to obtain multiple device sets, each device set including one or more devices; Real-time acquisition of parameter information of each of the devices in the device cluster; Based on the parameter information of the devices in each of the aforementioned device sets, an assessment is conducted using the corresponding health assessment model for each device set to obtain the assessment results for each working stage. The evaluation results of each work stage are compared with the corresponding stage thresholds to obtain multiple comparison results; Based on the multiple comparison results, an abnormal working stage is determined; based on the abnormal working stage, parameter information of the corresponding equipment in the equipment set is extracted; the comparison results corresponding to the abnormal working stage are used as weights; and the parameter information of the corresponding equipment in the equipment set is used as feature parameters. Multiplying the feature parameters by the weights yields the fault features; Based on the fault characteristics, a fault prediction model constructed using the kernel principal component analysis algorithm and the random forest algorithm is used to predict the faults of the sensing and control system, and the system fault prediction results are obtained.
2. The fault prediction method for a sensing and control system according to claim 1, characterized in that, The construction of the health assessment model includes the following steps: Historical data samples of each device set are obtained respectively. The historical data samples of each device set include historical device parameter information and the health status of the corresponding working stage. Based on historical data samples from each device set, an evaluation model based on the XGBoost algorithm is trained to obtain a health evaluation model for each device set.
3. The fault prediction method for a sensing and control system according to claim 1, characterized in that, It also includes the following steps: Obtain the fault training set and fault test set; Based on the fault training set, the kernel principal component analysis algorithm and the random forest algorithm are used to construct the model and obtain the initial prediction model. The initial prediction model is optimized using the fault test set to obtain a fault prediction model.
4. The fault prediction method for a sensing and control system according to claim 3, characterized in that, Based on the fault training set, a model is constructed using the kernel principal component analysis algorithm and the random forest algorithm to obtain an initial prediction model, including the following steps: The feature samples in the fault training set are normalized to obtain preprocessed feature samples; Random noise is added to the preprocessed feature samples to obtain new feature samples; The kernel principal component analysis algorithm is used to extract kernel principal components from the new feature samples to obtain a new feature sample matrix. Based on the new feature sample matrix and the fault training set, the pre-set random forest model is trained to obtain the initial prediction model.
5. The fault prediction method for a sensing and control system according to claim 1, characterized in that, It also includes the following steps: Based on the system fault prediction results, the faulty equipment is identified, and fault alert information is generated based on the faulty equipment.
6. A fault prediction device for a sensing and control system, characterized in that, include: The device partitioning module is used to partition the devices in the sensing and control system according to the working stage to obtain multiple device sets, each device set including one or more devices; The parameter information acquisition module is used to acquire parameter information of each device in the device set in real time. The evaluation module is used to evaluate each device set based on its parameter information and the corresponding health evaluation model, and to obtain the evaluation results for each working stage. The fault feature determination module is used to compare the evaluation results of each working stage with the corresponding stage threshold to obtain multiple comparison results; determine the abnormal working stage based on the multiple comparison results; extract the parameter information of the corresponding equipment in the equipment set based on the abnormal working stage; use the comparison result corresponding to the abnormal working stage as a weight; and use the parameter information of the corresponding equipment in the equipment set as feature parameters. Multiplying the feature parameters by the weights yields the fault features; The fault prediction module is used to predict the faults of the sensing and control system based on the fault characteristics and a fault prediction model constructed using the kernel principal component analysis algorithm and the random forest algorithm, so as to obtain the system fault prediction result.
7. An electronic device, characterized in that, include: Memory, used to store one or more programs; processor; When the processor executes the one or more programs, it implements the sensor and control system fault prediction method as described in any one of claims 1-5.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the fault prediction method for the sensing and control system as described in any one of claims 1-5.