Elevator evaluation system based on information physical system

Abstract

The invention provides an elevator evaluation system based on an information physical system. The elevator evaluation system comprises an elevator sensor system, a data preprocessing unit, a fault similarity rate processing unit, a running state recognition probability processing unit and an elevator comprehensive evaluation unit. The sensor system collects data such as load, speed and position of the lift car, and the data preprocessing unit processes the data to obtain formatted data. The fault similarity rate processing unit analyzes the formatted data and extracts a fault similarity rate vector. And the running state identification probability processing unit intelligently analyzes the formatted data to obtain a normal running identification probability vector. And the comprehensive evaluation unit evaluates the safety and reliability of the elevator according to the two vectors. Evaluation efficiency and accuracy of safety and reliability of the elevator can be improved, faults are effectively prevented, and safety of passengers is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of information-physical systems, and more specifically, to an elevator evaluation system based on information-physical systems. Background Art

[0002] Existing elevator evaluation systems mainly rely on regular manual inspections and maintenance. This traditional method often fails to monitor the operating status of elevators in real time, resulting in the inability to detect potential faults and safety hazards in a timely manner. In addition, manual inspections are inefficient and costly, and are limited by the experience and skill level of inspectors, making it difficult to achieve a comprehensive and accurate evaluation. In terms of technical principles, traditional systems lack effective data collection and analysis methods, and are unable to conduct in-depth quantitative analysis of the various components and operating status of elevators, making it impossible to accurately predict the failure trends and maintenance needs of elevators. With the development of cyber-physical system (CPS) technology, an elevator evaluation system that integrates sensors, data preprocessing, intelligent analysis, and comprehensive evaluation has become possible. This system can achieve real-time monitoring of the operating status of elevators and fault prediction, improving the safety and reliability of elevators.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: the existing elevator evaluation system cannot provide real-time fault diagnosis and prediction, lacks comprehensive quantitative analysis of elevator components and operating status, and cannot conduct comprehensive evaluation of elevator safety and reliability based on real-time data. These problems limit the efficiency and effectiveness of elevator maintenance and safety management, and increase the risk of elevator failures and accidents. Summary of the invention

[0004] The present invention provides an elevator evaluation system based on a cyber-physical system, comprising: The elevator sensor system is used to obtain the original physical data during the operation of the elevator. The data types include car load data, car speed data, car position data, shaft rail pressure data, car and shaft wall distance data, machine room motor temperature data, and machine room motor vibration data; A data preprocessing unit, used to process the original physical data to obtain formatted data of various elevator components and operating status; A fault similarity processing unit, used for obtaining a fault similarity vector presented by each component of the elevator according to formatted data of each component and operating status of the elevator; An operation status recognition probability processing unit is used to intelligently analyze the formatted data of each elevator component and operation status, so as to obtain a normal operation recognition probability vector corresponding to the elevator operation status; The elevator comprehensive evaluation unit conducts a comprehensive evaluation of the safety and reliability of the elevator based on the fault similarity vector and the normal operation identification probability vector.

[0005] Furthermore, the fault similarity processing unit includes: Component data segmentation module, component feature extraction module and fault feature comparison module; wherein the component data segmentation module is used to divide the formatted data of each component and operating status of the elevator according to the area and component type, and to segment the data according to the number of independent components divided, and to The data of each independent component is input into the component feature extraction module; The component feature extraction module is used to extract feature values ​​from component data, the extracted feature values ​​include car load change feature values ​​and guide rail wear feature values, and the obtained component feature values ​​are input into the fault feature comparison module; The fault feature comparison module is used to compare the component feature value with the data in the fault database to obtain The fault similarity vector of each independent component is obtained and provided to the elevator comprehensive evaluation unit.

[0006] Furthermore, the formatted data of the elevator components and operating status include component spatial position information, operating parameters, physical property data, temperature data and vibration data, wherein the operating parameters include speed and acceleration, and the physical property data includes material strength related data and component aging degree data.

[0007] Furthermore, the component characteristic values ​​include the load distribution percentage, relative wear degree and structural characteristic parameters of the component, wherein the load distribution percentage is calculated based on the ratio of the component load to the rated load, the relative wear degree is calculated based on the ratio of the component wear amount to the allowable wear amount, and the structural characteristic parameters are determined based on the geometric shape and structural characteristics of the component.

[0008] Furthermore, the fault feature comparison module compares the component feature value with the data in the fault database to obtain the fault similarity vector of the component. ,in for The serial number of the independent parts is The part number of to For components Relative to The similarity of different fault types is calculated as follows: , in, It is a component No. eigenvalues, It is the first Type of fault eigenvalues, is the total number of characteristic values; wherein, the fault database has Component characteristic value data of the fault type, including the name of the fault , Load distribution percentage data , relative wear degree data , similarity threshold , and by The fault similarity threshold vector composed of similarity thresholds of class faults .

[0009] Further, the operating state identification probability processing unit includes a data projection optimization module and an operating state intelligent identification module; The data projection optimization module is used to perform multi-angle projection on the formatted data of the elevator components and the operating status from the data preprocessing unit to obtain a projection map, perform image enhancement processing on the projection map, and obtain two-dimensional operating status projection data. The image enhancement processing adopts a histogram equalization algorithm; The operating status intelligent recognition module is used to intelligently recognize the two-dimensional operating status projection data, obtain the normal operating status recognition probability vector, and provide it to the elevator comprehensive evaluation unit; the two-dimensional operating status projection data includes two-dimensional coordinate data and operating parameter value data normalized to the [0,1] interval corresponding to the coordinate point.

[0010] Furthermore, the operating status intelligent identification module includes an input layer, a hidden layer and an output layer; The hidden layer includes a convolution layer, a pooling layer and a fully connected layer; wherein the input data of the input layer is two-dimensional running state projection data; The convolution layer has 20 convolution kernels, a convolution kernel size of 3×3, a step size of 1, and uses ReLU as an activation function to extract features from the input data from the input layer. For each input image, 20 feature maps can be extracted; The pooling layer adopts maximum pooling with a step size of 2; The fully connected layer is used to perform weighted summation on the features extracted by the convolution layer and the pooling layer to obtain the result of identifying and classifying the two-dimensional running state projection data; The output layer is the terminal neurons of the fully connected layer, and the total number of neurons is equal to the total number of normal operating state categories. ; Each output layer neuron represents a normal operating status category, and its output value is the normalized classification probability value; The operating state intelligent identification module uses the back propagation algorithm to train the network parameters before application. The training set includes Elevator operation data dataset with labels of normal operation status; The recognition results after training are statistically analyzed to obtain Classification identification threshold for normal operation status to , which constitutes the classification recognition threshold vector of normal operation status .

[0011] Furthermore, by inputting the two-dimensional operating state projection data into the operating state intelligent recognition module, the recognition probability relative to each type of normal operating state is obtained. to , which constitutes the normal operation recognition probability vector of the elevator operation state , the calculation formula is: , in, , is the weight connecting the i-th input feature and the j-th neuron in the fully connected layer, is the i-th feature of the input layer (such as two-dimensional coordinate data, normalized operating parameter value data, etc.), is the bias of the jth neuron. It refers to the recognition probability of the jth normal operating state. Its value indicates the possibility of the elevator being in the jth normal operating state. Its value range is between 0 and 1. It is calculated by the above formula. The normal operation identification probability vector T is constructed.

[0012] Furthermore, the safety and reliability of the elevator are comprehensively evaluated to obtain evaluation indicators, including: For m independent components, compare and evaluate them one by one, and sort the evaluation results into a list; The evaluation process is: The fault similarity vector and fault similarity threshold vector Compare, when hour ,Will Set to 0, where For components Structural characteristic parameters; The normal operation will identify the probability vector Classification recognition threshold vector for normal operation status Compare, when hour ,Will Set to 0; After the above processing, the evaluation result vector is calculated , the calculation formula is: , in, is the suspected fault assessment threshold for the jth type of fault, which is the critical value used to determine whether a component is suspected of having this type of fault based on the safety margin of the elevator under different fault modes, statistical analysis of historical fault data, and expert experience; is an evaluation index, which indicates the comprehensive suspicion degree of the component relative to the jth type of fault. When , it means that from the perspective of comprehensive fault similarity and normal operation recognition probability, the component has a tendency to be suspected of type j fault; when , it means that the component has no signs of type j failure from the current evaluation perspective.

[0013] Furthermore, it also includes an alarm unit; the alarm unit, for the suspected fault name evaluated, to The values ​​of items greater than 0 are sorted in real time and an alarm is issued on the display terminal.

[0014] The above-mentioned embodiments of the present invention have at least the following beneficial effects: the elevator evaluation system based on the information physical system of the present invention can monitor and analyze the key physical data of the elevator operation process in real time by integrating the elevator sensor system, data preprocessing unit, fault similarity processing unit, operation state identification probability processing unit and elevator comprehensive evaluation unit, so as to provide a detailed evaluation of each component and operation state of the elevator. The system can accurately identify the fault similarity vector and normal operation identification probability vector of the elevator, and then conduct a comprehensive evaluation of the safety and reliability of the elevator, which can effectively improve the accuracy and timeliness of elevator fault prediction.

[0015] In addition, the system can automatically extract the characteristic values ​​of elevator components through intelligent analysis and deep learning technology, and compare them with the data in the fault database to obtain the fault similarity vector. At the same time, the system uses image enhancement processing and intelligent recognition technology to project and identify the operating status of the elevator from multiple angles to obtain the normal operation recognition probability vector. The application of these technologies can not only improve the efficiency of elevator evaluation, but also reduce maintenance costs, enhance the safety of elevator use, and provide strong technical support for elevator maintenance and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which: Figure 1 A schematic diagram of the structure of an elevator evaluation system based on a cyber-physical system provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0017] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0018] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present invention can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0019] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0020] Reference below Figure 1 , Figure 1 This is a schematic diagram of the structure of an elevator evaluation system based on a cyber-physical system provided by an embodiment of the present invention. Figure 1 As shown, an elevator evaluation system 100 based on a cyber-physical system includes: The elevator sensor system 101 is used to obtain the original physical data during the operation of the elevator. The data types include car load data, car speed data, car position data, shaft guide rail pressure data, car and shaft wall distance data, machine room motor temperature data, and machine room motor vibration data; The data preprocessing unit 102 is used to process the original physical data to obtain formatted data of various components and operating status of the elevator; The fault similarity processing unit 103 is used to obtain the fault similarity vector presented by each component of the elevator according to the formatted data of each component and the running state of the elevator; The operation state identification probability processing unit 104 is used to intelligently analyze the formatted data of each component and operation state of the elevator, so as to obtain a normal operation identification probability vector corresponding to the elevator operation state; The elevator comprehensive evaluation unit 105 performs a comprehensive evaluation on the safety and reliability of the elevator according to the fault similarity vector and the normal operation identification probability vector.

[0021] It should be noted that the core of this elevator evaluation system is the integration of multiple units for real-time monitoring and evaluation of the elevator's operating status. The elevator sensor system is the data acquisition basis of the system, which is responsible for collecting the original physical data generated by the elevator during operation. These data include car load data, car speed data, car position data, etc., which are crucial for evaluating the safety and reliability of the elevator. Here, the elevator sensor system refers to a series of sensor devices installed inside the elevator, which can monitor and record the elevator's operating parameters in real time.

[0022] Specifically, the elevator sensor system may include a weight sensor to monitor the car load, a speed sensor to monitor the car speed, and a position sensor to determine the exact position of the car in the shaft. The data from these sensors are transmitted in real time to a data preprocessing unit, which formats the raw data into formatted data that can be used for further analysis. For example, the weight sensor can be set to a measurement range of 0-1000 kg, the speed sensor can be set to a measurement range of 0-5 m / s, and the position sensor can be accurate to the millimeter level.

[0023] Preferably, the data preprocessing unit can use a specific algorithm to filter out noise and abnormal values ​​to ensure that the data input to the fault similarity processing unit is accurate and reliable. For example, a Kalman filter can be used to reduce noise in speed data, or a threshold method can be used to exclude load data that exceeds a normal range.

[0024] Furthermore, the data preprocessing unit may also include a data normalization step to convert data of different dimensions into the same dimension to facilitate subsequent analysis and comparison. Through these optimization measures, the system can more accurately evaluate the operating status of the elevator and promptly detect potential fault risks.

[0025] In some embodiments, the fault similarity processing unit includes: Component data segmentation module, component feature extraction module and fault feature comparison module; wherein the component data segmentation module is used to divide the formatted data of each component and operating status of the elevator according to the area and component type, and to segment the data according to the number of independent components divided, and to The data of each independent component is input into the component feature extraction module; The component feature extraction module is used to extract feature values ​​from component data, the extracted feature values ​​include car load change feature values ​​and guide rail wear feature values, and the obtained component feature values ​​are input into the fault feature comparison module; The fault feature comparison module is used to compare the component feature value with the data in the fault database to obtain The fault similarity vector of each independent component is obtained and provided to the elevator comprehensive evaluation unit.

[0026] It should be noted that the fault similarity processing unit is a key component in the elevator evaluation system. It is responsible for calculating the fault similarity vector presented by each elevator component based on the formatted data of each elevator component and operating status. This vector can reflect the similarity between the elevator component and the known fault mode, thereby providing a basis for the fault diagnosis of the elevator. Here, the fault similarity vector refers to a set of multiple fault similarity values, each of which corresponds to an independent component of the elevator.

[0027] Specifically, the fault similarity processing unit includes a component data segmentation module, a component feature extraction module and a fault feature comparison module. The component data segmentation module is responsible for dividing the formatted data according to the region and component type, and then dividing the data according to the number of independent components divided. The component feature extraction module extracts feature values ​​from the segmented data, and these feature values ​​may include the feature values ​​of car load change, the feature values ​​of guide rail wear, etc. The fault feature comparison module compares these feature values ​​with the data in the fault database to obtain the fault similarity vector of each independent component.

[0028] Preferably, the component data segmentation module may use a specific algorithm, such as a clustering algorithm, to identify and segment data. The component feature extraction module may use statistical analysis methods to extract feature values ​​of car load changes, or use image processing technology to extract feature values ​​of guide rail wear.

[0029] Furthermore, the fault feature comparison module may use machine learning methods, such as support vector machines (SVM) or neural networks, to improve the accuracy of fault diagnosis. The parameter settings and algorithm selections of these modules can be adjusted according to the specific type and use environment of the elevator to achieve the best fault diagnosis effect.

[0030] In some embodiments, the formatted data of the elevator components and operating status include component spatial position information, operating parameters, physical property data, temperature data and vibration data, wherein the operating parameters include speed and acceleration, and the physical property data includes material strength related data and component aging degree data.

[0031] It should be noted that this sentence describes the specific content of the formatted data of each elevator component and operating status. The formatted data here refers to the structured and standardized data after being processed by the preprocessing unit, which provides the basis for subsequent fault analysis and operating status identification. The spatial position information of the component refers to the precise position of the component in the elevator system, while the operating parameters include dynamic data such as the speed and acceleration of the elevator.

[0032] Specifically, the formatted data of the elevator components and operating status includes the spatial position information of the components, which can be the coordinates or distance of the components relative to other parts of the elevator. Operating parameters include not only speed and acceleration, but may also include the opening and closing status of the elevator doors, the number of starts and stops of the elevator, etc. Physical property data may include data related to material strength, such as the yield strength of steel, and component aging data, such as the degree of wear of cables. Temperature data and vibration data refer to the temperature changes and vibration conditions of elevator components during operation, respectively, which are important parameters for evaluating the status of the elevator.

[0033] Preferably, the spatial position information of the components can be obtained by sensors installed on various parts of the elevator, which can provide accurate position data. The operating parameters can be monitored in real time by velocity sensors and acceleration sensors, which can be set to high sensitivity to capture small changes in the operation of the elevator. Physical property data can be obtained through regular maintenance inspections and non-destructive testing, such as using ultrasonic testing technology to evaluate the strength and aging of materials. Temperature data and vibration data can be monitored by thermal imagers and vibration analyzers, which can be set to automatically record and analyze data for further analysis in the data preprocessing unit. Through these refined operating steps, the status of the elevator can be evaluated more accurately and potential problems can be discovered in time.

[0034] In some embodiments, the component characteristic values ​​include the load distribution percentage, relative wear degree and structural characteristic parameters of the component, wherein the load distribution percentage is calculated based on the ratio of the component load to the rated load, the relative wear degree is calculated based on the ratio of the component wear amount to the allowable wear amount, and the structural characteristic parameters are determined based on the geometric shape and structural characteristics of the component.

[0035] It should be noted that this sentence elaborates on the specific content of the component characteristic values, which are key indicators for evaluating the condition of elevator components. Here, the load distribution percentage refers to the ratio between the actual load of the component and its rated load, and the relative wear degree refers to the ratio between the current wear of the component and its maximum allowable wear. The structural characteristic parameters involve the geometric shape and structural characteristics of the component, which are important parameters for evaluating the health of elevator components.

[0036] Specifically, component characteristic values ​​include the load distribution percentage, which is calculated by comparing the actual load of the component with its design rated load, and can be used to evaluate whether the component is overloaded. The relative degree of wear is determined by measuring the wear of the component and comparing it with the maximum allowable wear, which helps to identify the wear state of the component. Structural characteristic parameters are determined based on the geometry and structural characteristics of the component, and may include the size, shape, material properties, etc. of the component. These parameters are crucial to understanding the physical condition of the component and predicting its durability.

[0037] Preferably, the load distribution percentage can be measured by a load cell installed on the elevator and compared with the rated load provided in the elevator design documentation. The relative degree of wear can be evaluated through regular maintenance checks and the use of wear detection tools (such as wear gauges). The determination of structural characteristic parameters may involve the use of 3D scanning technology to obtain the precise geometry of the components, or the determination of the material properties of the components through material testing. The precise measurement and analysis of these parameters can provide a more accurate assessment of the condition of the elevator components, thereby improving the safety and reliability of the elevator. As an alternative, machine learning algorithms can also be used to analyze historical maintenance data to automatically predict the load distribution and wear degree of components, further optimizing the maintenance and assessment process of the elevator.

[0038] In some embodiments, the fault feature comparison module compares the component feature value with the data in the fault database to obtain the fault similarity vector of the component. ,in for The serial number of the independent parts is The part number of to For components Relative to The similarity of different fault types is calculated as follows: , in, It is a component No. eigenvalues, It is the first Type of fault eigenvalues, is the total number of characteristic values; wherein, the fault database has Component characteristic value data of the fault type, including the name of the fault , Load distribution percentage data , relative wear degree data , similarity threshold , and by The fault similarity threshold vector composed of similarity thresholds of class faults .

[0039] It should be noted that this sentence describes the working principle of the fault feature comparison module, that is, how to compare the component feature value with the data in the fault database to obtain the fault similarity vector. Here, the fault similarity vector is a set of values ​​that represent the similarity between the component and different fault types, while the fault database contains the feature value data of various faults, which are used to compare with the component feature values ​​monitored in real time. Fault similarity is a quantitative indicator used to measure the similarity between the current state of a component and a specific fault type.

[0040] Specifically, the fault feature comparison module compares the feature values ​​of the component with the feature values ​​in the fault database to calculate the similarity of each component with respect to each type of fault. This calculation involves comparing the difference between each feature value of each component and the feature value of the corresponding fault type in the fault database. The data in the fault database includes the name of the fault, load distribution percentage data, relative wear degree data, similarity threshold, etc. These data provide a benchmark for fault diagnosis, enabling the system to identify the similarity between the component status and known fault modes. The total number of feature values ​​in the calculation formula refers to the number of feature values ​​considered in the comparison process.

[0041] Preferably, the fault feature comparison module can use a machine learning algorithm, such as a nearest neighbor algorithm or a neural network, to improve the calculation accuracy of the fault similarity rate. In actual operation, a threshold can be set. When the calculated fault similarity rate exceeds this threshold, the system will consider that the component may have a corresponding fault risk.

[0042] Furthermore, the fault database can be regularly updated to include the latest fault cases and characteristic value data, thereby improving the diagnostic capabilities of the system. As an alternative, a cloud computing platform can also be used to store and process the fault database, which can utilize a larger data set and improve the accuracy and efficiency of fault diagnosis. Through these detailed operation steps, the system can more effectively monitor and evaluate the health of elevator components.

[0043] In some embodiments, the operating state identification probability processing unit includes a data projection optimization module and an operating state intelligent identification module; The data projection optimization module is used to perform multi-angle projection on the formatted data of the elevator components and the operating status from the data preprocessing unit to obtain a projection map, perform image enhancement processing on the projection map, and obtain two-dimensional operating status projection data. The image enhancement processing adopts a histogram equalization algorithm; The operating status intelligent recognition module is used to intelligently recognize the two-dimensional operating status projection data, obtain the normal operating status recognition probability vector, and provide it to the elevator comprehensive evaluation unit; the two-dimensional operating status projection data includes two-dimensional coordinate data and operating parameter value data normalized to the [0,1] interval corresponding to the coordinate point.

[0044] It should be noted that this sentence describes the composition and function of the operating state recognition probability processing unit, which is responsible for intelligently analyzing the formatted data of the elevator components and operating states to obtain the normal operation recognition probability vector corresponding to the elevator operating state. Here, the data projection optimization module refers to the module that performs multi-angle projection and image enhancement processing on the formatted data, and the operating state intelligent recognition module refers to the module that performs intelligent recognition on the projection data. They work together to obtain the probability vector of the normal operation of the elevator.

[0045] Specifically, the data projection optimization module projects the formatted data of each component and operating status of the elevator at multiple angles to obtain a projection map, and uses a histogram equalization algorithm to perform image enhancement processing on the projection map to obtain two-dimensional operating status projection data. These data include two-dimensional coordinate data and operating parameter value data normalized to the [0,1] interval corresponding to the coordinate points. The operating status intelligent recognition module intelligently recognizes these two-dimensional data and obtains a normal operating recognition probability vector, which provides the probability of the elevator being in a normal operating state.

[0046] Preferably, the data projection optimization module can set multiple projection angles to ensure that the operating status of the elevator is captured from different perspectives and enhance the robustness of the algorithm. The histogram equalization algorithm can adjust parameters, such as contrast-limited adaptive histogram equalization (CLAHE), to adapt to different lighting conditions and data characteristics. The operating status intelligent recognition module can use deep learning technology, such as convolutional neural network (CNN), to improve the accuracy of recognition. In this network, the number of convolution kernels of the convolution layer can be set to 20, the convolution kernel size can be set to 3×3, the step size can be set to 1, and the ReLU activation function can be used. The pooling layer can use maximum pooling with a step size of 2 to reduce the size of the feature map and extract key features. The fully connected layer performs weighted summation on the extracted features to obtain the recognition classification result. The total number of neurons in the output layer is equal to the total number of categories of normal operating status, and the output value of each neuron represents the normalized classification probability value of the corresponding category. Through these refined operation steps, the operating status of the elevator can be evaluated more accurately.

[0047] In some embodiments, the operating status intelligent identification module includes an input layer, a hidden layer, and an output layer; The hidden layer includes a convolution layer, a pooling layer and a fully connected layer; wherein the input data of the input layer is two-dimensional running state projection data; The convolution layer has 20 convolution kernels, a convolution kernel size of 3×3, a step size of 1, and uses ReLU as an activation function to extract features from the input data from the input layer. For each input image, 20 feature maps can be extracted; The pooling layer adopts maximum pooling with a step size of 2; The fully connected layer is used to perform weighted summation on the features extracted by the convolution layer and the pooling layer to obtain the result of identifying and classifying the two-dimensional running state projection data; The output layer is the terminal neurons of the fully connected layer, and the total number of neurons is equal to the total number of normal operating state categories. ; Each output layer neuron represents a normal operating status category, and its output value is the normalized classification probability value; The operating state intelligent identification module uses the back propagation algorithm to train the network parameters before application. The training set includes Elevator operation data dataset with labels of normal operation status; The recognition results after training are statistically analyzed to obtain Classification identification threshold for normal operation status to , which constitutes the classification recognition threshold vector of normal operation status .

[0048] It should be noted that this sentence describes in detail the structure and function of the operating status intelligent recognition module, which is the key part for intelligent recognition of two-dimensional operating status projection data. Here, the input layer refers to the first layer of the neural network, which receives the two-dimensional operating status projection data from the data preprocessing unit; the hidden layer contains multiple different layers, such as convolutional layers, pooling layers, and fully connected layers, which work together to extract features and classify them; the output layer is the last layer of the network, which provides the recognition probability of each normal operating status category.

[0049] Specifically, the input layer of the operating status intelligent recognition module receives two-dimensional operating status projection data, which includes two-dimensional coordinate data after image enhancement processing and operating parameter value data normalized to the interval [0,1]. The convolution layer in the hidden layer uses 20 3×3 convolution kernels with a step size of 1 and a ReLU activation function to extract features. The subsequent pooling layer uses maximum pooling with a step size of 2 to reduce the dimension of the feature map. The fully connected layer performs weighted summation on the features extracted by the convolution layer and the pooling layer to obtain the recognition and classification results of the two-dimensional operating status projection data. The output layer is composed of the terminal neurons of the fully connected layer, each neuron represents a normal operating status category, and the output value is the normalized classification probability value.

[0050] Preferably, the operating status intelligent recognition module uses a back-propagation algorithm to train network parameters before actual application, and the training set includes labels of multiple categories of normal operating status and corresponding elevator operation data sets. For the convolution layer, different numbers and sizes of convolution kernels can be selected to meet different feature extraction requirements. In addition to maximum pooling, the pooling layer can also use average pooling or other types of pooling techniques. The number of neurons in the fully connected layer can be adjusted according to the actual number of categories. The output value of each neuron in the output layer can be normalized by a softmax function to ensure that the sum of the probabilities of all categories is 1. In addition, dropout technology or L2 regularization can be used to prevent overfitting and improve the generalization ability of the model. Through these refined operation steps, the performance and accuracy of the intelligent recognition module can be improved.

[0051] In some embodiments, the two-dimensional operating state projection data is input into the operating state intelligent recognition module to obtain the recognition probability relative to each type of normal operating state. to , which constitutes the normal operation recognition probability vector of the elevator operation state , the calculation formula is: , in, , is the weight connecting the i-th input feature and the j-th neuron in the fully connected layer, is the i-th feature of the input layer (such as two-dimensional coordinate data, normalized operating parameter value data, etc.), is the bias of the jth neuron. It refers to the recognition probability of the jth normal operating state. Its value indicates the possibility of the elevator being in the jth normal operating state. Its value range is between 0 and 1. It is calculated by the above formula. The normal operation identification probability vector T is constructed.

[0052] It should be noted that this sentence describes how to obtain the normal operation recognition probability vector of the elevator operation state through the operation state intelligent recognition module. Here, the normal operation recognition probability vector refers to a set of multiple probability values, each probability value represents the possibility of the elevator being in a specific normal operation state. The fully connected layer refers to a layer in a neural network, in which each neuron is connected to all neurons in the previous layer, which is used to integrate feature information and output the final classification result.

[0053] Specifically, by inputting the two-dimensional running state projection data into the running state intelligent recognition module, the recognition probability relative to each type of normal running state can be obtained. These probability values ​​are calculated through the forward propagation of the neural network, which involves the calculation of weights and biases. Specifically, the output value of each output layer neuron is calculated through a weighted sum operation (including the weights and biases from the fully connected layer) and an activation function (such as softmax), which maps the input features to a probability distribution, indicating the probability of the elevator being in each normal running state.

[0054] Preferably, the specific formula for calculating the normal operation recognition probability can be further refined. For example, for each neuron in the fully connected layer, its output value can be calculated by the following formula: , in, is the recognition probability of the jth normal operating state, is the weight connecting the i-th input feature and the j-th neuron, is the value of the i-th input feature, is the bias of the jth neuron, and is the base of the natural logarithm. This formula ensures that the output value is between 0 and 1, and the sum of all output values ​​is 1, forming a valid probability distribution.

[0055] More specifically, the cross entropy loss function can be used to train the neural network, optimize the weights and biases, and make the probability distribution of the network output as close to the true label distribution as possible. In addition, various optimization algorithms, such as gradient descent, Adam optimizer, etc., can be used to update the network parameters.

[0056] Furthermore, in practical applications, data enhancement techniques can be used to expand the training set and improve the robustness of the model. These refined operation steps and alternatives help improve the accuracy and reliability of elevator operation status recognition.

[0057] In some embodiments, the comprehensive evaluation of the safety and reliability of the elevator to obtain the evaluation index includes: For m independent components, compare and evaluate them one by one, and sort the evaluation results into a list; The evaluation process is: The fault similarity vector and fault similarity threshold vector Compare, when hour ,Will Set to 0, where For components Structural characteristic parameters; The normal operation will identify the probability vector Classification recognition threshold vector for normal operation status Compare, when hour ,Will Set to 0; After the above processing, the evaluation result vector is calculated , the calculation formula is: , in, is the suspected fault assessment threshold for the jth type of fault, which is the critical value used to determine whether a component is suspected of having this type of fault based on the safety margin of the elevator under different fault modes, statistical analysis of historical fault data, and expert experience; is an evaluation index, which indicates the comprehensive suspicion degree of the component relative to the jth type of fault. When , it means that from the perspective of comprehensive fault similarity and normal operation recognition probability, the component has a tendency to be suspected of type j fault; when , it means that the component has no signs of type j failure from the current evaluation perspective.

[0058] It should be noted that this sentence describes how to conduct a comprehensive evaluation of the safety and reliability of the elevator and determine the evaluation indicators. Here, the evaluation indicators refer to a series of quantitative values ​​used to measure the comprehensive suspicion of elevator components relative to various types of faults. The suspected fault assessment threshold refers to a pre-set critical value used to determine whether a component is suspected of a certain type of fault.

[0059] Specifically, the evaluation process involves comparing and evaluating m independent components in turn, and sorting the evaluation results into a list. For each component, the fault similarity vector of the component is first compared with the fault similarity threshold vector. If the fault similarity of the component is lower than the corresponding threshold, the fault similarity of the component is set to 0. Next, the normal operation recognition probability vector is compared with the classification recognition threshold vector of the normal operation state. If the probability value is lower than the corresponding threshold, the probability value is set to 0. After processing, the evaluation result vector is obtained through a specific calculation formula, which reflects the comprehensive suspicion degree of the component relative to various types of faults.

[0060] Preferably, the calculation of the evaluation result vector can adopt the following formula: , in, is the evaluation index for the jth type of fault, is the failure similarity of the components, is the normal operation recognition probability, is the suspected fault assessment threshold for the jth type of fault. When , it indicates that the component has a tendency to be suspected of type j failure; when , it indicates that the component has no signs of type j failure.

[0061] More specifically, the suspected fault assessment threshold It can be determined based on the safety margin of the elevator under different failure modes, statistical analysis of historical failure data, expert experience, etc. For example, a dynamic threshold based on historical data can be set, which will be adjusted according to the frequency of use of the elevator, the aging of the components and the maintenance record.

[0062] Furthermore, the evaluation process can use machine learning methods, such as random forest or gradient boosting machine, to automatically determine the evaluation indicators and thresholds, and improve the accuracy and adaptability of the evaluation. These refined operation steps and alternatives help to achieve more accurate elevator component failure assessment.

[0063] In some embodiments, an alarm unit is also included; the alarm unit, based on the suspected fault name evaluated, to The values ​​of items greater than 0 are sorted in real time and an alarm is issued on the display terminal.

[0064] It should be noted that this sentence describes the function and operation of the alarm unit, which is an important part of the elevator evaluation system and is responsible for issuing real-time alarms based on the evaluation results. Here, the alarm unit refers to a module in the system that sorts the suspected faults in real time according to the values ​​of the items greater than 0 in the evaluation results based on the names of the suspected faults evaluated, and issues an alarm on the display terminal.

[0065] Specifically, the alarm unit receives the evaluation results from the comprehensive evaluation unit, which include the suspected fault evaluation indicators of each component. The alarm unit analyzes these indicators, identifies those components with evaluation indicators greater than 0 as suspected faults, and sorts them according to the size of the indicator value to determine the priority of the alarm. The alarm unit then displays the names of these suspected faulty components and the corresponding alarm levels on the display terminal so that maintenance personnel can take prompt action.

[0066] Preferably, the alarm unit can be set with multiple alarm levels, for example, it can be divided into three levels: warning, serious and emergency according to the value of the evaluation index. Each level of alarm can have different display colors and sound prompts, so that maintenance personnel can quickly identify the severity of the alarm. In addition, the alarm unit can also be integrated into the elevator maintenance management system to realize remote monitoring and alarm notification functions, such as push notifications to maintenance personnel via email, SMS or mobile applications.

[0067] More specifically, the alarm logic of the alarm unit can be further refined. For example, a dynamic threshold adjustment mechanism can be set to automatically adjust the alarm threshold according to the elevator's usage history and maintenance records to adapt to different usage environments and conditions.

[0068] Furthermore, the alarm unit can also provide auxiliary information for fault diagnosis, such as the possible cause of the faulty component, recommended maintenance measures and estimated repair time. These detailed operation steps and alternatives help improve the practicality and maintenance efficiency of the alarm unit.

[0069] The above-mentioned embodiments of the present invention have the following beneficial effects: The elevator evaluation system of the present invention can comprehensively evaluate the safety and reliability of the elevator by comprehensively using the elevator sensor system, data preprocessing unit, fault similarity processing unit, operating state identification probability processing unit and elevator comprehensive evaluation unit. The system can collect and process the original physical data during the operation of the elevator in real time, including key parameters such as car load, speed, position, etc., and extract the fault similarity vector and normal operation identification probability vector through intelligent analysis technology. This integrated evaluation method can significantly improve the accuracy of elevator fault diagnosis, reduce potential safety risks, extend the service life of the elevator, optimize maintenance plans, and reduce maintenance costs.

[0070] In addition, the system is also designed with an alarm unit that can issue real-time alarms based on the evaluation results to remind managers to pay attention to possible faulty components. This real-time monitoring and alarm mechanism can quickly respond to abnormal conditions of the elevator, take timely measures to avoid failures, and ensure the safety of passengers. Through this intelligent evaluation and alarm system, the efficiency and effectiveness of elevator management can be greatly improved, providing a strong guarantee for the safe operation of elevators.

[0071] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0072] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. An elevator evaluation system based on a cyber-physical system, characterized in that: include: The elevator sensor system is used to obtain the original physical data during the operation of the elevator. The data types include car load data, car speed data, car position data, shaft rail pressure data, car and shaft wall distance data, machine room motor temperature data, and machine room motor vibration data; A data preprocessing unit, used to process the original physical data to obtain formatted data of various elevator components and operating status; A fault similarity processing unit, used for obtaining a fault similarity vector presented by each component of the elevator according to formatted data of each component and operating status of the elevator; An operation status recognition probability processing unit is used to intelligently analyze the formatted data of each elevator component and operation status, so as to obtain a normal operation recognition probability vector corresponding to the elevator operation status; The elevator comprehensive evaluation unit conducts a comprehensive evaluation of the safety and reliability of the elevator based on the fault similarity vector and the normal operation identification probability vector.

2. The elevator evaluation system based on cyber-physical system according to claim 1, characterized in that: The fault similarity processing unit comprises: Component data segmentation module, component feature extraction module and fault feature comparison module; wherein the component data segmentation module is used to divide the formatted data of each component and operating status of the elevator according to the area and component type, and to segment the data according to the number of independent components divided, and to The data of each independent component is input into the component feature extraction module; The component feature extraction module is used to extract feature values ​​from component data, the extracted feature values ​​include car load change feature values ​​and guide rail wear feature values, and the obtained component feature values ​​are input into the fault feature comparison module; The fault feature comparison module is used to compare the component feature value with the data in the fault database to obtain The fault similarity vector of each independent component is obtained and provided to the elevator comprehensive evaluation unit.

3. The elevator evaluation system based on cyber-physical system according to claim 2, characterized in that: The formatted data of the elevator components and operating status include component spatial position information, operating parameters, physical property data, temperature data and vibration data, wherein the operating parameters include speed and acceleration, and the physical property data includes material strength related data and component aging degree data.

4. The elevator evaluation system based on cyber-physical system according to claim 3, characterized in that: The component characteristic values ​​include the load distribution percentage, relative wear degree and structural characteristic parameters of the component, wherein the load distribution percentage is calculated based on the ratio of the component load to the rated load, the relative wear degree is calculated based on the ratio of the component wear amount to the allowable wear amount, and the structural characteristic parameters are determined based on the geometric shape and structural characteristics of the component.

5. The elevator evaluation system based on cyber-physical system according to claim 4, characterized in that: The fault feature comparison module compares the component feature value with the data in the fault database to obtain the fault similarity vector of the component. ,in for The serial number of the independent parts is The part number of to For components Relative to The similarity of different fault types is calculated as follows: ， in, It is a component No. eigenvalues, It is the first Type of fault eigenvalues, is the total number of characteristic values; wherein, the fault database has Component characteristic value data of the fault type, including the name of the fault , Load distribution percentage data , relative wear degree data , similarity threshold , and by The fault similarity threshold vector composed of similarity thresholds of class faults .

6. The elevator evaluation system based on cyber-physical system according to claim 5, characterized in that: The operating state recognition probability processing unit includes a data projection optimization module and an operating state intelligent recognition module; The data projection optimization module is used to perform multi-angle projection on the formatted data of the elevator components and the operating status from the data preprocessing unit to obtain a projection map, perform image enhancement processing on the projection map, and obtain two-dimensional operating status projection data. The image enhancement processing adopts a histogram equalization algorithm; The operating status intelligent recognition module is used to intelligently recognize the two-dimensional operating status projection data, obtain the normal operating status recognition probability vector, and provide it to the elevator comprehensive evaluation unit; the two-dimensional operating status projection data includes two-dimensional coordinate data and operating parameter value data normalized to the [0,1] interval corresponding to the coordinate point.

7. The elevator evaluation system based on cyber-physical system according to claim 6, characterized in that: The operating status intelligent identification module includes an input layer, a hidden layer and an output layer; The hidden layer includes a convolution layer, a pooling layer and a fully connected layer; wherein the input data of the input layer is two-dimensional running state projection data; The convolution layer has 20 convolution kernels, a convolution kernel size of 3×3, a step size of 1, and uses ReLU as an activation function to extract features from the input data from the input layer. For each input image, 20 feature maps can be extracted; The pooling layer adopts maximum pooling with a step size of 2; The fully connected layer is used to perform weighted summation on the features extracted by the convolution layer and the pooling layer to obtain the result of identifying and classifying the two-dimensional running state projection data; The output layer is the terminal neurons of the fully connected layer, and the total number of neurons is equal to the total number of normal operating state categories. ; Each output layer neuron represents a normal operating status category, and its output value is the normalized classification probability value; The operating state intelligent identification module uses the back propagation algorithm to train the network parameters before application. The training set includes Elevator operation data dataset with labels of normal operation status; The recognition results after training are statistically analyzed to obtain Classification identification threshold for normal operation status to , which constitutes the classification recognition threshold vector of normal operation status .

8. The elevator evaluation system based on cyber-physical system according to claim 7, characterized in that: By inputting the two-dimensional operation status projection data into the operation status intelligent recognition module, the recognition probability relative to each type of normal operation status is obtained. to , which constitutes the normal operation recognition probability vector of the elevator operation state , the calculation formula is: ， in, , is the weight connecting the i-th input feature and the j-th neuron in the fully connected layer, is the i-th feature of the input layer (such as two-dimensional coordinate data, normalized operating parameter value data, etc.), is the bias of the jth neuron. It refers to the recognition probability of the jth normal operating state. Its value indicates the possibility of the elevator being in the jth normal operating state. Its value range is between 0 and 1. It is calculated by the above formula. The normal operation identification probability vector T is constructed.

9. The elevator evaluation system based on cyber-physical system according to claim 8, characterized in that: The safety and reliability of the elevator are comprehensively evaluated to obtain evaluation indicators, including: For m independent components, compare and evaluate them one by one, and sort the evaluation results into a list; The evaluation process is: The fault similarity vector and fault similarity threshold vector Compare, when hour ,Will Set to 0, where For components Structural characteristic parameters; The normal operation will identify the probability vector Classification recognition threshold vector for normal operation status Compare, when hour ,Will Set to 0; After the above processing, the evaluation result vector is calculated , the calculation formula is: ， in, is the suspected fault assessment threshold for the jth type of fault, which is the critical value used to determine whether a component is suspected of having this type of fault based on the safety margin of the elevator under different fault modes, statistical analysis of historical fault data, and expert experience; is an evaluation index, which indicates the comprehensive suspicion degree of the component relative to the jth type of fault. When , it means that from the perspective of comprehensive fault similarity and normal operation recognition probability, the component has a tendency to be suspected of type j fault; when , it means that the component has no signs of type j failure from the current evaluation perspective.

10. The elevator evaluation system based on cyber-physical system according to claim 9, characterized in that: It also includes an alarm unit; the alarm unit, based on the suspected fault name evaluated, to The values ​​of items greater than 0 are sorted in real time and an alarm is issued on the display terminal.

Images