An intelligent elevator status detection method

Abstract

The present invention discloses an intelligent elevator status detection method, which includes elevator status data acquisition, determination of change indicators, initial clustering of elevator status, optimization of initial clustering of elevator status, refinement of secondary clustering of elevator status, and elevator status detection. The present invention belongs to the field of data processing, and specifically refers to an intelligent elevator status detection method. This solution evaluates the benefits and costs through change indicators, and introduces a time-delay attention weight to dynamically balance the processing quality and response speed of elevator status features. When the system load is high or the response requirement is urgent, the timeliness of fault detection is preferentially guaranteed; thereby improving the accuracy of elevator status detection; introducing a segmented chaotic mapping and a non-linear search strategy to shorten the overall processing time of elevator status detection; combining the representative vector within the cluster and time continuity to optimize the secondary clustering of the historical elevator status set, and introducing a time penalty term to optimize the merging strategy to reduce the false abnormal alarm rate; thereby improving the effect of elevator status detection.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and specifically refers to an intelligent elevator state detection method. Background Art

[0002] The elevator operation detection method is a kind of technical means that collects and analyzes various state data of the elevator during operation, monitors the operation health status of the elevator in real time, and conducts fault early warning or state evaluation according to the analysis results. However, the general elevator state detection method has problems in comprehensively considering the calculation cost and recognition accuracy when processing the elevator operation state characteristics, which is not conducive to the real-time early warning of elevator faults and is prone to missing slight abnormalities and early fault signs; the general elevator state detection method has problems that the accidental impact signals of elevator collision and emergency stop cause large deviations in the global parameter estimation and misjudgment of the instantaneous noise of the elevator state. Summary of the Invention

[0003] In view of the above situation, to overcome the defects of the prior art, the present invention provides an intelligent elevator state detection method. Aiming at the problem that the general elevator state detection method has problems in comprehensively considering the calculation cost and recognition accuracy when processing the elevator operation state characteristics, which is not conducive to the real-time early warning of elevator faults and is prone to missing slight abnormalities and early fault signs, this solution evaluates the benefits and costs through change indicators, so as to achieve the optimal compromise between the accuracy and speed of fault feature extraction; and introduces the time-delay attention weight to dynamically balance the processing quality and response speed of elevator state characteristics. When the system load is high or the response requirement is urgent, it automatically tends to reduce the calculation amount and give priority to ensuring the timeliness of fault detection; thereby improving the accuracy of elevator state detection; aiming at the problem that the general elevator state detection method has problems that the accidental impact signals of elevator collision and emergency stop cause large deviations in the global parameter estimation and misjudgment of the instantaneous noise of the elevator state, this solution introduces a piecewise chaotic mapping and a non-linear search strategy to accelerate the optimization convergence speed and shorten the overall processing time of elevator state detection; combines the representative vector within the cluster and time continuity to perform secondary clustering optimization on the historical elevator state set, introduces a time penalty term to optimize the merging strategy, reduces the state jump frequency, and reduces the false alarm rate of false abnormalities; thereby improving the effect of elevator state detection.

[0004] The technical solution adopted by the present invention is as follows: An intelligent elevator state detection method provided by the present invention, the method includes the following steps:

[0005] Step S1: Obtain elevator state data;

[0006] Step S2: Determine change indicators;

[0007] Step S3: Perform primary clustering of elevator states;

[0008] Step S4: Optimize the primary clustering of elevator states;

[0009] Step S5: Secondary clustering refinement of elevator status;

[0010] Step S6: Elevator status detection.

[0011] Furthermore, in Step S1, the acquisition of elevator status data is to record the sensor signals during the operation of the elevator and segment them at fixed time intervals and additionally construct a half-length window and a double-length window to obtain a window set S; a window is a data segment obtained by segmenting continuously collected sensor signals by time segments; calculate the feature vector X for all windows; obtain the historical elevator status set through data preprocessing ; is the set of all feature vectors; and label the elevator operation status type for the historical elevator status set, and use the elevator operation status type as a data label, which is only used for cluster label selection.

[0012] Furthermore, in Step S2, the determination of the change index is to perform K-means clustering on the historical elevator status set with the number of clusters c increasing from 1 to ; calculate the variance percentage sum and after each increase in the number of clusters; the variance percentage is expressed as: ; where j is the cluster index; is the L2 norm; is the sample set of the j-th cluster; is the sample, i is the sample index; is the center of the j-th cluster; is the mean vector of the historical elevator status set; the sample is the data of the historical elevator status set; calculate the ratio of the number of adjacent clusters, and the formula used is: ; ; where, and are the variance percentage ratio and the clustering running time ratio respectively when the number of clusters changes from n to n + 1; and are the variance percentages of the number of clusters n and n + 1 respectively; the value range of n is from 1 to ; and are the clustering running times of the number of clusters n and n + 1 respectively; introduce the time-delay attention weight, and finally calculate the change index , and the formula used is: ; ; where, is the time-delay attention weight; is the maximum index weight; is the current clustering running time; is the maximum clustering running time.

[0013] Further, in step S3, when the change index converges in the initial elevator state clustering, the cluster addition ends, and the at this time is the optimal number of clusters of the historical elevator state set; the clustering result when the K-means clustering of the historical elevator state set converges based on the optimal number of clusters is used as the initial elevator state clustering, and step S5 is executed; if it reaches and the change index does not converge at this time, then step S4 is executed; if , then the change index converges, is the convergence determination threshold.

[0014] Further, in step S4, the optimization of the initial elevator state clustering specifically includes the following steps:

[0015] Step S41: Initialize the optimized individual position: Establish a parameter search space based on the maximum number of clusters, the maximum index weight, and the convergence determination threshold, with a total of three dimensions; define the obtained by running the K-means clustering on the historical elevator state set based on the parameters set according to the individual position as the individual fitness value; initialize the optimized individual position; introduce a piecewise linear chaotic map, and the piecewise linear chaotic map is expressed as: ; where x(·) is the value of the chaotic sequence; b is the number of chaotic map times; p is the piecewise threshold; The initial value of is a random value between 0 and 1;

[0016] Step S42: Sort the fitness values: Sort the fitness values from smallest to largest, and mark the top three individuals as , and in turn, and use g to traverse; use o to traverse the remaining individuals; adopt a non-linear decreasing factor a, and the formula used is: ; ; where m is the current optimization number; M is the maximum optimization number; is the random coefficient vector of the optimized individual g; is a random vector; is the Hadamard product;

[0017] Step S43: Update the individual position: The formula used is: ; ; and update the top three individuals of the fitness value based on the change index, which is expressed as: ; where is the candidate position of the optimized individual g for the optimized individual o; and are the positions of the optimized individual g and the individual o respectively; is the updated position of individual o; and are defined the same and are both random vectors, but they are independent of each other; , and are the optimized individuals , and for the candidate positions of the optimized individual o; is the updated position of the optimized individual g;

[0018] Step S44: Optimization iteration: Set the index threshold. When there is an individual fitness value lower than the index threshold, the optimization iteration ends. Based on the individual positions, re-determine the values of the maximum number of clusters, the maximum index weight, and the convergence determination threshold, and re-run the K-means clustering on the historical elevator state set. Judge the optimal number of clusters according to the change index to obtain the initial clustering result of the elevator state, and execute Step S5; if the maximum optimization number is reached, return to execute the initialization of the optimized individual positions; otherwise, return to execute the fitness sorting.

[0019] Further, in Step S5, the refinement of the secondary clustering of the elevator state takes the clusters of the primary clustering of the elevator state as the initial partition, arranges them in chronological order to obtain a cluster sequence. The chronological order refers to the timestamps of the samples in each cluster, and the earliest sample timestamp in the cluster is selected to represent the appearance order of the clusters; calculate the mean vector and the central sample for each cluster; evaluate the cluster quality and , and the used formula is: ; ; where Y is the feature vector of any sample in cluster C; select the representative vector , and the used formula is: ; furthermore, add a time continuity penalty term to define the optimization distance , and the used formula is: ; where I and J are two adjacent clusters; and are the representative vectors of clusters I and J respectively; is the time penalty coefficient; is the timestamp assigned to the last sample in cluster I; is the timestamp assigned to the first sample in cluster J; T is the timestamp of the total data assignment; Iteratively merge the adjacent clusters with the smallest optimization distance until the merging reaches the predetermined number of clusters to obtain the final clustering result of the historical elevator state set.

[0020] Further, in step S6, the elevator state detection is based on the final clustering result of the historical elevator state set. The label of the clustering center in each cluster is selected as the cluster label. The elevator state data is collected in real time and assigned to the corresponding cluster based on the Euclidean distance. The cluster label of the cluster where the real-time collected elevator state data is located is used as the elevator state detection result.

[0021] The beneficial effects achieved by the present invention using the above solution are as follows:

[0022] (1) Aiming at the problem that the general elevator state detection method has a comprehensive consideration of the calculation overhead and recognition accuracy when processing the elevator operation state characteristics, which is not conducive to the real-time early warning of elevator faults and is prone to missing slight anomalies and early fault signs. This solution evaluates the benefits and costs through change indicators, thereby achieving the optimal compromise between the accuracy and speed of fault feature extraction; and introduces the time-delay attention weight to dynamically balance the processing quality and response speed of elevator state characteristics. When the system load is high or the response requirement is urgent, it automatically tends to reduce the calculation amount and prioritize ensuring the timeliness of fault detection; thereby improving the accuracy of elevator state detection.

[0023] (2) Aiming at the problem that the general elevator state detection method has a large deviation in the global parameter estimation caused by the occasional impact signals of elevator collision and emergency stop, and misjudges the instantaneous noise of the elevator state. This solution introduces a piecewise chaotic mapping and a non-linear search strategy to accelerate the optimization convergence speed and shorten the overall processing time of elevator state detection; combines the representative vector within the cluster and time continuity to perform secondary clustering optimization on the historical elevator state set, introduces a time penalty term to optimize the merging strategy, reduces the state jump frequency, and reduces the false anomaly alarm rate; thereby improving the effect of elevator state detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic flowchart of an intelligent elevator state detection method provided by the present invention;

[0025] Figure 2 is a schematic flowchart of step S4.

[0026] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0028] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc. indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the system or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0029] Example 1, refer to Figure 1 , an intelligent elevator state detection method provided by the present invention, the method includes the following steps:

[0030] Step S1: Elevator state data acquisition: Collect elevator state data during the operation of the elevator and construct a historical elevator state set;

[0031] Step S2: Determine the change index: Perform K-means clustering on the historical elevator state set and define the change index of the clustering of the historical elevator state set;

[0032] Step S3: Initial clustering of elevator states: Determine whether the change index converges according to the change index. If the change index converges, determine the optimal number of clusters and complete the initial clustering, and execute Step S5. If the change index does not converge, execute Step S4;

[0033] Step S4: Optimization of the initial clustering of elevator states: Introduce a chaotic mapping and an individual position update mechanism to optimize the initial clustering to obtain the result of the initial clustering of elevator states, and execute Step S5;

[0034] Step S5: Refinement of the secondary clustering of elevator states: Based on the initial clustering, combine the representative vector within the cluster and the time continuity penalty to perform secondary refinement clustering of elevator states;

[0035] Step S6: Elevator state detection: Based on the final clustering result obtained from the secondary refinement clustering of elevator states, realize the online detection of the elevator operation state.

[0036] Example 2, refer to Figure 1 , this embodiment is based on the above embodiment. In Step S1, the acquisition of elevator state data is to record the sensor signals during the operation of the elevator and segment them at fixed time intervals to additionally construct a half-length window and a double-length window , the half-length window captures the short-term events of door opening and closing; the double-length window reflects the overall trend of the entire operation, and the window set S is obtained; the window is the data segment obtained by dividing the continuously collected sensor signal into time segments; the feature vector X is calculated for all windows, and each feature vector includes one or more dimensions of acceleration mean, variance, speed mean and maximum, motor current peak, door opening and closing times, acceleration mutation rate and current root mean square value; the historical elevator state set is obtained through data preprocessing ; is the set of all feature vectors; and the historical elevator state set is labeled with the elevator operation state type, and the elevator operation state type is used as a data label, which is only used for cluster label selection; the acceleration mutation rate J is expressed as: ; Current RMS value It is expressed as: ; Where t is the time index in the window; a(·) is the acceleration value at the corresponding moment; is the motor current; the elevator operation status types include normal operation, slight abnormality, moderate abnormality and severe abnormality; in order to avoid misjudgment in the actual labeling process, or due to the complexity of the elevator operation environment, some states are difficult to accurately define. The same sample vector will have different operation states in different environments. For example, the same vector in the descending state, ascending state, and rapid change state will have different operation states, but a single label cannot be used for judgment. Therefore, based on semi-supervised learning, clustering is performed through the characteristics and internal structure of the data itself, and hidden patterns and features in the data are mined to have a more comprehensive understanding of the elevator's operation status; and parameters are adjusted through a unique optimization method to ensure accuracy while improving processing time.

[0037] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, the change index is determined to be a historical elevator state set. Run K-means clustering, and increase the number of clusters c from 1 to ; Calculate the variance percentage and ; Percentage of variance It is expressed as: ; where j is the cluster cable; is the sample set of the jth cluster; is the L2 norm; is a sample, i is the sample index; is the center of the jth cluster; is the mean vector of the historical elevator state set; sample is the data of the historical elevator state set; the ratio of the number of adjacent clusters is calculated using the formula: ; ;in, and They are the variance percentage ratio and the clustering running time ratio when the number of clusters changes from n to n + 1; the value range of n is from 1 to ; and are the variance percentages of the number of clusters n and the number of clusters n + 1 respectively; and are the clustering running times of the number of clusters n and the number of clusters n + 1 respectively; to balance the occupation of computing resources and the response speed, the time delay attention weight is introduced, and finally the change index is calculated. The formula used is: ; ; where, is the time delay attention weight; is the maximum index weight; is the current clustering running time; is the maximum clustering running time; by comparing the clustering quality and the computing cost of the historical elevator state set, the best compromise point is found to ensure that it is neither over-segmented nor the elevator state is missed due to too few clusters.

[0038] Example 4, refer to Figure 1 In this example, based on the above example, in step S3, the initial clustering of the elevator state is when the change index converges and the addition of clusters ends. At this time, the is the optimal number of clusters of the historical elevator state set; the clustering result when the K-means clustering of the historical elevator state set converges based on the optimal number of clusters is used as the initial clustering of the elevator state, and step S5 is executed; if the change index does not converge when reaching , then step S4 is executed; if , then the change index converges, and is the convergence determination threshold.

[0039] By performing the above operations, for the problem that the general elevator state detection method has a comprehensive consideration of the computing overhead and the recognition accuracy when processing the elevator operation state characteristics, which is not conducive to the real-time early warning of elevator faults and is prone to missing slight anomalies and early fault signs, this solution evaluates the benefits and costs through the change index, so as to achieve the optimal compromise between the accuracy and speed of fault feature extraction; and introduces the time delay attention weight to dynamically balance the processing quality of elevator state characteristics and the response speed. When the system load is high or the response demand is urgent, it automatically tends to reduce the amount of calculation and give priority to ensuring the timeliness of fault detection; thereby improving the accuracy of elevator state detection.

[0040] Example 5, refer to Figure 1 and Figure 2 In this example, based on the above example, in step S4, the optimization of the initial clustering of the elevator state specifically includes the following steps:

[0041] Step S41: Initialize the optimized individual positions: Establish a parameter search space based on the maximum number of clusters, the maximum index weight, and the convergence determination threshold, with a total of three dimensions; Define the parameter set based on the individual positions to perform K-means clustering on the historical elevator state set for as the individual fitness value; Initialize the optimized individual positions; Introduce a piecewise linear chaotic map to accelerate the initial clustering optimization speed of elevator states; The formula used is: ; where x(·) is the numerical value of the chaotic sequence; b is the number of chaotic map iterations; p is the piecewise threshold; The initial value of is a random value between 0 and 1; Initialize the optimized individual positions and represent them as: ; where, is the initial position of the k-th dimension of the u-th optimized individual; and are the upper and lower limits of the k-th dimension respectively;

[0042] Step S42: Sort the fitness values: Sort the fitness values from smallest to largest, and mark the top three individuals as , and respectively, and use g to traverse; Use o to traverse the remaining individuals; Adopt a non-linear decreasing factor a, and the formula used is: ; ; where m is the current optimization iteration; M is the maximum optimization iteration; is the random coefficient vector of the optimized individual g; is a random vector, and each component is independently and uniformly sampled from [0,1]; is the Hadamard product;

[0043] Step S43: Update the individual positions: The formula used is: ; ; And update the top three individuals with the fitness values based on the change index, which is expressed as: ; where, is the candidate position of the optimized individual g for the optimized individual o; and are the positions of the optimized individual g and the individual o respectively; is the position of the individual o after update; and are defined in the same way, both are random vectors, but independent of each other; , and are the candidate positions of the optimized individuals , and for the optimized individual o respectively; is the position of the optimized individual g after update;

[0044] Step S44: Optimization iteration: Set the index threshold. When there is an individual fitness value lower than the index threshold, the optimization iteration ends. Based on the individual positions, re-determine the values of the maximum number of clusters, the maximum index weight, and the convergence determination threshold, and re-run the K-means clustering on the historical elevator state set. Determine the optimal number of clusters according to the change index to obtain the initial clustering result of the elevator state, and execute Step S5. If the maximum number of optimizations is reached, return to execute the initialization of the optimized individual positions. Otherwise, return to execute the fitness sorting. During the elevator operation, sudden impacts or signal jitters are inevitable. This optimization strategy can avoid the overall deviation caused by single abnormal samples and is more stable and reliable.

[0045] Embodiment Six. Refer to Figure 1 , this embodiment is based on the above embodiment. In Step S5, the refinement of the secondary clustering of the elevator state takes the clusters obtained from the initial clustering of the elevator state as the initial partition, arranges them in chronological order to obtain a cluster sequence. The chronological order refers to the timestamps of the samples in each cluster, and the earliest sample timestamp in the cluster is selected to represent the order of appearance of the clusters. Calculate the mean vector and the central sample for each cluster; Evaluate the cluster quality and , and the formula used is: ; ; where Y is the feature vector of any sample in cluster C; Select the representative vector , and the formula used is: ; Furthermore, add a time continuity penalty term to define the optimized distance , and the formula used is: ; where I and J are two adjacent clusters; and are the representative vectors of clusters I and J respectively; is the time penalty coefficient; is the timestamp assigned to the last sample in cluster I; is the timestamp assigned to the first sample in cluster J; T is the timestamp of the total data assignment; Iteratively merge the adjacent clusters with the smallest optimized distance until the number of clusters is 4 to obtain the final clustering result of the historical elevator state set. For the situation in the initial clustering where the door opening and closing may be over-segmented in a short period or the same motion state is over-segmented, through merging, ensure that each elevator state is continuously and completely expressed, reduce the state jump frequency, and reduce the misjudgment of instantaneous false abnormal states by maintenance personnel.

[0046] By performing the above operations, in view of the problems existing in the general elevator status detection method, such as the accidental impact signals of elevator collision and emergency stop causing large offsets in the estimation of global parameters and misjudging the instantaneous noise of the elevator status, this solution introduces a piecewise chaotic mapping and a non-linear search strategy to accelerate the optimization convergence speed and shorten the overall processing time of elevator status detection; combines the representative vector within the cluster and time continuity to perform secondary clustering optimization on the historical elevator status set, introduces a time penalty term to optimize the merging strategy, reduces the state jump frequency, and lowers the false alarm rate of abnormal alarms; thereby improving the elevator status detection effect.

[0047] Embodiment 7. Refer to Figure 1 , this embodiment is based on the above embodiment. In step S6, the elevator status detection is based on the final clustering result of the historical elevator status set. The label of the clustering center in each cluster is selected as the cluster label; the elevator status data is collected in real time and assigned to the corresponding cluster based on the Euclidean distance. The cluster label of the cluster where the real-time collected elevator status data is located is used as the elevator status detection result; at this time, the elevator status data for which the detection is completed is incorporated into the historical elevator status set, and the clustering result is updated at a fixed period.

[0048] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0049] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention.

[0050] The above describes the present invention and its implementation manners. This description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. Generally speaking, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments without creative efforts without departing from the purpose of the present invention, they should all fall within the protection scope of the present invention.

Claims

1. An intelligent elevator status detection method, characterized in that: The method includes the following steps: Step S1: Obtaining elevator status data: Collecting elevator status data during the operation of the elevator and constructing a historical elevator status set; Step S2: Determining change indicators: Performing K-means clustering on the historical elevator status set and defining change indicators for the clustering of the historical elevator status set; Step S3: Initial clustering of elevator status: Judging whether the change indicator converges according to the change indicator. If the change indicator converges, determining the optimal number of clusters and completing the initial clustering, and executing Step S5. If the change indicator does not converge, then execute Step S4; Step S4: Optimization of the initial clustering of elevator status: Introducing a chaotic mapping and an individual position update mechanism to optimize the initial clustering, obtaining the result of the initial clustering of elevator status, and executing Step S5; Step S5: Refinement of the secondary clustering of elevator status: On the basis of the initial clustering, combining the representative vector within the cluster and the time continuity penalty to perform secondary refinement clustering of elevator status; Step S6: Detection of elevator status: Based on the final clustering result obtained from the secondary refinement clustering of elevator status, realizing online detection of the elevator operation status; In step S2, the determination of the change index is performed on the historical elevator state set to run K-means clustering, with the number of clusters c incrementing from 1 to ; after each increment of the number of clusters, calculate the variance percentage and the clustering running time ; the variance percentage is expressed as: ; where j is the cluster index; is the sample set of the j-th cluster; is the L2 norm; is the sample, i is the sample index; is the center of the j-th cluster; is the mean vector of the historical elevator state set; the sample is the data of the historical elevator state set; calculate the ratio of the number of clusters between two adjacent clusters, and the formula used is: ; ; where and are the variance percentage ratio and the clustering running time ratio respectively when the number of clusters changes from n to n + 1; the value range of n is from 1 to ; and are the variance percentages of the number of clusters n and n + 1 respectively; and are the clustering running times of the number of clusters n and n + 1 respectively; introduce the time-delay attention weight, and finally calculate the change index , and the formula used is: ; ; where is the time-delay attention weight; is the maximum index weight; is the current clustering execution time; is the maximum clustering execution time.

2. The intelligent elevator status detection method according to claim 1, wherein: In step S3, when the change index converges in the initial clustering of elevator states, the addition of clusters ends, and the at this time is the optimal number of clusters for the historical elevator state set; the clustering result when performing K-means clustering convergence on the historical elevator state set based on the optimal number of clusters is used as the initial clustering of elevator states, and step S5 is executed; If it reaches and the change index does not converge, then step S4 is executed; if , then the change index converges, where is the convergence determination threshold value.

3. The intelligent elevator status detection method according to claim 2, characterized in that: In Step S4, the optimization of the initial clustering of elevator status specifically includes the following steps: Step S41: Initialize the position of the optimized individual: Establish a parameter search space based on the maximum number of clusters, the maximum index weight, and the convergence determination threshold, with a total of three dimensions; Define the obtained by performing K-means clustering on the historical elevator state set using the parameters set based on the individual position as the individual fitness value; Initialize the position of the optimized individual; Introduce a piecewise linear chaotic map, and the piecewise linear chaotic map is expressed as: ; where x(·) is the value of the chaotic sequence; b is the number of chaotic map iterations; p is the piecewise threshold; The initial value of is a random value between 0 and 1; The initialization of the position of the optimized individual is expressed as: ; where, is the initial position of the k-th dimension of the u-th optimized individual; and are the upper and lower limits of the k-th dimension, respectively; Step S42: Fitness value sorting: Sort in ascending order of fitness value, and mark the top three individuals as , and respectively, and traverse them using g; traverse the remaining individuals using o; adopt a non-linear decreasing factor a, and the formula used is: ; ; where m is the current optimization number; M is the maximum optimization number; is the random coefficient vector for optimizing individual g; is the random vector; is the Hadamard product; Step S43: Update the position of the individual: The formula used is: ; ; And update the top three individuals with the fitness value based on the change index, expressed as: ; Where, is the candidate position of the optimized individual g for the optimized individual o; and are the positions of the optimized individual g and the individual o respectively; is the position of the individual o after update; and are defined the same, both are random vectors, but independent of each other; 、 and are the candidate positions of the optimized individuals 、 and for the optimized individual o; is the position of the optimized individual g after update; Step S44: Optimization iteration: Setting an index threshold. When there is an individual fitness value lower than the index threshold, the optimization iteration ends. Based on the individual positions, re-determine the values of the maximum number of clusters, the maximum index weight, and the convergence determination threshold, and re-run K-means clustering on the historical elevator status set. Judging the optimal number of clusters according to the change indicator to obtain the result of the initial clustering of elevator status, and executing Step S5; If the maximum number of optimization times is reached, then return to execute the initialization of optimizing the individual positions; Otherwise, return to execute the fitness sorting.

4. An intelligent elevator status detection method according to claim 3, characterized in that: In step S5, the secondary clustering refinement of elevator states takes the clusters obtained from the primary clustering of elevator states as the initial partition, arranges them in chronological order to obtain a cluster sequence, where the chronological order refers to the timestamps of the samples in each cluster, and the earliest sample timestamp in the cluster is selected to represent the order of appearance of the clusters; calculate the mean vector for each cluster and the central sample ; evaluate the cluster quality and , and the formula used is: ; ; where Y is the feature vector of any sample in cluster C; select the representative vector , and the formula used is: ; then add a time continuity penalty term to define the optimized distance , and the formula used is: ; where I and J are two adjacent clusters; and are the representative vectors of cluster I and cluster J respectively; is the time penalty coefficient; is the timestamp assigned to the last sample in cluster I; is the timestamp assigned to the first sample in cluster J; T is the timestamp of the total data assignment; iteratively merge the adjacent clusters with the smallest optimized distance until merged to the predetermined number of clusters to obtain the final clustering result of the historical elevator state set.

5. The intelligent elevator status detection method according to claim 4, characterized in that: In step S1, the elevator state data acquisition is to record the sensor signals during the elevator operation at fixed time intervals segmented, and additionally construct a half-length window and a double-length window , to obtain a window set S; a window is a data segment obtained by segmenting the continuously collected sensor signals by time segments; calculate the feature vector X for all windows; obtain the historical elevator state set through data preprocessing ; is the set of all feature vectors X; and label the historical elevator state set with the elevator operation state type, and use the elevator operation state type as the data label, which is only used for cluster label selection.

6. The intelligent elevator status detection method according to claim 5, wherein: In Step S6, the detection of elevator status is based on the final clustering result of the historical elevator status set, and the label of the clustering center in each cluster is selected as the cluster label; Real-time collecting elevator status data, allocating it to the corresponding cluster based on the Euclidean distance, and taking the cluster label of the cluster where the real-time collected elevator status data is located as the elevator status detection result.

Images