State detection model training method and device and state detection method and device

Abstract

The invention discloses a state detection model training method and device and a state detection method and device. The method comprises the steps that all parts in an elevator are inquired; the part is configured with a triggering mode; in the running process of the elevator, running signals related to the components are collected according to the triggering mode; marking the running states of the parts of the elevator according to the running signals; extracting operation characteristics representing part characteristics of the elevator from the operation signals; and training a state detection model for parts of the elevator according to the operation characteristics and the operation state. According to the embodiment of the invention, the elevator parts are automatically monitored, the operation signals in the elevator are captured in time, the fault positioning accuracy is improved, the state detection model is trained according to the operation signals in various historical operation states so as to realize automatic auxiliary troubleshooting, the dependence on the personal experience of maintenance personnel is reduced during troubleshooting, the troubleshooting time is shortened, and the fault positioning efficiency is improved. Therefore, the troubleshooting efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of elevators, and in particular, to a training of a state detection model, a state detection method, and a device therefor. Background Art

[0002] There are many components in an elevator. In the case of long-term operation, the components of the elevator are subject to wear and are prone to failure.

[0003] Currently, the fault diagnosis of an elevator mainly relies on on-site maintenance personnel to export the fault signals recorded when a fault occurs from the elevator, and troubleshoot based on the fault signals to determine the cause of the fault.

[0004] However, it is difficult to reproduce the fault signals in the elevator, and it is easy to deviate when locating the cause of the fault. Moreover, the troubleshooting process mainly relies on the personal experience of the maintenance personnel, and there are large differences in the experience levels of the maintenance personnel. Some maintenance personnel lack experience accumulation, and the troubleshooting takes a long time, resulting in low efficiency of troubleshooting. Summary of the Invention

[0005] In view of this, the present invention provides a training of a state detection model, a state detection method, and a device therefor to improve the efficiency of elevator troubleshooting.

[0006] The first aspect of the present invention provides a method for training a state detection model, including:

[0007] Query each component in the elevator; the component is configured with a triggering method;

[0008] During the operation of the elevator, collect operation signals related to the component according to the triggering method;

[0009] Label the operation state of the component of the elevator for the operation signal;

[0010] Extract operation features characterizing the component characteristics of the elevator from the operation signals;

[0011] Train a state detection model for the component of the elevator according to the operation features and the operation state.

[0012] The second aspect of the present invention provides a state detection method, including:

[0013] Query each component in the elevator; the component is configured with a triggering method;

[0014] During the operation of the elevator, collect operation signals related to the component according to the triggering method;

[0015] Extract operation features characterizing the component characteristics of the elevator from the operation signals;

[0016] Load the state detection model trained according to the method described in the first aspect;

[0017] Input the operation characteristics into the state detection model to detect the operation state of the components of the elevator.

[0018] The third aspect of the present invention provides a training device for a state detection model, including:

[0019] A component query module for querying each component in the elevator; the component has a configured triggering method;

[0020] An operation signal acquisition module for acquiring operation signals related to the component according to the triggering method during the operation of the elevator;

[0021] An operation state annotation module for annotating the operation state of the components of the elevator for the operation signals;

[0022] An operation feature extraction module for extracting operation features characterizing the component characteristics of the elevator from the operation signals;

[0023] A state detection model training module for training a state detection model for the components of the elevator according to the operation features and the operation state.

[0024] The fourth aspect of the present invention provides a state detection device, including:

[0025] A component query module for querying each component in the elevator; the component has a configured triggering method;

[0026] An operation signal acquisition module for acquiring operation signals related to the component according to the triggering method during the operation of the elevator;

[0027] An operation feature extraction module for extracting operation features characterizing the component characteristics of the elevator from the operation signals;

[0028] A state detection model loading module for loading the state detection model trained according to the method described in Embodiment 1;

[0029] An operation state detection module for inputting the operation features into the state detection model to detect the operation state of the components of the elevator.

[0030] The fifth aspect of the present invention provides an electronic device, the electronic device includes:

[0031] At least one processor; and

[0032] A memory communicatively connected to the at least one processor; wherein,

[0033] The memory stores a computer program executable by the at least one processor. When the computer program is executed by the at least one processor, the at least one processor is enabled to execute the training method of the state detection model as described in the first aspect above or the state detection method as described in the second aspect above.

[0034] A sixth aspect of the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the training method of the state detection model as described in the first aspect above or the state detection method as described in the second aspect above.

[0035] A seventh aspect of the present invention provides a computer program product comprising a computer program which, when executed by a processor, implements the training method of the state detection model as described in the first aspect above or the state detection method as described in the second aspect above.

[0036] In this embodiment, each component in the elevator is queried; the component has a configured triggering mode; during the operation of the elevator, operation signals related to the component are collected according to the triggering mode; the operation states of the components of the elevator are labeled for the operation signals; operation features characterizing the component characteristics of the elevator are extracted from the operation signals; and a state detection model for the components of the elevator is trained based on the operation features and the operation states. This embodiment performs automated monitoring of elevator components, timely captures operation signals in the elevator, improves the accuracy when locating faults, trains a state detection model based on operation signals in various historical operation states, so as to achieve automatic auxiliary troubleshooting, reduce the dependence on the personal experience of maintenance personnel during troubleshooting, reduce the time spent on troubleshooting, and thus improve the efficiency of troubleshooting.

[0037] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0039] Figure 1 It is a flowchart of a training method for a state detection model provided in Embodiment 1 of the present invention.

[0040] Figure 2 It is a schematic diagram of triggering acquisition provided in the first embodiment of the present invention.

[0041] Figure 3 It is an example diagram of a triggering condition provided in the first embodiment of the present invention.

[0042] Figure 4 It is an example diagram of acquisition parameters provided in the first embodiment of the present invention.

[0043] Figure 5 It is an example diagram of pulse characteristics provided in the first embodiment of the present invention.

[0044] Figure 6 It is an example diagram of relative change characteristics provided in the first embodiment of the present invention.

[0045] Figure 7 It is an example diagram of clustering provided in the first embodiment of the present invention.

[0046] Figure 8 It is a flowchart of a state detection method provided in the second embodiment of the present invention.

[0047] Figure 9 It is a schematic structural diagram of a training device for a state detection model provided in the third embodiment of the present invention.

[0048] Figure 10 It is a schematic structural diagram of a state detection device provided in the fourth embodiment of the present invention.

[0049] Figure 11 It is a schematic structural diagram of an electronic device provided in the fifth embodiment of the present invention. Detailed implementation manners

[0050] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying 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.

[0051] It should be noted that the terms "first", "second", etc. in the description, claims and the above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here can cover implementations in sequences other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0052] Embodiment 1

[0053] See Figure 1 , which shows a flowchart of a method for training a state detection model provided in Embodiment 1 of the present invention. This method can be executed by a training device for the state detection model. The training of the state detection model and its state detection device can be implemented in the form of hardware and / or software, and the training device for the state detection model can be configured in an electronic device. As Figure 1 shown, the method includes:

[0054] Step 101, query each component in the elevator.

[0055] In different types of buildings, there are different transportation requirements for people, pets, goods, etc. Therefore, different types of elevators can be deployed in the building according to different transportation requirements. For example, passenger elevators, freight elevators, sightseeing elevators, and so on.

[0056] Generally, an elevator is a comprehensive system, and different types of elevators include different components.

[0057] In one example, the components configured for a certain type of elevator include: an elevator control system, call buttons distributed on each floor, a car (including car doors), a motor for towing the car (also known as a traction machine), a control cabinet, a speed limiter, a door operator, a car frame, car doors, counterweight rails, car rails, rail brackets, trailing cables, counterweight devices, compensating chains (ropes), landing doors, guiding devices for compensating chains (ropes), buffers, and so on.

[0058] In some types of elevators, components such as the traction machine, control cabinet, speed limiter, and trailing cables can be omitted.

[0059] These components can be divided into different sets according to their functions, thus forming each subsystem that supports the operation of the elevator. The elevator control system is connected to multiple systems of the elevator through a serial port or a serial clock line (SCL) in a wired manner. The elevator control system monitors each system and controls the operation of each subsystem, enabling the car to move in the hoistway and reach each floor of the building.

[0060] In one example, the controller includes a door system, a frequency conversion system, a call system, and a traction system. Among them, the door system is used to control the elevator doors. The elevator doors include the car door and the landing doors of each floor. The type of the car door is the same as that of the landing door and they open and close simultaneously. The frequency conversion system is used to control the frequency converter. The call system is used to control the logic of in-car calls (calls inside the car) and out-of-car calls (calls at the landing). The traction system is used to control the vertical movement (vertically upward or vertically downward) of the car in the hoistway.

[0061] The elevator manager or owner can choose whether to install an edge computing node on the elevator according to factors such as the load status of the elevator. If the choice is not to install an edge computing node, the controller maintains the original control logic and does not affect the normal operation of the elevator. If the choice is to install an edge computing node, a suitable computing device can be selected as the edge computing node of the elevator according to the requirements. The edge computing node and the original elevator control system are combined into a new elevator control system to redefine the operation logic of the elevator.

[0062] Generally, the edge computing node is a computing device with relatively strong computing power, such as a computer, a server, or an embedded device, etc. In addition, according to different intelligent services, a graphics processing unit (GPU) or an embedded neural network processor (NPU) can be equipped in the edge computing node.

[0063] The edge computing node refers to building a new service platform at the network edge close to the elevator, providing resources such as storage, computing, and network, and sinking some key business applications to the access network edge to reduce the bandwidth and latency losses caused by network transmission and multi-level forwarding. The position of the edge computing node is between the user and the cloud (server). Compared with the traditional cloud, it is closer to the user (data source), and has the characteristics of miniaturization, distribution, and being closer to the user. A large amount of data (such as audio data) does not need to be uploaded to the cloud for processing, and the data is processed at the network edge side, reducing the request response time, reducing the network bandwidth, and ensuring the security and privacy of the data.

[0064] In addition, the edge computing node can implement algorithm functions and model inference, communicate with the original elevator control system, and provide artificial intelligence (AI) and complex computing capabilities for the original elevator control system; the edge computing node can also communicate with the cloud, implement algorithm functions and model updates, and transfer functions such as function calls of the original elevator control system.

[0065] Step 102: During the operation of the elevator, collect operation signals related to components according to the triggering method.

[0066] During the operation of the elevator, each component generates one or more operation signals.

[0067] Some operation signals are pulse signals, and their values include low level (0) and high level (1). For example, a certain operation signal represents the opening and closing state of the car door. The operation signal with low level indicates that the car door is in the closed state, and the operation signal with high level indicates that the car door is in the open state, and so on.

[0068] Some operation signals are numerical signals. For example, a certain operation signal represents the opening and closing speed of the car door, thereby representing the opening and closing stages of the car door. Among them, the opening and closing stages usually include opening acceleration section, opening constant speed section, opening deceleration section, opening hold, closing acceleration section, closing constant speed section, closing deceleration section, closing hold constant speed section, and so on.

[0069] In this embodiment, some or all components in the elevator can be safely monitored. Since the operating states of the components in the elevator are usually manifested in multiple aspects, when separately monitoring the safety of a certain component in the elevator, the operation signals related to the component of the elevator can be queried in the elevator control system.

[0070] For example, when the car door of the elevator has faults such as pulley slipping, abnormal motor speed, and light curtain output loop failure, when safely monitoring the car door of the elevator, the vibration signal of the belt, motor speed, light curtain signal and other operation signals related to the operation of the car door of the elevator can be queried in the elevator control system.

[0071] In practical applications, the faults that occur in elevator components may involve multiple aspects of reasons. At this time, the content of the operation signals involved is relatively large. If all the content of the operation signals of the components in the elevator is collected simultaneously for fault diagnosis, it will cause a great consumption of resources such as bandwidth and storage. Therefore, in this embodiment, a triggering mode is used to collect partial content of the operation signals of the components in the elevator.

[0072] Then, during the operation of the elevator, the controllers associated with the same component in the elevator have been configured with the same triggering method. The controllers are monitored according to the triggering method, so as to collect the operation signals generated by the controllers that may be related to faults.

[0073] In a specific implementation, the triggering method includes a triggering condition and acquisition parameters. Among them, the triggering condition is the condition for triggering the acquisition of the operation signals related to the components of the elevator, and the acquisition parameters are the parameters describing how to acquire the operation signals related to the components of the elevator.

[0074] For the same component in the elevator, multiple triggering conditions can be set for it. The various triggering conditions can be in a parallel relationship or a series relationship, and this embodiment does not limit this.

[0075] For example, as Figure 2 shown, for component 1 of the elevator, triggering condition 1 can be set. When the relevant signal (i.e., the operation signal) changes, if triggering condition 1 is satisfied, data acquisition is performed on the relevant operation signals (such as signal A1, signal B1, signal C1); for component 2 of the elevator, triggering condition 2 can be set. When the relevant signal (i.e., the operation signal) changes, if triggering condition 2 is satisfied, data acquisition is performed on the relevant operation signals (such as signal A2, signal B2, signal C2); and so on. For component n of the elevator, triggering condition n can be set. When the relevant signal (i.e., the operation signal) changes, if triggering condition n is satisfied, data acquisition is performed on the relevant operation signals (such as signal An, signal Bn, signal Cn).

[0076] Generally, the triggering condition is set for the signals generated during the operation of the elevator. The signals indicated by the triggering condition (i.e., the signals to be matched with it) can be the operation signals related to the components of the elevator or the operation signals unrelated to the components of the elevator (for example, after the car door is closed, the speed of the car (operation signal) is collected. At this time, the triggering condition can be set for the opening and closing state of the car door (operation signal), and the speed of the car has nothing to do with the opening and closing state of the car door), and this embodiment does not limit this.

[0077] Exemplarily, the triggering condition includes a comparison relationship with a threshold value. For example, equal to, not equal to, greater than, less than, and so on.

[0078] Furthermore, the triggering condition includes at least one of the following types:

[0079] Valid, change hold, invalid low level (i.e., invalid 0), invalid high level (i.e., invalid 1).

[0080] Valid means that it is at a low level (0) when the triggering condition is not satisfied and at a high level (1) when the triggering condition is satisfied.

[0081] The change hold means that it is at a low level (0) when the trigger condition is not met, and remains at a high level (1) after the trigger condition is first met.

[0082] The invalid low level means that the trigger condition is unavailable and remains at a low level.

[0083] The invalid high level means that the trigger condition is unavailable and remains at a high level.

[0084] If each trigger condition forms a trigger judgment loop, then the valid level refers to the level when the trigger judgment loop is a conducting path, the invalid low level refers to the low level when the trigger judgment loop is open, and the invalid high level refers to the high level when the trigger judgment loop is open.

[0085] For example, Figure 3 As shown, in a certain trigger condition, the set trigger conditions include: "Signal A is change hold", "Signal B is invalid 0", "Signal C is invalid 0", "Signal D is invalid 0", "Signal E is invalid 0", "Signal F is invalid 0".

[0086] Set "Signal A is change hold", "Signal B is invalid 0", "Signal C is invalid 0", "Signal D is invalid 0", "Signal E is invalid 0", "Signal F is invalid 0" as the trigger judgment loop, where "Signal A is change hold", "Signal B is invalid 0", "Signal C is invalid 0", "Signal D is invalid 0", "Signal E is invalid 0" are in parallel and are in series with "Signal F is invalid 0".

[0087] Exemplarily, as Figure 4 shown, the acquisition parameters include at least one of the following:

[0088] The capture period (i.e., how often to check whether to trigger the acquisition of the running signal), the capture time ratio before triggering, the capture time ratio after triggering, the total capture time, and the delay time after triggering.

[0089] Of course, the above trigger conditions and acquisition parameters are only examples. When implementing this embodiment, other trigger conditions and acquisition parameters can be set according to the actual situation, and this embodiment does not limit this. In addition, in addition to the above trigger conditions and acquisition parameters, those skilled in the art can also adopt other trigger conditions and acquisition parameters according to actual needs, and this embodiment does not limit this either.

[0090] Then, during the operation of the elevator, when the trigger condition is determined, query the signal indicated by the trigger condition and match the signal indicated by the trigger condition with the trigger condition.

[0091] If the matching is successful, the operation signals related to the components are collected according to the corresponding collection parameters, and while ensuring the accuracy of the safety monitoring of the elevator components, the amount of information of the operation signals is reduced, thereby greatly reducing the consumption of resources such as bandwidth and storage.

[0092] Step 103: Label the operation status of the elevator components for the operation signals.

[0093] In this embodiment, for the same component of the elevator, the operation status represented by the component of the elevator for the operation signal can be labeled as a tag.

[0094] Among them, the operation status of the elevator components includes the normal status, and one or more categories of fault statuses (also known as abnormal statuses), etc.

[0095] Step 104: Extract the operation features representing the component characteristics of the elevator from the operation signals.

[0096] In practical applications, deep learning technologies (such as LSTM (Long Short-Term Memory), GRU (Gated Recurrent Unit), etc.), machine learning technologies, etc. can be used to statistically analyze each operation signal, and extract the operation features representing the component characteristics of the elevator from it for analyzing its operation status.

[0097] In an embodiment of the present invention, step 104 may include the following steps:

[0098] Step 1041: If the type of the operation signal is a pulse signal, extract at least one of the waveform feature, time domain feature, and frequency domain feature from the operation signal as the operation feature representing the component characteristics of the elevator.

[0099] If the type of a single operation signal is a pulse signal, at least one of the waveform feature, time domain feature, and frequency domain feature can be extracted from the operation signal as the operation feature representing the component characteristics of the elevator.

[0100] Among them, as Figure 5 shown, the waveform feature includes the time from the rising edge to the falling edge, that is, the duration in the high-level state (abbreviated as high-level time), the time from the falling edge to the rising edge, that is, the duration in the low-level state (abbreviated as low-level time), etc.

[0101] Furthermore, considering that the operation signals are collected with the elevator as the data source, according to the operation logic of the elevator itself, the occurrence time features corresponding to different operation signals should be included in the calculation to improve the accuracy of safety monitoring.

[0102] On the one hand, according to the operating logic of the elevator, the relative time when the operating signal changes from low level to high level (such as the change time point in Figure 5 ) is set as the waveform feature and used as the operating feature characterizing the component characteristics of the elevator.

[0103] On the other hand, according to the operating logic of the elevator, the relative time when the operating signal changes from high level to low level is set as the waveform feature and used as the operating feature characterizing the component characteristics of the elevator.

[0104] Among them, the relative time refers to the time relative to a certain time stamp (such as the time stamp when the trigger condition is met).

[0105] For example, when the elevator is in the idle state (trigger condition), it is stipulated that devices such as the lights and screens in the car enter the sleep state after 2 seconds. If the devices such as the lights and screens in the car enter the sleep state after 3 seconds (relative time), there may be a fault.

[0106] In addition, the time domain features and frequency domain features include obtaining frequency components and energy magnitudes through Fourier transform, calculating power spectral density to analyze the distribution of power in the frequency domain, performing wavelet transform to process pulse signals with abruptness and singularity, and so on.

[0107] Step 1042: If the type of the operating signal is a numerical signal, extract at least one of the statistical feature, change feature, and threshold feature from the operating signal as the operating feature characterizing the component characteristics of the elevator.

[0108] If the type of a single operating signal is a numerical signal, at least one of the statistical feature, change feature, and threshold feature can be extracted from the operating signal in the time domain as the operating feature characterizing the component characteristics of the elevator.

[0109] Among them, the statistical features include mean, variance, skewness, kurtosis, and so on.

[0110] Since the operating signal is triggered for acquisition, the operating signal itself does not depend on time alignment. There are changes in the operating signal, and its change features include numerical change rate, maximum value, minimum value, and so on.

[0111] Generally, when the operating signal itself usually reaches a certain value, it has an impact on the elevator. Therefore, the relationship between the operating signal and the threshold can be used as the threshold feature, and the threshold features include the time when the set threshold is reached, the speed of reaching the set threshold, and so on.

[0112] Step 1043: According to the operating logic of the elevator, calculate the relative change feature between the operating signals as the operating feature characterizing the component characteristics of the elevator.

[0113] In practical applications, due to the relationship of the operation logic of the elevator, the relative change between operation signals is also one of the occurrence phenomena of faults. Therefore, the relative change characteristics between two operation signals can be calculated as the operation characteristics representing the component characteristics of the elevator.

[0114] Since there are many relative change characteristics between two operation signals in the elevator, Pearson correlation coefficient and other methods can usually be used to screen out the relative change characteristics with a relatively high correlation with the faults of elevator components.

[0115] In specific implementation, two operation signals with a relative relationship can be screened out according to the operation logic of the elevator as the first target signal and the second target signal.

[0116] Statistical the relative time when the first target signal generates the first change trend as the first target time point, and after the first target time point, statistical the relative time when the second target signal generates the second change trend as the second target time point.

[0117] Wherein, the second change trend is the same as or opposite to the first change trend.

[0118] When they are opposite, if the first change trend is from low level to high level, the second change trend is from high level to low level; if the first change trend is from high level to low level, the second change trend is from low level to high level.

[0119] When they are the same, if the first change trend is from low level to high level, the second change trend is from low level to high level; if the first change trend is from high level to low level, the second change trend is from high level to low level.

[0120] Set the difference between the second target time point and the first target time point as the relative change characteristic, which is used as the operation characteristic representing the component characteristics of the elevator.

[0121] For example, as Figure 6 shown, when the car stops at a certain floor at time point A (i.e., the first target time point, such as Figure 6 the first dotted line from left to right in Figure 6 ), at this time, the relevant signal A (operation signal) changes from low level to high level. When the car door opens at time point B (i.e., the second target time point, such as

[0122] the second dotted line from left to right in Figure 6 ), the relevant signal B (operation signal) changes from high level to low level. At this time, the difference between time point A and time point B (i.e., the signal time gap) represents the door opening time after the car stops at the floor. If the specified door opening time is 2 seconds and the actual measured door opening time is 3 seconds, there may be a fault.Step 105: Detect the training state of the elevator components based on the operating characteristics and operating status.

[0123] In practical applications, a state detection model can be constructed based on deep learning technology, machine learning technology, etc. Using the operating characteristics of elevator components as samples and the operating status of elevator components as labels, the state detection model is supervised trained so that the state detection model has the ability to detect the operating status of elevator components based on the operating characteristics of elevator components.

[0124] In an embodiment of the present invention, step 105 may include the following steps:

[0125] Step 1051: Use the operating characteristics as sample points, and cluster the sample points according to the operating status to obtain multiple clusters.

[0126] In practical applications, most of the operating information is data collected when the elevator components are in normal state, with small deviations. Therefore, if there are data with obvious deviations, they are usually data collected when the elevator components are in fault state.

[0127] Therefore, in this embodiment, as Figure 7 shown, regarding the synchronously collected operating characteristics as sample points in a multi-dimensional vector space, using clustering algorithms such as K-means (K-Means) and DBSCAN (Density-Based Spatial Clustering of Applications with Noise), the sample points are unsupervised clustered with the operating status as a reference to obtain multiple clusters, thereby distinguishing the operating signals in the normal state from the operating signals in various fault states (i.e., abnormal categories).

[0128] Taking K-means as an example, the total number of the normal state and one or more categories of fault states can be counted. Using the operating characteristics as sample points, K-means clustering is performed on the sample points to obtain clusters that meet the total number, so as to classify the deviation degree of the operating signals.

[0129] Further, the process of using K-means to cluster the sample points is as follows:

[0130] S511: Initialize K clusters; K represents the total number, and each cluster has a center point.

[0131] S512: Traverse each sample point, and calculate the distance between the operating characteristics of the sample point and the center points of each cluster, such as the Euclidean distance, etc.

[0132] S513: Divide the sample point into the cluster with the smallest distance.

[0133] S514. When the sample points are re-divided and completed, traverse each cluster, calculate the average value of the running characteristics of all sample points within the cluster, and assign the average value to the center point of the cluster, thereby updating the center point.

[0134] S515. When the update of the center points of the clusters is completed, calculate the clustering metrics for the K clusters, such as SSE (Sum of Squared Errors), CH index (Calinski-Harabasz Index), DB index (Davies-Bouldin Index), and so on.

[0135] S516. Determine whether the clustering metrics meet the clustering objective, such as the clustering metric is less than a certain threshold, the fluctuation range of the clustering metric is less than a certain threshold, and so on; if so, determine that the K clusters have completed clustering; if not, return to S512.

[0136] Step 1052. Configure target weights for the sample points in each cluster.

[0137] In this embodiment, the target weights can be configured according to the distribution of the sample points in each cluster to distinguish the importance degrees of the normal state and the failure states of one or more categories.

[0138] In specific implementation, each cluster can be traversed, and methods such as marking and voting can be used to determine the clusters representing the failure states and the clusters representing the normal state.

[0139] Among them, marking means screening a number of typical running signals in the normal state, and the clusters where the sample points corresponding to these running signals are located are the clusters representing the normal state, and screening a number of typical running signals in the failure state, and the clusters where the sample points corresponding to these running signals are located are the clusters representing the failure state.

[0140] Voting means counting the number of sample points in each running state within the cluster, and this cluster represents the running state with the highest number.

[0141] Calculate the distance between the center points of the clusters representing the failure states and the center points of the clusters representing the normal state, such as the Euclidean distance, etc.

[0142] Configure the first candidate weights for the sample points in the clusters representing the failure states according to the distance; among them, the first candidate weights are positively correlated with the distance, that is, the greater the distance, the greater the first candidate weights, so as to highlight the failure states with more obvious anomalies.

[0143] In addition, the second candidate weights default set for the failure states can be queried, and the second candidate weights are usually empirical values.

[0144] Such as Figure 7As shown, the first candidate weight and the second candidate weight in the same fault state are subjected to polynomial regression to obtain the target weight of the sample points in the cluster representing the fault state.

[0145] Step 1053: Use the sample points with configured target weights and the operating states to train a decision tree to obtain a state detection model for the components of the elevator.

[0146] In this embodiment, the sample points with configured target weights and the corresponding operating states can be used to train a decision tree as the state detection model for the corresponding components of the elevator.

[0147] Generally, a decision tree contains a root node, several internal nodes, and several leaf nodes. The leaf nodes correspond to the operating states of the elevator components, and each of the other nodes corresponds to a test of an operating feature. The samples finally divided into the same leaf node have the same decision attributes. The average value of these samples can be calculated to achieve regression, and voting (selecting the category with the largest number of samples) can be performed on these samples to achieve classification.

[0148] In this embodiment, each component in the elevator is queried; the component has a configured triggering method; during the operation of the elevator, the operating signals related to the component are collected according to the triggering method; the operating states of the elevator components are labeled for the operating signals; the operating features representing the characteristics of the elevator components are extracted from the operating signals; and the state detection model for the elevator components is trained based on the operating features and the operating states. This embodiment performs automatic monitoring on the elevator components, timely captures the operating signals in the elevator, improves the accuracy when positioning faults, trains the state detection model based on the operating signals in various historical operating states to achieve automatic auxiliary troubleshooting, reduces the dependence on the personal experience of maintenance personnel during troubleshooting, reduces the time spent on troubleshooting, and thus improves the efficiency of troubleshooting.

[0149] Embodiment 2

[0150] See Figure 8 , which shows a flowchart of a state detection method provided in Embodiment 1 of the present invention. This method can be executed by a state detection device. The training of this state detection model and the state detection device can be implemented in the form of hardware and / or software. The state detection device can be configured in an electronic device. As Figure 8 shown, this method includes:

[0151] Step 801: Query each component in the elevator.

[0152] Among them, the component has a configured triggering method.

[0153] Step 802: During the operation of the elevator, collect the operating signals related to the component according to the triggering method.

[0154] In a specific implementation, the triggering method includes a triggering condition and acquisition parameters. Then, during the operation of the elevator, the signal indicated by the triggering condition is matched with the triggering condition. If the match is successful, the operation signal related to the component is acquired according to the acquisition parameters.

[0155] Among them, the triggering condition includes at least one of the following types:

[0156] Valid, change maintained, invalid low level, invalid high level;

[0157] Valid means low level when the triggering condition is not met and high level when the triggering condition is met;

[0158] Change maintained means low level when the triggering condition is not met and high level is maintained after the triggering condition is first met;

[0159] Invalid low level means the triggering condition is unavailable and the low level is maintained;

[0160] Invalid high level means the triggering condition is unavailable and the high level is maintained;

[0161] The acquisition parameters include at least one of the following:

[0162] Capture period, capture time ratio before triggering, capture time ratio after triggering, total capture time, delay time after triggering.

[0163] Step 803: Extract the operation characteristics representing the component characteristics of the elevator from the operation signals.

[0164] In an embodiment of the present invention, step 803 may include the following steps:

[0165] Step 8031: If the type of the operation signal is a pulse signal, extract at least one of the waveform characteristics, time-domain characteristics, and frequency-domain characteristics from the operation signal as the operation characteristics representing the component characteristics of the elevator.

[0166] Regarding the waveform characteristics, on the one hand, according to the operation logic of the elevator, the relative time when the operation signal changes from low level to high level can be set as the waveform characteristic as the operation characteristic representing the component characteristics of the elevator; on the other hand, according to the operation logic of the elevator, the relative time when the operation signal changes from high level to low level can be set as the waveform characteristic as the operation characteristic representing the component characteristics of the elevator.

[0167] Step 8032: If the type of the operation signal is a numerical signal, extract at least one of the statistical characteristics, change characteristics, and threshold characteristics from the operation signal as the operation characteristics representing the component characteristics of the elevator.

[0168] Step 8033: According to the operation logic of the elevator, calculate the relative change characteristics between operation signals as the operation characteristics representing the component characteristics of the elevator.

[0169] In a specific implementation, according to the operation logic of the elevator, two operation signals with a relative relationship are selected as the first target signal and the second target signal.

[0170] The relative time when the first target signal generates the first change trend is counted as the first target time point; after the first target time point, the relative time when the second target signal generates the second change trend is counted as the second target time point, where the second change trend is the same as or opposite to the first change trend.

[0171] The difference between the second target time point and the first target time point is set as the relative change characteristic, which is used as the operation characteristic representing the component characteristics of the elevator.

[0172] Step 804: Load the state detection model.

[0173] In this embodiment, the state detection model trained according to the method described in Embodiment 1 can be loaded at the time of elevator startup or other opportunities.

[0174] Step 805: Input the operation characteristics into the state detection model to detect the operation state of the components of the elevator.

[0175] In this embodiment, the operation characteristics can be input into the state detection model for processing, and the state detection model outputs the operation state of the components of the elevator.

[0176] If the operation state is a fault state, the operation signals that may be abnormal in this fault state, that is, the operation signals that may cause the fault, can be queried by means of table lookup or the like, and the maintenance suggestions configured for the operation information that may cause the fault can be queried by means of table lookup or the like, and the maintenance suggestions are pushed to the maintenance personnel.

[0177] This embodiment can integrate the management of trigger methods (such as trigger conditions, acquisition parameters), the notification of fault states (especially analysis and display), maintenance suggestions and other closed-loop management into a unified elevator control system. The unified elevator control system can manage the entire maintenance process of the elevator, and can manage the full life cycle of the maintenance of each elevator, and can always understand the history of potential hazards, the specific triggering situation, and the abnormal performance of specific operation signals.

[0178] In this embodiment, since Steps 801 - 804 are basically similar to those in Embodiment 1 in application, the description is relatively simple. For the relevant parts, refer to the corresponding descriptions in Embodiment 1, and this embodiment will not be elaborated here.

[0179] In this embodiment, each component in the elevator is queried; the component has a configured triggering method; during the operation of the elevator, operation signals related to the component are collected according to the triggering method; operation features characterizing the component characteristics of the elevator are extracted from the operation signals; a state detection model is loaded; and the operation features are input into the state detection model to detect the operation state of the component of the elevator. This embodiment performs automated monitoring on elevator components, captures operation signals in the elevator in a timely manner, improves the accuracy when locating faults, uses operation signals in various historical operation states to train the state detection model to automatically assist in troubleshooting, reduces the dependence on the personal experience of maintenance personnel during troubleshooting, reduces the time spent on troubleshooting, and thus improves the efficiency of troubleshooting.

[0180] Embodiment III

[0181] See Figure 9 , which shows a schematic structural diagram of a training device for a state detection model provided in Embodiment III of the present invention. As Figure 9 shown, the device includes:

[0182] A component query module 901, configured to query each component in the elevator; the component has a configured triggering method;

[0183] An operation signal acquisition module 902, configured to collect operation signals related to the component according to the triggering method during the operation of the elevator;

[0184] An operation state annotation module 903, configured to annotate the operation state of the component of the elevator for the operation signal;

[0185] An operation feature extraction module 904, configured to extract operation features characterizing the component characteristics of the elevator from the operation signal;

[0186] A state detection model training module 905, configured to train a state detection model for the component of the elevator according to the operation feature and the operation state.

[0187] In an embodiment of the present invention, the triggering method includes a triggering condition and acquisition parameters;

[0188] The operation signal acquisition module 902 includes:

[0189] A condition matching module, configured to match the signal indicated by the triggering condition with the triggering condition during the operation of the elevator;

[0190] A parameter acquisition module, configured to, if the matching is successful, collect operation signals related to the component according to the acquisition parameters;

[0191] Wherein, the triggering condition includes at least one of the following types:

[0192] Valid, change maintained, invalid low level, invalid high level;

[0193] Said "valid" means low level when the trigger condition is not met, and high level when the trigger condition is met;

[0194] Said "change maintained" means low level when the trigger condition is not met, and high level is maintained after the trigger condition is first met;

[0195] Said "invalid low level" means that the trigger condition is unavailable and the low level is maintained;

[0196] Said "invalid high level" means that the trigger condition is unavailable and the high level is maintained;

[0197] The acquisition parameters include at least one of the following:

[0198] Capture period, proportion of capture time before trigger, proportion of capture time after trigger, total capture time, delay time after trigger.

[0199] In an embodiment of the present invention, the operation feature extraction module 904 includes:

[0200] A pulse feature extraction module, which is used to extract at least one of waveform features, time-domain features, and frequency-domain features from the operation signal as the operation feature characterizing the component characteristics of the elevator if the type of the operation signal is a pulse signal;

[0201] A numerical feature extraction module, which is used to extract at least one of statistical features, change features, and threshold features from the operation signal as the operation feature characterizing the component characteristics of the elevator if the type of the operation signal is a numerical signal;

[0202] A relative feature extraction module, which is used to calculate the relative change feature between the operation signals according to the operation logic of the elevator as the operation feature characterizing the component characteristics of the elevator.

[0203] In an embodiment of the present invention, the pulse feature extraction module includes:

[0204] A first waveform feature extraction module, which is used to set the relative time when the operation signal changes from low level to high level as the waveform feature according to the operation logic of the elevator as the operation feature characterizing the component characteristics of the elevator;

[0205] A second waveform feature extraction module, which is used to set the relative time when the operation signal changes from high level to low level as the waveform feature according to the operation logic of the elevator as the operation feature characterizing the component characteristics of the elevator.

[0206] In one embodiment of the present invention, the relative feature extraction module includes:

[0207] A target signal screening module, configured to screen out two running signals with a relative relationship as a first target signal and a second target signal according to the operation logic of the elevator;

[0208] A first target time point statistics module, configured to count the relative time when the first target signal generates a first change trend as the first target time point;

[0209] A second target time point statistics module, configured to count the relative time when the second target signal generates a second change trend after the first target time point as the second target time point; the second change trend is the same as or opposite to the first change trend;

[0210] A relative change feature setting module, configured to set the difference between the second target time point and the first target time point as a relative change feature, as an operation feature characterizing the component characteristics of the elevator.

[0211] In one embodiment of the present invention, the state detection model training module 905 includes:

[0212] A clustering module, configured to cluster the sample points according to the operation state with the operation feature as the sample points to obtain multiple clusters;

[0213] A target weight configuration module, configured to configure target weights for the sample points in each of the clusters;

[0214] A decision tree training module, configured to train a decision tree using the sample points with the target weights configured and the operation state to obtain a state detection model of the components of the elevator.

[0215] In one embodiment of the present invention, the operation state includes a normal state and a fault state;

[0216] The clustering module includes:

[0217] A total quantity statistics module, configured to count the total quantities of the normal state and the fault state;

[0218] A K-means clustering module, configured to perform K-means clustering on the sample points with the operation feature as the sample points to obtain clusters that meet the total quantity;

[0219] The target weight configuration module includes:

[0220] A cluster determination module, configured to determine the cluster representing the fault state and the cluster representing the normal state;

[0221] A distance calculation module, configured to calculate the distance between the center points of the clusters representing the fault states and the center point of the cluster representing the normal state;

[0222] A first candidate weight configuration module, configured to configure first candidate weights for the sample points in the cluster representing the fault state according to the distance; the first candidate weight is positively correlated with the distance;

[0223] A second candidate weight query module, configured to query second candidate weights default-set for the fault state;

[0224] A polynomial regression module, configured to perform polynomial regression on the first candidate weight and the second candidate weight of the same fault state to obtain the target weight of the sample points in the cluster representing the fault state.

[0225] The training device of the state detection model provided by the embodiments of the present invention can execute the training method of the state detection model provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the training method of the state detection model.

[0226] Embodiment 4

[0227] See Figure 10 , which shows a structural schematic diagram of a state detection device provided by Embodiment 4 of the present invention. As Figure 10 shown, the device includes:

[0228] A component query module 1001, configured to query each component in the elevator; the component has a configured triggering mode;

[0229] An operation signal acquisition module 1002, configured to acquire operation signals related to the component according to the triggering mode during the operation of the elevator;

[0230] An operation feature extraction module 1003, configured to extract operation features representing the component characteristics of the elevator from the operation signals;

[0231] A state detection model loading module 1004, configured to load a state detection model trained according to the method described in Embodiment 1;

[0232] An operation state detection module 1005, configured to input the operation features into the state detection model to detect the operation state of the components of the elevator.

[0233] In an embodiment of the present invention, the triggering mode includes a triggering condition and acquisition parameters;

[0234] The operation signal acquisition module 1002 includes:

[0235] A condition matching module, configured to match the signal indicated by the trigger condition with the trigger condition during the operation of the elevator;

[0236] A parameter acquisition module, configured to, if the matching is successful, acquire the operation signals related to the component according to the acquisition parameters;

[0237] Wherein, the trigger condition includes at least one of the following types:

[0238] Valid, change maintained, invalid low level, invalid high level;

[0239] The "valid" means that when the trigger condition is not satisfied, it is at a low level, and when the trigger condition is satisfied, it is at a high level;

[0240] The "change maintained" means that when the trigger condition is not satisfied, it is at a low level, and after the trigger condition is first satisfied, it remains at a high level;

[0241] The "invalid low level" means that the trigger condition is unavailable and maintains a low level;

[0242] The "invalid high level" means that the trigger condition is unavailable and maintains a high level;

[0243] The acquisition parameters include at least one of the following:

[0244] Capture period, capture time ratio before trigger, capture time ratio after trigger, total capture time, delay time after trigger.

[0245] In an embodiment of the present invention, the operation feature extraction module 1003 includes:

[0246] A pulse feature extraction module, configured to, if the type of the operation signal is a pulse signal, extract at least one of waveform features, time-domain features, and frequency-domain features from the operation signal as the operation features characterizing the component characteristics of the elevator;

[0247] A numerical feature extraction module, configured to, if the type of the operation signal is a numerical signal, extract at least one of statistical features, change features, and threshold features from the operation signal as the operation features characterizing the component characteristics of the elevator;

[0248] A relative feature extraction module, configured to calculate the relative change features between the operation signals according to the operation logic of the elevator as the operation features characterizing the component characteristics of the elevator.

[0249] In an embodiment of the present invention, the pulse feature extraction module includes:

[0250] The first waveform feature extraction module is configured to set, according to the operation logic of the elevator, the relative time when the operation signal changes from a low level to a high level as a waveform feature, which is used as an operation feature characterizing the component characteristics of the elevator.

[0251] The second waveform feature extraction module is configured to set, according to the operation logic of the elevator, the relative time when the operation signal changes from a high level to a low level as a waveform feature, which is used as an operation feature characterizing the component characteristics of the elevator.

[0252] In an embodiment of the present invention, the relative feature extraction module includes:

[0253] The target signal screening module is configured to screen out two operation signals with a relative relationship as the first target signal and the second target signal according to the operation logic of the elevator.

[0254] The first target time point statistics module is configured to count the relative time when the first target signal generates a first change trend as the first target time point.

[0255] The second target time point statistics module is configured to, after the first target time point, count the relative time when the second target signal generates a second change trend as the second target time point; the second change trend is the same as or opposite to the first change trend.

[0256] The relative change feature setting module is configured to set the difference between the second target time point and the first target time point as a relative change feature, which is used as an operation feature characterizing the component characteristics of the elevator.

[0257] The state detection device provided by the embodiment of the present invention can execute the state detection method provided by any embodiment of the present invention, and has corresponding function modules and beneficial effects for executing the state detection method.

[0258] Embodiment 5

[0259] See Figure 11 , which shows a schematic structural diagram of an electronic device provided by an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0260] AsFigure 11 As shown, the electronic device 10 includes at least one processor 11 and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. Among them, the memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0261] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disc, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0262] The processor 11 can be various general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the training method or the state detection method of the state detection model.

[0263] In some embodiments, the training method or the state detection method of the state detection model can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the training method or the state detection method of the state detection model described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute the training method or the state detection method of the state detection model in any other appropriate way (for example, by means of firmware).

[0264] The various embodiments of the systems and techniques described above in this specification can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that receives data and instructions from, and transmits data and instructions to, a storage system, at least one input device, and at least one output device.

[0265] The computer program for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The computer program can be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

[0266] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0267] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0268] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0269] The computing system can include a client and a server. The client and the server are generally far from each other and typically interact through a communication network. The relationship between the client and the server is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0270] Example Six

[0271] The embodiment of the present invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the training method or the state detection method of the state detection model provided in any embodiment of the present invention.

[0272] In the process of implementing the computer program product, computer program code for performing the operations of the present invention can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network - including a local area network (LAN) or a wide area network (WAN) - or, alternatively, can be connected to an external computer (e.g., by using an Internet service provider to connect through the Internet).

[0273] It should be understood that the various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is imposed herein.

[0274] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A training method for a state detection model, characterized in that: include: Check the various components in the elevator; The component has been configured with a trigger mode; During the operation of the elevator, collecting operation signals related to the component according to the triggering method; marking the operating status of the components of the elevator on the operating signal; Extracting operation features characterizing component characteristics of the elevator from the operation signal; A state detection model is trained on components of the elevator according to the operation characteristics and the operation state.

2. The method according to claim 1, characterized in that The triggering mode includes triggering conditions and acquisition parameters; During the operation of the elevator, collecting the operation signal related to the component according to the triggering mode includes: During the operation of the elevator, matching the signal indicated by the trigger condition with the trigger condition; If the match is successful, the operation signal related to the component is collected according to the collection parameters; The trigger condition includes at least one of the following types: Valid, change hold, invalid low level, invalid high level; The validity refers to a low level when the trigger condition is not met, and a high level when the trigger condition is met; The change hold refers to a low level when the trigger condition is not met, and a high level after the trigger condition is met for the first time; The invalid low level means that the trigger condition is not available and maintains a low level; The invalid high level means that the trigger condition is not available and maintains a high level; The acquisition parameters include at least one of the following: Capture cycle, capture time ratio before trigger, capture time ratio after trigger, total capture time, delay time after trigger.

3. The method according to claim 1, characterized in that The extracting the operation feature characterizing the component characteristics of the elevator from the operation signal comprises: If the type of the operation signal is a pulse signal, extracting at least one of a waveform feature, a time domain feature, and a frequency domain feature from the operation signal as an operation feature characterizing the component characteristics of the elevator; If the type of the operation signal is a numerical signal, extracting at least one of a statistical feature, a change feature, and a threshold feature from the operation signal as an operation feature characterizing a component characteristic of the elevator; According to the operation logic of the elevator, relative change characteristics between the operation signals are calculated as operation characteristics characterizing the component characteristics of the elevator.

4. The method according to claim 3, characterized in that The step of extracting at least one of a waveform feature, a time domain feature, and a frequency domain feature from the operation signal as an operation feature characterizing the component characteristics of the elevator comprises: According to the operation logic of the elevator, the relative time when the operation signal changes from a low level to a high level is set as a waveform feature, which serves as an operation feature characterizing the component characteristics of the elevator; According to the operation logic of the elevator, the relative time when the operation signal changes from a high level to a low level is set as a waveform feature, which serves as an operation feature characterizing the component characteristics of the elevator.

5. The method according to claim 3, characterized in that: The step of calculating the relative change characteristics between the operation signals according to the operation logic of the elevator as the operation characteristics characterizing the component characteristics of the elevator includes: According to the operation logic of the elevator, two operation signals having a relative relationship are screened out as a first target signal and a second target signal; Counting the relative time when the first target signal generates a first change trend as a first target time point; After the first target time point, the relative time when the second target signal generates a second change trend is counted as the second target time point; the second change trend is the same as or opposite to the first change trend; The difference between the second target time point and the first target time point is set as a relative change feature as an operation feature characterizing the component characteristics of the elevator.

6. The method according to any one of claims 1 to 5, characterized in that The training state detection model of the elevator components according to the operation characteristics and the operation state includes: Taking the operation characteristics as sample points, clustering the sample points according to the operation status to obtain a plurality of clusters; configuring a target weight for the sample points in each of the clusters; A decision tree is trained using the sample points configured with the target weights and the operating status to obtain a status detection model for components of the elevator.

7. The method according to claim 6, characterized in that The operating state includes a normal state and a fault state; The operation characteristics are used as sample points, and the sample points are clustered according to the operation status to obtain multiple clusters, including: Counting the total number of the normal state and the fault state; Taking the operation characteristics as sample points, performing K-means clustering on the sample points to obtain clusters that meet the total number; The configuring target weights for the sample points in each of the clusters includes: Determining the cluster representing the fault state and the cluster representing the normal state; Calculating the distance between the center point of each cluster representing the fault state and the center point of each cluster representing the normal state; According to the distance, a first candidate weight is configured for the sample points in the cluster representing the fault state; the first candidate weight is positively correlated with the distance; querying a second candidate weight set by default for the fault state; The first candidate weight and the second candidate weight of the same fault state are subjected to polynomial regression to obtain a target weight of the sample point in the cluster representing the fault state.

8. A state detection method, characterized in that: include: Check the various components in the elevator; The component has been configured with a trigger mode; During the operation of the elevator, collecting operation signals related to the component according to the triggering method; Extracting operation features characterizing component characteristics of the elevator from the operation signal; Loading a state detection model trained according to any one of the methods of claims 1-7; The operation characteristics are input into the state detection model to detect the operation states of the components of the elevator.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the training method of the state detection model as described in any one of claims 1 to 7 or the state detection method as described in claim 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the training method of the state detection model according to any one of claims 1 to 7 or the state detection method according to claim 8 is implemented.

Images