Elevator inspection process optimization method

Abstract

The invention provides an elevator inspection process optimization method, which relates to the field of internet intelligent manufacturing, and comprises the following steps: collecting real-time sensor data corresponding to a target elevator, and preprocessing the real-time sensor data corresponding to each sensor; the digital twin elevator model corresponding to the target elevator is updated in real time, and the corresponding current operation state is output; to-be-detected point locations of virtual elevators in the digital twin elevator model at the current moment are determined, and all the to-be-detected point locations are mapped and coded into specific detection procedures; forming a plurality of candidate detection sequences according to the detection process, optimizing each candidate detection sequence, selecting an optimal detection sequence as a final detection sequence, and feeding back the final detection sequence to an inspector; acquiring current position information corresponding to the inspector and an inspection picture corresponding to the worn intelligent acquisition terminal; and according to the current position information of the inspector and the inspection picture, judging whether the inspection process currently executed by the inspector is correct or not.

Description

Technical Field

[0001] This invention relates to the field of Internet-based intelligent manufacturing technology, specifically a method for optimizing elevator inspection procedures. Background Technology

[0002] With the deep penetration of Internet-based intelligent manufacturing technology into the field of special equipment, elevators, as a key urban transportation infrastructure, are facing increasingly higher requirements for operational safety and inspection efficiency.

[0003] Current elevator inspection still has many technical shortcomings: traditional inspection relies on fixed procedures and does not dynamically adjust the plan according to the real-time operating status of the elevator, often resulting in the omission of high-risk components or the over-inspection of low-risk components, and it cannot adapt to the individualized safety needs of elevators.

[0004] The multi-sensor data processing stage has shortcomings. Sensor data from elevators in different dynamic ranges, such as acceleration and constant speed, are easily interfered with, resulting in issues such as time asynchrony, noise superposition, and data loss. This leads to a lack of reliable data support for subsequent inspection point location and process generation. Furthermore, the inspection sequence optimization often uses general algorithms without considering the spatial correlation and process dependencies of elevator inspections, resulting in low efficiency and poor execution rationality of the optimized sequence.

[0005] The lack of a real-time monitoring and error correction mechanism in the inspection process makes it easy for inspectors to make mistakes such as skipping steps or positional deviations. Traditional methods are unable to identify and provide timely guidance, which not only affects the quality of inspection but may also lead to safety risks.

[0006] Therefore, this paper presents an optimization method for elevator inspection procedures. Supported by Internet-based intelligent manufacturing, it uses refined preprocessing of sensor data, real-time digital twin updates to locate high-risk points, and multi-objective algorithms to generate the optimal detection sequence. Combined with real-time verification alarms, this method achieves intelligent, precise, and safe elevator inspection, meeting the high-quality supervision requirements of special equipment. Summary of the Invention

[0007] To address the aforementioned technical problems, the present invention aims to provide an elevator inspection process optimization method. Supported by Internet-based intelligent manufacturing, this method utilizes refined preprocessing of sensor data, real-time digital twin updates to locate high-risk points, and multi-objective algorithms to generate the optimal detection sequence. Coupled with real-time verification alarms, this method achieves intelligent, precise, and safe elevator inspection, meeting the high-quality supervision requirements of special equipment.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an elevator inspection process optimization method, comprising the following steps: Collect real-time sensor data corresponding to the target elevator and preprocess the real-time sensor data corresponding to each sensor. Based on the preprocessed real-time sensor data, the digital twin elevator model corresponding to the target elevator is updated in real time, and the current operating status of the target elevator is output. Based on the current operating status, determine the points to be detected in the virtual elevator in the digital twin elevator model at the current moment, and map and encode each point to be detected as a specific detection procedure. According to the detection process, several candidate detection sequences are formed. Based on a preset sequence optimization algorithm, each candidate detection sequence is optimized, and the optimal detection sequence is selected as the final detection sequence and fed back to the inspection personnel. When the inspectors receive the final test sequence, the system collects the inspectors' current location information and the test screen corresponding to the smart data acquisition terminal they are wearing. Based on the inspector's current location information and the inspection screen, it is determined whether the inspection procedure currently being performed by the inspector is correct. If incorrect, the inspector's current location and the inspection screen are visualized in the digital twin elevator model, and an alarm signal is issued.

[0009] Preferably, the preprocessing of the real-time sensor data corresponding to each sensor includes: Extract the operating status parameters corresponding to the target elevator, and based on the operating status parameters, divide the elevator operation process into four dynamic intervals: starting acceleration segment, constant speed operation segment, deceleration and stopping segment, and stationary segment; the operating status parameters include the elevator's real-time position data, running direction, and speed change rate. Differentiated time alignment strategies are adopted for different dynamic ranges to align the real-time sensor data corresponding to each sensor. The real-time sensor data after alignment is denoised and missing data is filled in.

[0010] Preferably, determining the detection point of the virtual elevator in the digital twin elevator model at the current moment based on the current operating state includes: Based on the current operating status of the digital twin elevator model and the real-time sensor data of each corresponding sensor, and based on the historical operating data stored in the database, a weighted fusion algorithm is used to calculate the dynamic risk index of each component in the digital twin elevator model. Based on the predefined risk factor weights corresponding to each component, the dynamic risk index is corrected to generate a comprehensive risk priority number, and all components to be inspected are sorted according to the risk priority number. Based on the risk priority ranking of each component, the component ranked in the preset position is selected as the high-risk component, and the specific geometric feature points corresponding to each high-risk component in the digital twin elevator model are determined as the corresponding detection points.

[0011] Preferably, the step of mapping and encoding each detection point into a specific detection process includes: Based on the current operating status of the digital twin elevator model corresponding to the target elevator and the corresponding historical operating data, a corresponding multi-dimensional dynamic point code is generated for each point to be detected. The dynamic point code includes at least a basic identity ID representing the component identity, a real-time risk label representing the real-time risk level, and a resource and environment label representing the site conditions. Based on the dynamic point code, and using a preset inspection process knowledge base and process matching strategy, the corresponding inspection process is matched for all dynamic point codes. The preset process matching strategy is as follows: In the inspection process knowledge base, search for a process template that perfectly matches the complete dynamic point code; If the match fails, the most suitable process template is selected from multiple candidate templates based on the value of one or more specific tags in the dynamic point code. If no suitable template is found, the corresponding smallest operation unit is selected from the preset process atomic operation library and dynamically combined according to the rules to generate a temporary detection process.

[0012] Preferably, the step of generating several candidate detection sequences through the detection process, optimizing each candidate detection sequence based on a preset sequence optimization algorithm, selecting the optimal detection sequence as the final detection sequence, and feeding it back to the inspection personnel includes: According to the detection process, the positions of each sub-step in the detection process are randomly changed to form several candidate detection sequences, which serve as the corresponding initial population. Calculate the process detection efficiency value and process detection rationality corresponding to each individual in the population; Within the first iteration, individuals with process detection efficiency values ​​greater than the first preset efficiency value are selected, and individuals with process detection efficiency values ​​greater than or equal to the second preset efficiency value and less than the first preset efficiency value are stored in the spare pool to form a spare population. After exceeding the first iteration, the current population is combined with the backup population to form a new population, and individuals in the current population whose process detection rationality is greater than or equal to the preset rationality are selected. The selected population was used to obtain offspring populations through a hybridization and variation strategy. After obtaining the offspring population, the calculation and screening of the process detection efficiency value and process detection rationality continue until the preset number of iterations is met; The offspring population is output as the final detection sequence and fed back to the testing personnel.

[0013] Preferably, storing individuals with process detection efficiency values ​​greater than or equal to a second preset efficiency value and less than a first preset efficiency value in a backup database to form a backup population includes: Individuals whose process inspection efficiency value is greater than or equal to the second preset efficiency value and less than the first preset efficiency value are selected. Based on the selected individuals, calculate the distance between each individual; Compare the current iteration number with the number of the first iteration: When the number of iterations in this iteration is less than or equal to the first percentage of the number of iterations in the first iteration, individuals are sorted in descending order of distance value and the second-highest percentage are selected and stored in the reserve database. When the current iteration number is greater than the first percentage of the first iteration number but less than or equal to the first iteration number, individuals in the top third percentage are selected and stored in the backup database after being sorted by distance value, and the second percentage is greater than the third percentage.

[0014] Preferably, the step of determining whether the inspection procedure currently being performed by the inspector is correct based on the inspector's current location information and the inspection screen, and visually displaying the inspector's current location and the inspection screen in the digital twin elevator model and issuing an alarm signal when incorrect, includes: Based on the final detection sequence, determine the node process information associated with each process node in the final detection sequence. The node process information includes the node's unique sequential number, theoretical execution area, and standard point characteristics. Based on the current location information of the inspectors and the inspection screen, determine the current location coordinates of the inspectors and the actual point features in the inspection screen; Find the last completed process node in the final detection sequence, and determine the next process node corresponding to the last process node as the next execution node; Check whether the current position coordinates match the theoretical execution area of ​​the next execution node, and whether the matching degree between the actual point features and the standard point features of the next execution node is greater than the preset matching threshold. If the location matches and the matching degree meets the standard, the inspection procedure currently being performed by the inspector is considered correct; if any one of these conditions is not met, the inspection procedure currently being performed by the inspector is considered incorrect. If the positions do not match, the theoretical execution position of the next execution node is marked with a yellow dynamically flashing icon in the digital twin elevator model, and the current position of the inspector is marked with a red solid icon. The two are connected by a red dashed line and the straight-line distance is marked. If the matching degree does not meet the standard, a red text bubble will pop up at the current position of the inspector in the digital twin elevator model, indicating the current process node to be executed and the actual process node to be executed.

[0015] Preferably, the hybridization variation strategy includes: Based on the preset elevator inspection space area and the pre-determined dependencies of the inspection process, the individuals in the selected population are divided into 3 associated process partitions. The processes in each associated process partition satisfy the conditions of spatial continuity and no pre-determined dependency conflicts. Individuals are selected from the population and paired up. Three independent first random numbers are randomly generated for the three associated process partitions of individual one. The value of the random number is between 0 and 1. If the first random number is greater than the preset random threshold, then the associated process partition of individual one corresponding to the first random number is swapped with the associated process partition of individual two. If the first random number is less than or equal to a preset random threshold, a second random number between 0 and 1 is randomly generated for the associated process partition of individual one corresponding to the first random number. If the second random number is greater than the first random number, the associated process partition of individual one corresponding to the second random number is mutated.

[0016] Compared with the prior art, the beneficial effects of the present invention are: Existing technologies often employ pre-set, fixed inspection procedures, ignoring the differences in real-time elevator operating conditions. This can lead to the omission of high-risk components or the over-inspection of low-risk components. This invention collects real-time sensor data from the elevator, updates the operating status in real time using a digital twin model, calculates the dynamic risk index of each component using a weighted fusion algorithm, and generates risk priorities based on pre-set risk factor weights, accurately locating high-risk components and their corresponding inspection points.

[0017] Differentiated time alignment strategies are adopted for different dynamic ranges to effectively solve the problem of time asynchrony of multi-sensor data. After denoising and missing data filling, compared with traditional simple data preprocessing, the authenticity and integrity of the data are greatly improved, providing reliable data support for subsequent detection processes and sequence optimization, and reducing inspection deviations caused by data errors. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a schematic diagram of a method for optimizing elevator inspection procedures. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0021] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0022] like Figure 1 As shown, an elevator inspection process optimization method includes the following steps: Real-time sensor data corresponding to the target elevator is collected and preprocessed. In this embodiment, eight types of sensors are deployed, including a traction machine temperature sensor (sampling frequency 1 time / 10s), a braking system pressure sensor (1 time / 5s), a door operator operating current sensor (1 time / 8s), a car position sensor (1 time / 2s), a running direction sensor (1 time / 1s), a speed sensor (1 time / 2s), a car vibration sensor (1 time / 5s), and an ambient temperature and humidity sensor (1 time / 30s). A 5G industrial gateway is used to transmit the sensor data to the cloud server in real time, with transmission latency controlled to ≤100ms and a data packet loss rate ≤0.1%.

[0023] It should be noted that the preprocessing of the real-time sensor data corresponding to each sensor includes: The operating status parameters corresponding to the target elevator are extracted. Based on these parameters, the elevator operation process is divided into four dynamic intervals: acceleration, constant speed, deceleration and stopping, and stationary. The operating status parameters include the elevator's real-time position data, direction of travel, and rate of change of speed. In this embodiment, the elevator's position, direction, and rate of change of speed data are read in real time using a built-in algorithm on the server. The elevator is divided into acceleration, constant speed, deceleration and stopping, and stationary intervals according to standards, and the interval determination results are updated every 2 seconds.

[0024] For example, during the initial acceleration phase: the rate of change of velocity > 0.5 m / s 2 And the position increases / decreases with time; uniform speed segment: rate of change of velocity ∈ [-0.1m / s] 2 0.1m / s 2 And the speed is stable; during deceleration and stopping: the rate of speed change is < -0.5m / s 2 And the speed approaches 0; stationary segment: speed = 0 and duration ≥ 3s.

[0025] Differentiated time alignment strategies are adopted for different dynamic ranges to align the real-time sensor data corresponding to each sensor. In this embodiment, a differentiated time alignment strategy is adopted, that is, different alignment periods are used for different intervals, including: Acceleration / Deceleration Stop: Due to rapid parameter changes, the alignment period is set to 5 seconds, meaning that every 5 seconds, all sensor data within this period are aligned according to timestamps, retaining the complete data set under each timestamp.

[0026] Uniform speed operation segment: parameters are stable, and the alignment period is set to 15s.

[0027] Static segment: Parameters change very little, and the alignment period is set to 60s.

[0028] The real-time sensor data after alignment is denoised and missing data is filled in.

[0029] In this embodiment, a wavelet transform denoising algorithm is used to remove high-frequency noise (such as electromagnetic interference noise from a current sensor) from the sensor data. For missing data, if the missing duration is less than or equal to the alignment period, linear interpolation is used to fill in the missing data (e.g., if the data in the 3rd second is missing within a 5-second alignment period, the average of the data in the 2nd and 4th seconds is used to fill in the missing data). If the missing duration is greater than the alignment period, historical data from the same period (historical data from the same time period and operating state) is called to fill in the missing data.

[0030] Based on the preprocessed real-time sensor data, the digital twin elevator model corresponding to the target elevator is updated in real time, and the current operating status of the target elevator is output. Based on the current operating status, determine the points to be detected in the virtual elevator in the digital twin elevator model at the current moment, and map and encode each point to be detected as a specific detection procedure. It should be noted that determining the detection points of the virtual elevator in the digital twin elevator model at the current moment, based on the current operating state, includes: Based on the current operating status of the digital twin elevator model and the real-time sensor data of each corresponding sensor, and based on the historical operating data stored in the database, a weighted fusion algorithm is used to calculate the dynamic risk index of each component in the digital twin elevator model. The formula for calculating the dynamic risk index is as follows: ; In the formula, The weights for the i-th type of sensor data are (e.g., traction machine temperature weight 0.3, braking system pressure weight 0.25). The standardized value of the i-th type of sensor data (range [0,1], the larger the value, the higher the risk). This corresponds to the dynamic risk index.

[0031] Based on the predefined risk factor weights corresponding to each component, the dynamic risk index is corrected to generate a comprehensive risk priority number, and all components to be inspected are sorted according to the risk priority number. The risk priority number is calculated as follows: ; In the formula, Risk priority number For risk factor weights.

[0032] Based on the risk priority ranking of each component, components ranked in a preset position are selected as high-risk components, and specific geometric feature points corresponding to each high-risk component in the digital twin elevator model are identified as corresponding detection points. In this embodiment, specific geometric feature points of high-risk components (such as the bearing position of the traction machine, the jaw contact point of the safety gear) are located in the digital twin elevator model, and these feature points are marked as detection points. Each point generates unique spatial coordinates (e.g., ...). , , ).

[0033] It should be noted that the mapping and encoding of each detection point into a specific detection procedure includes: Based on the current operating status of the digital twin elevator model corresponding to the target elevator and the corresponding historical operating data, a corresponding multi-dimensional dynamic point code is generated for each point to be detected. The dynamic point code includes at least a basic identity ID representing the component identity, a real-time risk label representing the real-time risk level, and a resource and environment label representing the site conditions. In this embodiment, when the dynamic point code includes a basic identity ID representing the component identity and a real-time risk label representing the real-time risk level, a 12-bit encoding structure is adopted, and the format is ID-RE. Among them, ID (6 digits): component identification code (such as traction machine code, 010001); R (3 digits): Real-time risk level (001 - low risk, 010 - medium risk, 100 - high risk); E (3 digits): On-site environment label (001 - normal temperature and humidity, 010 - high temperature environment, 100 - humid environment).

[0034] Based on the dynamic point code, and using a preset inspection process knowledge base and process matching strategy, the corresponding inspection process is matched for all dynamic point codes. It should be noted that the preset process matching strategy is as follows: In the inspection process knowledge base, a process template that perfectly matches the complete dynamic point code is searched. In this embodiment, the inspection process knowledge base (which stores 1,000+ standard process templates, covering the inspection processes of common components of mainstream elevator brands such as Otis, Mitsubishi, and Kone) is searched. If a template that is completely consistent with the dynamic point code exists (such as the code "010001-100-001" corresponding to the inspection template under high-risk and normal conditions of the traction machine), the template is directly called as the inspection process.

[0035] If the match fails, the most suitable process template is selected from multiple candidate templates based on the value of one or more specific tags in the dynamic point code. If no suitable template is found, the corresponding smallest operation unit is selected from the preset process atomic operation library and dynamically combined according to the rules to generate a temporary detection process.

[0036] In this embodiment, if the matching fails, a specific tag in the code (such as the "R=100" high-risk tag and the "E=010" high-temperature environment tag) is taken, and candidate templates containing the same tag (such as all "R=100" traction machine detection templates) are retrieved in the knowledge base. The matching degree between the candidate template and the current code is calculated by the cosine similarity algorithm, and the template with the highest matching degree (≥80%) is selected as the detection process; if the matching degree is <80%, the subsequent matching is performed.

[0037] The process atomic operation library (containing 200+ minimum operation units, such as "remove protective cover", "measure temperature", "record data", and "install protective cover") is invoked. Based on the component type, risk level, and environmental label in the dynamic point code, atomic operations are combined according to preset rules (such as adding "repeat measurement twice" for high-risk components and "cooling equipment pretreatment" for high-temperature environments) to generate temporary inspection processes. For example, for the point "010001-100-010" (traction machine, high risk, high-temperature environment), the processes "cooling equipment pretreatment, remove protective cover, measure bearing temperature (repeat twice), record data, and install protective cover" are combined as temporary processes.

[0038] According to the detection process, several candidate detection sequences are formed. Based on a preset sequence optimization algorithm, each candidate detection sequence is optimized, and the optimal detection sequence is selected as the final detection sequence and fed back to the inspection personnel. It should be noted that the process of generating several candidate detection sequences through the detection procedure, optimizing each candidate detection sequence based on a preset sequence optimization algorithm, selecting the optimal detection sequence as the final detection sequence, and feeding it back to the inspection personnel includes: According to the detection process, the positions of each sub-step in the detection process are randomly changed to form several candidate detection sequences, which serve as the corresponding initial population. The process detection efficiency value and process detection rationality corresponding to each individual in the population are calculated; in this embodiment, the process detection efficiency value... The calculation formula is: ; In the formula, As a weight for movement time, Total movement time is the cumulative time spent by inspectors moving from one process point to the next in the inspection sequence. Weighted by operation time. The total operation time is the cumulative standard operation time of all procedures in the detection sequence. The path repetition rate weight. Path repetition rate is the proportion of repeated movement paths in the total path count of the detected sequence.

[0039] Reasonableness of process inspection The calculation formula is: ; In the formula, To assess dependency compliance, the number of process pairs in the detection sequence that satisfy the "preceding-following" dependency relationship is compared. Dependency rules are extracted from the detection process knowledge base (e.g., "power off operation" is a preceding process of "circuit detection," and "removing the protective cover" is a preceding process of "bearing detection"). The number of process pairs Nright that satisfy the dependency rules in the sequence is compared with the total number of dependent process pairs Ntotal. If there is no dependency, then the value is 1; To ensure spatial logical compliance, the proportion of paths in the sequence that conform to the "floor / area continuity" principle is detected, and spatial areas are divided according to the elevator structure (e.g., low floors: 1-6 floors, middle floors: 7-12 floors, high floors: 13-18 floors). The number of path segments Sright (consecutive detections within the same region) or Stotal (orderly transition between adjacent regions) in the statistical sequence are compared with the total number of path segments. ; To assess risk matching compliance, the proportion of high-risk processes in the detection sequence is prioritized for detection. Processes are categorized into high-risk (RPN≥1.0), medium-risk (0.5≤RPN<1.0), and low-risk (RPN<0.5) based on their Risk Priority Number (RPN). The number of processes in the sequence with "high-risk first, medium-risk in the middle, and low-risk last" (Kright) is calculated and compared with the total number of processes (Ktotal). ; , , These are the corresponding weight values.

[0040] Within the first iteration, individuals with process detection efficiency values ​​greater than the first preset efficiency value are selected, and individuals with process detection efficiency values ​​greater than or equal to the second preset efficiency value and less than the first preset efficiency value are stored in the spare pool to form a spare population. The step of storing individuals whose process inspection efficiency value is greater than or equal to the second preset efficiency value and less than the first preset efficiency value in a backup database to form a backup population includes: Individuals whose process inspection efficiency value is greater than or equal to the second preset efficiency value and less than the first preset efficiency value are selected. Based on the selected individuals, calculate the distance between each individual; In this embodiment, the formula for calculating the distance between the individuals is as follows: ; In the formula, The distance between individuals in two detection sequences X and Y comprehensively reflects the degree of difference between the two detection sequences X and Y. To determine the process sequence difference, the process position difference between two sequences is calculated based on Hamming distance. Specifically, let the sequence be... , (n is the total number of processes, xi, yi are the process IDs of the i-th position), count the number of processes with different positions k (i.e. (number of i values) ; The risk sequence difference is calculated based on the distance between the risk priority ranking differences. Specifically, the risk priority number (RPN) of each process in the two sequences is extracted to generate the risk sequence. , ( For the RPN of the i-th process in sequence X, calculate the Spearman rank correlation coefficient between the two risk sequences. (Measure the consistency of the ranking, To be completely consistent, (for the exact opposite) Mapping ρ to the interval [0,1], the smaller ρ is, the better. The larger.

[0041] Compare the current iteration number with the number of the first iteration: When the number of iterations in this iteration is less than or equal to the first percentage of the number of iterations in the first iteration, individuals are sorted in descending order of distance value and the second-highest percentage are selected and stored in the reserve database. When the current iteration count is greater than the first percentage of the first iteration count but less than or equal to the first iteration count, individuals in the top third percentage, sorted by distance, are stored in the reserve pool. The second percentage is greater than the third percentage. In this embodiment, this strategy dynamically adjusts the selection ratio of the reserve population, achieving an iterative optimization transition from "diversity priority" to "quality focus," resolving the contradiction between "insufficient population diversity" and "slow convergence speed" in traditional fixed strategies. The setting of the second percentage being greater than the third percentage essentially dynamically adjusts the weight of "exploring new solutions" and "utilizing old solutions" during the iteration process: early stage emphasizes exploration (high proportion + high diversity), later stage emphasizes utilization (low proportion + focusing on quality). This balance enables the algorithm to stably output detection sequences that balance efficiency and rationality when facing different elevator scenarios (such as high / low floors, new / old elevators).

[0042] After exceeding the first iteration, the current population is combined with the backup population to form a new population, and individuals in the current population whose process detection rationality is greater than or equal to the preset rationality are selected. The selected population is used to obtain offspring populations through hybridization and variation strategies; It should be noted that the hybridization variation strategy includes: Based on the preset elevator inspection space area and the pre-determined dependencies of the inspection process, the individuals in the selected population are divided into 3 associated process partitions. The processes in each associated process partition satisfy the conditions of spatial continuity and no pre-determined dependency conflicts. Individuals are selected from the population and paired up. Three independent first random numbers are randomly generated for the three associated process partitions of individual one. The value of the random number is between 0 and 1. If the first random number is greater than the preset random threshold, then the associated process partition of individual one corresponding to the first random number is swapped with the associated process partition of individual two. If the first random number is less than or equal to a preset random threshold, a second random number between 0 and 1 is randomly generated for the associated process partition of individual one corresponding to the first random number. If the second random number is greater than the first random number, the associated process partition of individual one corresponding to the second random number undergoes mutation. Associative process partitions are divided according to the elevator inspection space area and its prerequisite dependencies to ensure that the process space within each partition is continuous and free of dependency conflicts. This avoids problems such as "cross-regional disorder" and "missing prerequisite processes" that do not conform to the actual operation logic of elevator inspection after hybridization and mutation, providing a compliant basis for subsequent gene exchange and mutation, and reducing the generation of invalid offspring sequences. The first random number determines whether a partition has hybridized, enabling differentiated gene exchange between different partitions and avoiding the destruction of high-quality process fragments caused by blind hybridization of the entire sequence. The second random number triggers mutation (and must satisfy "the second random number is greater than the first random number"), linking the mutation probability with the hybridization determination result. This preserves the stability of high-quality process sequences in the population while introducing new gene mutations to the population through random triggering, balancing the relationship between "utilizing existing high-quality solutions" and "exploring potential better solutions," and improving the global search capability for sequence optimization.

[0043] After obtaining the offspring population, the calculation and screening of the process detection efficiency value and process detection rationality continue until the preset number of iterations is met; The offspring population is output as the final detection sequence and fed back to the testing personnel.

[0044] When the inspectors receive the final test sequence, the system collects the inspectors' current location information and the test screen corresponding to the smart data acquisition terminal they are wearing. Based on the inspector's current location information and the inspection screen, it is determined whether the inspection procedure currently being performed by the inspector is correct. If incorrect, the inspector's current location and the inspection screen are visualized in the digital twin elevator model, and an alarm signal is issued.

[0045] It should be noted that, based on the inspector's current location information and the inspection screen, the system determines whether the inspection procedure currently being performed by the inspector is correct. If incorrect, the inspector's current location and the inspection screen are visualized in the digital twin elevator model, and an alarm signal is issued, including: Based on the final detection sequence, determine the node process information associated with each process node in the final detection sequence. The node process information includes the node's unique sequential number, theoretical execution area, and standard point characteristics.

[0046] Based on the current location information of the inspectors and the inspection screen, determine the current location coordinates of the inspectors and the actual point features in the inspection screen; Find the last completed process node in the final detection sequence, and determine the next process node corresponding to the last process node as the next execution node; Check whether the current position coordinates match the theoretical execution area of ​​the next execution node, and whether the matching degree between the actual point features and the standard point features of the next execution node is greater than the preset matching threshold. If the location matches and the matching degree meets the standard, the inspection procedure currently being performed by the inspector is considered correct; if any one of these conditions is not met, the inspection procedure currently being performed by the inspector is considered incorrect. If the positions do not match, the theoretical execution position of the next execution node is marked with a yellow dynamically flashing icon in the digital twin elevator model, and the current position of the inspector is marked with a red solid icon. The two are connected by a red dashed line and the straight-line distance is marked. If the matching degree does not meet the standard, a red text bubble will pop up at the current position of the inspector in the digital twin elevator model, indicating the current process node to be executed and the actual process node to be executed.

[0047] In this embodiment, if the location does not match or the matching degree is not up to standard, a dual-layer overlay display is used. The first layer is the location layer, marked with a red dynamically flashing icon to indicate the inspector's current location. The icon size is graded according to the positioning accuracy (2cm in diameter when the positioning error is ≤0.5m, and 3cm in diameter when the error is 0.5-1m). A yellow dotted line connects the current location and the "theoretical inspection location," indicating the straight-line distance between the two. The second layer is the image layer, displaying the real-time inspection image in a pop-up window next to the corresponding area of ​​the model. The lower right corner of the pop-up window indicates the image acquisition timestamp. At the same time, "standard point features" are selected with a green frame and "features missing / incorrect in the current image" are selected with a red frame, along with text prompts (such as "The current image does not match the machine room traction machine; the theoretical area should be the machine room area"). During visualization, the model view is automatically switched to the centered view of "current location - theoretical location," and the inspection management personnel can drag and zoom the model with the mouse to view detailed matching differences.

[0048] Optionally, in this embodiment, those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0049] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0050] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more electronic devices to execute all or part of the steps of the methods described in the various embodiments of this application.

[0051] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0052] In the several embodiments provided in this application, it should be understood that the disclosed application can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0053] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0054] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0055] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for optimizing elevator inspection procedures, characterized in that, The method includes: Collect real-time sensor data corresponding to the target elevator and preprocess the real-time sensor data corresponding to each sensor. Based on the preprocessed real-time sensor data, the digital twin elevator model corresponding to the target elevator is updated in real time, and the current operating status of the target elevator is output. Based on the current operating status, determine the points to be detected in the virtual elevator in the digital twin elevator model at the current moment, and map and encode each point to be detected as a specific detection procedure. According to the detection process, several candidate detection sequences are formed. Based on a preset sequence optimization algorithm, each candidate detection sequence is optimized, and the optimal detection sequence is selected as the final detection sequence and fed back to the inspection personnel. When the inspectors receive the final test sequence, the system collects their current location information and the test screen displayed on their smart data acquisition terminal. Based on the inspector's current location information and the inspection screen, it is determined whether the inspection procedure currently being performed by the inspector is correct. If incorrect, the inspector's current location and the inspection screen are visualized in the digital twin elevator model, and an alarm signal is issued.

2. The elevator inspection process optimization method according to claim 1, characterized in that, The preprocessing of real-time sensor data corresponding to each sensor includes: Extract the operating status parameters corresponding to the target elevator, and based on the operating status parameters, divide the elevator operation process into four dynamic intervals: starting acceleration segment, constant speed operation segment, deceleration and stopping segment, and stationary segment; the operating status parameters include the elevator's real-time position data, running direction, and speed change rate. Differentiated time alignment strategies are adopted for different dynamic ranges to align the real-time sensor data corresponding to each sensor. The real-time sensor data after alignment is denoised and missing data is filled in.

3. The elevator inspection process optimization method according to claim 2, characterized in that, The process of determining the points to be detected in the virtual elevator of the digital twin elevator model at the current moment, based on the current operating state, includes: Based on the current operating status of the digital twin elevator model and the real-time sensor data of each corresponding sensor, and based on the historical operating data stored in the database, a weighted fusion algorithm is used to calculate the dynamic risk index of each component in the digital twin elevator model. Based on the predefined risk factor weights corresponding to each component, the dynamic risk index is corrected to generate a comprehensive risk priority number, and all components to be inspected are sorted according to the risk priority number. Based on the risk priority ranking of each component, the component ranked in the preset position is selected as the high-risk component, and the specific geometric feature points corresponding to each high-risk component in the digital twin elevator model are determined as the corresponding detection points.

4. The elevator inspection process optimization method according to claim 3, characterized in that, The process of mapping and encoding each detection point into a specific detection procedure includes: Based on the current operating status of the digital twin elevator model corresponding to the target elevator and the corresponding historical operating data, a corresponding multi-dimensional dynamic point code is generated for each point to be detected. The dynamic point code includes at least a basic identity ID representing the component identity, a real-time risk label representing the real-time risk level, and a resource and environment label representing the site conditions. Based on the dynamic point code, and using a preset detection process knowledge base and process matching strategy, the corresponding detection process is matched for all dynamic point codes. The preset process matching strategy is as follows: In the inspection process knowledge base, search for a process template that perfectly matches the complete dynamic point code; If the match fails, the most suitable process template is selected from multiple candidate templates based on the value of one or more specific tags in the dynamic point code. If no suitable template is found, the corresponding smallest operation unit is selected from the preset process atomic operation library and dynamically combined according to the rules to generate a temporary detection process.

5. The elevator inspection process optimization method according to claim 4, characterized in that, The process of generating several candidate detection sequences according to the detection procedure, optimizing each candidate detection sequence based on a preset sequence optimization algorithm, selecting the optimal detection sequence as the final detection sequence, and feeding it back to the inspection personnel includes: According to the detection process, the positions of each sub-step in the detection process are randomly changed to form several candidate detection sequences, which serve as the corresponding initial population. Calculate the process detection efficiency value and process detection rationality corresponding to each individual in the population; Within the first iteration, individuals with process detection efficiency values ​​greater than the first preset efficiency value are selected, and individuals with process detection efficiency values ​​greater than or equal to the second preset efficiency value and less than the first preset efficiency value are stored in the spare pool to form a spare population. After exceeding the first iteration, the current population is combined with the backup population to form a new population, and individuals in the current population whose process detection rationality is greater than or equal to the preset rationality are selected. The selected population was used to obtain offspring populations through a hybridization and variation strategy. After obtaining the offspring population, the calculation and screening of the process detection efficiency value and process detection rationality continue until the preset number of iterations is met; The offspring population is output as the final detection sequence and fed back to the testing personnel.

6. The elevator inspection process optimization method according to claim 5, characterized in that, The step of storing individuals whose process inspection efficiency value is greater than or equal to the second preset efficiency value and less than the first preset efficiency value in a backup database to form a backup population includes: Individuals whose process inspection efficiency value is greater than or equal to the second preset efficiency value and less than the first preset efficiency value are selected. Based on the selected individuals, calculate the distance between each individual; Compare the current iteration number with the number of the first iteration: When the number of iterations in this iteration is less than or equal to the first percentage of the number of iterations in the first iteration, individuals are sorted in descending order of distance value and the second-highest percentage are selected and stored in the reserve database. When the current iteration number is greater than the first percentage of the first iteration number but less than or equal to the first iteration number, individuals in the top third percentage are selected and stored in the backup database after being sorted by distance value, and the second percentage is greater than the third percentage.

7. The elevator inspection process optimization method according to claim 6, characterized in that, The process involves determining whether the inspection procedure being performed by the inspector is correct based on the inspector's current location information and the inspection screen. If incorrect, the inspector's current location and the inspection screen are visualized in the digital twin elevator model, and an alarm signal is issued. This includes: Based on the final detection sequence, determine the node process information associated with each process node in the final detection sequence. The node process information includes the node's unique sequential number, theoretical execution area, and standard point characteristics. Based on the current location information of the inspectors and the inspection screen, determine the current location coordinates of the inspectors and the actual point features in the inspection screen; Find the last completed process node in the final detection sequence, and determine the next process node corresponding to the last process node as the next execution node; Check whether the current position coordinates match the theoretical execution area of ​​the next execution node, and whether the matching degree between the actual position features and the standard position features of the next execution node is greater than the preset matching threshold. If the location matches and the matching degree meets the standard, the inspection procedure currently being performed by the inspector is considered correct; if any one of these conditions is not met, the inspection procedure currently being performed by the inspector is considered incorrect. If the positions do not match, the theoretical execution position of the next execution node is marked with a yellow dynamically flashing icon in the digital twin elevator model, and the current position of the inspector is marked with a red solid icon. The two are connected by a red dashed line and the straight-line distance is marked. If the matching degree does not meet the standard, a red text bubble will pop up at the current position of the inspector in the digital twin elevator model, indicating the current process node to be executed and the actual process node to be executed.

8. The elevator inspection process optimization method according to claim 7, characterized in that, The hybridization variation strategy includes: Based on the preset elevator inspection space area and the pre-determined dependencies of the inspection process, the individuals in the selected population are divided into 3 associated process partitions. The processes in each associated process partition satisfy the conditions of spatial continuity and no pre-determined dependency conflicts. Individuals are selected from the population and paired up. Three independent first random numbers are randomly generated for the three associated process partitions of individual one. The value of the random number is between 0 and 1. If the first random number is greater than the preset random threshold, then the associated process partition of individual one corresponding to the first random number is swapped with the associated process partition of individual two. If the first random number is less than or equal to a preset random threshold, a second random number between 0 and 1 is randomly generated for the associated process partition of individual one corresponding to the first random number. If the second random number is greater than the first random number, the associated process partition of individual one corresponding to the second random number is mutated.

Images