An elevator operation state detection method and device, equipment, and storage medium

Abstract

The application provides an elevator operation state detection method and device, equipment and a storage medium. The method comprises the following steps: acquiring a first speed and a second speed, and determining a target speed based on the first speed and the second speed; acquiring a target acceleration; the target acceleration is an acceleration of a target elevator currently detected by an acceleration detection device; determining a target operation stage based on the target speed and the target acceleration; the target operation stage is an operation stage currently entered by the target elevator; acquiring a plurality of target distance information; each target distance information is distance information between a corresponding distance measuring sensor and the target elevator currently detected by the distance measuring sensor; and determining a current operation state of the target elevator based on the target operation stage, the target speed, the target acceleration and the plurality of target distance information. The application can improve the accuracy of elevator operation state detection.

Description

Technical Field

[0001] This application belongs to the field of elevator operation detection technology, and more specifically, relates to an elevator operation status detection method, device, equipment, and storage medium. Background Technology

[0002] As an indispensable vertical transportation tool in modern high-rise buildings, the safety and reliability of elevators are directly related to the life safety and property loss of passengers. Traditional elevator operation status detection mainly relies on periodic inspections or simple speed detection to determine the elevator's operating status. However, relying solely on speed sensors and corresponding alarm speed limits to detect the elevator's operating status yields inaccurate results.

[0003] Therefore, an accurate method for detecting elevator operating status is needed. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, equipment, and storage medium for detecting elevator operating status, so as to improve the accuracy of elevator operating status detection.

[0005] A first aspect of this application provides an elevator operation status detection method, applied to an elevator status detection system. The elevator status detection system includes: a control device, a first speed detection device, a second speed detection device, an acceleration detection device, and multiple distance sensors. The control device establishes communication connections with the first speed detection device, the second speed detection device, the acceleration detection device, and the multiple distance sensors. The elevator operation status detection method is executed by the control device and includes: A first speed and a second speed are acquired, and a target speed is determined based on the first speed and the second speed; the first speed is the speed of the target elevator currently detected by the first speed detection device; the second speed is the speed of the target elevator currently detected by the second speed detection device; the first speed detection device is installed on the car body of the target elevator; the second speed detection device is installed on the traction machine, which is used to provide power to the target elevator; Obtain the target acceleration; the target acceleration is the acceleration of the target elevator currently detected by the acceleration detection device. The target operating stage is determined based on the target speed and target acceleration; the target operating stage is the current operating stage of the target elevator. Acquire distance information for multiple targets; each target distance information is the distance between the corresponding ranging sensor and the target elevator, as detected by the corresponding ranging sensor. The current operating status of the target elevator is determined based on the target operating stage, target speed, target acceleration, and multiple target distance information.

[0006] A second aspect of this application provides an elevator operation status detection device. The elevator operation status detection device is installed in the control equipment of an elevator status detection system. The elevator status detection system further includes: a first speed detection device, a second speed detection device, an acceleration detection device, and multiple distance sensors. The control equipment establishes communication connections with the first speed detection device, the second speed detection device, the acceleration detection device, and the multiple distance sensors. The elevator operation status detection device includes: A speed acquisition module is used to acquire a first speed and a second speed, and determine a target speed based on the first speed and the second speed; the first speed is the speed of the target elevator currently detected by the first speed detection device; the second speed is the speed of the target elevator currently detected by the second speed detection device; the first speed detection device is installed on the car body of the target elevator; the second speed detection device is installed on the traction machine, which is used to provide power to the target elevator; The acceleration acquisition module is used to acquire the target acceleration; the target acceleration is the acceleration of the target elevator currently detected by the acceleration detection device. The operation phase determination module is used to determine the target operation phase based on the target speed and target acceleration; the target operation phase is the current operation phase of the target elevator. The distance acquisition module is used to acquire distance information for multiple targets; the distance information for each target is the distance information between the corresponding ranging sensor and the target elevator, which is detected by the corresponding ranging sensor. The operation status determination module is used to determine the current operation status of the target elevator based on the target operation stage, target speed, target acceleration, and multiple target distance information.

[0007] A third aspect of this application provides a control device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the elevator operation status detection method described above.

[0008] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the elevator operation status detection method described above.

[0009] The beneficial effects of the elevator operation status detection method, device, equipment, and storage medium provided in this application embodiment are as follows: This application employs a multi-detection device working collaboratively. A first speed detection device is installed on the target elevator car, and a second speed detection device is installed on the traction machine. By acquiring the first and second speeds detected by these two devices at different locations, and determining the target speed based on them, the error that may occur with a single detection device is reduced, improving the accuracy of speed detection. Simultaneously, by combining the target acceleration acquired by the acceleration detection device and the target distance information acquired by multiple ranging sensors, a comprehensive judgment of the elevator's operating status is made from multiple dimensions and parameters, further enhancing the reliability and comprehensiveness of the data and improving the accuracy of elevator operating status detection. Based on the acquired target speed and target acceleration, this application can determine the current target operating stage of the target elevator. Elevators have different operating characteristics in different operating stages, and accurately determining the operating stage is the foundation for further determining the elevator's operating status. By clearly defining the operating stage and combining the target speed, target acceleration, and multiple target distance information, the elevator's operating status at each stage can be analyzed more accurately, and abnormal states can be detected in a timely manner, such as abnormal speed during the start-up stage or abnormal acceleration during the deceleration stage, thereby effectively improving the accuracy of judging the elevator's operating status. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A flowchart illustrating an elevator operation status detection method according to an embodiment of this application; Figure 2 A schematic diagram of the operation of a target elevator provided in one embodiment of this application; Figure 3 This is a structural block diagram of an elevator operation status detection device provided in one embodiment of this application; Figure 4 This is a schematic block diagram of a control device provided in an embodiment of this application. Detailed Implementation

[0012] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0013] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0014] This application provides an embodiment of an elevator operation status detection method. This method is applied to an elevator status detection system, which includes: a control device, a first speed detection device, a second speed detection device, an acceleration detection device, and multiple distance sensors. The control device establishes communication connections with the first speed detection device, the second speed detection device, the acceleration detection device, and the multiple distance sensors. The elevator operation status detection method is executed by the control device. Figure 1 As shown, the elevator operation status detection method may include: S101-S105.

[0015] S101: Obtain a first speed and a second speed, and determine a target speed based on the first speed and the second speed; the first speed is the speed of the target elevator currently detected by the first speed detection device; the second speed is the speed of the target elevator currently detected by the second speed detection device; the first speed detection device is installed on the car body of the target elevator; the second speed detection device is installed on the traction machine, which is used to provide power to the target elevator.

[0016] Figure 2 This application provides a schematic diagram of the operation of a target elevator according to an embodiment; please refer to... Figure 2 In this embodiment, the first speed detection device can be a laser Doppler velocimeter, which can be installed on the target elevator car, specifically on the outer top of the target elevator. The target elevator refers to the elevator currently being detected and monitored, and is the object of the entire detection process. The second speed detection device can be a traction machine spindle encoder, specifically installed on the traction machine spindle. The traction machine is the power unit of the target elevator, which can drive the elevator car to rise and fall by driving components such as traction ropes, providing power support for the operation of the elevator.

[0017] In this embodiment, the elevator operation status detection method further includes: obtaining a target signal-to-noise ratio; the target signal-to-noise ratio is the signal-to-noise ratio when the first speed detection device detects the speed of the target elevator. Further, in this embodiment, the target speed can be determined based on a first speed and a second speed in the following manner: determining the weight of the first speed and the weight of the second speed based on the target signal-to-noise ratio; performing a weighted calculation based on the first speed, the weight of the first speed, the second speed, and the weight of the second speed to obtain the target speed.

[0018] In this embodiment, the laser Doppler velocimetry device can directly measure the speed of a moving object by measuring the frequency change of the reflected light after the laser shines on it, while the traction machine spindle encoder can indirectly calculate the speed by recording the rotation angle and rotation speed of the traction machine spindle.

[0019] The target signal-to-noise ratio (SNR) refers to the ratio of the effective signal strength to the noise signal strength when a laser Doppler velocimeter detects speed. A higher SNR indicates less interference and higher reliability in laser velocimetry results; conversely, a low SNR suggests potential measurement errors. If the target SNR is high, meaning laser velocimetry is more reliable, the weight of the first speed is greater, and the weight of the second speed is smaller. Conversely, if the target SNR is low, meaning laser velocimetry is unreliable, the weight of the first speed is reduced, and the weight of the second speed is increased. In this case, indirect speed measurement via the traction machine spindle encoder becomes more crucial.

[0020] In this embodiment, the first velocity weight and the second velocity weight can be determined based on a preset mapping relationship or a linear function based on the target signal-to-noise ratio. Specifically, the preset mapping relationship is the mapping relationship between the target signal-to-noise ratio and the first velocity weight, because the target signal-to-noise ratio is only related to the first velocity detection device, and the second velocity weight is determined based on the first velocity weight. The sum of the first velocity weight and the second velocity weight is 1. The preset mapping relationship can be determined based on multiple experiments.

[0021] S102: Obtain the target acceleration; the target acceleration is the acceleration of the target elevator currently detected by the acceleration detection device.

[0022] In this embodiment, the target acceleration can be a specific value obtained by the acceleration detection device after detecting the acceleration of the target elevator during its current movement. It reflects the rate and direction of change in the target elevator's current speed; for example, acceleration is positive when accelerating upwards and negative when decelerating upwards. The acceleration detection device can be an acceleration sensor, which is essentially a triaxial acceleration sensor, such as... Figure 2 As shown, the acceleration detection device can be installed inside the target elevator to detect the acceleration changes of the target elevator in real time during its operation.

[0023] S103: Determine the target operating stage based on the target speed and target acceleration; the target operating stage is the current operating stage of the target elevator.

[0024] In this embodiment, the target operating stage can be the specific operating state stage that the target elevator is currently in, such as the stationary stage, the starting acceleration stage, the constant speed stage, or the deceleration stage.

[0025] Specifically, determining the target's operational phase based on target velocity and target acceleration includes: In response to the target velocity being zero and the target acceleration being zero, the target's running phase is determined to be the stationary phase; In response to the target velocity being greater than zero, the target acceleration being greater than zero, and the target velocity increasing over time, the target running phase is determined to be the initiation acceleration phase. In response to the target velocity being greater than zero, the absolute value of the target acceleration being less than a preset acceleration threshold, and the change in the target velocity over time being less than a preset velocity change threshold, the target running phase is determined to be a uniform velocity phase. In response to the target velocity being greater than zero, the target acceleration being less than zero, and the target velocity decreasing over time, the target's running phase is determined to be the deceleration phase.

[0026] In this embodiment, the target speed is zero and the target acceleration is zero: this means that the current speed value of the elevator is 0 (no movement) and the acceleration value is also 0 (no trend of speed change). This is the condition for determining that the elevator is in a stationary phase.

[0027] Stationary Phase: One of the current operating phases of the target elevator, where the elevator is completely stationary, showing no speed or tendency to accelerate or decelerate. Acceleration Phase: One of the operating phases of the target elevator, referring to the process of the elevator starting from a stationary state and gradually increasing its speed from zero. Constant Speed ​​Phase: One of the operating phases of the target elevator, referring to the process of the elevator continuously running at a relatively stable speed with minimal speed changes. Deceleration Phase: One of the operating phases of the target elevator, referring to the process of the elevator gradually decreasing its speed from a stable operating state until it finally comes to a stop.

[0028] The target speed is greater than zero, the target acceleration is greater than zero, and the target speed increases over time: The elevator currently has a positive speed and positive acceleration (the speed is increasing), and the speed value is constantly increasing over time, which is the condition for determining the start-up acceleration phase.

[0029] The elevator is currently moving, but the magnitude (absolute value) of the acceleration is very small (below the set acceleration threshold), and the change in speed over time is also very small (below the set speed change threshold), meaning the speed remains basically stable. This is the condition for determining the uniform speed phase. The preset acceleration threshold and preset speed change threshold can be determined based on multiple experiments or experience.

[0030] The target velocity is greater than zero, the target acceleration is less than zero, and the target velocity decreases over time: The elevator currently has velocity (it is moving), but the acceleration is negative (the velocity is decreasing), and the velocity value is constantly decreasing over time. This is the condition for judging the deceleration phase.

[0031] Target operating phase: As mentioned above, this refers to the specific operating phase that the target elevator is currently in, including the stationary phase, acceleration phase, constant speed phase, or deceleration phase, etc.

[0032] S104: Acquire distance information for multiple targets; each target distance information is the distance between the corresponding ranging sensor and the target elevator, detected by the corresponding ranging sensor.

[0033] In this embodiment, as Figure 2 As shown, the ranging sensor can specifically be an ultrasonic ranging sensor, which can be installed on the inner wall of the elevator shaft. For example, multiple ranging sensors can be fixed near specific floors within the shaft or distributed vertically along the shaft, facing the car, to detect the real-time distance between themselves and the car. Multiple target distance information can characterize the current specific position of the target elevator within the shaft, and whether its trajectory during operation conforms to a preset path.

[0034] S105: Determine the current operating status of the target elevator based on the target operating stage, target speed, target acceleration, and multiple target distance information.

[0035] In this embodiment, the target running stage, target speed, target acceleration, and multiple target distance information can be input into a random forest model or a neural network model to obtain the current running state of the target elevator. It should be noted that the aforementioned random forest model or neural network model should be a trained model.

[0036] As can be seen from the above, the embodiments of this application employ a collaborative approach using multiple detection devices. A first speed detection device is installed on the target elevator car, and a second speed detection device is installed on the traction machine. By acquiring the first and second speeds detected by these two devices at different locations, and determining the target speed based on them, the potential errors of a single detection device are reduced, improving the accuracy of speed detection. Simultaneously, by combining the target acceleration acquired by the acceleration detection device with the target distance information acquired by multiple ranging sensors, a comprehensive judgment of the elevator's operating status is made from multiple dimensions and parameters, further enhancing the reliability and comprehensiveness of the data and improving the accuracy of elevator operating status detection. Based on the acquired target speed and target acceleration, the embodiments of this application can determine the current target operating stage of the target elevator. Elevators exhibit different operating characteristics at different operating stages; accurately determining the operating stage is fundamental to further determining the elevator's operating status. By clearly defining the operating stage and combining the target speed, target acceleration, and multiple target distance information, the elevator's operating status at each stage can be analyzed more accurately, and abnormal states can be detected in a timely manner, such as abnormal speed during the start-up phase or abnormal acceleration during the deceleration phase, thereby effectively improving the accuracy of judging the elevator's operating status.

[0037] In one embodiment of this application, the elevator status detection system further includes: multiple pressure sensors; the multiple pressure sensors are disposed at the bottom of the target elevator car to measure the pressure value at the bottom of the target elevator and obtain multiple target pressure data; wherein, the elevator operation status detection method further includes: Acquire pressure data from multiple targets; The determination of the target elevator's operating status, based on the target operating stage, target speed, target acceleration, and multiple target distance information, includes: In response to the fact that multiple target pressure data meet the normal pressure distribution conditions, the target operating stage, target speed, target acceleration, multiple target distance information and multiple target pressure data are input into the target random forest model, and the N decision trees in the target random forest model are called to make decisions to obtain the operating status of the target elevator; where N is a positive integer; the normal pressure distribution conditions include: the difference between the largest pressure data and the smallest pressure data among multiple target pressure data is less than a preset pressure threshold; In response to multiple target pressure data not meeting the normal pressure distribution conditions, the target operating stage, target speed, target acceleration, multiple target distance information, and multiple target pressure data are input into the target random forest model, and the M decision trees in the target random forest model are called to make decisions to obtain the operating status of the target elevator; where M is a positive integer, M is greater than N.

[0038] In this embodiment, as Figure 2 As shown, multiple pressure sensors, specifically four pressure sensors, are distributed at the four corners of the bottom of the target elevator to comprehensively capture the pressure distribution at the bottom of the elevator (only two pressure sensors are shown in the figure because...). Figure 2 This is a two-dimensional diagram with no depth; it can be understood as being in... Figure 2 The two pressure sensors shown are those visible from the current perspective (the two pressure sensors mentioned above obscure the pressure sensors behind them). The target pressure data is the pressure value measured by the pressure sensors at the corresponding location at the bottom of the elevator. The normal pressure distribution condition is a standard used to determine whether the pressure distribution at the bottom of the elevator is balanced. Specifically, it refers to the difference between the maximum and minimum pressure values ​​among multiple target pressure data being less than a preset pressure threshold. For example, if the maximum pressure is 1000N and the minimum is 800N, the difference of 200N is less than the preset pressure threshold of 500N, then the condition is met, indicating that the load distribution is relatively uniform (e.g., passengers are evenly standing). The preset pressure threshold can be determined based on the target elevator's manufacturer's instructions or set based on experience.

[0039] The target random forest model contains multiple decision trees. By combining the independent decisions of these trees, the accuracy of the judgment is improved. It's important to note that the target random forest model is a pre-trained model. When multiple target stress data points meet the normal stress distribution conditions, a smaller number of decision trees (N trees) can be used from the target random forest model to reduce data processing volume, increase processing speed, and save computational resources. N is a positive integer. Because the judgment difficulty is low under normal circumstances, excessive computational resources are not required. When multiple target stress data points do not meet the normal stress distribution conditions, a larger number of decision trees (M trees) can be used from the target random forest model. M is a positive integer, M>N, and M is not greater than the total number of decision trees in the target random forest model. Because abnormal situations (such as unbalanced loading or centroid shift) may be more complex, more decision trees are needed for detailed analysis to improve the reliability of the judgment.

[0040] In this embodiment, when multiple target pressure data meet the normal pressure distribution conditions, the target operating stage, target speed, target acceleration, multiple target distance information, and multiple target pressure data are input into the target random forest model, and a small number of decision trees (N trees) are called for decision-making. Since the load distribution is relatively uniform under normal conditions, the judgment difficulty is low. Calling fewer decision trees can reduce the consumption of computing resources and improve detection efficiency while ensuring judgment accuracy, thus achieving a fast and accurate determination of the target elevator's operating status. When multiple target pressure data do not meet the normal pressure distribution conditions, a larger number of decision trees (M trees, M>N and M is not greater than the total number of decision trees in the target random forest model) are called for decision-making. Under abnormal conditions (such as uneven load or center of gravity shift), the elevator's operating status may be more complex, requiring more in-depth analysis and judgment. Increasing the number of decision trees called can fully utilize the advantages of the random forest model. Through the independent decision-making of multiple trees and the comprehensive results, a more detailed analysis of abnormal conditions can be performed, thereby improving the reliability of elevator operating status judgment and timely detection of potential safety hazards in elevators under abnormal load conditions.

[0041] In one embodiment of this application, the target operating stage, target speed, target acceleration, multiple target distance information, and multiple target pressure data are input into a target random forest model, and M decision trees in the target random forest model are invoked to make decisions to obtain the operating state of the target elevator, including: Determine the target acceleration level corresponding to the target acceleration; Input the target's operational phase, target velocity, target acceleration, multiple target distance information, and multiple target pressure data into the target random forest model, and call the M decision trees in the target random forest model to make decisions, and obtain M decision results; Responding to whether the target operation phase is an acceleration phase or a deceleration phase, for each decision outcome, the weight of the decision tree is determined based on the number of samples of the first target sample and the number of samples of the second target sample in the target training sample set. The target training sample set is the training sample set used by the decision tree corresponding to the decision outcome. A training sample set contains multiple training samples, and a training sample includes: a sample parameter set, which includes: operation phase samples, velocity samples, acceleration samples, multiple sample distance information, and multiple sample pressure data; the first target sample is the operation phase sample in the target training set, and the second target sample is the acceleration sample in the target training set that belongs to the target acceleration level; the acceleration phase is an operation phase where the target velocity is greater than zero, the target acceleration is greater than zero, and the target velocity increases over time; the deceleration phase is an operation phase where the target velocity is greater than zero, the target acceleration is less than zero, and the target velocity decreases over time. The operating status of the target elevator is obtained by weighting each decision result and its corresponding weight.

[0042] In this embodiment, continuous acceleration values ​​are divided into different intervals, each interval corresponding to a level. The target acceleration is judged to determine which interval it belongs to, and the level corresponding to that interval is taken as the target acceleration level. The target training sample set refers to the historical data set used to train each decision tree, containing multiple training samples. Each sample contains a set of parameters: running phase sample, velocity sample, acceleration sample, multiple sample distance information, and multiple sample pressure data, as well as the known running status label corresponding to this set of parameters. The running status can be normal, minor fault, or fault.

[0043] In this embodiment, the first target sample refers to a sample in the training sample set whose operating phase is consistent with the current target elevator operating phase (i.e., the target operating phase). The second target sample is a sample in the training sample set whose acceleration belongs to the current target acceleration level (i.e., the target acceleration level).

[0044] Typically, when a random forest model outputs its final result, it uses the decision tree with the highest number of votes as the output. In this process, each decision tree has the same weight in the votes; that is, the random forest model obtains its output by weighting the votes. Because each decision tree in a random forest model selects its training data with replacement from a given training dataset, each decision tree should have a preferred application scenario. In other words, the more training data a decision tree has in a particular scenario or interval, the more reliable its output will be in that scenario or interval.

[0045] Based on this, the embodiments of this application consider that when the target elevator is in the start-up acceleration phase or deceleration phase, the acceleration changes significantly, and the stability requirements of the elevator are higher in these two phases, requiring more accurate assessment of the operating status. Therefore, when the target operating phase is determined to be the start-up acceleration phase or deceleration phase, the embodiments of this application determine the weight of the decision tree based on the number of samples of the first target sample and the number of samples of the second target sample in the target training sample set. That is, the more samples in the training set of the decision tree that are consistent with the current operating phase, the richer the experience of the tree in judging the current phase, and the higher the weight. Similarly, the more samples in the training set of the decision tree that are consistent with the current acceleration level, the more reliable the tree's judgment of the current acceleration state, and the higher the weight.

[0046] Specifically, the first adjustment weight can be determined based on the number of samples of the first target sample in the target training sample set, the second adjustment weight can be determined based on the number of samples of the second target sample in the target training sample set, the target adjustment weight can be determined based on the first adjustment weight and the second adjustment weight, and the default weight of the decision tree can be adjusted based on the target adjustment weight to determine the weight of the decision tree.

[0047] In this embodiment, the two functional relationships can be determined based on experience or multiple experiments. For example, they can be linear functional relationships, where the intercept and slope can be obtained through fitting or experimentation. The two functional relationships described above represent the relationship between the number of first target samples in the target training sample set and the first adjusted weight, and the relationship between the number of second target samples in the target training sample set and the second adjusted weight, respectively. The sum of the first and second adjusted weights is the third adjusted weight. The first and second adjusted weights are numerical values ​​with positive and negative signs because they are both determined based on the aforementioned linear functions, thus naturally containing signs; a positive sign indicates an increase, and a negative sign indicates a decrease. The default weight is 1 / the total number of decision trees in the target random forest model.

[0048] For example, suppose the target elevator is currently in the acceleration phase, and its acceleration level is detected to be medium (e.g., acceleration between 0.5-1.0 m / s²). In this case, it is necessary to call M decision trees (e.g., 3 trees: tree A, tree B, and tree C; this is just an example, and more decision trees should be used in actual applications) from the target random forest model to make decisions and assign weights to each tree.

[0049] The training sample set for tree A contains 1000 training samples, of which 800 are first-target samples (samples in the startup acceleration phase) and 600 are second-target samples (samples at the medium acceleration level). This indicates that tree A has very rich training data in the startup acceleration phase and medium acceleration scenarios, and has more experience in judging the current elevator state.

[0050] The training sample set for tree B contains 1000 training samples, of which 300 are for the first target and 200 are for the second target. This indicates that there is limited training data for tree B in the current scenario, and relevant experience is insufficient.

[0051] The training sample set for tree C contains 1000 training samples, of which 500 are first target samples and 400 are second target samples. Its experience richness is between that of tree A and tree B.

[0052] The more primary and secondary target samples there are in the training samples, the higher the weight of the decision tree. Therefore: Tree A has the highest weight (because it has the most relevant samples); Tree C has the second highest weight; and Tree B has the lowest weight.

[0053] Ultimately, the decision result of tree A carries more weight in the final judgment, while that of tree B carries the least weight. This weight adjustment allows the model to rely more on decision trees with extensive experience in similar scenarios during the startup acceleration phase, where stability is critical, thereby improving the accuracy of the judgment.

[0054] Specifically, the weights of the decision tree should be determined based on the ratio of the number of samples in the first target sample to the number of parameter sets in the decision tree, and the ratio of the number of samples in the second target sample to the number of parameter sets in the decision tree, since the number of parameter sets in the target training sample set corresponding to each decision tree may not be the same. More specifically, a first weight adjustment amount should be determined based on the ratio of the number of samples in the first target sample to the number of parameter sets in the decision tree, and the default weight of the decision tree should be increased based on the first weight adjustment amount to obtain the weight of the decision tree.

[0055] In this embodiment, the default weight of each decision tree is 1 / M. That is, when the target operation phase is not the startup acceleration phase or the deceleration phase, the weights of each decision tree in the target random forest model should be the same when performing weighted calculations, all being 1 / M. The process of determining the first weight adjustment amount based on the aforementioned two proportions will not be elaborated in this embodiment; those skilled in the art can set it based on multiple experiments or experience. By weighting and averaging the M decision results according to their respective weights, the final operation status judgment result can be obtained. The result can be the result with the highest proportion after weighted calculation, or the weighted calculation result can be directly used as the output, for example, 80% probability of normal operation and 20% probability of minor fault.

[0056] As can be seen from the above, this embodiment considers that the acceleration and deceleration phases are critical stages in the operation process with extremely high stability requirements, characterized by large and complex acceleration variations. Therefore, this embodiment designs a special decision weight determination method for these two stages, taking into account their unique characteristics in elevator operation. Unlike the traditional random forest model where all decision trees have the same weight, the above-mentioned targeted processing enables the model to more accurately assess the elevator's operating status during critical stages, promptly identify potential safety hazards, and effectively avoid misjudgments or omissions caused by large acceleration variations, thereby improving the accuracy and reliability of elevator operating status assessment. The more samples in the decision tree's training set that are consistent with the current operating stage, the richer the tree's experience in judging the current stage; the more samples that are consistent with the current acceleration level, the more reliable the tree's judgment of the current acceleration state. Determining weights based on these two factors allows the decision tree, which is better at handling the current scenario, to play a greater role in the decision-making process, thereby improving the overall decision accuracy. For example, if most of the samples in the training set of a decision tree were collected during the initial acceleration phase and when the acceleration is at the current level, then when the elevator is in a similar operating phase and acceleration level, the weight of the decision tree will be increased accordingly, and its decision result will have a greater impact on the final operating state judgment, thereby improving the accuracy of the judgment.

[0057] In one embodiment of this application, the process of determining the target random forest model includes: acquiring a total training data set and a total validation data set; training P initial decision trees based on the total training data set to obtain P trained decision trees; where P is a positive integer; P is less than N and P is less than M; performing performance evaluation on the P trained decision trees to obtain performance evaluation results; the performance evaluation results include: the average validation error of the P trained decision trees on the total validation data set, and the average split gain of each feature in the P trained decision trees; each feature includes: running stage, speed, acceleration, multiple distance information, and multiple stress data; determining the parameters to be adjusted and the corresponding values ​​of the parameters to be adjusted based on the performance evaluation results, and adjusting the values ​​of the parameters to be adjusted to obtain target training parameter values; training Q initial decision trees based on the target training parameter values ​​and the training dataset to obtain Q trained decision trees; and combining the P trained decision trees with the Q trained decision trees to obtain the target random forest model.

[0058] In this embodiment, in addition to the above-mentioned method of first training a batch of decision trees and then determining the training of the next batch of decision trees based on the performance of the previous batch of decision trees to finally obtain the target random forest model, the random forest can also be trained based on the conventional training method of the random forest model to obtain the target random forest model.

[0059] In this embodiment, the target random forest model refers to the final random forest model used to determine the elevator's operating status. It is composed of multiple decision trees and is an optimized model obtained through training, evaluation, and adjustment. The training data set is the historical data set used to train the decision trees, containing a large number of elevator operation samples. The validation data set is an independent data set used to evaluate the model's performance; it does not participate in training and its purpose is to test the model's ability to judge unknown data. The results of analyzing the performance of the P initial decision tree models, generated in the first stage of training, include two parts: average validation error and average split gain. The average empirical error is the average error rate of the P trees on the validation data, reflecting the overall accuracy of the model (the lower the error, the better the performance). The average split gain is the average gain value of each feature (such as speed, acceleration) when it is used as a splitting feature (a key feature used to divide samples) in the P trees. The higher the gain, the greater the influence (more important) of the feature on the decision.

[0060] In this embodiment, the parameters to be adjusted during the training of the decision tree are the parameters that can be optimized, such as the feature selection probability or the maximum depth of the tree, etc. The values of the parameters to be adjusted are the specific values of the feature selection probability and the maximum depth of the tree. The target training parameter values are the final parameter values obtained by optimizing the values of the parameters to be adjusted, and are used to guide the subsequent training of the decision tree. The Q initial decision trees and the new decision trees trained based on the target training parameter values are a supplement and optimization to the P initial trees, and are used to improve the overall performance of the model.

[0061] The construction of the target random forest model in this embodiment is an iterative process of preliminary exploration, evaluation and optimization, supplementary training, and combination and shaping. The specific steps are as follows: First, collect the total training data set and the total validation data set. Use the total training data set to train P initial decision trees (the number is small, P < N and P < M). This step is to quickly obtain a basic model and explore the basic laws of the data; performance evaluation can discover the shortcoming of the model. Use the total validation data set to evaluate the performance of the P trees: judge whether the model is accurate through the average validation error. If the error is high, it means that the model's judgment ability is poor; judge which features (such as speed, pressure data) are more important for decision-making through the average feature split gain. Features with low gain may be ignored, affecting the model's judgment. Secondly, optimize targeted according to the evaluation results, determine the parameters that need to be adjusted, such as increasing the weight of high-gain features and restricting the depth of the tree to avoid overfitting, to obtain the target training parameter values, and prepare for training better decision trees in the future. Thirdly, supplementary training to strengthen the model's ability. Use the target training parameter values and the total training data set to train Q new decision trees. Optimize according to the shortcoming found in the previous step. Finally, combine and shape to generate the final model. Combine the initially trained P trees with the supplementary trained Q trees to form the target random forest model. That is, the generation of subsequent decision trees is guided by the evaluation of the first P trees. In this embodiment, M, N, P, and Q can be set based on preferences, but basic decision-making requirements need to be met. The sum of P and Q is the total number of decision trees in the target random forest algorithm, that is, neither M nor N can exceed the sum of P and Q.

[0062] As can be seen from the above, the embodiments of this application adopt a method of first training a small number of initial decision trees (P trees) as the basic model to quickly explore the basic patterns of the data, avoiding the resource waste and inefficiency problems that may be caused by large-scale training at the beginning. On this basis, the model's shortcomings are identified through performance evaluation, and then the parameters are adjusted and supplemented (Q decision trees are trained) in a targeted manner. Finally, the two parts of decision trees are combined into a target random forest model, which can make full use of the training results at each step, gradually optimize the model performance, and make the final target random forest model more accurate and reliable. The performance of the trained P decision trees is evaluated using the validation data set to obtain key indicators such as average validation error and average split gain of each feature. The average validation error reflects the overall accuracy of the model. If the error is high, it indicates that the model's judgment ability is poor and further optimization is needed. The average split gain reveals the importance of each feature to the decision. Features with low gain may be ignored, affecting the model's judgment. Based on these evaluation results, the parameters to be adjusted and their corresponding values ​​are determined, and targeted optimization is carried out to enable the model to better adapt to the data characteristics and improve its ability to judge the elevator's operating status. For example, if the average split gain of a certain feature (such as acceleration) is found to be low, but this feature is important for judging the elevator's operating status based on practical experience, the weight of this feature in the decision tree can be increased by adjusting the relevant parameters, thereby enhancing the model's ability to utilize this feature.

[0063] In one embodiment of this application, the parameters to be adjusted include feature selection parameters and tree structure parameters; Based on the performance evaluation results, determine the parameters to be adjusted and their corresponding values, and adjust the values ​​of the parameters to be adjusted, including: In response to the fact that the average split gain of at least one feature is greater than a preset gain threshold, the feature selection parameter value corresponding to that feature is adjusted to an enhanced state, wherein the probability of that feature being selected as a splitting feature in the training of Q decision trees is increased by a first preset proportion. In response to the average verification error exceeding a preset error threshold, the maximum depth parameter value in the tree structure parameters is adjusted to a reduced state. The reduced state is that the maximum depth of Q decision trees is limited to a second preset ratio of the average depth of P initial decision trees.

[0064] In this embodiment, the tree structure parameter can be the maximum depth parameter, which is the maximum number of layers the decision tree is allowed to grow. Excessive depth may lead to overcomplexity, i.e., overfitting, where the tree only remembers the training data and has poor generalization ability; conversely, insufficient depth may fail to capture data patterns, i.e., underfitting. The feature selection parameter controls the probability of the decision tree selecting input features (such as running stage, speed, stress data, etc.) during training. For example, the higher the selection parameter of a feature, the greater the likelihood that it will be selected as a splitting feature during training. Splitting features are the key features used to divide samples.

[0065] In this embodiment, the average split gain is an indicator that measures the degree to which a feature improves the classification performance of the decision tree. A higher split gain indicates that when using that feature to divide samples, it can more significantly separate samples from different operating states. For example, a high gain for stress data means it is important for determining bias. The preset gain threshold is a preset critical value for determining the importance of a feature. If the average split gain of a feature exceeds this threshold, it indicates that its contribution to the decision is significant and requires close attention. The enhancement state is an adjustment result of the feature selection parameters, referring to increasing the probability of a specific feature, such as a feature with an average split gain exceeding the threshold, being selected as a splitting feature by a first preset proportion during the training of Q decision trees, thereby enhancing the feature's influence on the model.

[0066] In this embodiment, the average validation error is the average error rate of the P initial decision trees on the validation data set, reflecting the model's accuracy in predicting unknown data. The higher the error, the less reliable the model. The preset error threshold is a preset critical value for judging the model's accuracy. If the average validation error exceeds this threshold, it indicates that the model performance is substandard and the tree structure parameters need to be adjusted. The "reduction state" refers to the adjustment result of the maximum depth parameter in the tree structure parameters, specifically limiting the maximum depth of the Q decision trees to a second preset proportion of the average depth of the P initial decision trees to simplify the tree structure and avoid overfitting. The first and second preset proportions can be determined based on experience or multiple experiments.

[0067] The core of parameter tuning is to optimize the training rules for the subsequent Q decision trees based on the performance defects of the initially trained P decision trees. The specific process is as follows: Define the objects to be adjusted: The parameters to be adjusted are divided into feature selection parameters and tree structure parameters.

[0068] Feature selection parameter adjustment logic: If the average split gain of a certain feature is greater than the preset gain threshold, it means that the feature is crucial to the decision. In this case, the feature selection parameter value is set to an enhanced state, that is, when training Q trees, the probability of the feature being selected as the splitting feature is increased, for example, from 30% to 50%. The purpose is to make the final target random forest model pay more attention to high-value features and enhance its ability to capture key information.

[0069] Tree structure parameter adjustment logic: If the average validation error of P trees exceeds the preset error threshold, it indicates that the model's prediction accuracy for unknown data is poor, and overfitting may be present. Therefore, the maximum depth parameter is reduced, effectively limiting the maximum depth of the Q trees. For example, the original average depth of 10 layers is reduced to 8 layers. The aim is to simplify the tree structure, reduce overfitting of the model to the details of the training data, and improve generalization ability.

[0070] It should be noted that when the average split gain without features is greater than the preset gain threshold and the average verification error is less than or equal to the preset error threshold, it indicates that the training results of the first P initial decision trees are good. In this case, there is no need to intervene in the generation of the subsequent Q decision trees. That is, there is no need to determine the parameters to be adjusted and their corresponding values, and to adjust the values ​​of the parameters to be adjusted. Training can be performed based on the default values ​​or the parameters used when training the first P initial decision trees.

[0071] As can be seen from the above, the embodiment of this application can adjust the feature selection parameter corresponding to the identified high-value features to an improved state, increasing the probability that it will be selected as a splitting feature in the training of Q decision trees. For example, the probability of this feature being selected as a splitting feature is increased from 30% to 50%. This adjustment makes the model pay more attention to these key features during the subsequent construction of decision trees, and can more effectively use the information they contain to make decisions, thereby significantly improving the accuracy of the target random forest model in judging the elevator's operating status. In this embodiment of the application, based on the average validation error and the preset error threshold, it is possible to determine whether the model has an overfitting problem. The average validation error reflects the model's prediction accuracy for unknown data. When the average validation error of the P initial decision trees is greater than the preset error threshold, it indicates that the model performs well on the training data, but its prediction ability is poor when facing new and unseen data, and overfitting is likely to occur. This judgment method based on actual validation results provides a reliable basis for subsequent tree structure optimization. When it is determined that there is a risk of overfitting, this embodiment of the application adjusts the maximum depth parameter in the tree structure parameters to a reduced state, limiting the maximum depth of the Q decision trees. For example, the maximum depth can be limited to a second preset proportion of the average depth of the P initial decision trees, such as limiting the original average depth of 10 layers to 8 layers. By simplifying the tree structure, the overfitting of the target random forest model to the details of the training data is reduced, enabling the model to better capture the general patterns in the data, thereby improving the generalization ability of the target random forest model and enabling it to make accurate judgments when faced with elevator operation data in different scenarios.

[0072] Corresponding to the elevator operation status detection method in the above embodiment, Figure 3 This is a structural block diagram of an elevator operation status detection device according to an embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 3The elevator operation status detection device 20 is installed in the control equipment of the elevator status detection system. The elevator status detection system also includes: a first speed detection device, a second speed detection device, an acceleration detection device, and multiple distance sensors. The control equipment establishes communication connections with the first speed detection device, the second speed detection device, the acceleration detection device, and the multiple distance sensors. The elevator operation status detection device 20 includes: a speed acquisition module 21, an acceleration acquisition module 22, an operation stage determination module 23, a distance acquisition module 24, and an operation status determination module 25.

[0073] The speed acquisition module 21 is used to acquire a first speed and a second speed, and determine a target speed based on the first speed and the second speed; the first speed is the speed of the target elevator currently detected by the first speed detection device; the second speed is the speed of the target elevator currently detected by the second speed detection device; the first speed detection device is installed on the car body of the target elevator; the second speed detection device is installed on the traction machine, which is used to provide power to the target elevator; The acceleration acquisition module 22 is used to acquire the target acceleration; the target acceleration is the acceleration of the target elevator currently detected by the acceleration detection device. The operation phase determination module 23 is used to determine the target operation phase based on the target speed and target acceleration; the target operation phase is the current operation phase of the target elevator. The distance acquisition module 24 is used to acquire distance information of multiple targets; the distance information of each target is the distance information between the corresponding ranging sensor and the target elevator, which is detected by the corresponding ranging sensor. The operation status determination module 25 is used to determine the current operation status of the target elevator based on the target operation stage, target speed, target acceleration and multiple target distance information.

[0074] In one embodiment of this application, the elevator status detection system further includes: multiple pressure sensors; the multiple pressure sensors are disposed at the bottom of the target elevator car to measure the pressure value at the bottom of the target elevator and obtain multiple target pressure data; The elevator operation status detection device 20 also includes: a pressure acquisition module, used to acquire multiple target pressure data; The operating status determination module 25 is specifically used to respond to the fact that multiple target pressure data meet the normal pressure distribution conditions. It inputs the target operating stage, target speed, target acceleration, multiple target distance information and multiple target pressure data into the target random forest model, and calls N decision trees in the target random forest model to make decisions to obtain the operating status of the target elevator; where N is a positive integer; the normal pressure distribution conditions include: the difference between the largest pressure data and the smallest pressure data among multiple target pressure data is less than a preset pressure threshold; In response to multiple target pressure data not meeting the normal pressure distribution conditions, the target operating stage, target speed, target acceleration, multiple target distance information, and multiple target pressure data are input into the target random forest model, and the M decision trees in the target random forest model are called to make decisions to obtain the operating status of the target elevator; where M is a positive integer, M is greater than N.

[0075] In one embodiment of this application, the running state determination module 25 is further used to determine the target acceleration level corresponding to the target acceleration; Input the target's operational phase, target velocity, target acceleration, multiple target distance information, and multiple target pressure data into the target random forest model, and call the M decision trees in the target random forest model to make decisions, and obtain M decision results; Responding to whether the target operation phase is an acceleration phase or a deceleration phase, for each decision outcome, the weight of the decision tree is determined based on the number of samples of the first target sample and the number of samples of the second target sample in the target training sample set. The target training sample set is the training sample set used by the decision tree corresponding to the decision outcome. A training sample set contains multiple training samples, and a training sample includes: a sample parameter set, which includes: operation phase samples, velocity samples, acceleration samples, multiple sample distance information, and multiple sample pressure data; the first target sample is the operation phase sample in the target training set, and the second target sample is the acceleration sample in the target training set that belongs to the target acceleration level; the acceleration phase is an operation phase where the target velocity is greater than zero, the target acceleration is greater than zero, and the target velocity increases over time; the deceleration phase is an operation phase where the target velocity is greater than zero, the target acceleration is less than zero, and the target velocity decreases over time. The operating status of the target elevator is obtained by weighting each decision result and its corresponding weight.

[0076] In one embodiment of this application, the elevator operation status detection device 20 further includes: a target random forest model determination module, used to obtain a total set of training data and a total set of validation data; Based on the training data set, train P initial decision trees to obtain P trained decision trees; where P is a positive integer; P is less than N and P is less than M; The performance of the trained P decision trees is evaluated to obtain the performance evaluation results. The performance evaluation results include: the average validation error of the trained P decision trees on the total validation data set, and the average split gain of each feature in the trained P decision trees. Each feature includes: running stage, speed, acceleration, multiple distance information and multiple stress data. Based on the performance evaluation results, determine the parameters to be adjusted and their corresponding values, and then adjust the values ​​of the parameters to be adjusted to obtain the target training parameter values. Based on the target training parameter values ​​and the training dataset, train Q initial decision trees to obtain Q trained decision trees; The trained P decision trees are combined with the trained Q decision trees to obtain the target random forest model.

[0077] In one embodiment of this application, the target random forest model determination module is specifically used to adjust the feature selection parameter value corresponding to the feature to an enhanced state in response to the existence of at least one feature whose average split gain is greater than a preset gain threshold. The enhanced state is that the probability of the feature being selected as a splitting feature in the training of Q decision trees is increased by a first preset proportion. In response to the average verification error exceeding a preset error threshold, the maximum depth parameter value in the tree structure parameters is adjusted to a reduced state. The reduced state is that the maximum depth of Q decision trees is limited to a second preset ratio of the average depth of P initial decision trees.

[0078] In one embodiment of this application, the first speed detection device is a laser Doppler velocimeter, which is installed on the outer top of the target elevator; the second speed detection device is a traction machine spindle encoder, which is installed on the spindle of the traction machine. The elevator operation status detection device 20 further includes: a signal-to-noise ratio acquisition module, used to acquire a target signal-to-noise ratio; the target signal-to-noise ratio is the signal-to-noise ratio when the first speed detection device detects the speed of the target elevator; The speed acquisition module 21 is specifically used to determine the weight of the first speed and the weight of the second speed based on the target signal-to-noise ratio; The target speed is obtained by weighting the first speed, its weight, the second speed, and the weight of the second speed.

[0079] In one embodiment of this application, the target running phase further includes a stationary phase and a uniform speed phase; the running phase determination module 23 is specifically used to determine the target running phase as a stationary phase in response to the target velocity being zero and the target acceleration being zero. In response to the target velocity being greater than zero, the absolute value of the target acceleration being less than a preset acceleration threshold, and the change in target velocity over time being less than a preset velocity change threshold, the target running phase is determined to be a uniform velocity phase.

[0080] See Figure 4 , Figure 4 This is a schematic block diagram of a control device provided in one embodiment of this application. Figure 4The control device 300 shown in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memory 304 stores computer programs, including program instructions. The processors 301 execute the program instructions stored in the memory 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of each module / unit in the above-described device embodiments, for example... Figure 3 The functions of the velocity acquisition module 21, acceleration acquisition module 22, running phase determination module 23, distance acquisition module 24, and running status determination module 25 are shown.

[0081] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0082] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0083] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0084] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation method described in the elevator operation status detection method provided in the embodiments of this application, or they can execute the implementation method of the control device described in the embodiments of this application, which will not be repeated here.

[0085] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0086] The computer-readable storage medium can be an internal storage unit of the control device in any of the foregoing embodiments, such as a hard disk or memory of the control device. The computer-readable storage medium can also be an external storage device of the control device, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc., provided on the control device. Furthermore, the computer-readable storage medium can include both internal storage units and external storage devices of the control device. The computer-readable storage medium is used to store computer programs and other programs and data required by the control device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0087] Those skilled in the art will recognize that the modules / units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0088] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the control device and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed control devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules, units, or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or modules / units, or it may be an electrical, mechanical, or other form of connection.

[0090] The modules / units described as separate components may or may not be physically separate. Similarly, the components shown as modules / units may or may not be physical modules / units; they may be located in one place or distributed across multiple network modules / units. Some or all of the modules / units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0091] Furthermore, the functional modules / units in the various embodiments of this application can be integrated into one processing module / unit, or each module / unit can exist physically separately, or two or more modules / units can be integrated into one module / unit. The integrated modules / units described above can be implemented in hardware or in the form of software functional modules / units.

[0092] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting the operating status of an elevator, characterized in that, An elevator status detection system is applied, comprising: a control device, a first speed detection device, a second speed detection device, an acceleration detection device, multiple distance sensors, and multiple pressure sensors; the multiple pressure sensors are used to measure the pressure value at the bottom of the target elevator and obtain multiple target pressure data; the control device establishes communication connections with the first speed detection device, the second speed detection device, the acceleration detection device, and the multiple distance sensors respectively; the elevator operation status detection method is executed by the control device, and the elevator operation status detection method includes: A first speed, a second speed, and a target signal-to-noise ratio (SNR) are acquired, and the weights of the first speed and the second speed are determined based on the target SNR. A weighted calculation is performed based on the first speed, its weight, the second speed, and its weight to obtain the target speed. The target SNR is the SNR detected by the first speed detection device when detecting the speed of the target elevator. The first speed is the speed of the target elevator currently detected by the first speed detection device. The second speed is the speed of the target elevator currently detected by the second speed detection device. The first speed detection device is mounted on the elevator car. The second speed detection device is mounted on the traction machine, which provides power to the target elevator. Obtain the target acceleration; the target acceleration is the acceleration of the target elevator currently detected by the acceleration detection device; The target operating stage is determined based on the target speed and the target acceleration; the target operating stage is the current operating stage of the target elevator. Acquire distance information for multiple targets; each target distance information is the distance information between the corresponding ranging sensor and the target elevator, detected by the corresponding ranging sensor. In response to multiple target pressure data satisfying the normal pressure distribution condition, the target operating stage, the target speed, the target acceleration, the multiple target distance information, and the multiple target pressure data are input into the target random forest model, and N decision trees in the target random forest model are called to make decisions to obtain the operating status of the target elevator; where N is a positive integer; the normal pressure distribution condition includes: the difference between the largest and smallest pressure data among the multiple target pressure data is less than a preset pressure threshold; In response to multiple target pressure data not meeting the normal pressure distribution conditions, the target acceleration level corresponding to the target acceleration is determined; the target running stage, the target velocity, the target acceleration, the multiple target distance information, and the multiple target pressure data are input into a target random forest model, and M decision trees in the target random forest model are called to make decisions, resulting in M ​​decision results; in response to the target running stage being an acceleration phase or a deceleration phase, for each decision result, the weight of the decision tree is determined based on the number of samples of the first target sample and the number of samples of the second target sample in the target training sample set. The target training sample set is the training sample set used by the decision tree corresponding to the decision result, and a training sample set contains multiple training samples. The system includes: a sample parameter set, comprising: operation phase samples, speed samples, acceleration samples, multiple sample distance information, and multiple sample pressure data; the first target sample is the target training sample set, and the operation phase samples are the training samples of the target operation phase; the second target sample is the target training sample set, and the acceleration samples belong to the target acceleration level; the start-up acceleration phase is an operation phase where the target speed is greater than zero, the target acceleration is greater than zero, and the target speed increases over time; the deceleration phase is an operation phase where the target speed is greater than zero, the target acceleration is less than zero, and the target speed decreases over time; the target elevator's operating state is obtained by weighted calculation based on each decision result and its corresponding weight; where M is a positive integer, and M is greater than N.

2. The elevator operation status detection method as described in claim 1, characterized in that, The multiple pressure sensors are installed at the bottom of the target elevator car. The elevator operation status detection method further includes: Obtain the pressure data of the multiple targets.

3. The elevator operation status detection method as described in claim 1, characterized in that, The process of determining the target random forest model includes: Obtain the total training data set and the total validation data set; Based on the total training data set, P initial decision trees are trained to obtain P trained decision trees; where P is a positive integer; P is less than N, and P is less than M; The performance of the trained P decision trees is evaluated to obtain performance evaluation results. The performance evaluation results include: the average validation error of the trained P decision trees on the total validation data set, and the average split gain of each feature in the trained P decision trees. The features include: running stage, speed, acceleration, multiple distance information and multiple stress data. Based on the performance evaluation results, the parameters to be adjusted and their corresponding values ​​are determined, and the values ​​of the parameters to be adjusted are adjusted to obtain the target training parameter values. Based on the target training parameter values ​​and the training dataset, Q initial decision trees are trained to obtain Q trained decision trees; The trained P decision trees are combined with the trained Q decision trees to obtain the target random forest model.

4. The elevator operation status detection method as described in claim 3, characterized in that, The parameters to be adjusted include feature selection parameters and tree structure parameters; The step of determining the parameters to be adjusted and their corresponding values ​​based on the performance evaluation results, and then adjusting the values ​​of the parameters to be adjusted, includes: In response to the existence of at least one feature whose average split gain is greater than a preset gain threshold, the feature selection parameter value corresponding to that feature is adjusted to an enhanced state, wherein the probability of that feature being selected as a splitting feature in the training of the Q decision trees is increased by a first preset proportion. In response to the average verification error being greater than a preset error threshold, the maximum depth parameter value in the tree structure parameters is adjusted to a reduced state. The reduced state is that the maximum depth of the Q decision trees is limited to a second preset proportion of the average depth of the P initial decision trees.

5. The elevator operation status detection method as described in claim 1, characterized in that, The first speed detection device is a laser Doppler velocimeter, which is installed on the outer top of the target elevator; the second speed detection device is a traction machine spindle encoder, which is installed on the spindle of the traction machine.

6. The elevator operation status detection method as described in claim 1, characterized in that, The target operation phase also includes: a stationary phase and a uniform speed phase; The step of determining the target running stage based on the target velocity and the target acceleration includes: In response to the target velocity being zero and the target acceleration being zero, the target's running phase is determined to be the stationary phase; In response to the target velocity being greater than zero, the absolute value of the target acceleration being less than a preset acceleration threshold, and the change in target velocity over time being less than a preset velocity change threshold, the target running phase is determined to be the uniform velocity phase.

7. An elevator operation status detection device, characterized in that, For implementing the method as described in any one of claims 1 to 6, the elevator operating status detection device is disposed in the control equipment of the elevator status detection system, and the elevator status detection system further includes: a first speed detection device, a second speed detection device, an acceleration detection device, and a plurality of distance measuring sensors; the control equipment establishes communication connections with the first speed detection device, the second speed detection device, the acceleration detection device, and the plurality of distance measuring sensors respectively; the elevator operating status detection device includes: A speed acquisition module is used to acquire a first speed and a second speed, and determine a target speed based on the first speed and the second speed; the first speed is the speed of the target elevator currently detected by the first speed detection device; the second speed is the speed of the target elevator currently detected by the second speed detection device; the first speed detection device is installed on the car body of the target elevator; the second speed detection device is installed on the traction machine, which is used to provide power to the target elevator; An acceleration acquisition module is used to acquire the target acceleration; the target acceleration is the acceleration of the target elevator currently detected by the acceleration detection device. The operation phase determination module is used to determine the target operation phase based on the target speed and the target acceleration; the target operation phase is the current operation phase of the target elevator. The distance acquisition module is used to acquire distance information of multiple targets; the distance information of each target is the distance information between the corresponding ranging sensor and the target elevator, which is detected by the corresponding ranging sensor. The operation status determination module is used to determine the current operation status of the target elevator based on the target operation stage, the target speed, the target acceleration, and the multiple target distance information.

8. A control device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 6.

Images