Methods, systems and equipment for adaptive adjustment of braking force of elevator disc brakes

Abstract

This invention provides a method, system, and device for adaptive adjustment of braking force in elevator caliper disc brakes, relating to the field of braking control for elevator caliper disc brakes. The adjustment method includes: acquiring braking characteristic data of the brake in real time; after the brake is activated, calculating the real-time braking force and theoretical braking force based on the braking scenario and braking characteristic data, and obtaining a second braking force based on the theoretical braking force and a pre-trained neural network model; when the deviation between the real-time braking force and the second braking force exceeds a preset braking force threshold, calculating the theoretical compression of the spring based on the second braking force; and controlling the actual compression of the spring through a telescopic device based on the theoretical and real-time compression of the spring, so that the spring outputs a compensating braking force to adaptively adjust the braking force of the brake. Using this invention, the braking force of the brake can be adaptively adjusted, dynamically compensating for braking force attenuation and ensuring stable and reliable braking performance.

Description

Technical Field

[0001] This invention relates to the field of braking control for elevator disc brakes, and more specifically, to a method, system, and device for adaptive adjustment of braking force of an elevator disc brake. Background Technology

[0002] Existing elevator brake control technology mainly adopts a binary switch control mode, which realizes the "power on release brake, power off hold brake" function by directly switching the power supply or redundant circuits. It cannot dynamically adjust the braking force according to the elevator load, running speed and real-time status of the brake pads, which can easily lead to excessive or insufficient braking, affecting operational safety and ride comfort. Summary of the Invention

[0003] The present invention provides a method, system, device, storage medium and computer program product for adaptive adjustment of braking force of elevator caliper disc brake. By acquiring multi-dimensional braking characteristic data in real time and calculating the theoretical compression of the spring, the spring output is controlled by the telescopic device to compensate for the braking force, thereby adaptively adjusting the braking force of the brake. This can dynamically compensate for the attenuation of braking force and ensure stable and reliable braking performance.

[0004] According to a first aspect of the present invention, an embodiment of the present invention provides a method for adaptive adjustment of braking force of an elevator caliper disc brake, comprising: acquiring braking characteristic data of the brake in real time, the braking characteristic data including: real-time compression of the spring in a spring adjustment device, brake disc temperature, armature temperature, coil temperature, and magnetic flux of an electromagnet; after the brake is activated, calculating a real-time braking force and a theoretical braking force based on the braking scenario and the braking characteristic data, and obtaining a second braking force based on the theoretical braking force and a pre-trained neural network model; calculating a braking force deviation based on the real-time braking force and the second braking force; when the braking force deviation is greater than a preset braking force threshold, calculating the theoretical compression of the spring based on the second braking force; and controlling the actual compression of the spring through a telescopic device based on the theoretical compression of the spring and the real-time compression, so that the spring outputs a compensating braking force to adaptively adjust the braking force of the brake.

[0005] According to the above embodiments of the present invention, by real-time monitoring of the brake performance data and calculation of the theoretical braking force and braking force deviation, and by controlling the actual compression of the spring through a telescopic device based on the theoretical compression and real-time compression of the spring when the braking force deviation exceeds a preset threshold, the spring outputs a compensating braking force to adaptively adjust the braking force of the brake. Thus, braking force attenuation can be dynamically compensated, ensuring that braking performance is always maintained above the design safety threshold.

[0006] In some embodiments of the present invention, the adjustment method further includes: analyzing and diagnosing positioning brake faults based on the braking characteristic data.

[0007] In some embodiments of the present invention, analyzing and diagnosing brake faults based on the braking characteristic data includes: preprocessing the braking characteristic data, the preprocessing including: filtering, amplification, linearization, and spatiotemporal alignment; performing feature extraction and deep learning feature fusion on the preprocessed data to obtain feature fusion data; and locating brake faults based on the feature fusion data.

[0008] In some embodiments of the present invention, locating brake faults based on the feature fusion data includes: extracting fault features from the feature fusion data; inputting the fault features into the pre-trained neural network model to obtain brake fault information; and sending early warning information based on the fault information.

[0009] According to the above embodiments of the present invention, by extracting multimodal features through deep learning and locating brake faults based on multimodal features, a shift from "post-event maintenance" to "predictive maintenance" can be achieved, accurately locating specific clamping points and fault causes, thereby providing early warning of component performance degradation and reducing the risk of unplanned downtime.

[0010] In some embodiments of the present invention, the adjustment method further includes: recording and storing the acquired braking characteristic data, processed feature fusion data, and braking force adjustment result data, and updating the parameters in the digital twin model based on the brake according to the measured brake data after braking force adjustment.

[0011] The above-described embodiments of the present invention update the parameters of the digital twin model based on the brake using the adjusted measured data of the brake, ensuring dynamic synchronization between the model and the physical brake, forming a complete closed loop of "data acquisition - processing - decision-making - execution - model update". This enables dynamic linkage between the physical entity and the digital model, with real-time feedback of measured data ensuring the model always closely matches the actual operating state of the brake, avoiding a disconnect between theory and practice. An adaptive optimization mechanism is formed, continuously correcting the decision logic and execution parameters through closed-loop iteration, improving the accuracy and stability of braking force adjustment and control. Simultaneously, it provides data support for predictive maintenance, further enhancing the safety and reliability of elevator operation.

[0012] In some embodiments of the present invention, the adjustment method further includes: automatically adjusting the preset threshold of braking force based on the digital twin model, historical braking characteristic data, and measured brake data.

[0013] In some embodiments of the present invention, the adjustment method further includes: when the braking force deviation is greater than a preset braking force threshold, optimizing the PID control parameters of the PID control algorithm used to adjust the coil current; outputting the adjustment amount of the coil current according to the PID control parameters, and outputting a third braking force according to the adjustment amount of the coil current.

[0014] According to the above embodiments of the present invention, by optimizing the parameters of the PID control algorithm used to adjust the coil current, the braking force can be dynamically adjusted to adapt to the real-time load, running speed and brake disc temperature changes of the elevator equipment, avoiding the problems of "unable to brake" or "over-braking" and improving the safety of equipment operation.

[0015] In some embodiments of the present invention, the brake has an actuator that controls the telescopic device to control the actual compression of the spring.

[0016] According to a second aspect of the present invention, an embodiment of the present invention provides an adaptive braking force adjustment system for an elevator caliper disc brake. The system includes: a data acquisition module for acquiring braking characteristic data of the brake in real time, the braking characteristic data including: real-time compression of the spring in a spring adjustment device, brake disc temperature, armature temperature, coil temperature, and magnetic flux of an electromagnet; a braking force calculation module for calculating real-time braking force and theoretical braking force based on the braking scenario and the braking characteristic data after the brake is activated, and obtaining a second braking force based on the theoretical braking force and a pre-trained neural network model; a spring compression calculation module for calculating the braking force deviation based on the real-time braking force and the second braking force, and calculating the theoretical compression of the spring based on the second braking force when the braking force deviation is greater than a preset braking force threshold; and a braking force compensation module for controlling the actual compression of the spring through a telescopic device based on the theoretical compression and the real-time compression, so that the spring outputs a compensating braking force to adaptively adjust the braking force of the brake.

[0017] According to the above embodiments of the present invention, by real-time monitoring of the brake performance data and calculation of the theoretical braking force and braking force deviation, and by controlling the actual compression of the spring through a telescopic device based on the theoretical compression and real-time compression of the spring when the braking force deviation exceeds a preset threshold, the spring outputs a compensating braking force to adaptively adjust the braking force of the brake. Thus, braking force attenuation can be dynamically compensated, ensuring that braking performance is always maintained above the design safety threshold.

[0018] In some embodiments of the present invention, the adjustment system further includes a fault diagnosis module for analyzing and diagnosing positioning brake faults based on the braking characteristic data.

[0019] In some embodiments of the present invention, analyzing and diagnosing brake faults based on the braking characteristic data includes: preprocessing the braking characteristic data, the preprocessing including: filtering, amplification, linearization, and spatiotemporal alignment; performing feature extraction and deep learning feature fusion on the preprocessed data to obtain feature fusion data; and locating brake faults based on the feature fusion data.

[0020] In some embodiments of the present invention, locating brake faults based on the feature fusion data includes: extracting fault features from the feature fusion data; inputting the fault features into the pre-trained neural network model to obtain brake fault information; and sending early warning information based on the fault information.

[0021] According to the above embodiments of the present invention, by extracting multimodal features through deep learning and locating brake faults based on multimodal features, a shift from "post-event maintenance" to "predictive maintenance" can be achieved, accurately locating specific clamping points and fault causes, thereby providing early warning of component performance degradation and reducing the risk of unplanned downtime.

[0022] In some embodiments of the present invention, the adjustment system further includes a model update module, which is used to record and store the acquired braking characteristic data, processed feature fusion data, and braking force adjustment result data, and to update the parameters in the digital twin model based on the brake according to the measured brake data after braking force adjustment.

[0023] The above-described embodiments of the present invention update the parameters of the digital twin model based on the brake using the adjusted measured data of the brake, ensuring dynamic synchronization between the model and the physical brake, forming a complete closed loop of "data acquisition - processing - decision-making - execution - model update". This enables dynamic linkage between the physical entity and the digital model, with real-time feedback of measured data ensuring the model always closely matches the actual operating state of the brake, avoiding a disconnect between theory and practice. An adaptive optimization mechanism is formed, continuously correcting the decision logic and execution parameters through closed-loop iteration, improving the accuracy and stability of braking force adjustment and control. Simultaneously, it provides data support for predictive maintenance, further enhancing the safety and reliability of elevator operation.

[0024] In some embodiments of the present invention, the preset threshold of braking force is automatically adjusted based on the digital twin model, historical braking characteristic data, and measured brake data.

[0025] In some embodiments of the present invention, the adjustment system further includes a braking force adjustment module, which is used to optimize the PID control parameters of the PID control algorithm for adjusting the coil current when the braking force deviation is greater than a preset braking force threshold; and to output an adjustment amount of the coil current according to the PID control parameters, and output a third braking force according to the adjustment amount of the coil current.

[0026] According to the above embodiments of the present invention, by optimizing the parameters of the PID control algorithm used to adjust the coil current, the braking force can be dynamically adjusted to adapt to the real-time load, running speed and brake disc temperature changes of the elevator equipment, avoiding the problems of "unable to brake" or "over-braking" and improving the safety of equipment operation.

[0027] In some embodiments of the present invention, the brake has an actuator that controls the telescopic device to control the actual compression of the spring.

[0028] According to a third aspect of the present invention, an embodiment of the present invention provides a computer-readable storage medium having computer-readable instructions stored thereon, which, when executed by a processor, cause a computer to perform the following operations: the operations include the steps included in the adaptive adjustment method for braking force of an elevator caliper disc brake as described in any of the above embodiments.

[0029] According to a fourth aspect of the present invention, an embodiment of the present invention provides a computer device including a memory and a processor, wherein the memory is used to store one or more computer-readable instructions, wherein the one or more computer-readable instructions, when executed by the processor, can realize the adaptive adjustment method of braking force of an elevator caliper disc brake as described in any of the above embodiments.

[0030] According to a fifth aspect of the present invention, an embodiment of the present invention provides a computer program product including a computer program, which, when executed by a processor, implements the adaptive adjustment method for braking force of an elevator caliper disc brake as described in any of the above embodiments.

[0031] As described above, the adaptive braking force adjustment method, system, device, medium, and computer program product for elevator caliper disc brakes provided by the embodiments of the present invention adaptively adjust the braking force of the brake by real-time monitoring of the brake performance data and calculating the theoretical braking force and braking force deviation. Furthermore, when the braking force deviation exceeds a preset threshold, the actual compression of the spring is controlled by a telescopic device based on the theoretical and real-time compression of the spring, thereby enabling the spring to output compensating braking force to adaptively adjust the braking force of the brake. This dynamically compensates for braking force attenuation, ensuring that braking performance is always maintained above the design safety threshold. Attached Figure Description

[0032] Figure 1 This is a flowchart illustrating the adaptive adjustment method of braking force for an elevator caliper disc brake according to Embodiment 1 of the present invention.

[0033] Figure 2 This is a perspective view of the structure of the brake body and spring adjusting device of a multi-caliper disc brake according to an embodiment of the present invention;

[0034] Figure 3 yes Figure 2 Front view of the structure of the central brake body;

[0035] Figure 4 This is a first perspective schematic diagram of the structure of a multi-caliper disc brake, a displacement sensor, and a Hall sensor according to an embodiment of the present invention.

[0036] Figure 5 This is a second perspective schematic diagram of the structure of a multi-caliper disc brake, a displacement sensor, and a Hall sensor according to an embodiment of the present invention;

[0037] Figure 6 This is a three-dimensional schematic diagram of the structure of the rear brake block and spring of a multi-caliper disc brake according to an embodiment of the present invention;

[0038] Figure 7 This is a perspective view of the structure of a spring adjusting device according to an embodiment of the present invention;

[0039] Figure 8 This is a schematic diagram of the operation of a spring adjusting device according to an embodiment of the present invention;

[0040] Figure 9 This is a flowchart illustrating the adaptive adjustment method of braking force for an elevator caliper disc brake according to Embodiment 2 of the present invention.

[0041] Figure 10 This is a schematic diagram of the adaptive braking force adjustment system of an elevator caliper disc brake according to Embodiment 4 of the present invention.

[0042] Figure 11 This is a flowchart illustrating the adaptive adjustment method of braking force for an elevator caliper disc brake according to Embodiment 5 of the present invention.

[0043] The reference numerals in the attached drawings are explained as follows: 1-Brake body, 2-Spring adjustment device, 3-Displacement sensor, 4-Hall sensor, 11-Current controller, 12-Main braking device, 13-Electromagnet, 14-Intermediate engaging device, 15-Rear brake device, 16-Intermediate engaging brake block, 17-Rear brake block, 18-Bolt, 21-Spring, 22-Telescopic device sleeve, 23-Telescopic device, 24-Telescopic device top plate, 25-Fixing screw. Detailed Implementation

[0044] The various aspects of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. Well-known modules, units, and their connections, links, communications, or operations are not shown or described in detail. Furthermore, the described features, architectures, or functions can be combined in any way in one or more embodiments. Those skilled in the art should understand that the various embodiments described below are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. It will also be readily understood that the modules, units, or processing methods in the embodiments described herein and shown in the accompanying drawings can be combined and designed in various different configurations.

[0045] The following is a brief explanation of the terms used in this text.

[0046] GNN: Graph Neural Network, a deep learning model specifically designed to handle graph-structured data.

[0047] CNN: Convolutional Neural Network.

[0048] Identity Mapping: Identity mapping.

[0049] Global Average Pooling (GAP)

[0050] WKS: Wavelet Packet Decomposition.

[0051] IMF: Intrinsic Mode Function.

[0052] LVQ1: Learning Vector Quantization 1, learns the vector quantization algorithm.

[0053]

Example 1

[0054] Figure 1 This is a flowchart illustrating the adaptive adjustment method of braking force for an elevator caliper disc brake according to Embodiment 1 of the present invention.

[0055] like Figure 1 As shown, in Embodiment 1 of the present invention, the adaptive adjustment method of braking force of the elevator caliper disc brake may include at least the following steps S11, S12, S13, S14 and S15, which are described in detail below.

[0056] In step S11, the braking characteristic data of the brake is acquired in real time. The braking characteristic data includes: the real-time compression of the spring in the spring adjustment device, the brake disc temperature, the armature temperature, the coil temperature, and the magnetic flux of the electromagnet.

[0057] In some implementations, the spring compression is obtained by a displacement sensor, the temperature of the brake disc and armature is collected by a temperature sensor, and the magnetic flux reflecting the strength of the electromagnet's magnetic field is obtained by a magnetic flux sensor.

[0058] This invention provides a specific example of a data acquisition device for collecting multi-source data from a caliper disc brake, wherein the caliper disc brake is a multi-caliper disc brake, and the acquisition device includes: a displacement sensor, a Hall sensor, and a temperature sensor. It should be understood that this invention is not limited thereto, and other acquisition devices known in the art can be used to acquire multi-source raw datasets. The following describes... Figures 2 to 6 The example will be explained in detail.

[0059] Figure 2 This is a perspective view of the structure of the brake body and spring adjusting device of a multi-caliper disc brake according to an embodiment of the present invention; Figure 3 yes Figure 2 Front view of the structure of the central brake body; Figure 4 This is a first perspective schematic diagram of the structure of a multi-caliper disc brake, a displacement sensor, and a Hall sensor according to an embodiment of the present invention. Figure 5 This is a second perspective schematic diagram of the structure of a multi-caliper disc brake, a displacement sensor, and a Hall sensor according to an embodiment of the present invention; Figure 6 This is a three-dimensional schematic diagram of the structure of the rear brake block and spring of a multi-caliper disc brake according to an embodiment of the present invention.

[0060] like Figure 2 and 3 As shown, eight spring adjustment devices 2 are evenly distributed on the brake body 1, and are respectively installed on the brake front arm. Each spring adjustment device 2 is flexibly connected to the mounting hole of the brake body 1 through a stop pin soft connection, ensuring that there is no direct rigid contact between the spring and the brake body.

[0061] like Figures 4 to 6As shown, the current controller 11 is located on the outside of the brake body, near the mounting position of the electromagnet 13, and is connected to the electromagnet 13 via a cable, providing it with precisely regulated coil current. The main braking device 12 is the core load-bearing frame of the entire device. Eight mounting holes are evenly distributed on the brake arm, connecting the spring adjustment device 2 and the mounting slot of the intermediate engaging device 14 via a stop pin, and the bottom is connected to the rear brake device 15. The electromagnet 13 is located in the middle of the inner side of the main braking device 12, coaxial with the intermediate engaging device 14. The spring adjustment device 2 is located on the outside of the brake arm of the main braking device 12, with eight evenly distributed circumferentially. The intermediate engaging device 14 is located between the electromagnet 13 and the intermediate engaging brake block 16, and is disc-shaped. One end is connected to the intermediate engaging brake block 16, and the other end faces the iron core of the electromagnet 13. It is controlled by the magnetic field of the electromagnet 13, driving the intermediate engaging brake block 16 to achieve the release / engagement action. The rear brake assembly 15 is located behind the main brake assembly 12. It houses the rear brake block 17 and spring 21, forming a shell structure. It is rigidly connected to the main brake assembly 12 via bolts 18 and is symmetrically distributed on both sides of the brake disc along with the intermediate engagement device 14. The intermediate engagement brake block 16 is located on the front side of the brake disc, rigidly connected to the intermediate engagement device 14, and fits against the edge of the brake disc. It is controlled by the pressure of the spring adjusting device 2, adjusting the positive pressure on the brake disc. The spring 21 (e.g., spring D12) is located inside the spring adjusting device 2, with a high-carbon steel spiral structure. Its two ends abut against the top plate and the fixed seat of the spring adjusting device 2, respectively, and its extension and retraction are controlled by the extension and retraction device. The rear brake block 17 is located inside the rear brake assembly 15, fitting against the rear side of the brake disc, symmetrical to the intermediate engagement brake block 16, and together they clamp the brake disc.

[0062] The brake disc is made of high-strength cast iron / composite material, with heat dissipation fins and ventilation holes on the surface to improve thermal stability. The brake caliper body is a modular aluminum alloy structure, rigidly connected to the main controller and the suction block by bolts to achieve floating installation to adapt to brake disc sway. The electromagnet uses silicon steel sheet laminated iron core and enameled copper wire coil, and works with the current controller to achieve precise control of magnetic field strength. The bolts adopt an internal hexagon countersunk head design to ensure compact assembly and anti-loosening performance. The displacement sensor 3 is installed on the side wall of the brake caliper body or the outside of the sleeve of the spring adjustment device 2. It adopts a non-contact capacitive design to accurately monitor the brake block stroke. The Hall sensor 4 is integrated into the side wall of the electromagnet 13 to collect the magnetic field strength in real time. Temperature sensors (not shown in the figure) are installed on the three core heat-generating and heat-affected parts of the brake: the brake disc, the armature, and the electromagnet coil. For example, the brake disc generates a lot of heat due to friction with the brake block. The sensor is attached to its surface or embedded near the heat dissipation fins to monitor the temperature change caused by frictional heat to reflect the heat fade effect.

[0063] Furthermore, the displacement signal, Hall voltage signal, and temperature signal acquired by the aforementioned displacement sensor, Hall sensor, and temperature sensor are converted into digital signals through the analog signal conditioning circuit and analog-to-digital conversion module of the sensor interface and then transmitted to the actuator for subsequent processing.

[0064] The above-mentioned data acquisition device of the present invention, through the scientific arrangement of displacement sensor, Hall sensor and temperature sensor, can realize the real-time acquisition of multi-source data, which solves the problem that the wear state can only be indirectly inferred by periodically measuring brake gap or temperature, and cannot detect the performance degradation of components such as brake shoes and springs in real time, so that the failure is already threatening safety when it appears.

[0065] In step S12, after the brake is activated, the real-time braking force and theoretical braking force are calculated based on the braking scenario and the braking characteristic data, and the second braking force is obtained based on the theoretical braking force and the pre-trained neural network model.

[0066] In some implementations, the real-time braking force and the theoretical braking force are calculated in the following manner:

[0067] (1) Calculation of real-time braking force: Based on the principle of friction braking, the real-time braking force is calculated. (Unit: N) The core calculation logic is "brake block normal pressure × friction coefficient × effective radius of brake disc". At the same time, the synergistic effect of multiple brake blocks is considered. For example, in an example brake of the present invention, the brake contains 8 sets of spring adjustment devices, which correspond to 8 brake block action points. The real-time braking force of the brake is calculated by the following formula (1):

[0068] (1)

[0069] Where n is the total number of brake blocks involved in braking in the brake system. For example, in this embodiment, the brake includes 8 sets of spring adjustment devices, corresponding to 8 brake block action points (evenly distributed on the brake arm), then n=8. It is the normal force exerted by a single brake pad on the brake disc, which is derived from the spring compression of the spring adjustment device or directly collected by a pressure sensor. The relationship between the normal force and the compression is as follows: x and k are the spring constants. x is the spring compression (unit: m); T is the brake disc temperature obtained by the temperature sensor; μ(T) is the dynamic friction coefficient, which is obtained by looking up the "temperature-friction coefficient" mapping table; r is the effective radius of the brake disc, which is a factory-calibrated fixed value, such as 0.3m.

[0070] (2) Calculate the theoretical braking force (i.e. the first optimal braking force): Under normal deceleration and braking scenarios, the goal is to make the elevator decelerate at a constant speed. (For example, take 0.5 m / s², which meets the "comfort-type deceleration" requirement in GB / T 10058-2023 "Elevator Technical Conditions") and stop smoothly at the floor. At this time, the theoretical braking force is calculated by the following formula (2). :

[0071] (2)

[0072] Where m is the load. It is the weight of the elevator car itself. This is a fixed value specified before leaving the factory, such as 1000 kg; This is the normal braking deceleration, a fixed value of 0.5 m / s², to balance safety and comfort; R is the traction sheave radius, a core parameter of the elevator, and R is a fixed value, for example, R=0.4 m; i is the elevator transmission ratio, i.e., the speed transmission relationship between the traction machine and the car, and i is a fixed value, for example, i=2:1; r is the effective radius of the brake disc, which is a factory-calibrated fixed value, for example, 0.3 m.

[0073] In emergency braking scenarios, such as when the elevator overspeeds (e.g., exceeds 115% of the rated speed) or malfunctions, it must decelerate at the maximum safe speed. (For example, take 1.5 m / s², which meets the "upper limit of emergency braking deceleration" requirement in GB / T 7588.1-2020 "Safety Code for Elevator Manufacturing and Installation Part 1: Passenger Elevators and Freight Elevators") Rapid braking, at this time the theoretical braking force is calculated by the following formula (3). :

[0074] (3)

[0075] in, θ is the emergency braking deceleration, a fixed value of 1.5 m / s², to ensure rapid braking without causing impact on the car; g is the gravitational acceleration (taken as 9.8 m / s²); θ is the elevator shaft tilt angle (for a vertical elevator, θ = 90°, sinθ = 1; for an inclined elevator, calculate sinθ by substituting the actual angle).

[0076] Furthermore, after calculating and / or After determining the theoretical braking force, a pre-trained neural network model is used to further optimize the theoretical braking force, resulting in the optimized theoretical optimal braking force target value. Specifically, firstly, the load m, speed v, brake disc temperature T, and historical braking deviation are input into the pre-trained neural network model, which outputs the optimal braking force correction coefficient γ (its value ranges from 0.9 to 1.1). Secondly, the second braking force (i.e., the final theoretical optimal braking force) is calculated. : ,or Where β is the safety redundancy coefficient, β=1.2 in normal braking scenarios (20% redundancy reserved to cope with load fluctuations), and β=1.5 in emergency braking scenarios (50% redundancy reserved to ensure effective braking in case of failure).

[0077] In step S13, the braking force deviation is calculated based on the real-time braking force and the second braking force.

[0078] In step S14, when the braking force deviation is greater than the preset braking force threshold, the theoretical compression of the spring is calculated based on the second braking force.

[0079] The preset braking force threshold is related to the elevator's rated power, braking scenario, and the status of key components. Its value is based on the rated power, taking 5% of the rated braking force under normal braking conditions and 3% of the rated braking force under emergency braking conditions. When the brake pad wear exceeds the initial thickness by 30% or the spring elasticity coefficient drops by more than 10%, the normal and emergency braking thresholds are increased to 7% and 5% respectively. Furthermore, the preset braking force threshold is automatically calibrated and fine-tuned every 24 hours in conjunction with the digital twin model and historical data.

[0080] In some implementations, the preset threshold of braking force is automatically adjusted based on the digital twin model, historical braking characteristic data, and measured brake data. The historical braking characteristic data includes: 1. Historical braking force related data: real-time braking force, theoretical braking force (normal / emergency scenarios), and historical records of the second braking force within a preset time period (e.g., the last 30 days), as well as the deviation values ​​of each braking force; 2. Historical operating condition matching data: historical correlation records of operating condition parameters such as elevator load m, running speed v, brake disc temperature T, and elevator shaft tilt angle θ during braking; 3. Component status historical data: historical trend of brake pad wear, records of spring elastic coefficient decay, and historical data of magnetic flux related to electromagnet coil aging; 4. Fault correlation historical data: historical fault types, braking force status at the time of fault occurrence, records of threshold adjustment corresponding to the fault, and compensation effects; 5. Threshold adjustment historical data: adjustment time, adjustment range, adjustment reason, and braking performance feedback data after adjustment of the preset braking force threshold in the past; 6. Environmental and usage frequency data: historical ambient temperature at which braking occurs, statistics of daily elevator braking frequency, and braking load distribution data for different time periods (e.g., peak / off-peak periods).

[0081] The digital twin model refers to a digital virtual model built on the physical entity of the brake and dynamically synchronized with the physical device. By mapping the structure, performance parameters and operating status of the physical brake, it realizes the simulation, monitoring and prediction of the braking process, and provides accurate virtual simulation and parameter benchmarks for adaptive adjustment of braking force. The digital twin model contains information covering three core dimensions: First, brake structure and component parameter information, including digital parameters of physical structures such as brake disc (material, heat dissipation fin design, diameter, etc.), brake caliper (aluminum alloy modular structure dimensions, installation coordinates, etc.), spring adjustment devices (e.g., the distribution of 8 adjustment devices, spring stiffness coefficient, telescopic device drive method, etc. in this embodiment), and electromagnet (silicon steel core parameters, number of coil turns); Second, real-time operating status information, including dynamic data such as spring compression, brake disc / armature / coil temperature, electromagnet magnetic flux, brake gap, response time, and coil electromagnetic force collected and synchronized by displacement, temperature, and Hall sensors; Third, historical and decision-making correlation information, including fault types and compensation correction coefficients in the historical fault case library (CBR), optimal braking force data inferred by neural network, PID optimization parameters, and friction coefficient mapping table (correspondence between temperature and friction coefficient) and braking force safety threshold range related to braking force calculation.

[0082] Furthermore, the application of this digital twin model can achieve the following effects in the adaptive adjustment of braking force: (1) Automatic calibration of the preset braking force threshold: calling the structural parameters, real-time status data and historical braking deviation data in the digital twin model, and dynamically adjusting the threshold through a weighted algorithm to ensure that the threshold adapts to the aging trend of the equipment and changes in working conditions, and avoids the threshold being too loose or too tight. (2) Calculation of braking force deviation and residual: the digital twin model outputs the theoretical braking force simulation value {Y}_{sim} based on the real-time collected braking characteristic data (such as spring compression, temperature, magnetic flux, etc.). (3) Calculation of optimal braking force and spring compression: the digital twin model provides physical structural reference parameters (effective radius r of brake disc, radius R of traction wheel, transmission ratio i) and historical decision data (optimal braking force correction coefficient γ under the same working conditions) to assist neural network reasoning. (4) Virtual verification of adjustment effect: after the spring adjustment device adjusts the compression, the digital twin model simulates the braking process based on the updated physical parameters (such as the adjusted spring compression Δx, real-time temperature T, etc.) and predicts the compensated braking force curve. (5) Dynamic update of model parameters: The gradient descent algorithm is used to adjust the model parameters θ. (Where η is the learning rate, ranging from 0.01 to 0.1,) (For the loss function gradient); Sensor noise is suppressed by adaptive Kalman filtering, the filter gain is dynamically adjusted, and the state equation and covariance matrix are updated; The historical failure case library (CBR) is called to match similar failure scenarios, quickly correct the model parameters, ensure that the digital twin model and the physical brake are dynamically synchronized, and correct parameter drift caused by component aging.

[0083] In a further embodiment, when the braking force deviation is greater than a preset braking force threshold, the PID control parameters of the PID control algorithm used to adjust the coil current are optimized; the adjustment amount of the coil current is output according to the PID control parameters, and a third braking force is output according to the adjustment amount of the coil current.

[0084] In this embodiment, based on real-time braking force Second braking force The result calculates the braking force deviation value and compares it with a set threshold. If the braking force deviation value is less than the set threshold, the status quo is maintained, and the brake is waited for to act again. When the braking force deviation value is greater than the set threshold, the PID control parameters of the PID control algorithm used to adjust the coil current are optimized. In this embodiment, the PID parameters are optimized by combining a genetic algorithm with real-time data. Specifically, the parameters are first set... (1.0-5.0) (0.1-2.0) The initial range is (0.05-1.0), and then optimization is performed with the goal of minimizing braking force deviation. For example, if a brake disc temperature of 250℃ causes the braking force deviation to exceed a preset threshold, then displacement and temperature data are extracted and processed using a genetic algorithm to select (retain parameters with small deviations) and perform crossover (e.g., ...). =3.1 and 3.3 cross to get =3.2), mutation (fine-tuning) =0.4 is 0.42), and the adaptation parameters are output after iteration (e.g. =3.2、 =1.0、 =0.45). And calculate the theoretical spring compression according to the following formula (4). :

[0085] (4)

[0086] in, It is the second braking force (i.e., the final theoretical optimal braking force); R is the radius of the traction wheel; μ is the friction coefficient between the brake pads and the brake disc; k is the spring stiffness coefficient; r is the effective radius of the brake disc.

[0087] Then, based on the calculated theoretical spring compression... The spring compression of the spring adjustment device is dynamically adjusted to compensate for the braking force.

[0088] In step S15, based on the theoretical compression and real-time compression of the spring, the actual compression of the spring is controlled by the telescopic device to enable the spring to output compensating braking force to adaptively adjust the braking force of the brake. The compensating braking force, as a physical supplementary force generated by adjusting the spring compression, can be used to offset the braking force attenuation caused by brake pad wear, spring fatigue, and thermal fade, bringing the real-time braking force closer to the second braking force. Specifically, when the deviation between the real-time braking force and the second braking force exceeds a preset threshold, the compensating braking force is activated. The magnitude of the compensating braking force = second braking force - current real-time braking force (deviation value). Then, according to the formula ΔF = k × Δx (where ΔF is the compensating braking force, k is the spring stiffness coefficient, and Δx is the spring compression adjustment value), a precise match between the compensating braking force and the spring compression adjustment value is achieved.

[0089] In this embodiment, a high-carbon steel helical spring is used, and the preload is infinitely adjustable through a telescopic device. The spring adjustment device, in conjunction with a displacement sensor, monitors the spring compression in real time and feeds it back to the control system. Thus, the dynamic adjustment mechanism of the spring directly corresponds to "adaptive pressure adjustment," achieving precise compensation of braking force through closed-loop feedback.

[0090] In some embodiments, the brake includes an actuator that controls the telescopic device to control the actual compression of the spring. Optionally, the actuator can be an electric push rod (driven by a servo motor), a pneumatic servo valve, or a hydraulic servo valve to adapt to the braking force requirements under different operating conditions. Specifically, in electric mode, the motor receives braking force adjustment commands from the main controller and drives the push rod to extend and retract through precise speed and torque control, pushing the spring to adjust the contact pressure between the brake pad and the brake disc, thereby changing the braking force. In pneumatic mode, the pressure and flow rate of compressed air are adjusted according to the controller commands to push the spring using air pressure. In hydraulic mode, the internal liquid pressure is adjusted according to the controller commands to push the spring.

[0091] Furthermore, the actuator serves not only as the mechanism for performing the basic braking action of the brake, but also as an adjustment mechanism adapted to the braking force requirements of different working conditions. It integrates a dual-mode electric push rod (servo motor driven) and a pneumatic servo valve, specifically performing adjustment actions on three types of structures: First, it adjusts related components of the brake caliper (such as the brake block bracket). During braking, it pushes the brake block closer to the brake disc to achieve "brake engagement," and when releasing the brake, it pulls the brake block away from the brake disc to release the brake. It can also finely adjust the contact force and speed of the brake block. Second, it adjusts in conjunction with the spring adjustment device to assist in controlling the movement speed and stroke of the top plate of the telescopic device, ensuring the accuracy of the spring compression adjustment to optimize braking force compensation. Third, it coordinates with the electromagnet adjustment. Through the current controller, it dynamically adjusts the coil current based on the magnetic field strength collected by the Hall sensor, thereby adjusting the magnetic field strength of the electromagnet to ensure the stability of the electromagnet's attraction force and guarantee the braking force.

[0092] The relationship between the magnetic flux density B of the electromagnet and the coil current I is as follows: ,in, Vacuum permeability ( H / m), Let be the relative permeability of the iron core (approximately 1000 for silicon steel sheets), n be the number of coil turns, and l be the magnetic circuit length. According to this formula, the magnetic field strength B can be stably adjusted by precisely controlling the coil current I.

[0093] Furthermore, as mentioned above, this invention employs an adaptive PID control algorithm to adjust the coil current, achieving precise control of the magnetic field strength, thereby outputting a third braking force. Exemplarily, this invention provides an example of adjusting the coil current using an adaptive PID control algorithm, specifically including the following steps: Step 1. The Hall sensor collects the magnetic field strength B_measured every 10ms; Step 2. The control system calculates the target magnetic field strength B_target; Step 3. Calculate the magnetic field deviation e = B_target - B_measured, where... Step 4. Calculate the current adjustment I_adjust using a PID algorithm, specifically, the magnetic field strength collected in real time by the Hall sensor. ,in, (proportion coefficient) (Integral coefficient) and The differential coefficients are obtained through optimization using a genetic algorithm, with initial values ​​of 1.2, 0.3, and 0.1 respectively; Step 5. Correct the coil current I_new = I_old + I_adjust; Step 6. The current controller drives the power amplifier circuit to output the adjusted current; Step 7. The magnetic field strength changes with the current, and the system continuously monitors and adjusts the coil current.

[0094] In this example, step 5 will also apply a safety limit to the corrected current according to the following formula (5):

[0095] (5)

[0096] in, (Safety lower limit) (Safety limit).

[0097] Meanwhile, this invention provides a specific example of a spring adjusting device for controlling the compression of a spring, which is described below in conjunction with... Figure 7 and Figure 8 The example will be explained in detail.

[0098] Figure 7 This is a perspective view of the structure of a spring adjusting device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the operation of a spring adjusting device according to an embodiment of the present invention.

[0099] like Figure 7 As shown, the spring adjustment device 2 includes: a spring 21, a telescopic device sleeve 22, a telescopic device 23, a telescopic device top plate 24, and a fixing screw 25. The spring 21 provides static elastic force, supplying the braking force of the brake; the telescopic device sleeve 22 protects the internal structure (i.e., the telescopic device) and guides the extension and retraction of the spring; the telescopic device 23 and the telescopic device top plate 24 can be controlled by electric / pneumatic means to increase or decrease the spring compression to transmit the braking force; the fixing screw 25 connects the telescopic device sleeve 22 and the telescopic device 23. This spring adjustment device allows for dynamic adjustment of the braking force. Specifically, eight spring adjustment devices, evenly distributed on the brake arm and softly connected by stop pins, serve as the core actuators. When the elevator braking action is initiated, data such as spring compression, brake disc temperature, and electromagnet flux are collected by displacement, temperature, and Hall effect sensors. Combined with the aforementioned method for calculating the deviation between the real-time braking force and the theoretical optimal braking force value based on elevator load and speed, braking force adjustment is triggered when the braking force deviation exceeds a preset threshold.

[0100] like Figure 8 As shown, before the spring adjustment device is activated, spring 21 is in its initial state, unaffected by external force, and maintains its original length. After the spring adjustment device is activated, when it is necessary to increase / decrease the braking force, the telescopic device pushes the top plate 24 of the telescopic device forward / backward, causing the spring 21 to increase / decrease its compression, resulting in greater / less pressure from the spring, thereby increasing / decreaseing the braking force. When the spring compression is increased, the spring force is increased, that is, the pressure on the friction plate is increased, thereby increasing the friction force of the clamp disc (i.e., the braking force).

[0101] Optionally, based on the calculated theoretical spring compression... The dynamic adjustment of the spring compression of the spring regulating device to compensate for braking force specifically includes: when increased braking force is needed, the telescopic device pushes the top plate forward, increasing the spring compression and generating greater pressure, thereby enhancing braking friction (i.e., braking force); when reduced braking force is needed, the top plate retracts, reducing spring compression and decreasing braking friction. Finally, the digital twin model parameters are updated, waiting for the brake to activate again; if the braking force deviation is less than the set threshold (the set braking force safety threshold), the status quo is maintained, waiting for the brake to activate again. This forms a closed loop of "sensing-diagnosis-adjustment-update," achieving adaptive braking force control and ensuring stable braking performance.

[0102] In a further embodiment, the adjustment method further includes: recording and storing the acquired braking characteristic data, processed feature fusion data, and braking force adjustment result data, and updating the parameters in the digital twin model based on the brake according to the measured brake data after braking force adjustment.

[0103] In a further embodiment, the adjustment method further includes: analyzing and diagnosing positioning brake faults based on the braking characteristic data.

[0104] In some implementations, analyzing and diagnosing brake faults based on the braking characteristic data includes: preprocessing the braking characteristic data, the preprocessing including: filtering, amplification, linearization, and spatiotemporal alignment; performing feature extraction and deep learning feature fusion on the preprocessed data to obtain feature fusion data; and locating brake faults based on the feature fusion data.

[0105] The process of locating brake faults based on the feature fusion data may include: extracting fault features from the feature fusion data; inputting the fault features into the pre-trained neural network model to obtain brake fault information; and sending early warning information based on the fault information.

[0106] The adaptive braking force adjustment method described in Embodiment 1 of this invention uses displacement, temperature, and electromagnetic force sensors to monitor the brake's condition around the clock, collecting real-time data on mechanical, thermal, clearance, and dynamic characteristics. Intelligent algorithms are used to perform in-depth data mining and feature analysis, accurately diagnosing potential faults such as brake shoe wear, spring fatigue, coil aging, and motion jamming. Furthermore, a trend prediction model is used to assess the health status of components, achieving a shift from "reactive maintenance" to "predictive maintenance." Thus, it can not only issue warnings accurate to the specific clamp position and cause, but also compensate and adjust the braking force online through adaptive control mechanisms (such as pressure regulating devices and genetic PID algorithms), forming a closed-loop management system from "perception" to "diagnosis" to "execution."

[0107]

Example 2

[0108] Figure 9 This is a flowchart illustrating the adaptive adjustment method for braking force of an elevator caliper disc brake according to Embodiment 2 of the present invention. This adjustment method is implemented through a control system.

[0109] like Figure 9 As shown, in Embodiment 2 of the present invention, the adaptive adjustment method of the braking force of the elevator caliper disc brake may include at least the following steps S201, S202, S203, S204, S205, S206, S207, S208, S209, S210, S211 and S212, which are described in detail below.

[0110] Step S201: Control system loads the trained neural network model.

[0111] Step S202: Control the system and initialize the PID control parameters. In this embodiment, steps S201 and S202 can be executed simultaneously.

[0112] Step S203: Wait for the brake to activate.

[0113] Step S204: After the brake is activated, multi-source data such as spring compression, brake disc temperature, armature temperature, and magnetic flux strength of the electromagnet are collected simultaneously through displacement sensor, temperature sensor, magnetic flux sensor, etc.

[0114] Step S205 involves data fusion and processing of the aforementioned multi-source data. Specifically, the multi-source data undergoes filtering, amplification, linearization preprocessing, spatiotemporal alignment, and deep learning feature fusion to obtain feature fused data. By combining spatiotemporal alignment technology and deep learning algorithms, multi-dimensional data features such as mechanical, thermal, and gap characteristics are fused to improve the accuracy of subsequent fault diagnosis and prediction.

[0115] Step S206 involves calculating key parameters such as the action gap, response time, coil electromagnetic force, and armature temperature. In this embodiment, steps S205 and S206 can be executed simultaneously.

[0116] Step S207 involves jointly analyzing the feature fusion data obtained in step S205 and the key parameter data calculated in step S206 using an improved GNN and a Bayesian framework to diagnose and locate brake faults, such as brake pad wear and coil aging. The improved GNN effectively captures fault propagation characteristics, while the Bayesian framework outputs the fault probability of various faults through uncertainty quantification. Furthermore, when the inference result triggers a fault warning, a fault warning signal is generated; otherwise, braking force commands are directly output according to subsequent steps.

[0117] In this embodiment, the core of fault identification using an improved graph neural network lies in treating equipment components as "nodes" and physical connections as "edges" to capture fault propagation (such as bearing wear → rotor vibration). Specifically, this includes steps 1) to 4):

[0118] Step 1) Construct the graph structure: Nodes = Sensor monitoring points (such as bearings, coils), Edges = Physical connections of components, Features = Multi-dimensional sensor data.

[0119] Step 2) Spatiotemporal attention weighting: assign high weights to faulty nodes (such as worn bearings).

[0120] Step 3) Graph Convolution Message Passing: Update the current node features using the features of neighboring nodes using the following formula (6):

[0121] (6)

[0122] in, The updated node features are represented by N(u); N(u) represents the neighbors of node u. Node features; W is the convolution kernel weight, b is the bias. The average weight of the neighbors.

[0123] Step 4) Two-level classification: First determine "health / failure", then further subdivide the category (such as wear / aging).

[0124] In this embodiment, the classification of caliper brake fault types (such as "brake pad wear" and "excessive brake clearance") is achieved with high accuracy using an optimized connection weight method, specifically implemented through the following algorithm steps:

[0125] Step a. Calculate the distance (Euclidean distance, used to select the winning neuron) using the following formula (7):

[0126] (7)

[0127] Where, x j Input features (such as IMF features filtered by WKS); w ij The connection weights between the input layer and the competition layer; d i The smaller the value, the more similar the neuron is to the input features, and the more likely it is to become the "winning neuron".

[0128] Step b. Implement LVQ1 weight update (optimize classification boundary) using the following formulas (8) and (9):

[0129] , (8)

[0130] , (9)

[0131] Where η is the learning rate (usually ranging from 0.1 to 0.5). The input sample contains the actual fault labels; The fault label is the one corresponding to the winning neuron.

[0132] Through the above update rules, when the labels are consistent, the weights are adjusted towards the input features (enhancing the matching degree); when the labels are inconsistent, the weights are adjusted away from the input features, gradually optimizing the classification boundary and improving the classification accuracy of fault types. Therefore, by adopting LVQ, the early warning can be accurate to "which type of fault corresponds to which component" (e.g., "coil aging, corresponding to electromagnet 13"), thereby reducing unnecessary maintenance.

[0133] Furthermore, the "braking force adjustment" of this invention needs to formulate a strategy based on the fault type. For example, "excessive brake clearance" requires priority adjustment of the brake pad stroke, and "heat fade" requires correction of the friction coefficient. Therefore, the fault type output by LVQ can provide a basis for "differentiated adjustment" and avoid "one-size-fits-all" compensation errors.

[0134] In this embodiment, the quantification of uncertainty using the Bayesian framework specifically includes the following steps 1) to 4):

[0135] Step 1) Set the prior probability P(θ): Set the prior probability of the fault type θ based on historical data (e.g., the prior probability of coil aging). ).

[0136] Step 2) Calculate the likelihood function P(D|θ): Given the fault type θ, calculate the probability of the occurrence of monitoring data D (such as temperature, gap, electromagnetic force, etc.) and use Gaussian distribution to model this probability.

[0137] Step 3) Update the posterior probability P(θ|D): Update the posterior probability P(θ|D) according to the following Ye Bayes formula (10):

[0138] (10)

[0139] Where P(D) is the evidence factor, .

[0140] Step 4) Output the fault type θ with the highest posterior probability as the diagnostic result.

[0141] By using a Bayesian framework to quantify uncertainty, we can resolve diagnostic errors caused by data noise and sample imbalance, while outputting the precise probability of occurrence for various faults, thus improving the reliability and interpretability of diagnostic results.

[0142] Step S208: Calculate the theoretically optimal braking force (i.e., the final theoretically optimal braking force / second braking force in the aforementioned Example 1) through neural network reasoning.

[0143] Step S209: Calculate the deviation between the real-time braking force and the theoretical optimal braking force, i.e., the braking force deviation, and determine whether the braking force deviation value is greater than a set threshold. When the braking force deviation value is less than or equal to the set threshold, return to step S203 and wait for the brake to act; when the braking force deviation value is greater than the set threshold, execute step S210.

[0144] Step S210: Optimize PID control parameters. Specifically, when the braking force deviation is greater than a preset braking force threshold, optimize the PID control parameters of the PID control algorithm used to adjust the coil current; output the adjustment amount of the coil current according to the PID control parameters, and output a third braking force according to the adjustment amount of the coil current.

[0145] In this embodiment, the PID parameters are optimized using a genetic algorithm combined with real-time data. Specifically, the parameters are first set... (1.0-5.0) (0.1-2.0) The initial range is (0.05-1.0), and then optimization is performed with the goal of minimizing braking force deviation. For example, if a brake disc temperature of 250℃ causes the braking force deviation to exceed a preset threshold, then displacement and temperature data are extracted and processed using a genetic algorithm to select (retain parameters with small deviations) and perform crossover (e.g., ...). =3.1 and 3.3 cross to get =3.2), mutation (fine-tuning) =0.4 is 0.42), and the adaptation parameters are output after iteration (e.g. =3.2、 =1.0、 =0.45). After the PID control parameters are initialized, when the braking force deviation exceeds the threshold, the PID control parameters are optimized through a genetic algorithm to output a precise coil current adjustment. Combined with the spring compression adjustment, this ultimately achieves precise output of the third braking force and braking force compensation. Thus, by focusing on the "stable electromagnetic field strength," the response speed and basic accuracy of braking actions (brake release / brake engagement) can be indirectly guaranteed. At the same time, the PID control parameters can offset interference such as magnetic flux fluctuations and coil aging by dynamically correcting the coil current.

[0146] Step S211: Dynamically adjust the spring compression. Specifically, calculate the theoretical spring compression based on the aforementioned formula (4). Then, based on the calculated theoretical spring compression... The spring compression of the spring adjustment device is dynamically adjusted to compensate for the braking force. In this embodiment, after optimizing the PID control parameters, the coil current adjustment is immediately output based on the PID control parameters. At the same time, the theoretical spring compression is calculated in conjunction with the third braking force. The spring adjustment device pushes / retracts the top plate through the telescopic device according to the compression adjustment value to complete the braking force compensation, forming a seamless connection between "parameter optimization → physical adjustment".

[0147] Step S212, update the parameters of the digital twin model. The following functions can be achieved by applying the digital twin model: (1) Automatic calibration of the preset threshold of braking force: call the structural parameters, real-time status data and historical braking deviation data in the digital twin model, and dynamically adjust the threshold through a weighted algorithm to ensure that the threshold is adapted to the aging trend of the equipment and the change of working conditions, and avoid the threshold being too loose or too strict. (2) Calculation of braking force deviation and residual: the digital twin model outputs the theoretical braking force simulation value {Y}_{sim} based on the real-time collected braking characteristic data (such as spring compression, temperature, magnetic flux, etc.). (3) Calculation of optimal braking force and spring compression: the digital twin model provides physical structural benchmark parameters (effective radius r of brake disc, radius R of traction wheel, transmission ratio i) and historical decision data (optimal braking force correction coefficient γ under the same working conditions) to assist neural network reasoning. (4) Virtual verification of adjustment effect: after the spring adjustment device adjusts the compression, the digital twin model simulates the braking process based on the updated physical parameters (such as the adjusted spring compression Δx, real-time temperature T, etc.) and predicts the braking force curve after compensation. (5) Dynamic update of model parameters: The gradient descent algorithm is used to adjust the model parameters θ. (Where η is the learning rate, ranging from 0.01 to 0.1,) (For the loss function gradient); Sensor noise is suppressed by adaptive Kalman filtering, the filter gain is dynamically adjusted, and the state equation and covariance matrix are updated; The historical failure case library (CBR) is called to match similar failure scenarios, quickly correct the model parameters, ensure that the digital twin model and the physical brake are dynamically synchronized, and correct parameter drift caused by component aging.

[0148] The adaptive braking force adjustment method described in Embodiment 2 of this invention integrates sensor data such as displacement, temperature, and magnetic flux through multi-source data fusion technology. Combined with a neural network algorithm, it analyzes the brake caliper's operating status in real time, and on this basis, achieves dynamic adjustment of braking current and adaptive control of the engagement force. Specifically, by combining a genetic algorithm to optimize PID parameters and a digital twin model, the braking force is dynamically adjusted in real time to adapt to changes in the real-time load, operating speed, and brake disc temperature of equipment such as elevators and cranes. This allows for adaptation to various complex working conditions, avoiding problems of "inability to brake" or "over-braking," thereby ensuring stable braking performance, improving equipment operating safety, and solving the problems of braking force attenuation, low control accuracy, delayed fault diagnosis, and high maintenance costs caused by brake pad wear, structural component jamming, and thermal fade effects in traditional brakes. Simultaneously, through online compensation and adjustment by the adaptive adjustment mechanism, manual intervention and spare parts replacement frequency are reduced, lowering the total life-cycle maintenance cost and improving product economy and market competitiveness.

[0149]

Example 3

[0150] Embodiment 3 of the present invention provides a braking force adaptive throttle method based on a digital twin model, which specifically includes the following steps:

[0151] Step S31, sliding window division: Divide the data window with a fixed period of 5 seconds, and collect the measured data D (displacement, temperature, brake force, etc.) within the window. The value of the fixed period can be flexibly set according to actual needs.

[0152] Step S32, Residual Calculation: Compare the residuals of the digital twin model simulation output Y_sim with the measured data Y_meas. ,like (Threshold) triggers the calibration of the digital twin model. Y_sim may include, but is not limited to, the following data: 1. Braking force related data: real-time braking force, theoretical optimal braking force, etc.; 2. Spring adjustment device status data: spring compression, spring force; 3. Electromagnet core status data: electromagnet flux, coil current; 4. Temperature related data: brake disc temperature, armature temperature / coil temperature; 5. Braking response and clearance data: brake clearance, braking response time; 6. Brake block and friction related data: brake block wear, friction coefficient μ(T).

[0153] Step S33, define the residual loss function (optimization objective), as shown in the following formula (11):

[0154] (11)

[0155] Where n is the amount of data within the window; θ is the model parameter to be optimized, such as the elasticity coefficient k.

[0156] Step S34, parameter optimization: Adjust the model parameters θ (such as spring elastic coefficient k and friction coefficient μ) using the gradient descent algorithm.

[0157] Step S35, noise suppression: Minimize the loss function L(θ) through gradient descent, update the model parameters θ, and then use adaptive Kalman filtering (noise suppression) to dynamically adjust the filter gain to filter sensor noise (such as abnormal data caused by electromagnetic interference).

[0158] In this embodiment, the gradient descent parameters are updated using the following formula (12):

[0159] (12)

[0160] Where η is the learning rate, taking a value between 0.01 and 0.1. This is the gradient of the loss function.

[0161] The specific steps for achieving state and measurement updates using adaptive Kalman filtering are as follows:

[0162] (1) State prediction: Where F_k is the state transition matrix, B_k is the control matrix, and u_k is the control input;

[0163] (2) Covariance prediction: Where Q_k is the process noise covariance;

[0164] (3) Kalman gain calculation: Where H_k is the observation matrix and R_k is the measurement noise covariance;

[0165] (4) Status update: Where z_k is the measured data;

[0166] (5) Covariance update: , where I is the identity matrix.

[0167] Step S36, Knowledge Base Assisted Optimization: Call the Historical Fault Case Library (CBR), and call the preset correction coefficient by matching similar fault scenarios (such as spring fatigue, brake shoe wear, etc.) to shorten the parameter convergence time. For example, calling the correction coefficient corresponding to the spring fatigue case can shorten the convergence time by 40%.

[0168] Step S37: Adaptively adjust the application scenario.

[0169] For example, in scenario one: brake shoe wear leads to increased braking clearance → changes in measured displacement data → residual e > threshold → gradient descent to adjust model clearance parameters → synchronization of digital twin model with physical equipment → ensuring fault diagnosis accuracy;

[0170] For scenario two: Spring fatigue leads to a decrease in the elastic coefficient → a decrease in the measured brake force data → residual e > threshold → call the spring fatigue case in the CBR knowledge base → quickly correct the elastic coefficient parameter → shorten the model convergence time by 40%.

[0171] The adaptive braking force method of Embodiment 3 of this invention can achieve the following technical effects: (1) Real-time data drives the digital twin model update, ensuring dynamic synchronization between the model and physical equipment (such as parameter drift caused by brake shoe wear and spring fatigue). Specifically, multi-dimensional measured data (displacement, temperature, and brake force) are collected in a sliding window time sequence. The simulation and measured deviations are quantified by combining the residual loss function. The key model parameters such as elastic coefficient and friction coefficient are iteratively optimized by the gradient descent algorithm to ensure real-time synchronization between the digital twin model and the physical braking equipment. In addition, the adaptive Kalman filter is superimposed to dynamically suppress sensor noise such as electromagnetic interference, effectively reducing the impact of data interference on parameter optimization, keeping the model residual within the 5% threshold, and significantly reducing the braking force adjustment error, providing accurate data and model support for fault diagnosis and adaptive adjustment. (2) Through the closed-loop design of the whole process of "data acquisition - residual detection - parameter optimization - noise suppression - case assistance", the automatic calibration and parameter update of the model can be completed without manual intervention, reducing manual maintenance costs.

[0172]

Example 4

[0173] Figure 10 This is a schematic diagram of the adaptive braking force adjustment system of an elevator caliper disc brake according to Embodiment 4 of the present invention.

[0174] like Figure 10 As shown, the adjustment system includes: a data acquisition module 310, a braking force calculation module 320, a spring compression calculation module 330, a braking force compensation module 340, a braking force adjustment module 350, a fault diagnosis module 360, and a model update module 370.

[0175] The data acquisition module 310 is used to acquire the braking characteristic data of the brake in real time. The braking characteristic data includes: the real-time compression of the spring in the spring adjustment device, the brake disc temperature, the armature temperature, the coil temperature, and the magnetic flux of the electromagnet.

[0176] The braking force calculation module 320 is used to calculate the real-time braking force and theoretical braking force based on the braking scenario and the braking characteristic data after the brake is activated, and to obtain the second braking force based on the theoretical braking force and the pre-trained neural network model.

[0177] The spring compression calculation module 330 is used to calculate the braking force deviation based on the real-time braking force and the second braking force, and when the braking force deviation is greater than the braking force preset threshold, to calculate the theoretical compression of the spring based on the second braking force.

[0178] The preset threshold of braking force is automatically adjusted based on the digital twin model, historical braking characteristic data, and actual brake measurement data.

[0179] The braking force compensation module 340 is used to control the actual compression of the spring through the telescopic device based on the theoretical compression and the real-time compression of the spring, so that the spring outputs a compensating braking force to adaptively adjust the braking force of the brake.

[0180] In this embodiment, the brake has an actuator that controls the telescopic device to control the actual compression of the spring.

[0181] The braking force adjustment module 350 is used to optimize the PID control parameters of the PID control algorithm for adjusting the coil current when the braking force deviation is greater than the braking force preset threshold; and to output the adjustment amount of the coil current according to the PID control parameters, and output a third braking force according to the adjustment amount of the coil current.

[0182] By optimizing the parameters of the PID control algorithm used to adjust the coil current, the braking force can be dynamically adjusted to adapt to the real-time load, running speed, and brake disc temperature changes of the elevator equipment, avoiding problems such as "inability to brake" or "over-braking" and improving the safety of equipment operation.

[0183] The fault diagnosis module 360 ​​is used to analyze and diagnose positioning brake faults based on the braking characteristic data.

[0184] In some implementations, analyzing and diagnosing brake faults based on the braking characteristic data includes: preprocessing the braking characteristic data, the preprocessing including: filtering, amplification, linearization, and spatiotemporal alignment; performing feature extraction and deep learning feature fusion on the preprocessed data to obtain feature fusion data; and locating brake faults based on the feature fusion data.

[0185] In a further embodiment, locating brake faults based on the feature fusion data includes: extracting fault features from the feature fusion data; inputting the fault features into the pre-trained neural network model to obtain brake fault information; and sending early warning information based on the fault information.

[0186] By extracting multimodal features through deep learning and locating brake faults based on these features, a shift from "post-event maintenance" to "predictive maintenance" can be achieved. This allows for precise location of specific clamp positions and fault causes, thereby providing early warnings of component performance degradation and reducing the risk of unplanned downtime.

[0187] The model update module 370 is used to record and store the acquired braking characteristic data, processed feature fusion data, and braking force adjustment result data, and to update the parameters in the digital twin model based on the brake according to the measured data of the brake after braking force adjustment.

[0188] Updating the parameters of the digital twin model based on the brake using the adjusted measured brake data ensures dynamic synchronization between the model and the physical brake, forming a complete closed loop of "data acquisition - processing - decision-making - execution - model update". This enables dynamic linkage between the physical entity and the digital model, with real-time feedback of measured data ensuring the model always closely matches the actual operating state of the brake, avoiding a disconnect between theory and practice. An adaptive optimization mechanism is formed, continuously refining the decision logic and execution parameters through closed-loop iteration, improving the accuracy and stability of braking force adjustment and control. Simultaneously, it provides data support for predictive maintenance, further enhancing the safety and reliability of elevator operation.

[0189] The adjustment system described in Embodiment 4 of this invention collects mechanical, thermal, clearance, and dynamic characteristic data in real time, and uses intelligent algorithms to perform in-depth data mining and feature analysis to accurately diagnose potential faults such as brake shoe wear, spring fatigue, coil aging, and motion jamming. Furthermore, it assesses the health status of components through a trend prediction model, achieving a shift from "reactive maintenance" to "predictive maintenance." Thus, it can not only issue warnings accurate to the specific clamping position and cause, but also compensate and adjust braking force online through adaptive control mechanisms (such as pressure regulating devices and genetic PID algorithms), forming a closed-loop management system from "sensing" to "diagnosis" to "execution."

[0190] Specifically, the dynamic adjustment of braking force in the adaptive braking force adjustment system of Embodiment 4 of the present invention is achieved through a four-step closed-loop system of "perception-analysis-execution-verification", and the specific implementation method is as follows:

[0191] 1. Sensing stage: Simultaneously collect multi-dimensional data such as brake pad contact surface pressure, spring compression, and temperature, and then output the optimal spring compression based on the neural network model built on historical working condition data.

[0192] 2. Analysis phase: The model training adopts reinforcement learning algorithm to continuously optimize the compensation strategy. When the system detects that the measured braking force is lower than the safety threshold (80% of the rated braking force), the expansion amount is calculated and dynamic compensation is triggered.

[0193] 3. Execution phase: The spring adjustment device uses a telescopic structure (one of electric, pneumatic, or hydraulic) to control the movement of the top plate. Taking electric drive as an example, its control logic is as follows: When the braking force is insufficient, the servo motor rotates forward, pushing the top plate forward, and the spring compression increases to 0.25-0.35mm; when the braking force is overloaded, the servo motor reverses, the top plate moves backward, and the spring compression decreases to 0.25-0.35mm.

[0194] 4. Closed-loop verification and optimization stage: After each adjustment, the system immediately verifies the braking force, simulates the adjustment effect through a digital twin model, predicts the braking force curve after compensation, and automatically records the adjustment parameters to optimize the neural network weights and achieve closed-loop self-learning.

[0195] In one exemplary embodiment, the circuit composition of the adaptive braking force adjustment system of the elevator caliper disc brake of Embodiment 4 of the present invention includes:

[0196] 1. Power module: Adopts a wide voltage input design, supports AC-DC conversion, and provides a stable DC power supply for the system;

[0197] 2. Brake coil drive module: integrates a dual-mode actuator of electric push rod or pneumatic servo valve for dynamic adjustment of braking force;

[0198] 3. Sensor Interface Module: Equipped with analog signal conditioning circuitry and analog-to-digital conversion module, it can acquire multi-dimensional sensor data such as displacement, temperature, and magnetic flux in real time;

[0199] 4. Control Unit: Based on a high-performance microprocessor, it integrates a neural network acceleration module to achieve multi-source data fusion and intelligent decision-making.

[0200]

Example 5

[0201] Figure 11 This is a schematic flowchart of an adaptive braking force adjustment method for an elevator caliper disc brake according to Embodiment 5 of the present invention. This adjustment method is implemented using the adjustment system of Embodiment 4.

[0202] like Figure 11 As shown, in Embodiment 5 of the present invention, the adaptive adjustment method of braking force of the elevator caliper disc brake may include at least the following steps S41, S42, S43, S44, S45, S46, S47 and S48, which are described in detail below.

[0203] Step S41 involves monitoring the brake block stroke and collecting the spring compression using a displacement sensor, collecting the temperature of the brake disc, armature, and coil using a temperature sensor, and obtaining the magnetic flux of the electromagnet using a Hall sensor, thereby simultaneously collecting three key data types: displacement, temperature, and magnetic flux.

[0204] Step S42 involves performing spatiotemporal alignment processing on the multi-source data collected in step S41. In this embodiment, time synchronization is achieved based on dynamic time warping, and spatial registration is achieved using the least squares method, ensuring a unified spatiotemporal reference for the data (i.e., unifying the time reference and spatial coordinates of the data).

[0205] This invention provides a spatiotemporal alignment technique to solve the data misalignment problem. This spatiotemporal alignment technique can unify sensor data from different installation locations to the same spatiotemporal reference, with a synchronization error of less than 5ms and a spatial registration accuracy of 0.1mm. Specifically, the spatiotemporal alignment technique includes the following steps: (1) Time synchronization: Based on dynamic time warping (DTW), heterogeneous time series are mapped to a 100ms interval time axis; (2) Spatial registration: A three-dimensional global coordinate system of the device is established, and the local coordinates of the sensor are unified to the global coordinate system using the least squares method; (3) Data completion: For low sampling frequency data (such as temperature), cubic spline interpolation is used to generate equally spaced data (such as 100Hz).

[0206] For example, consider two original sequences: (Length is m) (With a length of n), for example, the first time series X represents raw sensor data (such as a displacement sequence), and the second time series Y represents raw data from another sensor (such as a temperature sequence). The DTW distance, used to measure the similarity of time series, can be calculated using the following method:

[0207] First, construct the distance matrix: 3D matrix D, matrix elements Defined as a distance metric between sequence elements, i.e., matrix elements. (Euclidean distance is used here, but other distance metrics such as Manhattan distance can also be selected as needed).

[0208] Secondly, constrain the optimal path: find the path from the starting point (1,1) to the ending point (m,n). The path constraint conditions are met: .

[0209] Then, the DWT distance between the two original sequences is calculated according to the following DTW distance formula (13):

[0210] (13)

[0211] Where K is the length of path γ (satisfying) ; γ represents the distance to the k-th point on the path; minγ means "take the value with the minimum cumulative distance among all legal paths".

[0212] Furthermore, the least squares method is used for spatial registration to achieve fitting between local and global coordinates. The specific process is as follows:

[0213] Let the local coordinates be (xi, yi), the global coordinates be (Xi, Yi), and the fitting function be... , The optimal registration parameters are solved by minimizing the following loss function (14):

[0214] (14)

[0215] Where N is the number of coordinate samples, and a, b, c, d, e, and f are the registration parameters to be optimized.

[0216] Step S43 involves performing feature fusion calculations on the spatial registration data obtained in step S42, and simultaneously extracting fault features from the fused features. In this embodiment, feature fusion is achieved by extracting multimodal features through deep learning and weighting them.

[0217] For example, the deep learning algorithm architecture is: multi-branch CNN + cross-channel attention mechanism + residual connection (used to solve the gradient vanishing problem). The specific steps of using this deep learning algorithm to achieve feature fusion include: (1) multi-modal input: input multi-source data such as displacement, temperature gradient (spatial dimension), and electromagnetic force into different CNN branches respectively; (2) parallel feature extraction: extract local features of each modality data through 3×3 convolution kernels, and perform dimensionality reduction through max pooling operation; (3) cross-channel attention weighting: assign high weights to fault-sensitive features (such as bearing fault frequency features); (4) residual fusion: fuse multi-branch features through Identity Mapping to generate a global feature vector.

[0218] The cross-channel attention weight calculation specifically includes the following steps: (1) Set the feature map (C is the number of channels, H is the height, and W is the width), and global average pooling is performed using the following formula (15): For each channel c, the average value of all pixels in that channel is calculated, compressing the spatial dimension (from H×W to 1):

[0219] (15)

[0220] in, F(c,h,w) represents the output value of the c-th channel in the feature map after global average pooling; c represents the "channel index" of the feature map, used to locate a specific feature channel; F(c,h,w) represents the "local feature pixel value at a specific location" on the feature map, where (c,h,w) corresponds to the three-dimensional coordinates of "channel-height-width", and the value reflects the feature intensity of the corresponding channel at that spatial location; H represents the "vertical height" of the feature map, that is, the number of pixels in the vertical direction of the feature map; W represents the "horizontal width" of the feature map, that is, the number of pixels in the horizontal direction of the feature map; H×W represents the "total number of pixels" of a single channel of the feature map.

[0221] (2) Weight generation: Channel weights are generated through a two-layer fully connected network and a sigmoid activation function. The calculation formula is shown in the following formula (16):

[0222] (16)

[0223] Where W represents the final generated "channel weight vector"; σ represents the sigmoid activation function; W1 and W2 are the weight matrices of the two fully connected layers, respectively; b1 and b2 are the "bias vectors" of the corresponding two fully connected layers, respectively; R C The dimension identifier represents the weight vector W, where R represents the real number field and C represents the vector dimension, which is equal to the number of channels in the feature map. That is, the weight vector contains C elements, which correspond one-to-one with the feature channels.

[0224] (3) Weighted features: The generated channel weights are multiplied element-wise with the original feature map using the following formula (17) to obtain the weighted feature map:

[0225] (17)

[0226] in, The weighted feature map is represented by W; W represents the channel weight vector, where the larger the weight value, the more critical the feature of the corresponding channel; ⊙ represents the element-wise multiplication operator; F represents the original feature map, which is then processed by multi-branch CNN in parallel extraction and max pooling dimensionality reduction to form a multimodal feature set.

[0227] Step S44: Input the fault features extracted in step S43 into the neural network for inference. If the inference result triggers the fault warning condition, generate a fault warning signal; if no warning is needed, execute step S45 and directly output the braking force command.

[0228] In this embodiment, the brake is first diagnosed to check for faults or potential faults. If a fault related to braking force is found, such as jamming or insufficient braking force, then the braking force is adjusted.

[0229] In one exemplary implementation, when the coil ages or the spring force weakens and the extension exceeds the standard range (0.25mm-0.35mm), the system issues a warning "Insufficient braking force detected" and performs fault location. At the same time, it needs to adaptively adjust according to relevant parameters. At this time, the sensor collects data (temperature, gap, electromagnetic force, etc.) and uses data fusion to feed back the status parameters to the pressure regulating device for adaptive adjustment to increase the braking force.

[0230] Step S45: Output braking force command.

[0231] Step S46: The braking force command drives the actuator to dynamically adjust the braking force. The actuator can be an electric, pneumatic, or hydraulic actuator.

[0232] Step S47 involves recording and storing the collected raw data, processed feature data, and adjustment command results. In this embodiment, steps S46 and S47 are executed simultaneously.

[0233] In a further implementation, by combining historical fault classification data accumulated in the data recording and storage, the LVQ algorithm can deeply mine and statistically analyze the characteristic change trends of various faults. For example, it can output quantitative results such as "the growth curve of brake pad wear with the number of equipment operations" and "the change law of brake clearance with the running time," providing data and model support for subsequent accurate predictive maintenance. For example, it can provide a 30-day advance warning that "the brake pad wear will reach the safety limit" and "timely replacement is required to avoid brake failure." This enables a shift from "reactive maintenance" to "predictive maintenance," accurately locating specific caliper faults and causes, providing early warnings of component performance degradation, and reducing the risk of unplanned downtime. This solves the problems of low efficiency and high subjectivity in existing maintenance models that rely primarily on periodic manual inspections and experience-based fault judgment.

[0234] Step S48 updates the parameters of the digital twin model based on the measured data after the actuator adjustment, ensuring that the model is dynamically synchronized with the physical actuator. This forms a complete closed loop of "data acquisition - processing - decision-making - execution - model update".

[0235] In one exemplary implementation, when the coil ages or the spring force weakens, causing the extension / retraction to exceed the standard range (0.25mm-0.35mm), the system issues a warning "Insufficient braking force detected" and performs fault location. Simultaneously, sensors collect multi-dimensional data such as temperature, gap, and electromagnetic force, and after data fusion processing, feed back the status parameters to the pressure regulating device. The pressure regulating device then adaptively adjusts based on the feedback parameters to increase the braking force.

[0236] In other implementations, the system achieves dynamic compensation of braking force through the following closed-loop process:

[0237] Step 1) The sensor monitors the brake force (braking force) data in real time.

[0238] Step 2) Compare the braking force data monitored in real time by the sensor with the preset safety threshold. When the measured braking force value is lower than the designed safety threshold (e.g., 80% of the rated braking force), the braking force adjustment is triggered.

[0239] Step 3) Neural network calculates the optimal braking force compensation amount: The neural network takes "real-time state (temperature, load, etc.) + operating condition constraints (safe range, etc.) + error target (braking force deviation)" as input, and combines parameters such as "network structure, training parameters, and prior mapping table" to accurately calculate the optimal compensation amount to offset "wear + thermal fade", thereby realizing adaptive control of braking force.

[0240] Step 4) The adjustment system issues a command to control the spring adjustment device to adjust the spring compression: the braking force compensation is achieved by adjusting the spring compression, and the two are linearly positively correlated. ,in, is the braking force compensation amount (unit: N); k is the spring stiffness coefficient (unit: N / mm). This is the spring compression adjustment value (unit: mm).

[0241] Step 5) Feed the measured braking force data back to the system to verify the compensation effect.

[0242] Step 6) Continuously optimize compensation parameters using a digital twin model to ensure that the braking force is always maintained above the safety threshold. The compensation parameters include, but are not limited to: 1. Core parameters for braking force calculation and compensation, including spring stiffness coefficient (k), dynamic friction coefficient (μ(T)), effective brake disc radius (r), safety redundancy coefficient (β), and braking force correction coefficient (γ). 2. Compensation execution parameters for the spring adjustment device, including theoretical spring compression (Δx), spring compression adjustment range (Δxmin, Δxmax), and telescopic device response speed. 3. PID control optimization parameters, including PID proportional coefficient Kp, PID integral coefficient Ki, PID derivative coefficient Kd, and coil current safety range. 4. Fault and threshold-related compensation trigger parameters, including preset safety threshold for braking force, brake pad wear threshold, and thermal fade judgment temperature reference. 5. Digital twin model calibration parameters, including model residual threshold, digital twin update cycle, and historical fault compensation coefficients.

[0243] Step 7) The system compensates for brake force reduction caused by brake pad wear, spring fatigue, or thermal fade by dynamically adjusting the spring compression (adjustment range: 0.25-0.35mm). For example, when the brake force reduction exceeds 5%, the system automatically activates the compensation mechanism to ensure that the braking performance is always maintained within the design safety threshold (80%-100% of rated power).

[0244] The adaptive braking force adjustment method described in Embodiment 5 of this invention, through multi-source data fusion and neural network construction of an adaptive braking control system, achieves dynamic optimization of braking force and fault early warning, and can be applied to scenarios requiring high-precision braking control, such as elevators and cranes. Specifically, the adjustment system realizes optimal braking force reasoning under complex working conditions by loading a neural network model, and forms an adaptive adjustment mechanism by combining dynamic optimization of PID parameters and real-time updates of the digital twin model. Simultaneously, the data acquisition module monitors multi-dimensional parameters such as displacement, temperature, and magnetic flux in real time. After filtering and amplification, the neural network completes fault diagnosis and braking force optimization, and finally, the actuator dynamically adjusts the spring compression, ensuring that the braking performance is always maintained above the design safety threshold.

[0245] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software combined with a hardware platform. Based on this understanding, all or part of the technical solution of the present invention that contributes to the background art can be embodied in the form of a computer software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0246] Correspondingly, embodiments of the present invention also provide a computer-readable storage medium storing computer-readable instructions or programs thereon. When executed by a processor, the computer-readable instructions or programs cause a computer to perform the following operations, which include the steps included in the adjustment method described in any of the above embodiments, and will not be repeated here. The storage medium may include, for example, an optical disc, a hard disk, a floppy disk, flash memory, magnetic tape, etc.

[0247] Furthermore, embodiments of the present invention also provide a computer device including a memory and a processor. The memory is used to store one or more computer-readable instructions or programs, wherein the one or more computer-readable instructions or programs, when executed by the processor, can implement the adjustment method described in any of the above embodiments. The computer device may be, for example, a server, a desktop computer, a laptop computer, a tablet computer, etc.

[0248] This invention also provides a computer program product including a computer program containing program code for executing the adjustment method shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the adjustment method provided in the embodiments of this disclosure.

[0249] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on a user's computing device, partially on a user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0250] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Therefore, the scope of protection of the present invention should be determined by the claims.

Claims

1. A method for adaptive adjustment of braking force of a brake of an elevator dog clutch, characterized in that, The adjusting method comprises: real-time acquisition of brake characteristic data of the brake, the brake characteristic data comprising: real-time compression amount of each spring in the eight spring adjusting devices, brake disc temperature, armature temperature, coil temperature, magnetic flux of the electromagnet; after the brake operates, real-time brake force and theoretical brake force are calculated according to the brake scene and the brake characteristic data; a brake force correction coefficient is acquired according to the brake characteristic data, historical brake deviation, historical decision data in the same working condition provided based on a digital twin model of the brake, and a pre-trained neural network model, and a second brake force is calculated according to the brake force correction coefficient and the theoretical brake force; a brake force deviation is calculated according to the real-time brake force and the second brake force; when the brake force deviation is greater than a brake force preset threshold, a theoretical compression amount of the spring is calculated according to the second brake force; the actual compression amount of the spring is cooperatively controlled by the telescopic device according to the theoretical compression amount of the spring and the real-time compression amount, so that the spring outputs a compensation brake force to adaptively adjust and compensate the brake force of the brake.

2. The conditioning method of claim 1, wherein, The adjusting method further comprises: analyzing and diagnosing the brake fault according to the brake characteristic data.

3. The conditioning method of claim 2, wherein, The analyzing and diagnosing of the brake fault according to the brake characteristic data comprises: preprocessing the brake characteristic data, the preprocessing comprising: filtering, amplifying, linearizing, and time-space alignment; feature extraction and deep learning feature fusion are performed on the preprocessed data to obtain feature fusion data; the brake fault is located according to the feature fusion data.

4. The conditioning method of claim 3, wherein, The locating of the brake fault according to the feature fusion data comprises: extracting fault features from the feature fusion data; inputting the fault features into the pre-trained neural network model to acquire brake fault information; sending a warning message according to the fault information.

5. The conditioning method of claim 1, wherein, The adjusting method further comprises: storing the acquired brake characteristic data, processed feature fusion data, and brake force adjustment result row data record, and updating parameters in the digital twin model based on the brake according to the measured data of the brake after the brake force is adjusted.

6. The conditioning method of claim 5, wherein, The adjusting method further comprises: automatically adjusting the brake force preset threshold according to the digital twin model and historical brake characteristic data and measured data of the brake.

7. The conditioning method of claim 1, wherein, The adjusting method further comprises: when the brake force deviation is greater than the brake force preset threshold, optimizing PID control parameters of a PID control algorithm for adjusting coil current; outputting an adjustment amount of the coil current according to the PID control parameters, and outputting a third brake force according to the adjustment amount of the coil current.

8. The conditioning method of claim 1, wherein, The brake comprises an execution mechanism, which controls the telescopic device to control the actual compression amount of the spring.

9. A brake force self-adapting regulation system of an elevator dog clutch brake, characterized in that The adjusting system comprises: a data acquisition module, configured to acquire real-time brake characteristic data of the brake, the brake characteristic data comprising: real-time compression amount of each spring in the eight spring adjusting devices, brake disc temperature, armature temperature, coil temperature, magnetic flux of the electromagnet; The brake force calculation module is configured to calculate a real-time brake force and a theoretical brake force according to a brake scene and the brake characteristic data after the brake actuation, obtain a brake force correction coefficient according to the brake characteristic data, a historical brake deviation, historical decision data in a similar working condition provided based on a digital twin model of the brake, and a pre-trained neural network model, and calculate a second brake force according to the brake force correction coefficient and the theoretical brake force. The spring compression amount calculation module is configured to calculate a brake force deviation according to the real-time brake force and the second brake force, and calculate a theoretical compression amount of the spring according to the second brake force when the brake force deviation is greater than a preset brake force threshold. The brake force compensation module is configured to cooperatively control an actual compression amount of the spring through the telescopic device according to the theoretical compression amount of the spring and the real-time compression amount, so as to adaptively adjust and compensate the brake force of the brake by the spring output compensation brake force.

10. A computer-readable storage medium storing computer-readable instructions, wherein, The computer readable instructions are executed by the processor to implement the adjustment method of any one of claims 1-8.

11. A computer device comprising a memory and a processor, The memory stores computer readable instructions, characterized in that, The processor executes the computer readable instructions to implement the adjustment method of any one of claims 1-8.

12. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the adjustment method of any one of claims 1-8.

Images