A method, storage medium and system for overload protection of a hoist

Abstract

The present invention relates to the technical field of hoist safety protection, and particularly relates to a method, a storage medium and a system for hoist overload protection. The method proposed by the present invention collects parameters such as torque and vibration frequency through multi-modal sensors, and transmits them to the central control unit after edge computing processing; a three-dimensional mathematical model is established to calculate the theoretical maximum safety torque; a two-level early warning mechanism is set up. The first-level early warning reduces the speed and starts the intelligent lubrication system, and the second-level early warning cuts off the power supply and brakes after data analysis; a prediction model is established using the LSTM neural network to generate maintenance work orders. The system includes modules such as multi-modal perception and edge computing, and an energy feedback device. Moreover, the program stored in the storage medium proposed by the present invention can implement the corresponding method. The present invention breaks through the limitation of single detection, dynamically calculates the threshold value, realizes predictive maintenance and energy recovery, and improves the operation safety and reliability of the hoist.

Description

Technical Field

[0001] The present invention relates to the technical field of gate hoist safety protection, and in particular to a gate hoist overload protection method, storage medium and system. Background Art

[0002] In water conservancy projects, the gate hoist is the core equipment for controlling the opening and closing of gates. Its operating safety directly affects the flood control, irrigation and ecological regulation functions of water conservancy projects. The traditional gate hoist overload protection method mainly relies on current threshold detection. Patent CN115425616B determines the overload state by monitoring the motor current.

[0003] However, this type of method has significant defects. It only relies on current signals and cannot perceive hidden faults such as wear and jamming of mechanical components in real time, resulting in delayed or misjudgment of overload warnings. The protection threshold is set based on a fixed water level difference and historical data and cannot adapt to dynamic working conditions. It is prone to insufficient protection or false shutdown. It can only respond passively after an overload occurs and cannot predict potential overload risks through historical data. Equipment maintenance relies on manual inspections, which is inefficient and costly.

[0004] In the existing technology, some solutions have attempted to introduce vibration detection or temperature monitoring, but have not formed an intelligent decision-making system that integrates multiple parameters, and have not solved the technical problems of energy recovery and mechanical and electrical coordinated protection. Therefore, there is an urgent need for an innovative solution that integrates multimodal perception, dynamic modeling, hierarchical protection and predictive maintenance to improve the safety and reliability of the operation of the gate hoist. Summary of the invention

[0005] The present invention proposes a method, a storage medium and a system for overload protection of a gate hoist. The purpose of the present invention is to provide a brightness measurement method which can improve measurement accuracy and stability and is easy to operate.

[0006] The technical solution of the present invention is as follows: A method for overload protection of a gate hoist comprises the following steps:

[0007] S1, synchronously collect the real-time torque T, vibration frequency f, motor temperature θ and gate inclination α of the gate hoist during operation through torque sensor, vibration sensor, temperature sensor and inclinometer, and transmit the original signal to the central control unit after noise reduction and filtering through edge computing module;

[0008] S2. Based on historical operation data and ANSYS finite element simulation, a three-dimensional mathematical model of "water level difference-opening and closing force-mechanical loss" is established to calculate the theoretical maximum safe torque T_safe=f(Δh,L,μ) of the gate hoist under the current working conditions in real time, where Δh is the water level difference on both sides of the gate, L is the gate opening and closing stroke, and μ is the friction coefficient of the mechanical parts;

[0009] S3. When the real-time torque T ≥ 0.8T_safe, a first-level warning is triggered. The central control unit controls the hoist speed to drop to 60% of the rated speed and starts the intelligent lubrication system to lubricate the transmission components.

[0010] When T ≥ T_safe and the vibration frequency f exceeds the normal operating condition threshold f > f0 + 3σ, a second-level warning is triggered, where f0 is the average vibration frequency statistically obtained based on the historical normal operation data of the hoist, and σ is the standard deviation of the vibration frequency under normal operation. The temperature θ and the inclination angle α data are analyzed synchronously. If θ exceeds 110% of the rated temperature or the fluctuation range of α > 5°, it is determined as "abnormal overload", and the motor power supply is immediately cut off and the mechanical braking device is started.

[0011] S4. Use the LSTM neural network to learn the historical overload data, equipment operation duration, water temperature, and sediment content, establish an overload fault prediction model. When the predicted overload probability within the next 2 hours > 70%, a maintenance work order is automatically generated and pushed to the operation and maintenance terminal.

[0012] It should be noted that the pressure sensor monitors the oil pressure P in the lubrication pipeline in real time. When P < 80% of the rated pressure, it is judged as "lubrication failure", and the standby oil pump is triggered to start.

[0013] Based on the gate opening and closing frequency, the oil injection cycle is dynamically adjusted. The oil injection cycle formula is T_oil = k×(1 + Δh / H_max), where k is the basic cycle and H_max is the historical maximum water level difference, so as to realize lubrication on demand and reduce the overload risk caused by frictional loss.

[0014] A hoist overload protection system includes:

[0015] A multi-modal perception module, integrating a torque sensor with an accuracy of ±0.5%FS, a vibration acceleration sensor with a range of ±50g, an infrared temperature sensor with a resolution of 0.1°C, and a MEMS inclination sensor with an accuracy of ±0.1°, is used to collect the hoist operation parameters in real time.

[0016] An edge computing unit uses an FPGA chip to perform real-time FFT spectrum analysis on the sensor signals, extracts the main frequency, harmonic components, and kurtosis value of the vibration signal, and identifies the early fault characteristics of the gearbox and bearing components.

[0017] A central control unit is equipped with a deep learning inference engine, built-in the above-mentioned dynamic load model and LSTM prediction model, and supports real-time simulation of the hoist operation state through digital twin technology to predict the fatigue degree of mechanical components.

[0018] The actuator includes a variable frequency drive, an electromagnetic clutch, and a hydraulic braking device. When receiving an overload warning signal, it sequentially performs three-level protection actions of "speed reduction - clutch separation - mechanical braking", and the response time ≤ 200 ms.

[0019] A non-transitory computer-readable storage medium stores a computer program thereon. When the program is executed by a processor, it implements the intelligent overload protection method for the hoist described above, including receiving multi-modal sensor data and performing spatio-temporal alignment processing, invoking pre-trained mechanical load prediction models and fault classification models, generating hierarchical control instructions, and outputting them to the actuator.

[0020] A hoist overload protection system further includes a permanent magnet generator coaxially connected to the hoist motor. When detecting an overload torque (T > T_safe), it switches the motor to the power generation mode through a control circuit, converts the inertial kinetic energy into electrical energy, and stores it in a super capacitor;

[0021] The super capacitor is electrically connected to a hydraulic buffer device. When the stored energy reaches a threshold value, it drives a hydraulic piston to push out a mechanical limit block to forcibly fix the position of the gate. At the same time, it balances the grid load through an energy management system to avoid instantaneous power-off impact.

[0022] The beneficial effects of the present invention are as follows:

[0023] The present invention breaks through the limitation of single current detection. By fusing and analyzing parameters such as torque, vibration, temperature, and inclination angle, it can identify potential faults of mechanical components in advance and solve the problem of protection lag.

[0024] Based on real-time working condition parameters, water level difference, stroke, and friction coefficient, it dynamically calculates the safety torque threshold to avoid protection failure or misoperation caused by a fixed threshold.

[0025] By recovering the overload kinetic energy with a permanent magnet generator to drive the hydraulic braking device, it reduces mechanical shock and improves the system reliability. At the same time, it balances the grid load, combining economy and safety.

[0026] Using an LSTM neural network to predict the overload risk, it realizes the transformation from passive protection to active prevention and reduces the unplanned shutdown rate. Description of the Drawings

[0027] The following further elaborates on the present invention in detail with reference to the drawings and specific embodiments.

[0028] Figure 1 It is a flowchart of an embodiment of the present invention. Detailed Embodiments

[0029] Next, in combination with the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0030] Embodiment 1

[0031] As Figure 1 shown, this embodiment proposes a method for overload protection of a hoist, including the following steps:

[0032] S1. Synchronously collect the real-time torque T, vibration frequency f, motor temperature θ, and gate inclination α during the operation of the hoist through a torque sensor, a vibration sensor, a temperature sensor, and an inclinometer. After the original signal is denoised and filtered by the edge computing module, it is transmitted to the central control unit;

[0033] S2. Based on historical operation data and ANSYS finite element simulation, establish a three-dimensional mathematical model of "water level difference - hoisting force - mechanical loss", and calculate the theoretical maximum safe torque T_safe = f(Δh, L, μ) of the hoist under the current working condition in real time, where Δh is the water level difference on both sides of the gate, L is the hoisting stroke of the gate, and μ is the friction coefficient of mechanical components;

[0034] S3. When the real-time torque T ≥ 0.8T_safe, trigger a first-level warning. The central control unit controls the rotation speed of the hoist to drop to 60% of the rated speed, and starts the intelligent lubrication system to add lubricant to the transmission components;

[0035] When T ≥ T_safe and the vibration frequency f exceeds the normal working condition threshold f > f0 + 3σ, trigger a second-level warning, where f0 is the average value of the vibration frequency statistically obtained based on the historical normal operation data of the hoist, and σ is the standard deviation of the vibration frequency under normal operation. Synchronously analyze the temperature θ and inclination α data. If θ exceeds 110% of the rated temperature or the fluctuation range of α > 5°, it is determined as "abnormal overload", and the motor power supply is immediately cut off and the mechanical braking device is started;

[0036] S4. Use the LSTM neural network to learn historical overload data, equipment operation duration, water temperature, and sediment content, establish an overload fault prediction model. When the predicted overload probability within the next 2 hours > 70%, automatically generate a maintenance work order and push it to the operation and maintenance terminal.

[0037] It should be noted that the intelligent lubrication system includes a pressure sensor to monitor the oil pressure P in the lubrication pipeline in real time. When P < 80% of the rated pressure, it is judged as "lubrication failure", and the standby oil pump is triggered to start;

[0038] Dynamically adjust the oil injection period based on the opening and closing frequency of the gate. The oil injection period formula is T_oil = k×(1 + Δh / H_max), where k is the basic period and H_max is the historical maximum water level difference, so as to achieve lubrication on demand and reduce the overload risk caused by frictional losses.

[0039] An overload protection system for a hoist, comprising:

[0040] A multi-modal perception module, integrating a torque sensor (accuracy ±0.5%FS), a vibration acceleration sensor (range ±50g), an infrared temperature sensor (resolution 0.1°C), and a MEMS inclinometer (accuracy ±0.1°), is used to collect the operation parameters of the hoist in real time;

[0041] An edge computing unit uses an FPGA chip to perform real-time FFT spectrum analysis on the sensor signals, extracts the main frequency, harmonic components, and kurtosis value of the vibration signal, and identifies the early fault characteristics of components such as the gearbox and bearings;

[0042] A central control unit is equipped with a deep learning inference engine, an built-in dynamic load model and an LSTM prediction model, which supports real-time simulation of the hoist operation state through digital twin technology and predicts the fatigue degree of mechanical components;

[0043] An actuator includes a variable frequency speed regulator, an electromagnetic clutch, and a hydraulic braking device. When receiving an overload warning signal, it sequentially performs three-level protection actions of "speed reduction - clutch separation - mechanical braking", and the response time ≤ 200ms.

[0044] A non-transitory computer-readable storage medium stores a computer program, which when executed by a processor implements the intelligent overload protection method for the hoist described above, including receiving multi-modal sensor data and performing spatio-temporal alignment processing, calling a pre-trained mechanical load prediction model and a fault classification model, generating hierarchical control instructions and outputting them to the actuator.

[0045] An overload protection system for a hoist further includes a permanent magnet generator coaxially connected to the hoist motor. When detecting an overload torque (T > T_safe), the motor is switched to the power generation mode through a control circuit, and the inertial kinetic energy is converted into electrical energy and stored in a super capacitor;

[0046] The super capacitor is electrically connected to a hydraulic buffer device. When the stored energy reaches the threshold, it drives the hydraulic piston to push out the mechanical limit block to forcibly fix the gate position, and at the same time balances the grid load through an energy management system to avoid instantaneous power-off impact.

[0047] In this embodiment, S1 is real-time multi-dimensional parameter perception. The real-time torque T, vibration frequency f, motor temperature θ, and gate inclination α during the operation of the hoist are synchronously collected through a torque sensor, a vibration sensor, a temperature sensor, and an inclinometer. After the original signal is denoised and filtered by an edge computing module (such as wavelet transform denoising), it is transmitted to the central control unit through an industrial Ethernet to ensure the accuracy and real-time nature of the data.

[0048] S2 is the construction of a dynamic load model. Based on historical operation data (including torque data under different water level differences, opening and closing strokes, and mechanical states) and ANSYS finite element simulation, a three-dimensional mathematical model of "water level difference - opening and closing force - mechanical loss" is established to calculate the theoretical maximum safety torque in real time: T_safe = f(Δh, L, μ);

[0049] Among them, Δh is the water level difference on both sides of the gate, which is collected in real time by an ultrasonic water level gauge, L is the opening and closing stroke of the gate, which is fed back in real time by an encoder; μ is the friction coefficient of mechanical components, which is dynamically corrected according to the lubrication state and the degree of component wear.

[0050] S3 is the overloading risk classification and assessment. First-level warning (warning adjustment): When the real-time torque T ≥ 0.8T_safe, it is determined as a mild overload. The central control unit controls the frequency converter to reduce the hoist speed to 60% of the rated speed, and at the same time starts the intelligent lubrication system to inject lubricant into transmission components such as the gearbox and bearings to reduce friction loss;

[0051] Second-level warning (emergency protection): When T ≥ T_safe and the vibration frequency f exceeds the normal working condition threshold (f > f0 + 3σ, f0 is the normal main frequency, σ is the standard deviation of historical vibration data), the temperature θ and the inclination α are analyzed synchronously: If θ exceeds 110% of the rated temperature or the fluctuation range of α > 5°, it is determined as "abnormal overload", and the motor power supply is immediately cut off, the power transmission is separated through an electromagnetic clutch, and the hydraulic braking device is started to force a stop, with a response time ≤ 200 ms.

[0052] S4 is the predictive maintenance decision-making. The LSTM neural network is used to learn environmental parameters such as historical overload data, equipment operation duration, water temperature, and sediment content to establish an overload fault prediction model. When the model predicts that the overload probability within the next 2 hours > 70%, a maintenance work order including the fault type and component location is automatically generated and pushed to the operation and maintenance terminal through the 4G / 5G network to guide preventive maintenance;

[0053] Furthermore, the intelligent lubrication system includes a pressure sensor to monitor the oil pressure P in the lubrication pipeline in real time. When P < 80% of the rated pressure, the standby oil pump is triggered to start to ensure lubrication reliability; the oil injection cycle is dynamically adjusted according to the opening and closing frequency of the gate, and the formula is: T_oil = k×(1 + Δh / H_max)

[0054] Among them, k is the basic period (default 8 hours), and H_max is the historical maximum water level difference, to achieve lubrication on demand to reduce the risk of friction overload.

[0055] A hoist overload protection system, which includes a multi-modal perception module integrating high-precision sensors: a torque sensor (accuracy ±0.5%FS), a vibration acceleration sensor (range ±50g), an infrared temperature sensor (resolution 0.1°C), and a MEMS inclinometer (accuracy ±0.1°), to collect the operating parameters of the hoist in real time, covering multi-dimensional states such as mechanical stress, vibration characteristics, temperature, and gate attitude.

[0056] Edge computing unit

[0057] Use an FPGA chip to implement real-time signal processing, extract the main frequency, harmonic components, and kurtosis value of the vibration signal through FFT spectrum analysis, identify early fault characteristics such as gearbox cracks and bearing roller wear, and reduce the computing pressure on the central control unit.

[0058] Central control unit

[0059] Equipped with a deep learning inference engine, with a built-in dynamic load model and an LSTM prediction model, supporting digital twin technology: real-time simulate the mechanical stress distribution of the hoist through a virtual simulation model, predict the fatigue degree of components such as gears and wire ropes, and provide data support for hierarchical protection decision-making.

[0060] Actuator

[0061] Includes a variable frequency speed regulator, an electromagnetic clutch, and a hydraulic braking device, and sequentially executes three-level protection actions of "speed reduction - clutch separation - mechanical braking" after receiving instructions from the central control unit, forming a collaborative protection mechanism of "flexible adjustment + rigid braking".

[0062] Furthermore, the system also includes a permanent magnet generator coaxially connected to the hoist motor: when an overload torque (T > T_safe) is detected, the control circuit switches the motor to the power generation mode, converts the inertial kinetic energy into electrical energy and stores it in the super capacitor; the super capacitor drives the hydraulic buffer device to push out the mechanical limit block to forcibly fix the gate position, and at the same time balances the grid load through the energy management system to avoid instantaneous power-off shock.

[0063] A non-transitory computer-readable storage medium stores a computer program, and when the program is executed by a processor, it realizes the following functions:

[0064] Receive multi-modal sensor data and perform spatio-temporal alignment processing to ensure the time synchronization of different sensor data;

[0065] Call the pre-trained mechanical load prediction model and fault classification model to calculate T_safe in real time and evaluate the overload risk level;

[0066] Generate hierarchical control instructions (such as speed reduction, braking, work order push) and output them to the actuator to achieve intelligent closed-loop control.

[0067] In this embodiment, the applicable scenarios are:

[0068] Taking a 2×500kN hoist of a certain reservoir as an example, with a motor power of 110kW, a maximum designed water level difference of the gate of 15m, the system described in the present invention is configured for overload protection.

[0069] Method implementation steps:

[0070] S1 is data acquisition and preprocessing. The torque sensor collects torque signals in real time, the vibration sensor monitors the vibration frequency of the gearbox, the infrared temperature sensor scans the motor temperature, and the MEMS inclinometer detects the inclination angle of the gate. The edge computing unit performs a 50Hz low-pass filter on the original signal, removes the power frequency interference, and then transmits it to the central control unit.

[0071] S2 is the calculation of the dynamic load model. Based on the training of 2000 groups of historical operation data, the friction coefficient correction formula is obtained: μ = 0.08 + 0.005Δh + 0.001L. When the real-time water level difference Δh = 10m and the hoisting stroke L = 8m, it is calculated that T_safe = 4500N⋅m.

[0072] S3 is hierarchical warning and protection. When it is detected that T = 3600N⋅m (0.8T_safe) and the vibration frequency is normal, the central control unit controls the speed to be reduced to 60%, that is, 300rpm, and starts the lubrication system to inject oil, and the oil injection cycle is automatically adjusted to 12 hours;

[0073] When T = 4800N⋅m (>T_safe) and the vibration kurtosis value suddenly increases to 3.5, exceeding the normal threshold of 2.0, and at the same time the motor temperature rises to 85℃, which is 113% of the rated temperature of 75℃, the system determines abnormal overload, immediately cuts off the power supply, the electromagnetic clutch disengages, and the hydraulic braking device completes braking within 150ms.

[0074] S4 is predictive maintenance. The LSTM model, based on the data of the past week, water temperature of 25℃, sediment content of 50kg / m³, and equipment operation duration of 400 hours, predicts that the overload probability in the next 2 hours is 75%, automatically generates a work order to prompt "Insufficient lubrication of the gearbox, it is recommended to repair", and the operation and maintenance personnel execute preventive maintenance after confirmation through the remote terminal.

[0075] System Hardware Collaboration: The multi-modal perception module collects data in real time. The edge computing unit completes the spectral analysis of vibration signals. For example, if an abnormal gear meshing frequency is identified, the central control unit combines the digital twin model to judge the gear wear degree, triggers a secondary warning and starts the energy feedback device. The permanent magnet generator converts the motor kinetic energy into electrical energy, with an energy storage efficiency ≥ 90%. The super capacitor drives the hydraulic piston to push out the limit block to ensure that the gate stops moving within 10 cm. At the same time, the redundant electrical energy is fed back to the power grid through the inverter to avoid voltage fluctuations.

[0076] Implementation of Storage Medium: The computer program is written in C++ language, integrated with the TensorFlow deep learning framework, and stored in an industrial-grade SSD hard disk. After the program starts, it reads the sensor data in real time, calls the pre-trained LSTM model, with the training data volume ≥ 100,000 groups, to predict the overload probability. The control instructions are sent to the actuator through the ModbusTCP protocol to achieve millisecond-level response.

Claims

1. A method for overload protection of a hoist, characterized in that, It includes the following steps: S1. Synchronously collect the real-time torque T, vibration frequency f, motor temperature θ, and gate inclination α during the operation of the hoist through a torque sensor, vibration sensor, temperature sensor, and inclinometer. After denoising and filtering the original signals through an edge computing module, transmit them to the central control unit; S2. Based on historical operation data and ANSYS finite element simulation, establish a three-dimensional mathematical model of "water level difference - hoisting force - mechanical loss", and calculate the theoretical maximum safe torque T_safe = f(Δh, L, μ) of the hoist under the current working conditions in real time, where Δh is the water level difference on both sides of the gate, L is the hoisting stroke of the gate, and μ is the friction coefficient of mechanical components; S3. When the real-time torque T ≥ 0.8T_safe, trigger a first-level warning. The central control unit controls the hoist speed to drop to 60% of the rated speed and starts the intelligent lubrication system to add lubricant to the transmission components; When T ≥ T_safe and the vibration frequency f exceeds the normal working condition threshold f > f0 + 3σ, trigger a second-level warning, where f0 is the average vibration frequency statistically obtained based on the historical normal operation data of the hoist, and σ is the standard deviation of the vibration frequency under normal operation. Synchronously analyze the temperature θ and inclination α data. If θ exceeds 110% of the rated temperature or the fluctuation range of α > 5°, it is determined as "abnormal overload", and immediately cut off the motor power supply and start the mechanical braking device; S4. Use the LSTM neural network to learn historical overload data, equipment operation duration, water temperature, and sediment content, establish an overload fault prediction model, and when the predicted overload probability within the next 2 hours > 70%, automatically generate a maintenance work order and push it to the operation and maintenance terminal.

2. A method for overload protection of a hoist according to claim 1, characterized in that, The intelligent lubrication system includes: A pressure sensor monitors the oil pressure P in the lubrication pipeline in real time. When P < 80% of the rated pressure, it is judged as "lubrication failure" and triggers the start of the standby oil pump; Dynamically adjust the oil injection cycle based on the gate hoisting frequency. The oil injection cycle formula is T_oil = k×(1 + Δh / H_max), where k is the basic cycle and H_max is the historical maximum water level difference, to achieve lubrication on demand and reduce the overload risk caused by friction loss.

3. A system for overload protection of a hoist, characterized in that, It includes: A multi-modal perception module integrates a torque sensor with an accuracy of ±0.5%FS, a vibration acceleration sensor with a range of ±50g, an infrared temperature sensor with a resolution of 0.1°C, and a MEMS inclinometer with an accuracy of ±0.1° to collect the operation parameters of the hoist in real time; An edge computing unit uses an FPGA chip to perform real-time FFT spectrum analysis on the sensor signals, extracts the main frequency, harmonic components, and kurtosis values of the vibration signals, and identifies the early fault characteristics of gearboxes and bearing components; A central control unit is equipped with a deep learning inference engine, and the three-dimensional mathematical model and overload fault prediction model described in claim 1 are built in. It supports real-time simulation of the hoist operation state through digital twin technology and predicts the fatigue degree of mechanical components; An actuator includes a variable frequency speed regulator, an electromagnetic clutch, and a hydraulic braking device. When receiving an overload warning signal, it sequentially performs three-level protection actions of "speed reduction - clutch separation - mechanical braking", and the response time ≤ 200ms.

4. A hoist overload protection system according to claim 3, characterized in that, It also includes a permanent magnet generator coaxially connected to the opening and closing motor. When it is detected that the overload torque T > T_safe, the motor is switched to the power generation mode through the control circuit, and the inertial kinetic energy is converted into electrical energy and stored in the super capacitor. The super capacitor is electrically connected to the hydraulic buffer device. When the stored energy reaches the threshold, it drives the hydraulic piston to push out the mechanical limit block to forcibly fix the position of the gate. At the same time, it balances the grid load through the energy management system to avoid instantaneous power-off impact.

5. A non-transitory computer-readable storage medium storing a computer program thereon, characterized in that, When the program is executed by the processor, it implements the method for overload protection of a hoist described in any one of claims 1-2, including: Receiving multi-modal sensor data and performing spatio-temporal alignment processing; Invoking the pre-trained mechanical load prediction model and fault classification model; Generating hierarchical control instructions and outputting them to the actuator.

Images