Electric drive-spring series brake system and method for mine-used monorail crane

Abstract

The application discloses a mine-used monorail crane electric drive-spring series brake system and method, and relates to the technical field of industrial control. The system comprises the following steps: collecting the real-time vehicle speed, track slope and vehicle load of the mine-used monorail crane as real-time running state parameters, inputting the parameters into a target brake force calculation unit built in a controller to obtain the target brake force required at present; generating a pulse width modulation driving signal to drive a servo motor to convert the rotary motion into linear motion via a ball screw transmission mechanism, push the piston rod to extend and compress the butterfly spring connected in series, make the butterfly spring generate elastic force and act on the brake shoe, thereby establishing the brake clamping force; during the braking process, the position and pressure collaborative closed-loop control strategy is executed based on the clamping force data, piston rod position data and motor rotation angle data collected by the controller in real time. The application effectively improves the brake response speed and safety of the mine-used monorail crane under variable load conditions.

Description

Technical Field

[0001] This invention relates to the field of industrial control technology, specifically to an electric drive-spring tandem braking system and method for mining monorail cranes. Background Technology

[0002] With the continuous improvement of automation and intelligence in coal mining, the safety and reliability of the braking system of mine monorails, as key equipment for underground auxiliary transportation, are receiving increasing attention. Currently, mine monorails mostly employ purely mechanical braking, hydraulic braking, or electromagnetic braking.

[0003] However, traditional braking methods are mostly open-loop control, which cannot actively adjust the braking force according to the dynamic changes of real-time vehicle speed, gradient and load. This results in insufficient braking force under dangerous conditions such as heavy-load downhill or excessive braking under light load, affecting braking stability and safety. Hydraulic or electromagnetic braking systems have a large response delay, making it difficult to achieve continuous and precise adjustment of braking force. This can easily cause braking shock, uneven wear of brake shoes or even brake failure, which restricts the safe operation of mining monorails in complex and harsh underground environments. Summary of the Invention

[0004] This invention provides an electric drive-spring series braking system and method for mining monorail cranes, aiming to solve the technical problems of existing braking systems that cannot dynamically match braking force according to real-time working conditions and lack position and pressure coordinated closed-loop control, resulting in low braking accuracy, slow response and large impact.

[0005] In view of the above problems, the present invention provides an electric drive-spring series braking system and method for mining monorail cranes.

[0006] In a first aspect, the present invention provides an electric drive-spring tandem braking system for mine monorail hoists, comprising: The target braking force calculation module is used to collect the real-time speed, track gradient and vehicle load of the mining monorail as real-time operating status parameters. The real-time operating status parameters are input into the target braking force calculation unit built into the controller to calculate the target braking force required at the moment. An electromechanical spring brake actuator module is used to generate a pulse width modulation drive signal based on the target braking force, drive a servo motor to convert the rotary motion into linear motion via a ball screw transmission mechanism, push the piston rod to extend and compress the series-connected disc spring, so that the disc spring generates elastic force and acts on the brake shoe, thereby establishing a braking clamping force. The position-pressure coordinated closed-loop control module is used to execute a coordinated closed-loop control strategy for position and pressure during braking, based on the clamping force data, piston rod position data, and motor rotation angle data collected in real time by the controller. The coordinated closed-loop control strategy takes the target position of the piston rod and the target pressure of the spring as control targets. It adjusts the drive direction and output torque of the servo motor in real time by adjusting the duty cycle of the pulse width modulation signal, so that the actual position of the piston rod and the actual pressure of the spring synchronously approach their respective target values, until the elastic force generated by the disc spring stabilizes within the preset deviation tolerance range of the target braking force.

[0007] Secondly, the present invention provides an electric drive-spring series braking method for mine monorail hoists, comprising: The real-time speed, track gradient, and vehicle load of the mining monorail are collected as real-time operating status parameters. These real-time operating status parameters are then input into the target braking force calculation unit built into the controller to calculate the target braking force required at the moment. Based on the target braking force, a pulse width modulation drive signal is generated, which drives the servo motor to convert the rotational motion into linear motion through the ball screw transmission mechanism. This pushes the piston rod to extend and compress the series-connected disc spring, causing the disc spring to generate elastic force and act on the brake shoe, thereby establishing a braking clamping force. During braking, a coordinated closed-loop control strategy for position and pressure is executed based on the clamping force data, piston rod position data, and motor rotation angle data collected in real time by the controller. The coordinated closed-loop control strategy takes the target position of the piston rod and the target pressure of the spring as the control targets. By adjusting the duty cycle of the pulse width modulation signal in real time, the driving direction and output torque of the servo motor are coordinated to make the actual position of the piston rod and the actual pressure of the spring synchronously approach their respective target values ​​until the elastic force generated by the disc spring stabilizes within the preset deviation tolerance range of the target braking force.

[0008] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention provides an electric-drive-spring tandem braking system and method for mining monorail cranes. By collecting real-time vehicle speed, track gradient, and vehicle load as inputs, the controller's built-in target braking force calculation unit automatically obtains the required braking force. A servo motor, via a ball screw drive, compresses a series of disc springs to generate elastic braking clamping force, achieving adaptive matching of braking force to operating conditions. Furthermore, this invention employs a coordinated closed-loop control strategy of position and pressure, adjusting the pulse width modulation signal in real time to coordinate the control of the piston rod position and spring pressure, ensuring that the actual braking force remains stable within the allowable deviation of the target value. This invention effectively solves the problems of existing braking systems' inability to dynamically adjust braking force according to real-time operating conditions, inaccurate braking force control, and lack of force-position coordinated closed-loop control, improving the braking response speed, control accuracy, and operational safety of mining monorail cranes under varying gradient and load conditions. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the electric drive-spring series braking system for mining monorail cranes provided in an embodiment of the present invention. Figure 2 A schematic diagram of the mechanical execution structure of the electric drive-spring series braking system for a mining monorail crane provided in an embodiment of the present invention; Figure 3 A schematic diagram of the spring brake assembly structure of the electric drive-spring series braking system for a mining monorail crane provided in an embodiment of the present invention; Figure 4 This is a schematic flowchart of an electric drive-spring series braking method for a mining monorail provided in an embodiment of the present invention. The components represented by each number in the attached diagram are explained below: Target braking force calculation module 11, electromechanical spring brake execution module 12, position-pressure coordinated closed-loop control module 13.

[0010] Servo motor 1, ball screw transmission mechanism 2, disc spring 3, piston 4, piston rod 5, spring seat 6, pressure sensor 7, displacement sensor 8, encoder 9. Detailed Implementation

[0011] This invention provides an electric drive-spring series braking system and method for mining monorail cranes, which addresses the technical problems of existing braking systems that cannot dynamically match braking force according to real-time working conditions and lack position and pressure coordinated closed-loop control, resulting in low braking accuracy, slow response and large impact.

[0012] Example 1, as Figure 1 As shown, the present invention provides an electric drive-spring series braking system for mining monorail cranes, the system comprising: The target braking force calculation module 11 is used to collect the real-time speed, track gradient and vehicle load of the mining monorail as real-time operating status parameters, and input the real-time operating status parameters into the target braking force calculation unit built into the controller to calculate the target braking force required at the moment.

[0013] In this embodiment of the invention, the real-time speed, track gradient, and vehicle load of the mining monorail are collected as real-time operating status parameters. These parameters are then input into the target braking force calculation unit built into the controller to calculate the required target braking force. The braking safety of the mining monorail depends on the accurate calculation of the target braking force. However, during its operation, the real-time speed, track gradient, vehicle load, and target braking force exhibit a complex nonlinear relationship. Traditional calculation methods based on empirical formulas cannot cover various complex operating conditions such as no-load, full-load, uphill, and downhill conditions, and are prone to excessive deviations in braking force calculation, leading to problems such as excessively abrupt or insufficient braking. Therefore, an offline-trained feedforward neural network model is required to achieve accurate output of the target braking force under multiple operating conditions. Simultaneously, incremental model updates compensate for calculation errors caused by changes in operating conditions and component wear during long-term operation, ensuring the stability of braking accuracy.

[0014] In one embodiment, the target braking force calculation module 11 is further configured to: The real-time operating status parameters are input into the target braking force calculation unit built into the controller to calculate the currently required target braking force, including: The collected real-time vehicle speed, track gradient, and vehicle load are used as input parameters and input to the input layer node of the offline-trained feedforward neural network model. The output layer node of the feedforward neural network model outputs the target braking force. The feedforward neural network model is trained offline by collecting historical operating data and historical braking data of the mining monorail crane under various working conditions, including unloaded, fully loaded, uphill, and downhill conditions. The trained feedforward neural network model is then deployed to the target braking force calculation unit of the controller. During operation, the controller periodically compares the actual braking effect with the target braking force output by the feedforward neural network model. When the deviation between the actual braking distance and the theoretical braking distance exceeds a preset deviation threshold, the controller triggers an incremental update of the feedforward neural network model and uses a backpropagation algorithm to fine-tune the network weights.

[0015] First, the collected real-time vehicle speed, track gradient, and vehicle load are used as input parameters and input to the input layer node of the offline-trained feedforward neural network model. The output layer node of the feedforward neural network model outputs the target braking force. The feedforward neural network model is trained offline by collecting historical operating data and historical braking data of the mining monorail crane under various working conditions, including unloaded, fully loaded, uphill, and downhill conditions. The trained feedforward neural network model is then deployed to the target braking force calculation unit of the controller.

[0016] Among them, the feedforward neural network model is an artificial neural network in which signals are transmitted unidirectionally from the input layer to the output layer without feedback loops. It can fit the nonlinear relationship between input and output through training and is suitable for multi-parameter collaborative prediction scenarios. Offline training refers to training the neural network model by collecting historical data before the actual underground operation of the mining monorail crane, so that it has a basic braking force prediction capability. After training, the initial network weights are fixed, eliminating the need for real-time training and reducing the load on the controller. Input layer nodes refer to the input end of the neural network model, used to receive external input parameters. Here, there are 3 input nodes, including real-time vehicle speed, track gradient, and vehicle load. Output layer nodes refer to the output end of the neural network model, used to output the prediction result. Here, there is 1 output node, corresponding to the target braking force.

[0017] First, historical operating data and historical braking data of the mining monorail crane under various typical working conditions such as no load, full load, uphill, and downhill were collected. The dataset came from two authoritative sources: online data collection from underground actual vehicles and ground braking bench tests. The historical operating data included three types of input parameters: real-time vehicle speed, track gradient, and vehicle load under different working conditions. The historical braking data included the actual braking force required to achieve safe braking under the corresponding working conditions, which served as label data for model training.

[0018] Secondly, the collected historical data was divided into training and testing sets in a 7:3 ratio. The constructed feedforward neural network model adopted a single hidden layer structure, with 3 neurons in the input layer, 12 neurons in the hidden layer, and 1 neuron in the output layer. The ReLU activation function was used for the hidden layer, and a linear activation function was used for the output layer. The adaptive moment estimation optimization algorithm was used for model training, with the mean squared error function as the loss function, and the convergence threshold of the loss function was set to 10. -4 The model is considered to have converged when the loss function value is below the convergence threshold for 50 consecutive iterations and does not decrease significantly. The training set is used for iterative updates of model weights, and the test set is used to verify the model's prediction accuracy. If the prediction error of the test set exceeds the preset 2% accuracy threshold, the number of neurons in the hidden layer is adjusted and training is restarted.

[0019] Finally, after the feedforward neural network model training converges and the prediction accuracy meets the target, the feedforward neural network model is solidified and deployed to the target braking force calculation unit built into the controller, so that it can receive three types of operating status parameters in real time: vehicle speed, track gradient, and vehicle load, and quickly output the target braking force matching the working condition.

[0020] For example, assuming the rated load of the mining monorail is 10t, four typical working conditions are set, and historical data are collected as follows: Working condition 1 is unloaded uphill: speed 3m / s, gradient 5°, load 0t, corresponding to a safe braking force of 8kN; Working condition 2 is fully loaded uphill: speed 3m / s, gradient 5°, load 10t, corresponding to a safe braking force of 15kN; Working condition 3 is unloaded downhill: speed 3m / s, gradient -5°, load 0t, corresponding to a safe braking force of 10kN; Working condition 4 is fully loaded downhill: speed 3m / s, gradient -5°, load 10t, corresponding to a safe braking force of 22kN. A total of 1000 sets of historical data from the above four working conditions and other similar working conditions were input into the constructed feedforward neural network model for offline training. This model employs a single hidden layer structure: 3 neurons in the input layer, 12 neurons in the hidden layer, and 1 neuron in the output layer. The activation function for the hidden layer is the ReLU function, and the activation function for the output layer is a linear function. The model is trained using an adaptive moment estimation optimization algorithm, with mean squared error as the loss function. The loss function is considered effective when the loss function value is below 10 for 50 consecutive iterations. -4 When the model converges; after training, the prediction error of the model on the test set is ≤2%, which meets the braking control accuracy requirements, and the converged model is deployed to the target braking force calculation unit of the controller.

[0021] Secondly, during operation, the controller periodically compares the actual braking effect with the target braking force output by the feedforward neural network model. When the deviation between the actual braking distance and the theoretical braking distance exceeds a preset deviation threshold, the controller triggers an incremental update of the feedforward neural network model and uses a backpropagation algorithm to fine-tune the network weights.

[0022] The actual braking effect refers to the actual braking effect achieved after the mining monorail crane performs braking operations. Here, the actual braking distance is used as the evaluation index; the closer the braking distance is to the theoretical value, the better the braking effect. The theoretical braking distance refers to the theoretical safe braking distance calculated based on the target braking force output by the feedforward neural network model and the overall dynamic characteristics of the monorail crane. The preset deviation threshold is a pre-set allowable deviation in braking distance. If the deviation between the actual braking distance and the theoretical braking distance exceeds this threshold, it indicates a decrease in model prediction accuracy, requiring an update. For example, the deviation threshold can be set to 0.5m.

[0023] Incremental updates refer to fine-tuning the network weights using the backpropagation algorithm only for the parts with large prediction errors, without retraining the entire neural network model, thus balancing accuracy and efficiency. Backpropagation is a neural network training algorithm that calculates the error between the predicted output and the actual output, propagates it back to each layer of nodes, adjusts the network weights, and reduces the prediction error. Network weights are the coefficients connecting the nodes in each layer of the neural network, determining the degree of influence of the input parameters on the output result; adjusting the weights can optimize the model's prediction accuracy.

[0024] Specifically, the controller collects real-time data on the speed, track gradient, and vehicle load of the mining monorail, using these as input parameters and feeding them into the input layer nodes of the deployed feedforward neural network model. The model outputs the required target braking force through signal transmission between the input layer, hidden layer, and output layer, serving as the core control basis for braking operations. The controller periodically compares the actual braking distance with the theoretical braking distance, calculating the deviation. If the deviation is less than or equal to a preset deviation threshold, the model maintains its current network weights and continues to output the target braking force normally. If the deviation exceeds the preset deviation threshold, an incremental update of the feedforward neural network model is triggered. Using a backpropagation algorithm, the weights of each network layer are fine-tuned based on the deviation. After the update, the model continues to be used for calculating the target braking force.

[0025] For example, after the feedforward neural network model is deployed, during a certain braking cycle: real-time collected parameters: vehicle speed 3m / s, gradient -5°, load 10t, the model outputs a target braking force of 22.3kN, and the theoretical braking distance calculated based on this braking force is 6.0m; after the monorail performs the braking operation, the actual measured braking distance is 6.6m, the deviation value = 6.6-6.0=0.6m, exceeding the preset deviation threshold of 0.5m; the controller triggers an incremental update of the model, using the backpropagation algorithm, and fine-tunes the network weights of each layer of the model based on the deviation value of 0.6m, such as adjusting the weights of the nodes corresponding to the load parameters to enhance the influence of the load on the braking force; after the update is completed, the same parameters are collected again: vehicle speed 3m / s, gradient -5°, load 10t, the feedforward neural network model outputs a target braking force adjusted to 22.8kN, the corresponding theoretical braking distance is 6.5m, the braking operation is performed again, the actual braking distance is 6.55m, the deviation value is 0.05m, which is lower than the deviation threshold, and the model resumes normal operation.

[0026] In this embodiment of the invention, an offline-trained feedforward neural network model achieves accurate fitting of real-time vehicle speed, track gradient, vehicle load, and target braking force, ensuring the accuracy of target braking force calculation under different operating conditions. By periodically comparing braking effects and triggering incremental model updates, prediction deviations caused by component wear and changes in operating conditions during long-term operation are compensated for, ensuring the long-term stability of braking force calculation accuracy. Real-time training of the neural network model is unnecessary; incremental updates are only performed when the deviation exceeds a threshold, reducing the controller's operating load. Simultaneously, the fine-tuning method of the backpropagation algorithm balances the efficiency and accuracy of model updates. Accurate target braking force provides a reliable basis for establishing subsequent braking clamping force and collaborative closed-loop control, avoiding problems of excessive or insufficient braking, reducing brake shoe wear, and improving the safety and reliability of the mine monorail braking system.

[0027] The electromechanical spring brake actuator module 12 is used to generate a pulse width modulation drive signal according to the target braking force, drive the servo motor to convert the rotational motion into linear motion through the ball screw transmission mechanism, push the piston rod to extend and compress the series-connected disc spring, so that the disc spring generates elastic force and acts on the brake shoe, thereby establishing a braking clamping force.

[0028] In embodiments of the present invention, such as Figure 2 The diagram shows the mechanical execution structure of the electric drive-spring series braking system for a mining monorail crane provided in this embodiment of the invention. When braking begins, the drive servo motor 1 drives the ball screw transmission mechanism 2 to convert rotational motion into linear motion, pushing the piston 4. The piston rod 5 extends and actively compresses the disc spring 3, causing the disc spring 3 to generate elastic force that acts on the brake shoes. The function of the electromechanical spring brake execution module 12 is to convert the target braking force into actual braking clamping force. The key to this conversion process is the precise control of the servo motor 1's movement, thereby driving the disc spring 3 to compress and generate a matching elastic force.

[0029] Since the current, speed, and angle of servo motor 1 change dynamically with the load, and the duty cycle of the pulse width modulation signal directly determines the motor's driving force and response speed, the traditional fixed duty cycle driving method cannot simultaneously meet the needs of rapid pressure build-up at the start of braking and the need for smooth adjustment when approaching the target braking force, easily leading to problems such as excessively slow or excessively fast pressure build-up. Therefore, it is necessary to establish a dynamic response model of the motor and solve for the optimal duty cycle through a quadratic programming algorithm to achieve dynamic adaptive generation of the pulse width modulation drive signal, ensuring that the braking clamping force quickly and smoothly approaches the target braking force.

[0030] In one embodiment, the electromechanical spring brake actuator module 12 is further configured to: The process of generating a pulse width modulation drive signal based on the target braking force includes: Using the current, speed, and angle of the servo motor as state variables and the duty cycle of the pulse width modulation signal as the control variable, a dynamic response model of the motor in the prediction time domain is established. In each control cycle, the controller uses a quadratic programming algorithm to solve a finite-time domain optimization problem based on the motor dynamic response model, minimizing the deviation between the piston rod displacement at the end of the predicted time domain and the target compression. Specifically, when the difference between the target braking force and the current disc spring pressure is greater than a preset rapid pressure build-up threshold, it is determined that the current braking initiation stage is in place, and the preset maximum duty cycle is automatically output. When the current disc spring pressure is close to the target braking force, the controller automatically reduces the duty cycle of the pulse width modulation signal. The first optimal duty cycle obtained from the solution is used as the pulse width modulation drive signal for the current cycle.

[0031] First, using the current, speed, and rotation angle of servo motor 1 as state variables and the duty cycle of the pulse width modulation signal as the control variable, a dynamic response model of the motor in the prediction time domain is established. The state variables are parameters describing the operating state of servo motor 1, specifically the current, speed, and rotation angle. The current reflects the output torque of servo motor 1, the speed reflects its rotational speed, and the rotation angle reflects its rotational angle, which is directly related to the displacement of piston rod 5. These three factors together determine the dynamic response characteristics of the motor.

[0032] The control variables are parameters used to adjust the operating state of servo motor 1, specifically the duty cycle of the pulse width modulation (PWM) signal. A larger duty cycle results in a higher driving voltage for servo motor 1, leading to greater output torque and speed; conversely, a smaller duty cycle results in a lower driving voltage, lower output torque, and lower speed. The prediction time domain refers to a pre-set time interval, such as 0.1s, used to predict the dynamic response of servo motor 1. Within this interval, the changing trend of the state variables of servo motor 1 is predicted, providing a basis for duty cycle optimization. The dynamic response model of servo motor 1 is constructed based on its dynamic equations, describing the dynamic relationship between state variables and control variables. It can predict the operating state of the motor and the displacement changes of piston rod 5 under different duty cycles.

[0033] Specifically, the parameters of servo motor 1, such as rated power, rated speed, and rated current, are defined. Combined with the transmission ratio and efficiency of ball screw transmission mechanism 2, a dynamic equation for servo motor 1 is established, relating the relationship between motor current, speed, rotation angle, and piston rod displacement 5. Using the current, speed, and rotation angle of servo motor 1 as state variables and the duty cycle of pulse width modulation signal as control variables, these are substituted into the dynamic equation to construct a motor dynamic response model in the prediction time domain. This motor dynamic response model can output the piston rod displacement value at the end of the prediction time domain under different duty cycles. The motor dynamic response model is then solidified into electromechanical spring brake execution module 12 as the basis for subsequent duty cycle optimization solutions.

[0034] For example, following the previous setting of a 10t rated load for a mining monorail crane, a servo motor 1 with a rated power of 1.5kW, a rated speed of 3000r / min, and a rated current of 5A is selected. The transmission ratio of the ball screw transmission mechanism 2 is 1:5, representing that for every 1 revolution of the servo motor 1, the piston rod 5 moves 5mm. The prediction time domain is set to 0.1s, and a motor dynamic response model is established: when the pulse width modulation signal duty cycle is 80%, the motor dynamic response model predicts that at the end of 0.1s, the motor rotation angle is 30° and the piston rod 5 displacement is 2.5mm; when the duty cycle is 30%, the motor dynamic response model predicts that at the end of 0.1s, the motor rotation angle is 10° and the piston rod 5 displacement is 0.83mm. This clarifies the correspondence between the duty cycle and the piston rod 5 displacement, providing support for subsequent optimization solutions.

[0035] Secondly, within each control cycle, the controller uses a quadratic programming algorithm to solve a finite-time domain optimization problem based on the motor dynamic response model, minimizing the deviation between the displacement of the piston rod 5 at the end of the predicted time domain and the target compression amount. Specifically, when the difference between the target braking force and the current pressure of the disc spring 3 is greater than a preset rapid pressure build-up threshold, it is determined that the current braking initiation stage is underway, and the preset maximum duty cycle is automatically output. When the current pressure of the disc spring 3 is close to the target braking force, the controller automatically reduces the duty cycle of the pulse width modulation signal.

[0036] Among them, the quadratic programming algorithm is an optimization algorithm for solving the extrema of a quadratic function under constraints. It is suitable for the optimization requirement of minimizing the displacement deviation of piston rod 5 in this scenario. It can quickly solve for the optimal duty cycle, balancing optimization accuracy and computational efficiency. The finite-time domain optimization problem refers to the problem of finding the optimal control variables within a preset prediction time domain, with the optimization objective being to minimize the deviation between the displacement of piston rod 5 at the end of the prediction time domain and the target compression amount. The constraints are that the current and speed of servo motor 1 do not exceed their rated values. The target compression amount refers to the required compression amount of piston rod 5 obtained by back-calculating the nonlinear mapping model of disc spring 3 based on the target braking force. The rapid pressure build-up threshold is a pre-set braking force difference threshold, such as 5kN, used to determine the braking stage. When the difference between the target braking force and the current pressure of disc spring 3 is greater than this threshold, it is determined to be the braking start stage, requiring rapid pressure build-up. The maximum duty cycle is a pre-set maximum drive duty cycle of servo motor 1, such as 80%, used in the braking start stage to ensure that piston rod 5 is quickly pushed out and disc spring 3 is compressed to achieve rapid pressure build-up.

[0037] Specifically, within each control cycle, the controller acquires the target compression of piston rod 5 corresponding to the current target braking force, as well as the actual pressure of the current disc spring 3; calculates the difference between the target braking force and the current pressure of disc spring 3, and determines the braking stage: if the difference is greater than the preset rapid pressure build-up threshold, it is determined to be the braking initiation stage, and the preset maximum duty cycle is directly output; if the difference is less than or equal to the rapid pressure build-up threshold, it is determined to be the smooth adjustment stage, and the quadratic programming algorithm is entered into the solution process; based on the motor dynamic response model, with the goal of minimizing the deviation between the actual displacement of piston rod 5 and the target compression at the end of the predicted time domain, and with the rated current and rated speed of servo motor 1 as constraints, the quadratic programming algorithm is used to solve the finite time domain optimization problem, and a series of duty cycle candidate values ​​are obtained in the predicted time domain; when the current pressure of disc spring 3 is close to the target braking force, the duty cycle candidate value is automatically reduced to avoid braking shock caused by excessively rapid pressure build-up.

[0038] For example, the target braking force is 22kN, corresponding to a target compression of 8mm for piston rod 5. The preset rapid pressure build-up threshold is 5kN, the maximum duty cycle is 80%, the control cycle is 0.01s, and the prediction time domain is 0.1s: Braking initiation stage: The current actual pressure of the disc spring 3 is 15kN, and the difference between this and the target braking force of 22kN is 7kN > 5kN, thus determining it as the braking initiation stage. The controller directly outputs the maximum duty cycle of 80%, driving the servo motor 1 to rotate rapidly, pushing the piston rod 5 to extend rapidly; Smooth adjustment stage: After 3 control cycles, i.e., 0.03s... The actual pressure of the disc spring 3 rises to 19kN, and the difference between it and the target braking force is 3kN≤5kN. The quadratic programming algorithm is then used to solve the problem. Based on the motor dynamic response model, candidate duty cycle values ​​within the predicted time domain of 0.1s are obtained: 45%, 42%, 40%, etc. Since the current pressure is close to the target value, the optimal candidate value is automatically adjusted to 38%. After running for two more control cycles, the actual pressure of the disc spring 3 rises to 21.5kN, and the difference between it and the target braking force is 0.5kN. The quadratic programming algorithm then determines the optimal duty cycle to be 30%, ensuring that the pressure steadily approaches the target value.

[0039] Furthermore, the first optimal duty cycle obtained from the solution is used as the pulse width modulation drive signal for the current cycle. The optimal duty cycle is the duty cycle value that minimizes the displacement deviation of piston rod 5, obtained through a quadratic programming algorithm, and is the most suitable motor drive parameter within the current control cycle. The pulse width modulation drive signal for the current cycle refers to the pulse width modulation signal generated using the optimal duty cycle as a parameter to drive servo motor 1. After this signal is input to the servo motor 1 driver, it can control the motor to run at a preset speed and torque.

[0040] Specifically, from a series of candidate duty cycle values ​​obtained by the quadratic programming algorithm, the first optimal duty cycle is selected. This is the duty cycle that best suits the current control cycle. The electromechanical spring brake actuator module 12 generates a corresponding pulse width modulation drive signal based on this optimal duty cycle. The frequency of this signal is set to 1kHz to match the drive requirements of the servo motor 1. The pulse width modulation drive signal is output to the servo motor 1 driver, driving the servo motor 1 to run at the speed and torque corresponding to the optimal duty cycle, thereby driving the ball screw transmission mechanism 2 to move.

[0041] For example, in the first control cycle of the smooth adjustment phase, the first optimal duty cycle obtained by the quadratic programming algorithm is 38%, generating a pulse width modulation drive signal with a frequency of 1kHz and a duty cycle of 38%, which is output to the servo motor 1 driver. After receiving the signal, the servo motor 1 rotates at the corresponding speed, and the rotational motion is converted into linear motion through the ball screw transmission mechanism 2, pushing the piston rod 5 to extend, so that the pressure of the disc spring 3 rises smoothly from 19kN to 20.2kN. In the next control cycle, the optimal duty cycle of 30% is obtained, generating the corresponding pulse width modulation drive signal, and continuing to push the pressure to approach the target value of 22kN, ensuring that the pressure build-up is smooth and without shock.

[0042] Finally, the servo motor is driven to convert the rotary motion into linear motion via the ball screw transmission mechanism 2, pushing the piston rod 5 to extend and compress the series-connected disc spring 3. This causes the disc spring 3 to generate elastic force, which acts on the brake shoes, thereby establishing a braking clamping force. The ball screw transmission mechanism 2 is a precision transmission component that converts the rotary motion of the servo motor into the linear motion of the piston rod 5. It features high transmission efficiency, high positioning accuracy, and low wear, and can accurately convert the motor rotation angle into the extension displacement of the piston rod 5.

[0043] Specifically, after receiving the pulse width modulation drive signal, the servo motor 1 rotates according to the speed and torque set by the signal, and transmits the rotational motion to the ball screw transmission mechanism 2. The ball screw transmission mechanism 2 accurately converts the rotational motion of the servo motor 1 into the linear extension motion of the piston rod 5 through the meshing motion of the screw and nut, maintaining a linear correspondence between displacement and rotation angle during the transmission process. During the extension of the piston rod 5, its end contacts one end of the series-connected disc spring 3 and applies a thrust, pushing the disc spring 3 to compress axially, so that the disc spring 3 generates an elastic force proportional to the compression amount. The other end of the disc spring 3 is connected to the brake shoe, and the elastic force is transmitted to the brake shoe through the transmission structure, pushing the brake shoe to clamp towards the direction of the mine monorail. The piston rod 5 is continuously pushed to extend and compress the disc spring 3 until the braking clamping force corresponding to the elastic force generated by the disc spring 3 matches the target braking force, thus completing the establishment of the braking clamping force.

[0044] For example, after receiving a pulse width modulation drive signal with a duty cycle of 30%, the servo motor 1 rotates at a constant speed of 30% of its rated speed. The rotational motion is converted into linear motion of the piston rod through the ball screw transmission mechanism 2. For every revolution of the motor, the piston rod 5 extends by 5mm. As the piston rod 5 continues to extend, it gradually compresses the series-connected disc spring 3. When the initial compression is 0mm, the elastic force is 0kN. When the compression reaches 4mm, the elastic force is 11kN. When the compression reaches 8mm, the elastic force is 22kN, which is precisely matched with the target braking force. At this time, the elastic force of the disc spring 3 is transmitted to the brake shoe, pushing the brake shoe to clamp with the monorail, establishing a braking clamping force of 22kN, which provides stable force support for subsequent braking actions.

[0045] In this embodiment of the invention, through precise pulse width modulation drive signal generation and mechanical transmission control, the rapid establishment and smooth adjustment of braking clamping force are achieved, effectively avoiding braking shock and braking lag. Utilizing the high-precision transmission characteristics of the ball screw transmission mechanism 2, the precise conversion of the servo motor's rotary motion into the linear motion of the piston rod 5 is ensured, improving the control accuracy of the piston rod 5's displacement and the compression of the disc spring 3, thus guaranteeing precise matching of the braking clamping force. The dynamic adaptation of the series-connected disc spring 3 with the pulse width modulation drive signal achieves adjustable elastic force, enabling the braking clamping force to precisely match the target braking force under different working conditions. This adapts to the complex and varied operating requirements of the mine monorail hoist, reducing the potential for failure in traditional braking mechanisms and providing stable support for the safe braking of the mine monorail hoist.

[0046] The position-pressure coordinated closed-loop control module 13 is used to execute a coordinated closed-loop control strategy for position and pressure during braking based on the clamping force data, piston rod position data, and motor rotation angle data collected in real time by the controller. The coordinated closed-loop control strategy takes the target position of the piston rod and the target pressure of the spring as control targets. It adjusts the drive direction and output torque of the servo motor in real time by adjusting the duty cycle of the pulse width modulation signal, so that the actual position of the piston rod and the actual pressure of the spring synchronously approach their respective target values ​​until the elastic force generated by the disc spring stabilizes within the preset deviation tolerance range of the target braking force.

[0047] In embodiments of the present invention, such as Figure 3The diagram shows the spring brake assembly structure of the electric drive-spring series braking system for a mining monorail crane provided in an embodiment of the present invention. During braking, the controller collects clamping force data from pressure sensor 7, position data from piston rod 5 from displacement sensor 8, and motor rotation angle data from encoder 9 in real time, and executes a coordinated closed-loop control strategy for position and pressure. This coordinated closed-loop control strategy uses the target position of piston rod 5 and the target pressure of disc spring 3 as control targets. It adjusts the duty cycle of the pulse width modulation signal in real time to coordinate the adjustment of the driving direction and output torque of servo motor 1, so that the actual position of piston rod 5 and the actual pressure of spring synchronously approach their respective target values ​​until the elastic force generated by disc spring 3 stabilizes within the preset deviation tolerance range of the target braking force.

[0048] The position of piston rod 5 and the pressure of disc spring 3 are two interconnected and mutually influential parameters. If independent control is used for position and pressure closed loops, control conflicts are likely to occur, leading to overshoot and oscillations during braking force establishment, affecting braking smoothness and accuracy. Furthermore, the underground operation of mine monorail hoists is complex, and the robustness and anti-interference capabilities of control strategies differ between conventional and severe braking scenarios. A single control algorithm cannot adapt to all operating conditions. Therefore, corresponding control algorithms must be adopted for different scenarios. Through coordinated closed-loop control, the actual position of piston rod 5 and the actual pressure of disc spring 3 are synchronized to approach the target value, suppressing braking force overshoot and oscillations, ensuring a smooth and accurate braking process, and adapting to the needs of different underground operating conditions.

[0049] In one embodiment, the position-pressure coordinated closed-loop control module 13 is further configured to: The coordinated closed-loop control strategy for execution position and pressure includes: In conventional braking scenarios, a sliding mode control algorithm is used to implement a coordinated closed-loop control strategy for position and pressure. The sliding mode control algorithm defines a coordinated sliding surface function, and the independent variables of the coordinated sliding surface function include piston rod position deviation and spring pressure deviation. The piston rod position deviation is the difference between the target compression and the actual compression, and the spring pressure deviation is the difference between the target pressure and the actual pressure. The sliding mode control algorithm integrates the piston rod position deviation and spring pressure deviation into a single cooperative error variable through a weighted summation method. The weight coefficients in the weighted summation are dynamically adjusted by the controller based on the currently collected track gradient and vehicle load, and the pulse width modulation correction signal is calculated with the cooperative error variable approaching zero as the control objective. The sliding mode control algorithm introduces a saturation function to replace the ideal sign function to eliminate the chattering phenomenon inherent in sliding mode control, and sets the boundary layer thickness of the saturation function to be positively correlated with the tangential stiffness coefficient of the disc spring under the current compression.

[0050] First, in a conventional braking scenario, a sliding mode control algorithm is used to implement a coordinated closed-loop control strategy for position and pressure. The sliding mode control algorithm defines a coordinated sliding surface function, whose independent variables include the position deviation of the piston rod 5 and the pressure deviation of the disc spring 3. The position deviation of the piston rod 5 is the difference between the target compression and the actual compression, and the pressure deviation of the disc spring 3 is the difference between the target pressure and the actual pressure. A conventional braking scenario refers to a braking scenario where track conditions are relatively stable and system parameter uncertainties are small, such as braking conditions in smooth underground tunnels, without significant pollution, and with stable loads.

[0051] The constant sliding mode control algorithm is a nonlinear control algorithm that, by designing a sliding mode surface function, allows the system state to converge to the target value along the sliding mode surface. It features fast response, strong robustness, and insensitivity to parameter perturbations, making it suitable for coordinated control in conventional braking scenarios. The coordinated sliding mode surface function is the core function of the sliding mode control, used to integrate the position deviation of piston rod 5 and the pressure deviation of disc spring 3 to construct a single control input, avoiding conflicts between independent position and pressure control. Its independent variables are two deviation parameters, and its output is the core reference quantity for coordinated control. The position deviation of piston rod 5 refers to the difference between the target compression and the actual compression of the piston rod, i.e., position deviation = target compression - actual compression. A positive difference indicates insufficient actual compression, and a negative difference indicates excessive actual compression. The pressure deviation of disc spring 3 refers to the difference between the target pressure and the actual pressure of the spring, i.e., pressure deviation = target pressure - actual pressure. A positive difference indicates insufficient actual pressure, and a negative difference indicates excessive actual pressure.

[0052] Specifically, the controller uses pressure sensor 7 to collect the actual compression of piston rod 5 and the actual pressure of disc spring 3 in real time. Combining this with the target compression and target pressure, it calculates the position deviation of piston rod 5 and the pressure deviation of disc spring 3, respectively. A cooperative sliding surface function is defined, using the position deviation of piston rod 5 and the pressure deviation of disc spring 3 as independent variables, to construct a linearly fused sliding surface function. The function expression is as follows: Where s is the output value of the cooperative sliding surface. This represents the positional deviation of piston rod 5. The pressure deviation of the disc spring 3. , The initial weighting coefficients are all set to 1. The calculated cooperative sliding surface output value s is used as the input to the sliding mode control algorithm for subsequent control quantity calculation.

[0053] For example, the target braking force of 22kN corresponds to the target pressure of the disc spring 3 of 22kN and the target compression of the piston rod 5 of 8mm. Under normal braking conditions, the pressure sensor 7 collects the actual compression of the piston rod 5 of 7.2mm and the actual pressure of the disc spring 3 of 20kN in real time; the deviation is calculated as the positional deviation of the piston rod 5. Pressure deviation of disc spring 3 Substitute the cooperative sliding surface function The output value of the sliding surface, s=2.8, is obtained and used as the input basis for subsequent sliding mode control.

[0054] Secondly, the sliding mode control algorithm integrates the position deviation of the piston rod 5 and the pressure deviation of the butterfly spring 3 into a single cooperative error variable through a weighted summation method. The weight coefficients in the weighted summation are dynamically adjusted by the controller according to the currently collected track slope and vehicle load, and the pulse width modulation correction signal is calculated with the cooperative error variable approaching zero as the control target.

[0055] The collaborative error variable refers to a single error variable obtained by weighted summation of the piston rod position deviation and the spring pressure deviation. This variable serves as the input for sliding mode control, avoiding conflicts between independent position and pressure control and achieving coordinated convergence. The weighting coefficient is the proportional coefficient used in the weighted summation process to adjust the influence weight of the piston rod 5 position deviation and the disc spring 3 pressure deviation in the collaborative error variable. It can be dynamically adjusted according to the operating conditions to achieve control priority switching under different operating conditions. Linear interpolation refers to the adjustment method of the weighting coefficient under intermediate operating conditions. Through linear calculation, it achieves a continuous and smooth change in weight from one operating condition to another, avoiding control fluctuations caused by abrupt weight changes.

[0056] Specifically, the independent variables based on the cooperative sliding surface function ( , The variables are weighted and summed to form a single collaborative error variable, expressed as: Where e is the cooperative error variable, This is the weighting coefficient for the positional deviation. is the weighting coefficient for pressure deviation, and The controller dynamically adjusts the weighting coefficients in real time based on the currently collected track gradient and vehicle load: For heavy-load conditions with steep gradients (gradient ≥ 8°, load ≥ 8t): the weighting of pressure deviation is increased. Prioritize ensuring rapid braking force establishment to avoid insufficient braking; for lightly loaded conditions on level roads (gradient ≤ 2°, load ≤ 2t): increase the weighting of positional deviation. Prioritize smooth braking and avoid braking shock; for intermediate operating conditions (slope 2°~8°, load 2t~8t): use linear interpolation to calculate based on the actual values ​​of slope and load. , The specific values ​​are used to achieve continuous and smooth changes in weights; with the cooperative error variable e approaching zero as the control target, the pulse width modulation (PWM) correction signal is calculated in combination with the sliding mode control algorithm to adjust the drive parameters of servo motor 1.

[0057] For example, with a target compression of 8mm and a target pressure of 22kN, the weight adjustment and calculation of the cooperative error variable are explained under three working conditions: Heavy load condition with a large gradient: track gradient 10°, load 10t, set... , Real-time data collection Cooperative error variables Light load condition on level track: track gradient 0°, load 0t, settings Real-time data collection Cooperative error variables Intermediate operating conditions: track gradient 5°, load 5t, obtained through linear interpolation. , Collaborative error variables This achieves a smooth transition of weights.

[0058] Finally, the sliding mode control algorithm introduces a saturation function to replace the ideal sign function to eliminate the chattering phenomenon inherent in sliding mode control. The boundary layer thickness of the saturation function is set to be positively correlated with the tangential stiffness coefficient of the disc spring 3 under the current compression. The saturation function is a continuous nonlinear function whose output changes linearly with the error within the boundary layer and has a fixed value outside the boundary layer. It is used to replace the ideal sign function and eliminate the chattering phenomenon inherent in sliding mode control. The ideal sign function is the function used in standard sliding mode control, and its output is a step switch, only 1, -1, or 0, which easily leads to high-frequency oscillation of the control quantity, causing fluctuations in braking clamping force and piston rod chattering.

[0059] Chattering refers to the frequent fluctuations in the servo motor drive signal caused by the step switching of the sign function in standard sliding mode control, which in turn causes the piston rod 5 to chatter and the pressure of the disc spring 3 to fluctuate, affecting braking smoothness. Boundary layer thickness refers to the thickness of the transition interval set in the saturation function, used to achieve a smooth transition of the control quantity. Within the boundary layer, the control quantity changes linearly with the cooperative error variable, while outside the boundary layer, the control quantity remains constant, and the thickness can be adaptively adjusted. Tangential stiffness coefficient refers to the instantaneous stiffness of the disc spring 3 under the current compression, reflecting the change in elastic force when the compression of the disc spring 3 changes slightly. Different compression amounts result in different tangential stiffness coefficients; the larger the compression amount, the larger the stiffness coefficient.

[0060] Specifically, a saturation function is used to replace the ideal sign function in sliding mode control. The expression for the saturation function is: Where δ is the boundary layer thickness and sign(e) is the sign function; the boundary layer thickness δ is set to be positively correlated with the tangential stiffness coefficient of the butterfly spring 3 under the current compression, so as to achieve adaptive adjustment: a large tangential stiffness coefficient means that the butterfly spring 3 is relatively stiff, such as when the compression is ≥6mm: a small change in position can cause a large fluctuation in pressure, so a larger boundary layer thickness is set, δ=0.5~0.8, to provide sufficient smooth transition and avoid chattering; a small tangential stiffness coefficient means that the butterfly spring 3 is relatively soft, such as when the compression is <6mm: the pressure is less sensitive to changes in position, so a smaller boundary layer thickness is used, δ=0.2~0.5, to balance smoothness and response speed; the output of the saturation function is substituted into the sliding mode control law to calculate the pulse width modulation correction signal, which is used to adjust the duty cycle of the servo motor 1 in real time, so as to make the cooperative error variable approach zero.

[0061] For example, when the actual compression of the piston rod 5 is 7.2mm ≥ 6mm, the tangential stiffness coefficient of the disc spring 3 is large, and the boundary layer thickness is set accordingly. ;because saturation function output Substituting the values ​​into the sliding mode control law, the PWM correction signal is calculated. Increasing the duty cycle of servo motor 1 pushes piston rod 5 to extend and the pressure of disc spring 3 to rise. As the control progresses, the cooperative error variable e decreases to 0.4. saturation function output The control quantity changes linearly to avoid abrupt switching; when the compression of piston rod 5 increases to 7.8mm, the tangential stiffness coefficient further increases, the boundary layer thickness is adjusted to 0.7, e decreases to 0.2, and the saturation function output... The control quantity continues to be smoothly adjusted until e approaches zero, without any chattering.

[0062] The coordinated closed-loop control strategy for execution position and pressure also includes: In severe braking scenarios, an active disturbance rejection control algorithm is adopted as an alternative to the sliding mode control algorithm. The active disturbance rejection control algorithm includes three functional units: a tracking differentiator, an extended state observer, and a nonlinear state error feedback law. The tracking differentiator ensures a smooth transition between the target position of the piston rod and the target pressure of the spring, avoiding abrupt changes in the target value that could impact the braking system. The expansion state observer estimates the total disturbance in the braking system in real time. The total disturbance includes the nonlinear stiffness characteristics of the disc spring, the time-varying characteristics of the friction coefficient between the track and the brake shoe, and the external impact load. The estimated total disturbance value is then compensated in the feedforward channel. The nonlinear state error feedback law generates a nonlinear combination of control outputs based on the error between the transition target value output by the tracking differentiator and the actual state estimate output by the extended state observer.

[0063] First, in severe braking scenarios, an active disturbance rejection control (ADRC) algorithm is adopted as an alternative to the sliding mode control algorithm. The ADRC algorithm comprises three functional units: a tracking differentiator, an extended state observer, and a nonlinear state error feedback law. Severe braking scenarios refer to braking situations with complex and variable track conditions, significant uncertainty in system parameters, or strong external disturbances, including conditions such as wet trackways, tracks contaminated with mud and sand, or significant fatigue decay of springs. The ADRC algorithm is a nonlinear robust control algorithm that does not require a precise system model. Through the tracking differentiator, extended state observer, and nonlinear state error feedback law, it achieves target value tracking and disturbance suppression, making it suitable for severe braking scenarios.

[0064] The tracking differentiator ensures a smooth transition between the target position of the piston rod and the target pressure of the spring, preventing abrupt changes in the target value from impacting the braking system. The tracking differentiator is the first functional unit of the active disturbance rejection control algorithm, used to ensure a smooth transition between the target position of the piston rod 5 and the target pressure of the disc spring 3, avoiding abrupt changes in the target value from impacting the braking system and ensuring a smooth control process. The transition process refers to the smooth curve output by the tracking differentiator, gradually transitioning from the current actual value to the target value, rather than a step jump, used to mitigate braking shocks caused by sudden changes in the target value.

[0065] Specifically, the controller acquires the target compression of piston rod 5 and the target pressure of disc spring 3, and inputs them into the tracking differentiator of the active disturbance rejection control algorithm. The tracking differentiator automatically generates a smooth transition target value based on the difference between the target value and the current actual value, and uses an exponential decay or linear transition method to make the transition target value gradually approach the final target value from the current actual value, avoiding step changes. The transition target value output by the tracking differentiator is used as the input of the subsequent extended state observer and nonlinear state error feedback law to ensure that the control process is smooth and shock-free.

[0066] For example, in the initial stage of braking, the actual compression of piston rod 5 is 0mm and the actual pressure of disc spring 3 is 0kN. If the target values ​​of 8mm and 22kN are directly input, it will cause servo motor 1 to accelerate suddenly, resulting in braking shock. After receiving the target values, the tracking differentiator generates a smooth transition curve: the transition target value of piston rod 5 starts from 0mm and gradually increases at a rate of 0.5mm / control cycle, reaching 8mm after 16 control cycles; the pressure transition target value of disc spring 3 starts from 0kN and gradually increases at a rate of 1.375kN / control cycle, reaching 22kN after 16 control cycles. The tracking differentiator outputs the transition target value in real time to avoid step changes in the target value, alleviate braking shock, and adapt to the smooth braking requirements of wet tunnels.

[0067] The extended state observer estimates the total disturbance in the braking system in real time. This total disturbance includes the nonlinear stiffness characteristics of the disc spring 3, the time-varying characteristics of the friction coefficient between the track and the brake shoes, and external impact loads. The estimated total disturbance value is then compensated in the feedforward channel. The extended state observer is the second functional unit of the active disturbance rejection control algorithm. It estimates the total disturbance in the braking system in real time and compensates for the estimated disturbance value in the feedforward channel, offsetting the impact of the disturbance on the control effect and improving the system's anti-interference capability. The total disturbance is the sum of various interference factors affecting the control accuracy in the braking system, including the nonlinear stiffness characteristics of the disc spring 3, the time-varying characteristics of the friction coefficient between the track and the brake shoes, and external impact loads. Feedforward channel compensation refers to substituting the estimated total disturbance value into the control law to correct the control quantity in advance, offsetting the deviation caused by the disturbance and avoiding an increase in position and pressure deviations due to the disturbance.

[0068] Specifically, the extended state observer collects real-time state parameters such as the actual position of piston rod 5, the actual pressure of disc spring 3, and the rotation angle of servo motor 1, and combines them with the transition target value output by the tracking differentiator. Through the algorithm logic of the extended state observer, the total disturbance in the braking system is estimated in real time, including the combined effects of factors such as the nonlinear stiffness of disc spring 3, the time-varying friction coefficient, and external impacts, and the disturbance estimate is output. The estimated total disturbance value is substituted into the feedforward channel to compensate for the subsequent generated control quantity in advance, offsetting the deviation caused by the disturbance and ensuring the coordinated convergence of position and pressure.

[0069] For example, under normal circumstances, the compression of piston rod 5 by 7mm corresponds to a pressure of 20kN for butterfly spring 3. Due to track mud and sand contamination, the friction coefficient decreases, resulting in an actual pressure of only 18.5kN for butterfly spring 3, generating a pressure deviation of 1.5kN, which is part of the total disturbance. The expansion state observer collects state parameters in real time and estimates the total disturbance value to be 1.5kN. This disturbance estimate is compensated for by the feedforward channel. When calculating the control quantity, an additional corresponding PWM correction signal is added to increase the output torque of servo motor 1, pushing piston rod 5 to continue extending, so that the pressure of butterfly spring 3 quickly rises from 18.5kN to 20kN, thus offsetting the deviation caused by the disturbance.

[0070] The nonlinear state error feedback law generates a nonlinear combination of control outputs based on the error between the transitional target value output by the tracking differentiator and the actual state estimate output by the extended state observer. This nonlinear state error feedback law is the third functional unit of the active disturbance rejection control algorithm. It generates a nonlinear combination of control outputs based on the error between the transitional target value output by the tracking differentiator and the actual state estimate output by the extended state observer, achieving rapid error convergence. The transitional target value refers to the target value output by the tracking differentiator that represents a smooth transition. The actual state estimate refers to the estimated value output by the extended state observer, including the actual position of the piston rod 5 after disturbance compensation and the actual pressure of the disc spring 3, which is more accurate than the directly acquired actual values. The nonlinear combination control output refers to the control output generated using a nonlinear function based on the magnitude of the error. A large error results in a large adjustment range for the control output, while a small error results in a small adjustment range, balancing convergence speed and stability.

[0071] Specifically, the error between the transition target value output by the tracking differentiator and the actual state estimate output by the extended state observer is calculated, including position error and pressure error. A nonlinear state error feedback law is used to nonlinearly combine the position error and pressure error to generate a control output. The adjustment range of the control output is nonlinearly related to the magnitude of the error; the larger the error, the larger the adjustment range, which speeds up convergence; the smaller the error, the smaller the adjustment range, which ensures stability. The generated nonlinear control output is converted into a pulse width modulation correction signal and output to the driver of servo motor 1. The driving direction and output torque of servo motor 1 are adjusted so that the actual position of piston rod 5 and the actual pressure of disc spring 3 synchronously approach the transition target value, and finally reach the final target value.

[0072] For example, the transition target value of piston rod 5 is 7mm, the transition target value of pressure of disc spring 3 is 20kN, the actual state estimate of piston rod 5 is 6.5mm, and the actual state estimate of pressure of disc spring 3 is 18.5kN. Calculation errors: position error = 7mm - 6.5mm = 0.5mm, pressure error = 20kN - 18.5kN = 1.5kN. The nonlinear state error feedback law generates a nonlinear combined control quantity based on the magnitude of the error: due to the large error, the control quantity adjustment amplitude is large, corresponding to a 5% increase in the duty cycle of the PWM correction signal. After receiving the correction signal, servo motor 1 increases the output torque, pushing piston rod 5 to extend rapidly. After one control cycle, the actual position of piston rod 5 rises to 6.8mm, and the pressure of disc spring 3 rises to 19.5kN. The error is calculated again: position error 0.2mm, pressure error 0.5mm. The nonlinear feedback law generates a smaller amplitude control quantity, which continues to be adjusted until the error approaches zero, achieving smooth convergence.

[0073] In this embodiment of the invention, a collaborative closed-loop control strategy is employed to synchronize the actual position of the piston rod 5 with the actual pressure of the disc spring 3, effectively suppressing overshoot and oscillation during the braking force build-up process and improving the smoothness of the braking process. In conventional braking scenarios, the sliding mode control algorithm, combined with dynamic weight adjustment of the collaborative error variable, achieves adaptive switching of control priorities under different operating conditions, balancing braking force build-up speed and braking smoothness. The adaptive boundary layer setting introduced by the saturation function completely eliminates the chattering phenomenon inherent in sliding mode control, balancing control response speed and smoothness, and adapting to conventional operating conditions with stable tracks and small parameter fluctuations. In severe braking scenarios, the active disturbance rejection control algorithm mitigates the step impact of the target value through a tracking differentiator, estimates and compensates for the total disturbance in real time through an extended state observer, and achieves rapid error convergence through a nonlinear state error feedback law, improving the system's anti-interference capability and robustness, and effectively coping with disturbances caused by severe operating conditions such as track contamination, external impacts, and spring fatigue. The adaptive switching between the two control algorithms enables the position-pressure coordinated closed-loop control to fully adapt to the complex and ever-changing braking scenarios of mine monorails, ensuring that the braking clamping force remains stable within the preset deviation tolerance range of the target braking force. This improves the accuracy, reliability, and adaptability of the braking system, providing a strong guarantee for the safe braking of mine monorails.

[0074] The system provided in this embodiment of the invention is also used for: A nonlinear mapping model between the compression and elastic force of a butterfly spring is built into the controller. The nonlinear mapping model is used to reverse the target compression of the piston rod based on the target braking force, and the target compression is used as the input reference for the coordinated closed-loop control strategy of position and pressure. The nonlinear mapping model uses a radial basis function neural network to positively fit the compression-elastic force characteristic curve of the butterfly spring under different combinations. The different combinations include single-piece stacked combination, multi-piece paired combination, and a combination of stacked and paired combinations. In high-precision braking scenarios, the input layer of the radial basis function neural network is the compression of a butterfly spring, the hidden layer uses Gaussian radial basis functions to partition the input space into local responses, and the output layer is the elastic force of the butterfly spring. During the braking preparation phase, the controller calls the nonlinear mapping model, substitutes the target braking force as the desired output into the radial basis function neural network, and solves the compression of the disc spring in the input layer through an iterative optimization algorithm. The obtained compression is then used as the target compression of the piston rod.

[0075] First, a nonlinear mapping model between the compression and elastic force of the disc spring 3 is built into the controller. The nonlinear mapping model is used to reverse the target compression of the piston rod 5 based on the target braking force, and the target compression is used as the input reference for the coordinated closed-loop control strategy of position and pressure. The nonlinear mapping model uses a radial basis function neural network to positively fit the compression-elastic force characteristic curve of the disc spring 3 under different combination methods. The different combination methods include single-piece stacking combination, multi-piece pairing combination, and a combination of stacking and pairing.

[0076] Among them, the nonlinear mapping model is a computational model used to describe the nonlinear correspondence between the compression and elastic force of the butterfly spring 3, supporting both forward calculation and backward solving. The radial basis function (RBF) neural network is a local approximation feedforward neural network that has high accuracy and fast convergence speed in fitting nonlinear static mappings, and is suitable for fitting the mechanical properties of butterfly springs. The butterfly spring 3 can be combined in three typical structures: single-piece stacked combination, multi-piece paired combination, and a hybrid combination of stacked and paired combinations.

[0077] Specifically, a nonlinear mapping model of compression and elastic force of the disc spring 3 is pre-built and solidified inside the controller. The model uses a radial basis function neural network as the fitting structure. The neural network is trained using bench test data under different combinations of disc springs 3 to positively fit the full-stroke compression-elastic force characteristic curve. The model can cover various structural forms such as single piece, pair, and hybrid combination to ensure versatility.

[0078] Secondly, in high-precision braking scenarios, the input layer of the radial basis function neural network is the compression of the butterfly spring 3, the hidden layer uses Gaussian radial basis functions to partition the input space into local responses, and the output layer is the elastic force of the butterfly spring 3. The Gaussian radial basis function is an activation function for the hidden layer of a radial basis function neural network, possessing local response characteristics, which can improve the accuracy of nonlinear fitting. High-precision braking scenarios refer to heavy-load, steep-gradient, and passenger-carrying transportation conditions with strict requirements for braking force error and braking smoothness, suitable for situations where the controller has sufficient computing resources and high back-calculation accuracy is required.

[0079] Specifically, in high-precision braking scenarios, a fixed structure for the radial basis function neural network is defined: Input layer: 1 neuron, with the input variable being the compression amount of the butterfly spring 3; Hidden layer: using Gaussian radial basis functions as activation functions to partition the local response space of the compression amount input; Output layer: 1 neuron, with the output variable being the elastic force of the butterfly spring 3; After the network is trained, it can accurately output the corresponding elastic force according to any compression amount.

[0080] Finally, during the braking preparation phase, the controller invokes the nonlinear mapping model, substituting the target braking force as the desired output into the radial basis function neural network. An iterative optimization algorithm is used to reverse-engineer the compression of the disc spring 3 in the input layer, and the calculated compression is used as the target compression of the piston rod 5. The iterative optimization algorithm is a numerical solution method that approximates the optimal solution through successive iterations, used to calculate the compression from the elastic force. The reverse-engineering uses the target braking force as the desired output to calculate the corresponding input compression. The target compression refers to the theoretical displacement of the piston rod 5 obtained by reverse-engineering the nonlinear mapping model, serving as the position control reference for the position-pressure coordinated closed loop.

[0081] Specifically, during the braking preparation phase, the controller calls the trained nonlinear mapping model; the target braking force is used as the expected output value of the radial basis function neural network; with the goal of minimizing the deviation between the network output elastic force and the target braking force, an iterative optimization algorithm is used to solve the compression amount of the disc spring 3 corresponding to the input layer; the compression amount after iterative convergence is determined as the target compression amount of the piston rod 5; and the target compression amount is input into the position-pressure coordinated closed-loop control module as the position control reference.

[0082] In this embodiment of the invention, a radial basis function neural network is used to achieve high-precision fitting of the strong nonlinear characteristics of butterfly springs 3 with different combinations, solving the problem of large errors in traditional analytical models; it achieves accurate reverse solution of the target braking force to the target compression of the piston rod 5, providing a high-precision position reference for coordinated closed-loop control; it is compatible with various combinations of butterfly springs 3, improving the versatility and adaptability of the braking system; it reduces the braking force reference error, suppresses control overshoot and pressure oscillation, and improves the overall braking accuracy and operational stability.

[0083] The nonlinear mapping model also includes alternative solutions: In scenarios requiring rapid response, a piecewise analytical function is used as an alternative to the radial basis function neural network. The piecewise analytical function divides the compression-elasticity characteristic curve of the disc spring into three intervals: a linear segment, a hardening segment, and a near-flattening segment, based on the number of stacked discs, the number of pairs, and the geometric parameters of a single disc spring. The linear segment uses a linear function expression to fit the linear relationship between compression and elasticity, while the hardening segment uses a cubic function expression to fit the nonlinear increasing relationship between compression and elasticity. The controller directly calculates the target compression of the piston rod by selecting the inverse function analytical expression of the corresponding segment interval based on the segment interval where the target braking force is located, without the need for reverse solving through iterative optimization algorithms.

[0084] First, in scenarios requiring rapid response, a piecewise analytical function is used as an alternative to the radial basis function neural network. This piecewise analytical function divides the compression-elasticity characteristic curve of the disc spring 3 into three intervals: a linear segment, a hardening segment, and a near-flattening segment, based on the number of stacked discs, the number of paired discs, and the geometric parameters of a single disc spring 3. The linear segment uses a linear function expression to fit the linear relationship between compression and elasticity, while the hardening segment uses a cubic function expression to fit the nonlinear increasing relationship between compression and elasticity. Rapid response scenarios refer to situations with high requirements for braking delay and control calculation speed, such as emergency braking, frequent starts and stops, and short-distance braking. These scenarios require low latency and high real-time performance in the control system, and are suitable for situations where controller computing resources are limited or rapid response is required.

[0085] The piecewise analytical function refers to dividing the compression-elasticity characteristic curve of the butterfly spring 3 into multiple intervals according to its mechanical properties. Each interval is fitted with an independent analytical expression, and the function model can be directly calculated in both forward and reverse directions. The linear segment represents the initial compression interval of the butterfly spring 3, where the compression and elasticity increase approximately linearly, and the stiffness remains essentially constant. The hardening segment represents the intermediate compression interval of the butterfly spring 3, where the elasticity increases rapidly and non-linearly with increasing compression, and the stiffness gradually increases. The segment approaching flattening refers to the period when the butterfly spring 3 approaches its maximum compression, nearing a flattened state, the elasticity increase becomes more gradual, and the stiffness increases sharply. The number of stacked plates refers to the number of individual butterfly springs 3 stacked in the same direction. The number of interlocking groups refers to the number of groups of butterfly springs 3 interlocked in opposite directions. The geometric parameters of a single piece include structural parameters such as the inner diameter, outer diameter, free height, and thickness of the individual butterfly spring 3.

[0086] Specifically, the controller identifies the critical points of the spring's mechanical characteristic range based on the number of stacked plates, the number of paired groups, and the geometric parameters of each plate in the disc spring 3. The compression-elasticity characteristic curve is uniformly divided into three continuous intervals: a linear segment, a hardening segment, and a near-flattening segment. For each interval, a fitting analytical expression is established: Linear segment: a linear function is used, where the elastic force changes linearly with the compression; Hardening segment: a cubic function is used, where the elastic force increases non-linearly with the compression, reflecting the stiffness hardening characteristic; Near-flattening segment: a non-linear function is used to fit the flattening characteristics, conforming to the mechanical changes of the disc spring 3 in its near-flattening state.

[0087] Secondly, the controller directly calculates the target compression of piston rod 5 by selecting the inverse function analytical expression of the corresponding segment interval based on the segment interval where the target braking force is located, without needing to solve it backwards through an iterative optimization algorithm. The inverse function analytical expression is a direct calculation formula for calculating the compression by reverse calculation of the braking force, derived from the forward function of each segment.

[0088] Specifically, the controller obtains the current target braking force and determines the segment interval to which it belongs; it calls the pre-stored inverse function analytical expression corresponding to the segment; it substitutes the target braking force into the inverse function and directly calculates the compression amount of the corresponding disc spring 3; it uses the calculation result as the target compression amount of the piston rod 5 and inputs it into the position-pressure coordinated closed-loop control module, without the need for iterative optimization algorithms throughout the process.

[0089] In this embodiment of the invention, a piecewise analytical function is used instead of a neural network, eliminating the iterative optimization process, reducing computational load, and increasing the speed of solving the target compression amount, thus meeting the low latency requirements of fast response scenarios. Piecewise fitting is performed according to linear segments, hardening segments, and near-flattening segments, taking into account both the nonlinear characteristics of the butterfly spring 3 and computational simplicity, achieving rapid computation while ensuring sufficient control accuracy. Direct solution via inverse function simplifies computational logic and reduces the computational load on the controller, which is beneficial for improving system real-time performance and stability. The piecewise segmentation and expression can be adaptively adjusted according to the number of stacked pieces, the number of paired groups, and geometric parameters, making it compatible with various combinations of butterfly spring 3 structures and highly versatile.

[0090] The system provided in this embodiment of the invention is also used in the braking release step: When the controller receives the brake release command, it sends a reverse pulse width modulation drive signal to the servo motor, which drives the servo motor to rotate in the opposite direction. The piston rod is pulled back through the ball screw transmission mechanism, so that the disc spring is synchronously reset with the piston rod and releases its elastic potential energy. During the piston rod retraction process, the speed control mode is switched to use closed-loop speed control to keep the piston rod retraction speed at a preset constant speed value, thus avoiding rigid collision between the piston rod and the spring seat. The system receives piston rod position data from the displacement sensor in real time. When the piston rod position data indicates that the piston rod has returned to the initial mechanical origin, the power supply to the servo motor is cut off and the electromagnetic brake is triggered to lock the motor shaft.

[0091] First, when the controller receives a brake release command, it sends a reverse pulse width modulation (PWM) drive signal to servo motor 1, driving servo motor 1 to rotate in the opposite direction. This rotation, via ball screw transmission mechanism 2, pulls the piston rod back, causing the disc spring 3 to synchronously reset with the piston rod 5 and release its elastic potential energy. The brake release command refers to the control command received by the controller to release the brake and release the brake shoes, issued by the vehicle control system. The reverse pulse width modulation (PWM) drive signal is a PWM signal with the opposite phase to the braking pressure, used to control the reverse rotation of servo motor 1. The reverse rotation of servo motor 1 means that the motor rotor rotates in the opposite direction to the braking pressure, changing the driving action from pushing to pulling. The release of elastic potential energy means that the compressed disc spring 3 gradually extends as the piston rod 5 retracts, releasing the stored elastic potential energy and returning to a free state.

[0092] Specifically, the controller receives the braking release command from the host computer or the operating terminal; the controller generates and outputs a reverse pulse width modulation drive signal according to the current state of the servo motor 1; the reverse signal drives the servo motor 1 to rotate in the opposite direction, and the power is converted into linear tension through the ball screw transmission mechanism 2; the tension pulls the piston rod 5 to retract backward, and the series disc spring 3 extends and resets synchronously with the piston rod, gradually releasing the elastic potential energy.

[0093] For example, the current braking state of the mining monorail is as follows: the piston rod 5 is compressed by 8mm, and the pressure of the disc spring 3 is 22kN. After receiving the brake release command, the controller outputs a reverse PWM drive signal with a duty cycle of 40%; the servo motor 1 changes from forward to reverse, the ball screw transmission mechanism 2 converts the rotational power into a pulling force, pulling the piston rod 5 from the 8mm position to retract, the disc spring 3 gradually extends, the elastic potential energy is gradually released, and the braking clamping force begins to decrease.

[0094] Secondly, during the piston rod retraction process, the system switches to speed control mode, employing closed-loop speed control to maintain the piston rod's retraction speed at a preset constant value, preventing rigid collisions between the piston rod 5 and the spring seat 6. Speed ​​control mode means that the servo motor 1 uses a constant rotational speed as the control target, rather than position or torque, ensuring a smooth retraction speed. Closed-loop speed control involves real-time acquisition of the motor speed or the displacement speed of the piston rod 5, comparing it with the target speed, and dynamically adjusting the PWM duty cycle. The preset constant speed value is a pre-set safe retraction speed for the piston rod 5, avoiding excessively fast impacts and excessively slow retractions leading to inefficiency. Rigid collisions refer to the direct impact of the piston rod 5's end with the spring seat 6 during high-speed retraction, causing impact, noise, and structural damage.

[0095] Specifically, the moment the piston rod 5 begins to retract, the controller automatically switches from position / torque mode to speed control mode; with a preset constant speed value as the control target, the speed signal of the displacement sensor 8 is collected in real time; the reverse PWM duty cycle is adjusted through speed closed loop to keep the retraction speed of the piston rod 5 stable at the preset constant value; the uniform retraction reduces motion impact and avoids rigid collision between the piston rod 5 and the spring seat 6.

[0096] For example, the preset constant retraction speed of piston rod 5 is 2mm / s. During the retraction process, displacement sensor 8 provides real-time feedback on the speed: if the instantaneous speed rises to 2.3mm / s, the controller reduces the reverse PWM duty cycle to 35% to reduce the speed of servo motor 1; if the speed drops to 1.7mm / s, the controller increases the duty cycle to 42% to increase the speed; finally, the speed is stabilized at around 2mm / s for uniform retraction, without impact or overspeed.

[0097] Finally, the position data of the piston rod 5 fed back by the displacement sensor 8 is received in real time. When the position data of the piston rod 5 indicates that the piston rod 5 has returned to the initial mechanical origin, the power supply to the servo motor is cut off and the electromagnetic brake is triggered to lock the motor shaft. The displacement sensor 8 is a sensing component installed on the brake actuator to detect the actual position of the piston rod 5 in real time. The initial mechanical origin refers to the mechanical zero position where the piston rod 5 is fully retracted and the disc spring 3 is completely free, with no compression and no braking clamping force. The electromagnetic brake is a power-off brake mechanism built into the servo motor 1. It is released when energized and locked when de-energized, used to lock the motor shaft when stationary. Locking the motor shaft is achieved by fixing the motor rotor with the friction force of the electromagnetic brake, preventing the piston rod 5 from shifting under external force.

[0098] Specifically, the controller receives position data of piston rod 5 from displacement sensor 8 in real time; determines in real time whether the position is equal to the initial mechanical origin; when it detects that piston rod 5 has returned to the initial mechanical origin, it immediately cuts off the drive power of servo motor 1; after servo motor 1 is powered off, electromagnetic brake automatically loses power and engages, clamping and locking the motor shaft, so that piston rod 3 remains at the origin position.

[0099] For example, the piston rod 5 retracts at a constant speed, and the displacement sensor 8 provides real-time feedback on the position: 7mm, 5mm, 2mm, 0mm. After the controller detects that the position is 0mm, it immediately cuts off the power supply to the servo motor 1; the electromagnetic brake changes from the energized release state to the de-energized lock state, firmly locking the motor shaft, and the piston rod 5 is fixed at the zero position and no longer moves, so that it will not deviate due to the reaction force of the disc spring 3 or vibration, and the braking is completed.

[0100] In this embodiment of the invention, reverse PWM drive combined with uniform speed retraction is used to achieve smooth energy release of the disc spring 3, avoiding vehicle jerking caused by sudden release of the brake shoes; speed closed-loop control ensures a constant retraction speed, fundamentally avoiding rigid collision between the piston rod 5 and the spring seat 6, reducing mechanical impact and component wear; precise detection of the initial mechanical origin and electromagnetic brake locking ensure that the piston rod 5 is fully reset and without slippage, ensuring stable and consistent braking gap, and improving the service life and operational safety of the mining monorail braking system.

[0101] The system provided in this embodiment of the invention is also used in the power failure emergency braking step: When the controller detects that the power supply voltage is lower than the preset power failure threshold and the duration exceeds the preset anti-shake time, it determines that the system has experienced a power failure and immediately triggers the electromagnetic brake of the servo motor to release through software logic. After the electromagnetic brake is released, the compressed disc spring actively pushes the piston rod to extend by relying on the preload, which drives the brake shoe to quickly clamp the track of the mining monorail until the deformation of the disc spring reaches the limit position of the mechanical limit device.

[0102] First, when the controller detects that the power supply voltage is lower than the preset power-off threshold and the duration exceeds the preset anti-shake time, it determines that a power-off fault has occurred in the system, and immediately triggers the release of the electromagnetic brake of the servo motor through software logic. The power-off threshold is the voltage critical value for determining a power-off fault. For example, it is set to 70% of the rated working voltage of the system. If it is lower than this value, the power supply is regarded as abnormal. The anti-shake time is a delay threshold set to avoid misjudgment due to instantaneous voltage fluctuations. For example, it is set to 20 milliseconds. Only when the voltage is lower than the threshold and lasts for more than the anti-shake time, a real power-off is determined. A power-off fault refers to an abnormal state where the system power supply voltage is insufficient and cannot maintain normal operation. Software logic triggering means that the controller judges through an internal program and automatically outputs a control instruction without manual operation.

[0103] Specifically, the controller continuously collects the system power supply voltage and compares it with the preset power-off threshold in real time. The determination condition is: power supply voltage < 70% of the rated voltage, and the duration > 20 milliseconds of anti-shake time. When the above conditions are met, the controller officially determines a power-off fault, and immediately issues a trigger instruction through the internal software logic to control the electromagnetic brake action of the servo motor 1.

[0104] Exemplarily, if the rated working voltage is DC24V, then the power-off threshold = 24V × 70% = 16.8V, and the anti-shake time is 20ms. The controller monitors the voltage in real time: at a certain moment, the voltage drops to 15.5V and remains at 25ms > 20ms, meeting the power-off determination conditions. The controller immediately determines that a power-off fault has occurred and triggers the electromagnetic brake action of the servo motor 1 through software logic.

[0105] Secondly, after the electromagnetic brake is released, the disc spring 3 in the compressed state actively pushes the piston rod 5 to extend by relying on the pre-tightening force,带动 the brake shoe to quickly clamp the track of the mine monorail crane until the deformation amount of the disc spring 3 reaches the limit position of the mechanical limit device. The electromagnetic brake refers to the power-off action mechanism supporting the servo motor 1, which loosens when powered on and releases when powered off, providing the movement condition for the energy storage braking of the disc spring 3. The pre-tightening force refers to the elastic acting force stored by the compression of the disc spring 3 in the braking state, which is the active power source for power-off emergency braking. The mechanical limit device is a mechanical stop block inside the braking actuator, used to limit the maximum extension position of the piston rod 5 and prevent the disc spring 3 from being over-compressed. The limit position refers to the maximum extension position of the piston rod 5 limited by the mechanical limit device, which is also the maximum safe deformation position of the disc spring 3.

[0106] Specifically, after the controller is triggered, the electromagnetic brake is de-energized and released, releasing the locking constraint on the motor shaft; during the braking process, the disc spring 3, which is in a compressed state, actively pushes the piston rod outward by relying on its stored preload; the piston rod 5 drives the brake shoes to move quickly, clamping the mine monorail and generating an emergency braking clamping force; the piston rod 5 continues to extend until the deformation of the disc spring 3 reaches the limit position of the mechanical limit device, and the brake remains locked.

[0107] For example, before the power failure, the system is in a braking state: the disc spring 3 is compressed to 8mm, storing a preload of 22kN, and the electromagnetic brake is energized and released. After the power failure is detected, the electromagnetic brake is de-energized and released, and the motor shaft is fully unlocked; the disc spring 3, through the preload, forcefully pushes the piston rod 5 forward, causing the brake shoes to quickly clamp the rail; the piston rod 5 continues to extend to 10mm, the disc spring 3 reaches its maximum safe deformation, the brake shoes maintain a constant clamping force, and the monorail comes to a complete stop, achieving emergency braking in the event of a power failure.

[0108] In this embodiment of the invention, the dual determination of voltage threshold and anti-shake time avoids false braking caused by instantaneous voltage fluctuations, thus improving the accuracy of fault diagnosis; emergency braking is achieved by relying on the energy storage of the butterfly spring 3, which does not require external power supply, has the highest reliability, and is suitable for extreme working conditions such as power outages in the mine; the electromagnetic brake acts immediately upon power failure, responds quickly, and the brake shoes quickly clamp the rail, effectively preventing slippage and sliding accidents; the mechanical limit device limits the deformation of the butterfly spring 3, avoids overload damage to the butterfly spring 3, and ensures the long-term reliability of the emergency braking mechanism.

[0109] Through the specific implementation methods described above, the embodiments of the present invention achieve the following technical effects: This invention provides an electric-spring series braking system and method for mine monorail cranes. Through real-time sensing of operating parameters and intelligent target braking force calculation, electric-spring series electromechanical actuation pressure build-up, position-pressure dual-variable collaborative closed-loop control, high-precision and fast-response dual-mode disc spring 3-nonlinear mapping, smooth and shock-free braking relief, and automatic emergency braking in the entire process, it solves the problems of traditional mine monorail crane braking systems, such as response lag, coarse braking force adjustment, control conflicts, susceptibility to impact oscillations, weak anti-interference ability, and power failure safety. This addresses shortcomings and other issues; it achieves accurate solution of braking force based on multiple working conditions, synchronous convergence of piston rod 5 position and disc spring 3 pressure, effectively suppressing pressure overshoot and oscillation during braking, and balancing control stability under normal working conditions with strong robustness under harsh working conditions; at the same time, it avoids mechanical rigid collisions by uniform speed reset, and achieves reliable emergency braking without power by relying on disc spring 3 energy storage, improving the control accuracy, response speed, working condition adaptability and intrinsic safety level of the entire braking system, and can stably adapt to complex transportation working conditions with heavy load, slope change, strong interference and high safety requirements in underground wells.

[0110] Example 2, as Figure 4 As shown, this invention provides an electric drive-spring series braking method for mine monorail cranes, the method comprising: S100: Collect the real-time speed, track gradient, and vehicle load of the mining monorail as real-time operating status parameters, input the real-time operating status parameters into the target braking force calculation unit built into the controller, and calculate the target braking force required at the moment.

[0111] Step S100 in the method provided in this embodiment of the invention includes: The real-time operating status parameters are input into the target braking force calculation unit built into the controller to calculate the currently required target braking force, including: The collected real-time vehicle speed, track gradient, and vehicle load are used as input parameters and input to the input layer node of the offline-trained feedforward neural network model. The output layer node of the feedforward neural network model outputs the target braking force. The feedforward neural network model is trained offline by collecting historical operating data and historical braking data of the mining monorail crane under various working conditions, including unloaded, fully loaded, uphill, and downhill conditions. The trained feedforward neural network model is then deployed to the target braking force calculation unit of the controller. During operation, the controller periodically compares the actual braking effect with the target braking force output by the feedforward neural network model. When the deviation between the actual braking distance and the theoretical braking distance exceeds a preset deviation threshold, the controller triggers an incremental update of the feedforward neural network model and uses a backpropagation algorithm to fine-tune the network weights.

[0112] S200: Generate a pulse width modulation drive signal according to the target braking force, drive the servo motor to convert the rotational motion into linear motion through the ball screw transmission mechanism, push the piston rod to extend and compress the series-connected disc spring, so that the disc spring generates elastic force and acts on the brake shoe, thereby establishing a braking clamping force.

[0113] Step S200 in the method provided in this embodiment of the invention includes: The process of generating a pulse width modulation drive signal based on the target braking force includes: Using the current, speed, and angle of the servo motor as state variables and the duty cycle of the pulse width modulation signal as the control variable, a dynamic response model of the motor in the prediction time domain is established. In each control cycle, the controller uses a quadratic programming algorithm to solve a finite-time domain optimization problem based on the motor dynamic response model, minimizing the deviation between the piston rod displacement at the end of the predicted time domain and the target compression. Specifically, when the difference between the target braking force and the current disc spring pressure is greater than a preset rapid pressure build-up threshold, it is determined that the current braking initiation stage is in place, and the preset maximum duty cycle is automatically output. When the current disc spring pressure is close to the target braking force, the controller automatically reduces the duty cycle of the pulse width modulation signal. The first optimal duty cycle obtained from the solution is used as the pulse width modulation drive signal for the current cycle.

[0114] S300: During braking, a coordinated closed-loop control strategy for position and pressure is executed based on the clamping force data, piston rod position data, and motor rotation angle data collected in real time by the controller. The coordinated closed-loop control strategy takes the target position of the piston rod and the target pressure of the spring as the control targets. By adjusting the duty cycle of the pulse width modulation signal in real time, the driving direction and output torque of the servo motor are coordinated to make the actual position of the piston rod and the actual pressure of the spring synchronously approach their respective target values ​​until the elastic force generated by the disc spring stabilizes within the preset deviation tolerance range of the target braking force.

[0115] Step S300 in the method provided in this embodiment of the invention includes: The coordinated closed-loop control strategy for execution position and pressure includes: In conventional braking scenarios, a sliding mode control algorithm is used to implement a coordinated closed-loop control strategy for position and pressure. The sliding mode control algorithm defines a coordinated sliding surface function, and the independent variables of the coordinated sliding surface function include piston rod position deviation and spring pressure deviation. The piston rod position deviation is the difference between the target compression and the actual compression, and the spring pressure deviation is the difference between the target pressure and the actual pressure. The sliding mode control algorithm integrates the piston rod position deviation and spring pressure deviation into a single cooperative error variable through a weighted summation method. The weight coefficients in the weighted summation are dynamically adjusted by the controller based on the currently collected track gradient and vehicle load, and the pulse width modulation correction signal is calculated with the cooperative error variable approaching zero as the control objective. The sliding mode control algorithm introduces a saturation function to replace the ideal sign function to eliminate the chattering phenomenon inherent in sliding mode control, and sets the boundary layer thickness of the saturation function to be positively correlated with the tangential stiffness coefficient of the disc spring under the current compression.

[0116] The coordinated closed-loop control strategy for execution position and pressure also includes: In severe braking scenarios, an active disturbance rejection control algorithm is adopted as an alternative to the sliding mode control algorithm. The active disturbance rejection control algorithm includes three functional units: a tracking differentiator, an extended state observer, and a nonlinear state error feedback law. The tracking differentiator ensures a smooth transition between the target position of the piston rod and the target pressure of the spring, avoiding abrupt changes in the target value that could impact the braking system. The expansion state observer estimates the total disturbance in the braking system in real time. The total disturbance includes the nonlinear stiffness characteristics of the disc spring, the time-varying characteristics of the friction coefficient between the track and the brake shoe, and the external impact load. The estimated total disturbance value is then compensated in the feedforward channel. The nonlinear state error feedback law generates a nonlinear combination of control outputs based on the error between the transition target value output by the tracking differentiator and the actual state estimate output by the extended state observer.

[0117] The method provided in this embodiment of the invention further includes: A nonlinear mapping model between the compression and elastic force of a butterfly spring is built into the controller. The nonlinear mapping model is used to reverse the target compression of the piston rod based on the target braking force, and the target compression is used as the input reference for the coordinated closed-loop control strategy of position and pressure. The nonlinear mapping model uses a radial basis function neural network to positively fit the compression-elastic force characteristic curve of the butterfly spring under different combinations. The different combinations include single-piece stacked combination, multi-piece paired combination, and a combination of stacked and paired combinations. In high-precision braking scenarios, the input layer of the radial basis function neural network is the compression of a butterfly spring, the hidden layer uses Gaussian radial basis functions to partition the input space into local responses, and the output layer is the elastic force of the butterfly spring. During the braking preparation phase, the controller calls the nonlinear mapping model, substitutes the target braking force as the desired output into the radial basis function neural network, and solves the compression of the disc spring in the input layer through an iterative optimization algorithm. The obtained compression is then used as the target compression of the piston rod.

[0118] The nonlinear mapping model also includes alternative solutions: In scenarios requiring rapid response, a piecewise analytical function is used as an alternative to the radial basis function neural network. The piecewise analytical function divides the compression-elasticity characteristic curve of the disc spring into three intervals: a linear segment, a hardening segment, and a near-flattening segment, based on the number of stacked discs, the number of pairs, and the geometric parameters of a single disc spring. The linear segment uses a linear function expression to fit the linear relationship between compression and elasticity, while the hardening segment uses a cubic function expression to fit the nonlinear increasing relationship between compression and elasticity. The controller directly calculates the target compression of the piston rod by selecting the inverse function analytical expression of the corresponding segment interval based on the segment interval where the target braking force is located, without the need for reverse solving through iterative optimization algorithms.

[0119] The method provided in this embodiment of the invention further includes a braking release step: When the controller receives the brake release command, it sends a reverse pulse width modulation drive signal to the servo motor, which drives the servo motor to rotate in the opposite direction. The piston rod is pulled back through the ball screw transmission mechanism, so that the disc spring is synchronously reset with the piston rod and releases its elastic potential energy. During the piston rod retraction process, the speed control mode is switched to use closed-loop speed control to keep the piston rod retraction speed at a preset constant speed value, thus avoiding rigid collision between the piston rod and the spring seat. The system receives piston rod position data from the displacement sensor in real time. When the piston rod position data indicates that the piston rod has returned to the initial mechanical origin, the power supply to the servo motor is cut off and the electromagnetic brake is triggered to lock the motor shaft.

[0120] The method provided in this embodiment of the invention further includes a power failure emergency braking step: When the controller detects that the power supply voltage is lower than the preset power failure threshold and the duration exceeds the preset anti-shake time, it determines that the system has experienced a power failure and immediately triggers the electromagnetic brake of the servo motor to release through software logic. After the electromagnetic brake is released, the compressed disc spring actively pushes the piston rod to extend by relying on the preload, which drives the brake shoe to quickly clamp the track of the mining monorail until the deformation of the disc spring reaches the limit position of the mechanical limit device.

[0121] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An electric drive-spring series braking system for mining monorail hoists, characterized in that, The system includes: The target braking force calculation module is used to collect the real-time speed, track gradient and vehicle load of the mining monorail as real-time operating status parameters. The real-time operating status parameters are input into the target braking force calculation unit built into the controller to calculate the target braking force required at the moment. An electromechanical spring brake actuator module is used to generate a pulse width modulation drive signal based on the target braking force, drive a servo motor to convert the rotary motion into linear motion via a ball screw transmission mechanism, push the piston rod to extend and compress the series-connected disc spring, so that the disc spring generates elastic force and acts on the brake shoe, thereby establishing a braking clamping force. The position-pressure coordinated closed-loop control module is used to execute a coordinated closed-loop control strategy for position and pressure during braking, based on the clamping force data, piston rod position data, and motor rotation angle data collected in real time by the controller. The coordinated closed-loop control strategy takes the target position of the piston rod and the target pressure of the spring as control targets. It adjusts the drive direction and output torque of the servo motor in real time by adjusting the duty cycle of the pulse width modulation signal, so that the actual position of the piston rod and the actual pressure of the spring synchronously approach their respective target values, until the elastic force generated by the disc spring stabilizes within the preset deviation tolerance range of the target braking force.

2. The electric drive-spring series braking system for mine monorail cranes according to claim 1, characterized in that, The real-time operating status parameters are input into the target braking force calculation unit built into the controller to calculate the currently required target braking force, including: The collected real-time vehicle speed, track gradient, and vehicle load are used as input parameters and input to the input layer node of the offline-trained feedforward neural network model. The output layer node of the feedforward neural network model outputs the target braking force. The feedforward neural network model is trained offline by collecting historical operating data and historical braking data of the mining monorail crane under various working conditions, including unloaded, fully loaded, uphill, and downhill conditions. The trained feedforward neural network model is then deployed to the target braking force calculation unit of the controller. During operation, the controller periodically compares the actual braking effect with the target braking force output by the feedforward neural network model. When the deviation between the actual braking distance and the theoretical braking distance exceeds a preset deviation threshold, the controller triggers an incremental update of the feedforward neural network model and uses a backpropagation algorithm to fine-tune the network weights.

3. The electric drive-spring series braking system for mine monorail cranes according to claim 1, characterized in that, Generating a pulse width modulation drive signal based on the target braking force includes: Using the current, speed, and angle of the servo motor as state variables and the duty cycle of the pulse width modulation signal as the control variable, a dynamic response model of the motor in the prediction time domain is established. In each control cycle, the controller uses a quadratic programming algorithm to solve a finite-time domain optimization problem based on the motor dynamic response model, minimizing the deviation between the piston rod displacement at the end of the predicted time domain and the target compression. Specifically, when the difference between the target braking force and the current disc spring pressure is greater than a preset rapid pressure build-up threshold, it is determined that the current braking initiation stage is in place, and the preset maximum duty cycle is automatically output. When the current disc spring pressure is close to the target braking force, the controller automatically reduces the duty cycle of the pulse width modulation signal. The first optimal duty cycle obtained from the solution is used as the pulse width modulation drive signal for the current cycle.

4. The electric drive-spring series braking system for mine monorail cranes according to claim 1, characterized in that, The strategy of coordinated closed-loop control of position and pressure includes: In conventional braking scenarios, a sliding mode control algorithm is used to implement a coordinated closed-loop control strategy for position and pressure. The sliding mode control algorithm defines a coordinated sliding surface function, and the independent variables of the coordinated sliding surface function include piston rod position deviation and spring pressure deviation. The piston rod position deviation is the difference between the target compression and the actual compression, and the spring pressure deviation is the difference between the target pressure and the actual pressure. The sliding mode control algorithm integrates the piston rod position deviation and spring pressure deviation into a single cooperative error variable through a weighted summation method. The weight coefficients in the weighted summation are dynamically adjusted by the controller based on the currently collected track gradient and vehicle load, and the pulse width modulation correction signal is calculated with the cooperative error variable approaching zero as the control objective. The sliding mode control algorithm introduces a saturation function to replace the ideal sign function to eliminate the chattering phenomenon inherent in sliding mode control, and sets the boundary layer thickness of the saturation function to be positively correlated with the tangential stiffness coefficient of the disc spring under the current compression.

5. The electric drive-spring series braking system for mining monorails according to claim 4, characterized in that, The coordinated closed-loop control strategy for execution position and pressure also includes: In severe braking scenarios, an active disturbance rejection control algorithm is adopted as an alternative to the sliding mode control algorithm. The active disturbance rejection control algorithm includes three functional units: a tracking differentiator, an extended state observer, and a nonlinear state error feedback law. The tracking differentiator ensures a smooth transition between the target position of the piston rod and the target pressure of the spring, avoiding abrupt changes in the target value that could impact the braking system. The expansion state observer estimates the total disturbance in the braking system in real time. The total disturbance includes the nonlinear stiffness characteristics of the disc spring, the time-varying characteristics of the friction coefficient between the track and the brake shoe, and the external impact load. The estimated total disturbance value is then compensated in the feedforward channel. The nonlinear state error feedback law generates a nonlinear combination of control outputs based on the error between the transition target value output by the tracking differentiator and the actual state estimate output by the extended state observer.

6. The electric drive-spring series braking system for mining monorail cranes according to claim 1, characterized in that, Also includes: A nonlinear mapping model between the compression and elastic force of a butterfly spring is built into the controller. The nonlinear mapping model is used to reverse the target compression of the piston rod based on the target braking force, and the target compression is used as the input reference for the coordinated closed-loop control strategy of position and pressure. The nonlinear mapping model uses a radial basis function neural network to positively fit the compression-elastic force characteristic curve of the butterfly spring under different combinations. The different combinations include single-piece stacked combination, multi-piece paired combination, and a combination of stacked and paired combinations. In high-precision braking scenarios, the input layer of the radial basis function neural network is the compression of a butterfly spring, the hidden layer uses Gaussian radial basis functions to partition the input space into local responses, and the output layer is the elastic force of the butterfly spring. During the braking preparation phase, the controller calls the nonlinear mapping model, substitutes the target braking force as the desired output into the radial basis function neural network, and solves the compression of the disc spring in the input layer through an iterative optimization algorithm. The obtained compression is then used as the target compression of the piston rod.

7. The electric drive-spring series braking system for mine monorail cranes according to claim 6, characterized in that, The nonlinear mapping model also includes alternatives: In scenarios requiring rapid response, a piecewise analytical function is used as an alternative to the radial basis function neural network. The piecewise analytical function divides the compression-elasticity characteristic curve of the disc spring into three intervals: a linear segment, a hardening segment, and a near-flattening segment, based on the number of stacked discs, the number of pairs, and the geometric parameters of a single disc spring. The linear segment uses a linear function expression to fit the linear relationship between compression and elasticity, while the hardening segment uses a cubic function expression to fit the nonlinear increasing relationship between compression and elasticity. The controller directly calculates the target compression of the piston rod by selecting the inverse function analytical expression of the corresponding segment interval based on the segment interval where the target braking force is located, without the need for reverse solving through iterative optimization algorithms.

8. The electric drive-spring series braking system for mining monorail cranes according to claim 1, characterized in that, It also includes brake release steps: When the controller receives the brake release command, it sends a reverse pulse width modulation drive signal to the servo motor, which drives the servo motor to rotate in the opposite direction. The piston rod is pulled back through the ball screw transmission mechanism, so that the disc spring is synchronously reset with the piston rod and releases its elastic potential energy. During the piston rod retraction process, the speed control mode is switched to use closed-loop speed control to keep the piston rod retraction speed at a preset constant speed value, thus avoiding rigid collision between the piston rod and the spring seat. The system receives piston rod position data from the displacement sensor in real time. When the piston rod position data indicates that the piston rod has returned to the initial mechanical origin, the power supply to the servo motor is cut off and the electromagnetic brake is triggered to lock the motor shaft.

9. The electric drive-spring series braking system for mine monorail cranes according to claim 1, characterized in that, It also includes emergency braking procedures in the event of a power outage: When the controller detects that the power supply voltage is lower than the preset power failure threshold and the duration exceeds the preset anti-shake time, it determines that the system has experienced a power failure and immediately triggers the electromagnetic brake of the servo motor to release through software logic. After the electromagnetic brake is released, the compressed disc spring actively pushes the piston rod to extend by relying on the preload, which drives the brake shoe to quickly clamp the track of the mining monorail until the deformation of the disc spring reaches the limit position of the mechanical limit device.

10. An electric drive-spring series braking method for mine monorail cranes, characterized in that, The method, applied to the electric drive-spring tandem braking system for mining monorails as described in any one of claims 1-9, comprises: The real-time speed, track gradient, and vehicle load of the mining monorail are collected as real-time operating status parameters. These real-time operating status parameters are then input into the target braking force calculation unit built into the controller to calculate the target braking force required at the moment. Based on the target braking force, a pulse width modulation drive signal is generated, which drives the servo motor to convert the rotational motion into linear motion through the ball screw transmission mechanism. This pushes the piston rod to extend and compress the series-connected disc spring, causing the disc spring to generate elastic force and act on the brake shoe, thereby establishing a braking clamping force. During braking, a coordinated closed-loop control strategy for position and pressure is executed based on the clamping force data, piston rod position data, and motor rotation angle data collected in real time by the controller. The coordinated closed-loop control strategy takes the target position of the piston rod and the target pressure of the spring as the control targets. By adjusting the duty cycle of the pulse width modulation signal in real time, the driving direction and output torque of the servo motor are coordinated to make the actual position of the piston rod and the actual pressure of the spring synchronously approach their respective target values ​​until the elastic force generated by the disc spring stabilizes within the preset deviation tolerance range of the target braking force.

Images