Adaptive coordinated control system and method for multiple permanent magnet synchronous motors of scraper conveyor
By employing three permanent magnet synchronous motors in a distributed drive system on the scraper conveyor, combined with dynamic load monitoring and intelligent prediction, adaptive and coordinated control of motor power is achieved, solving the problems of low efficiency and stability under traditional drive methods, and improving the operational stability and energy efficiency of the equipment.
Patent Information
- Application Number
- PCT/CN2025/128412
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-03
- Filing Date
- 2025-10-17
- Publication Date
- 2026-11-12
AI Technical Summary
Traditional scraper conveyors are inefficient, energy-intensive, and lack control precision. Their complex and variable load characteristics lead to uneven power distribution in the motor, affecting the stability of equipment operation.
The system employs a distributed drive system with three permanent magnet synchronous motors. Power is distributed in real time through bus communication. Combined with dynamic load monitoring, intelligent timing prediction, and adaptive power allocation, the system utilizes an improved adaptive Kalman filter and a fuzzy-PID controller to optimize power distribution and achieve closed-loop control of motor torque and speed.
It improves the operational stability and energy efficiency of scraper conveyors, adapts to complex and ever-changing underground coal mine conditions, extends equipment service life, and ensures stable operation in the event of sensor failure.
Smart Images

Figure CN2025128412_12112026_PF_FP_ABST
Abstract
Description
Adaptive Cooperative Control System and Method for Multiple Permanent Magnet Synchronous Motors in Scraper Conveyor Technical Field
[0001] This invention relates to the field of motor control technology, and in particular to an adaptive cooperative control system and method for multiple permanent magnet synchronous motors in a scraper conveyor. Background Technology
[0002] As a key piece of equipment in underground coal mining faces, scraper conveyors play a crucial role in transporting coal mined by the mining machine to the transport roadways. Traditional scraper conveyors mostly use asynchronous motors for drive; however, this drive method has significant drawbacks, such as low efficiency, high energy consumption, and insufficient control precision.
[0003] Permanent magnet synchronous motors (PMSMs) have become an ideal choice for scraper conveyor drive systems due to their high efficiency, high power density, and good speed regulation performance. However, the load characteristics of scraper conveyors are complex and variable, which often leads to uneven power distribution in the motors, seriously affecting the operating efficiency and stability of the equipment. Summary of the Invention
[0004] Purpose of the invention: The purpose of this invention is to provide an adaptive and coordinated control system and method for multiple permanent magnet synchronous motors in a scraper conveyor, so as to realize the coordinated control of the motor power of the scraper conveyor and improve the operational stability and energy efficiency of the scraper conveyor.
[0005] Technical solution: An adaptive and coordinated control system for multiple permanent magnet synchronous motors in a scraper conveyor, comprising a permanent magnet synchronous unit, a dynamic load monitoring unit, an intelligent timing prediction unit, and an adaptive power allocation unit;
[0006] The permanent magnet synchronous unit consists of three permanent magnet synchronous motors that are connected to the central controller via bus communication, respectively driving the head, middle and tail of the scraper conveyor;
[0007] The dynamic load monitoring unit collects the tension force in the head, middle, and tail areas of the scraper conveyor. Coal flow and the motor speed of each permanent magnet synchronous motor and torque The comprehensive load index L is generated by fusing i=1, 2, 3 using an improved adaptive Kalman filter. total The intelligent time-series prediction unit includes a constructed long short-term memory network, based on a comprehensive load index L. total It uses past load time-series data of a set duration as input to predict the load change sequence for a future period of time;
[0008] The adaptive power allocation unit adjusts the power distribution of each permanent magnet synchronous motor based on the load prediction deviation and the winding temperature of each permanent magnet synchronous motor through a fuzzy-PID controller, thereby controlling the final output torque and speed of each permanent magnet synchronous motor.
[0009] Furthermore, the dynamic load monitoring unit includes a pressure sensor, a Coriolis flow meter, and a motor status monitoring unit;
[0010] The pressure sensor is installed inside the hydraulic cylinder of the scraper chain tensioning device and is used to detect the tension force of the scraper chain. The Coriolis flow meter is installed at the discharge port of the scraper conveyor to detect the coal flow rate. The motor status monitoring unit monitors the motor speed in real time using a controller. Torque and winding temperature.
[0011] An adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor, employing any one of the aforementioned adaptive cooperative control systems for multiple permanent magnet synchronous motors in a scraper conveyor, includes the following steps:
[0012] S1, the dynamic load monitoring unit collects raw load data once per second, the raw load data including tension force. Coal flow rotational speed Torque The observation vector Z is composed of the original load data. k After preprocessing and multi-source signal fusion using an improved adaptive Kalman filter algorithm, the predicted regional load value L is generated. zone,i :
[0013]
[0014] Where, k j (t) represents the dynamic weights, j = 1, 2, 3, which are calculated and updated in real time using the entropy weight method;
[0015] Based on the regional load forecast value L zone,i Further generate the comprehensive load index L total ;
[0016]
[0017] S2, the long short-term memory network of the intelligent time series prediction unit inputs past load time series data of a preset duration and outputs a predicted load change sequence for a future period of time;
[0018] S3, the adaptive power matching unit adjusts power according to the load prediction deviation ΔL tThe temperature rise ΔT of each permanent magnet synchronous motor is used to calculate the power adjustment factor α using fuzzy control rules, and the target power P is generated. target ;
[0019] The power allocation calculation formula is as follows:
[0020]
[0021] For multi-permanent magnet motor drive systems, L zone,i This represents the regional load forecast value, based on the load forecast value L for each region. zone,i Dynamically allocate total power demand P total ;β i The motor health factor is determined by the motor temperature rise ΔT. i and efficiency η i Dynamic adjustment;
[0022]
[0023] in, The temperature rise attenuation factor is determined experimentally; e is the natural constant.
[0024] S4, the permanent magnet drive unit, based on the target power command, realizes closed-loop control of the torque and speed of each permanent magnet synchronous motor through the FOC controller.
[0025] Furthermore, in step S1, the improved adaptive Kalman filter algorithm uses the observed residual ∈ at the current time to... k Dynamically adjust the observation noise matrix R k The updated formula is:
[0026]
[0027] Where γ is the adaptive coefficient, R k-1 Let T represent the observation noise at the previous time step, and let T represent the matrix transpose.
[0028] Furthermore, in step S3, based on the fuzzy-PID controller, the input load prediction deviation ΔL is... t The temperature rise ΔT of each permanent magnet synchronous motor, the dynamic output power adjustment factor α, and thus the reference power P are generated. base :
[0029] P base =α·P rated Among them, P rated This refers to the rated power of the motor.
[0030] The PID controller is based on the reference power P base With actual power P actual The error was calculated as follows:
[0031] e(t) = P base -P actual ;
[0032] e(t) represents the instantaneous deviation between the reference power and the actual output power. A control signal is generated by defuzzifying multiple fuzzy control rules and using the centroid method to adjust the output power of the permanent magnet drive unit in real time. The target power calculation formula is as follows:
[0033]
[0034] Among them, K p K i K d These are the proportional, integral, and derivative gain coefficients, which were optimized through on-site debugging.
[0035] Furthermore, the load forecast deviation ΔL t =L predicted -L actual The universe of discourse is [-100%, +100%]; the universe of discourse for temperature rise ΔT is [0℃, 80℃].
[0036] Furthermore, under conditions of sudden load increase, the intelligent timing prediction unit predicts that the load will continue to increase in the future, and the adaptive power allocation unit increases the power adjustment factor α until the permanent magnet synchronous motor outputs maximum power; under conditions of sudden load decrease, the intelligent timing prediction unit predicts that the load will continue to decrease in the future, reduces the power adjustment factor α, and reduces the motor power output.
[0037] Furthermore, in step S3, the fuzzy control rules include:
[0038] If the load prediction deviation is "positive" and the motor temperature rise is "low", then the power adjustment factor is "large increase".
[0039] If the load prediction deviation is "positive" and the motor temperature rise is "medium", then the power adjustment factor is "medium increase".
[0040] If the load prediction deviation is "positive" and the motor temperature rise is "high", then the power adjustment factor is "small increase".
[0041] If the load prediction deviation is "moderate" and the motor temperature rise is "low", then the power adjustment factor is "medium increase".
[0042] If the load prediction deviation is "positive" and the motor temperature rise is "medium", then the power adjustment factor is "small increase".
[0043] If the load prediction deviation is "neutral" and the motor temperature rise is "high temperature", then the power adjustment factor is "maintain".
[0044] If the load prediction deviation is "zero" and the motor temperature rise is "low", then the power adjustment factor is "small increase".
[0045] If the load prediction deviation is "zero" and the motor temperature rise is "normal", then the power adjustment factor is "maintained".
[0046] If the load prediction deviation is "zero" and the motor temperature rise is "high", then the power adjustment factor is "small decrease".
[0047] If the load prediction deviation is "negative medium" and the motor temperature rise is "low temperature", then the power adjustment factor is "small decrease".
[0048] If the load prediction deviation is "negative medium" and the motor temperature rise is "medium temperature", then the power adjustment factor is "medium reduction".
[0049] If the load prediction deviation is "negative medium" and the motor temperature rise is "high temperature", then the power adjustment factor is "large reduction".
[0050] If the load prediction deviation is "negative large" and the motor temperature rise is "low low", then the power adjustment factor is "medium reduction".
[0051] If the load prediction deviation is "negative large" and the motor temperature rise is "medium temperature", then the power adjustment factor is "large decrease".
[0052] If the load prediction deviation is "large negative" and the motor temperature rise is "high temperature", then the power adjustment factor is "large decrease".
[0053] Furthermore, in step S4, the target power command P target Converted to direct-axis voltage V by the FOC controller d Cross-axis voltage V q The adjustment amount is used by the SVPWM module to generate a PWM signal to drive the inverter and control the motor output torque T. e With rotational speed ω.
[0054] Furthermore, it also includes an anomaly handling mechanism, which dynamically adjusts the anomaly trigger threshold based on sliding window statistics. The anomaly threshold for temperature rise ΔT is determined by adding three times the standard deviation to the mean value over the most recent hour; the comprehensive load index L... total The abnormal threshold is determined by adding twice the standard deviation to the mean of the most recent hour; when a sensor malfunctions, the system switches to digital twin virtual sensor mode, which uses the prediction value of the long short-term memory network and the physical model to jointly compensate for the missing sensor data.
[0055] Compared with the prior art, the significant advantages of this invention are as follows:
[0056] 1. The control system of this invention adopts three permanent magnet synchronous motors to drive the scraper conveyor in a distributed manner. Power is distributed in real time through bus communication, which optimizes the power concentration problem of traditional single motor drive. The explosion-proof design and fault tolerance mechanism of the whole system improve the applicability of the underground coal mine environment and ensure continuous and stable operation when the sensor fails.
[0057] 2. The control method of this invention integrates multiple types of sensors and an improved adaptive Kalman filter algorithm to enhance load monitoring capabilities under complex operating conditions;
[0058] 3. The control method of the present invention is based on the prediction of load trends by neural networks and combined with fuzzy control to dynamically adjust the power distribution, quickly adapt to sudden load changes and reduce motor losses; it integrates neural networks and permanent magnet drive technology to realize coordinated control of motor power of scraper conveyor, improve the operation stability and energy efficiency of scraper conveyor, extend the service life of equipment, and adapt to the complex and ever-changing working environment in coal mines. Attached Figure Description
[0059] Figure 1 is a structural block diagram of the system of the present invention;
[0060] Figure 2 is a flowchart of the dynamic load monitoring unit of the present invention;
[0061] Figure 3 is a structural diagram of the Long Short-Term Memory (LSTM) network of the present invention;
[0062] Figure 4 is a schematic diagram of the fuzzy-PID controller of the present invention. Detailed Implementation
[0063] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0064] The adaptive cooperative control system for multiple permanent magnet synchronous motors of a scraper conveyor provided in a typical embodiment of the present invention, as shown in Figure 1, includes a permanent magnet synchronous unit, a dynamic load monitoring unit, an intelligent timing prediction unit, and an adaptive power allocation unit.
[0065] The permanent magnet synchronous unit consists of three permanent magnet synchronous motors installed at the head, middle, and tail of the scraper conveyor, driving the conveyor at their respective points. Each of the three motors is connected to the central controller via a bus communication system, transmitting real-time data such as speed, torque, and temperature, and enabling coordinated power distribution.
[0066] The permanent magnet synchronous motor described in this embodiment is an explosion-proof permanent magnet synchronous motor (PMSM) with a built-in field-oriented controller (FOC), abbreviated as FOC controller. The FOC controller generates PWM signals through a space vector pulse width modulation (SVPWM) algorithm to drive the inverter to output three-phase voltage, thereby controlling the torque and speed of the permanent magnet synchronous motor.
[0067] The FOC controller is based on a three-loop control structure, including a current loop, a speed loop, and a position loop.
[0068] The current loop uses Clarke transformation to convert the three-phase currents into i in the stationary coordinate system. α i β Then, through Park transform, it is decomposed into direct-axis components i d and cross-axis component i q This achieves decoupled control. The speed loop adjusts the quadrature-axis current reference value i based on the deviation between the target speed and the feedback speed. q_ref The position loop generates a velocity reference value ω based on the deviation between the target position and the feedback position. ref .
[0069] As shown in Figure 2, the dynamic load monitoring unit monitors loads based on a pressure sensor, a Coriolis flow meter, and a motor status monitoring unit. The scraper conveyor is divided into three areas: head, middle, and tail. The dynamic load monitoring unit collects the raw load data, i.e., tension, from each of these areas. Coal flow motor speed and torque The original load data (i=1,2,3) are fused using an improved adaptive Kalman filter to generate a comprehensive load index L. total .
[0070] The pressure sensor is installed inside the hydraulic cylinder of the scraper conveyor tensioning device to detect the tension force of the scraper conveyor. For example, an explosion-proof pressure sensor with a range of 0-10MPa and an accuracy of ±0.5% is selected.
[0071] The Coriolis flow meter is installed at the discharge port of the scraper conveyor to detect the coal flow rate. A Coriolis flow meter with an accuracy of ±1.0% is preferred.
[0072] The motor condition monitoring unit monitors the motor speed in real time with the help of the controller. Torque The motor condition monitoring unit includes a Hall sensor and a temperature sensor to monitor winding temperature.
[0073] Hall sensors are installed at the drive end of each permanent magnet synchronous motor to collect speed data. and torque An explosion-proof temperature sensor (PT100, range -20℃ to 150℃) is installed near the windings of the permanent magnet synchronous motor to monitor ΔT in real time. i .
[0074] The intelligent time-series prediction unit includes a constructed Long Short-Term Memory (LSTM) network, based on a comprehensive load index L. total It uses past load time series data of a set duration as input to predict the load change sequence for a future period.
[0075] Long Short-Term Memory (LSTM) networks consist of an input layer, hidden layers, and an output layer. The input layer receives timing data (tension) of a preset duration. Coal flow motor speed and torque The output layer outputs a predicted load sequence for future durations. The hidden layers consist of a three-layer LSTM network, each containing a suitable number of hidden units. The training data covers a variety of typical operating conditions, and the model is trained using the Adam optimizer and a weighted mean squared error (W-MSE) loss function to ensure that the prediction error meets the requirements of engineering applications.
[0076] The adaptive power allocation unit adjusts the power distribution of each permanent magnet synchronous motor based on the load prediction deviation and the winding temperature of each permanent magnet synchronous motor through a fuzzy-PID controller, thereby controlling the final output torque and speed of each permanent magnet synchronous motor.
[0077] Another typical embodiment of the present invention provides an adaptive cooperative control method for multiple permanent magnet synchronous motors of a scraper conveyor, which is implemented based on the adaptive cooperative control system for multiple permanent magnet synchronous motors of a scraper conveyor described above.
[0078] The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor provided in this embodiment mainly includes the following steps:
[0079] Step 1: Data Acquisition and Fusion;
[0080] The dynamic load monitoring unit collects raw load data once per second, including tension. Coal flow motor speed and torque The observation vector Z is composed of the original load data. k After preprocessing and multi-source signal fusion using an improved adaptive Kalman filter algorithm, the predicted regional load value L is generated. zone,i ,
[0081]
[0082] Where, k j (t) represents the dynamic weights, j = 1, 2, 3, which are calculated and updated in real time using the entropy weight method;
[0083] Based on the regional load forecast value Lzone,i Further generate the comprehensive load index L total ;
[0084]
[0085] The improved adaptive Kalman filter algorithm uses the observed residuals at the current time step ∈ k Dynamically adjust the observation noise matrix R k The updated formula is:
[0086]
[0087] Where γ is the adaptive coefficient, R k-1 Let T represent the observation noise at the previous time step, and let T represent the matrix transpose.
[0088] Step 2, load timing prediction;
[0089] As shown in Figure 3, the Long Short-Term Memory (LSTM) network of the intelligent time series prediction unit takes into account past load time series data of a preset duration and outputs a predicted load change sequence for a future period of time.
[0090] Based on the comprehensive load index L generated in step one total Based on historical data, the intelligent time-series prediction unit predicts future load changes using a Long Short-Term Memory (LSTM) network, as follows:
[0091] (1) Input and output definitions:
[0092] Input: L from the past 60 seconds total Time series data {L total (t-59),…,L total (t)};
[0093] Output: Predicted sequence for the next 30 seconds
[0094] (2) LSTM network configuration:
[0095] The network has a three-layer structure with 128 hidden units in each layer. It captures temporal dependencies through gating mechanisms (input gate, forget gate, output gate).
[0096] The training data covers no-load, light-load, and heavy-load operating conditions, and is divided into training set, validation set, and test set in a 7:2:1 ratio.
[0097] (3) Training and optimization:
[0098] Loss function: Weighted mean squared error (W-MSE), with the weight increased by 50% during periods of sudden load changes;
[0099] Optimizer: Adam (learning rate 0.001), using early stopping to prevent overfitting.
[0100] (4) Online update and prediction applications:
[0101] The model parameters are updated incrementally every 24 hours to adapt to changes in downhole working conditions.
[0102] Predicted value Used to calculate load deviation Dynamic power adjustment for driving step three.
[0103] Step 3, Dynamic Power Adjustment;
[0104] In this step, based on the load prediction value and motor temperature, the input is fuzzified by the adaptive power allocation unit to obtain the fuzzy output of the power adjustment amount. After defuzzification, the fuzzy power adjustment amount is converted into a specific control signal. The fuzzy-PID controller adjusts the power distribution according to the load prediction deviation and the temperature rise of the permanent magnet synchronous motor.
[0105] The input to the fuzzy-PID controller is:
[0106] 1) Load forecast deviation ΔL t ΔL t =L predicd -L actual , Domain of discourse [-100%, +100%].
[0107] 2) Temperature rise ΔT is the increase in temperature of each permanent magnet synchronous motor winding relative to the ambient temperature or initial temperature, with a domain of [0℃, 80℃].
[0108] As shown in Figure 4, dynamic power adjustment is achieved based on a fuzzy-PID controller, with the input load prediction deviation ΔL. t The temperature rise ΔT, the dynamic output power adjustment factor α, and thus the reference power P are generated. base and generate the target power command P target .
[0109] P base =α·P rateb ;
[0110] Among them, P rated This refers to the rated power of the motor.
[0111] PID is based on reference power P base With actual power P actual The error was calculated as follows:
[0112] e(t) = P base -P actual ;
[0113] e(t) represents the instantaneous deviation between the reference power and the actual output power. A control signal is generated through multiple fuzzy control rule bases and centroid defuzzification, adjusting the output power of the permanent magnet drive unit in real time. The target power P... target The calculation formula is:
[0114]
[0115] Among them, K p K i K d These are the proportional, integral, and derivative gain coefficients, which were optimized through on-site debugging.
[0116] Under conditions of sudden load increase, the intelligent timing prediction unit predicts that the load will continue to increase in the future, and the adaptive power allocation unit increases the power adjustment factor α until the permanent magnet synchronous motor outputs maximum power; under conditions of sudden load decrease, the intelligent timing prediction unit predicts that the load will continue to decrease in the future, and decreases the power adjustment factor α to reduce the motor power output.
[0117] The fuzzy sets and fuzzy control rules involved in fuzzification are described below:
[0118] ΔL t The fuzzy set is: {Negative Large (NB), Negative Medium (NM), Zero (ZO), Positive Medium (PM), Positive Large (PB)};
[0119] The fuzzy set of ΔT is: {low temperature (LT), medium temperature (HM), high temperature (HT)};
[0120] The fuzzy set of power adjustment factor α is: {large decrease (DB), medium decrease (DM), small decrease (DS), maintain (NC), small increase (US), medium increase (UM), large increase (UB)}.
[0121] The fuzzy control rule base contains 15 fuzzy control rules, employing the Mamdani inference method and centroid defuzzification. The fuzzy control rules include:
[0122] If the load prediction deviation is "positive" and the motor temperature rise is "low", then the power adjustment factor is "large increase".
[0123] If the load prediction deviation is "positive" and the motor temperature rise is "medium", then the power adjustment factor is "medium increase".
[0124] If the load prediction deviation is "positive" and the motor temperature rise is "high", then the power adjustment factor is "small increase".
[0125] If the load prediction deviation is "moderate" and the motor temperature rise is "low", then the power adjustment factor is "medium increase".
[0126] If the load prediction deviation is "positive" and the motor temperature rise is "medium", then the power adjustment factor is "small increase".
[0127] If the load prediction deviation is "neutral" and the motor temperature rise is "high temperature", then the power adjustment factor is "maintain".
[0128] If the load prediction deviation is "zero" and the motor temperature rise is "low", then the power adjustment factor is "small increase".
[0129] If the load prediction deviation is "zero" and the motor temperature rise is "normal", then the power adjustment factor is "maintained".
[0130] If the load prediction deviation is "zero" and the motor temperature rise is "high", then the power adjustment factor is "small decrease".
[0131] If the load prediction deviation is "negative medium" and the motor temperature rise is "low temperature", then the power adjustment factor is "small decrease".
[0132] If the load prediction deviation is "negative medium" and the motor temperature rise is "medium temperature", then the power adjustment factor is "medium reduction".
[0133] If the load prediction deviation is "negative medium" and the motor temperature rise is "high temperature", then the power adjustment factor is "large reduction".
[0134] If the load prediction deviation is "negative large" and the motor temperature rise is "low low", then the power adjustment factor is "medium reduction".
[0135] If the load prediction deviation is "negative large" and the motor temperature rise is "medium temperature", then the power adjustment factor is "large decrease".
[0136] If the load prediction deviation is "large negative" and the motor temperature rise is "high temperature", then the power adjustment factor is "large decrease".
[0137] Based on this, the power allocation calculation formula is as follows:
[0138]
[0139] For multi-permanent magnet motor drive systems, L zone,i This represents the regional load forecast value, based on the load forecast value L for each region. zone,i Dynamically allocate total power demand P total ;β i The motor health factor is determined by the motor temperature rise ΔT. i and efficiency η i Dynamic adjustment:
[0140]
[0141] in, is the temperature rise attenuation factor, which was calibrated experimentally; e is the natural constant.
[0142] Step four, closed-loop control execution;
[0143] The permanent magnet drive unit is based on the target power command P target Converted to direct-axis voltage command V by the FOC controller d Cross-axis voltage command V q The SVPWM module generates a PWM signal to drive the inverter and adjust the motor torque T. e With rotational speed ω.
[0144] Furthermore, the adaptive cooperative control method for multiple permanent magnet synchronous motors of scraper conveyors described in this invention also includes an anomaly handling mechanism, specifically a dynamic threshold protection mechanism and sensor fault tolerance.
[0145] Specifically, the dynamic threshold protection mechanism dynamically adjusts the abnormal trigger threshold based on sliding window statistics. The abnormal threshold for temperature rise ΔT is determined by adding three times the standard deviation to the mean value over the most recent hour, i.e.: μ T +3σ T μ T σ is the average value over the most recent hour. T The standard deviation is denoted as L. The overall load index is L. total The anomaly threshold is determined by adding twice the standard deviation to the mean of the most recent hour, i.e.: μ L +2σ L μ L L indicates the last hour total The mean of σ reflects the baseline level of the load; L For the corresponding time period L total The standard deviation represents the load fluctuation range; exceeding the threshold triggers load reduction protection.
[0146] The sensor fault tolerance mechanism is that when a sensor fails, the system switches to a digital twin virtual sensor mode, which uses the prediction value of the Long Short-Term Memory Network (LSTM) and the physical model (such as the thermal model of the motor) to jointly compensate for the missing sensor data, thus ensuring control continuity.
Claims
1. A kind of scraper conveyor multi-permanent magnet synchronous motor adaptive collaborative control system, it is characterized in that, It includes a permanent magnet synchronous unit, a dynamic load monitoring unit, an intelligent timing prediction unit, and an adaptive power allocation unit; The permanent magnet synchronous unit consists of three permanent magnet synchronous motors that are connected to the central controller via bus communication, respectively driving the head, middle and tail of the scraper conveyor; The dynamic load monitoring unit collects the tension of each area of the head, middle and tail of the scraper conveyor Coal flow rate motor speed of each permanent magnet synchronous motor and torque i = 1, 2, 3, the improved adaptive Kalman filter fusion generates a comprehensive load index L total ; the intelligent time sequence prediction unit includes a constructed long short-term memory network, and predicts a load change sequence in a future period of time based on the comprehensive load index L total with load time sequence data in a past set time length as input The adaptive power allocation unit adjusts the power distribution of each permanent magnet synchronous motor based on the load prediction deviation and the winding temperature of each permanent magnet synchronous motor through a fuzzy-PID controller, thereby controlling the final output torque and speed of each permanent magnet synchronous motor.
2. The self-adaptive cooperative control system of the scraper conveyor multi-permanent magnet synchronous machines according to claim 1, characterized in that, The dynamic load monitoring unit includes a pressure sensor, a Coriolis flow meter, and a motor status monitoring unit. The pressure sensor is installed in a hydraulic cylinder of a scraper chain tensioning device, and is used for detecting the tension of the scraper chain The Coriolis force flowmeter is installed at the discharge port of a scraper conveyor for detecting the coal flow The motor state monitoring unit monitors motor speed in real time with the aid of a controller torque and winding temperature.
3. A method for adaptive cooperative control of a plurality of permanent magnet synchronous machines of a flight conveyor, characterized in that The adaptive cooperative control system for multiple permanent magnet synchronous motors of scraper conveyors as described in any one of claims 1-2 includes the following steps: S1, the dynamic load monitoring unit collects raw load data every second, the raw load data including tension force Coal flow rate RPM torque The observation vector Z is composed of the original load data k , and the regional load prediction value L is generated by preprocessing and multi-source signal fusion using an improved adaptive Kalman filtering algorithm zone,i : wherein, k j (t) is a dynamic weight, j = 1, 2, 3, which is updated in real time by the entropy weight method; According to the regional load prediction value L zone,i , a comprehensive load index L total is further generated; S2, the long short-term memory network of the intelligent time series prediction unit inputs past load time series data of a preset duration and outputs a predicted load change sequence for a future period of time; S3, the adaptive power allocation unit calculates a power adjustment factor a according to the load prediction deviation AL t and the temperature rise AT of each permanent magnet synchronous motor, and generates a target power P by a fuzzy control rule target ; The power allocation calculation formula is as follows: For multi-permanent magnet motor drive systems, L zone,i This represents the regional load forecast value, based on the load forecast value L for each region. zone,i Dynamically allocate total power demand P total ;β i The motor health factor is determined by the motor temperature rise ΔT. i and efficiency η i Dynamic adjustment; in, The temperature rise attenuation factor is determined experimentally; e is the natural constant. S4, the permanent magnet drive unit, based on the target power command, realizes closed-loop control of the torque and speed of each permanent magnet synchronous motor through the FOC controller.
4. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 3, characterized in that, In step S1, the improved adaptive Kalman filter algorithm uses the observed residual ∈ at the current time step... k Dynamically adjust the observation noise matrix R k The updated formula is: where γ is an adaptive coefficient, R k-1 denotes the observation noise at the previous time, and T denotes the matrix transpose.
5. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 3, characterized in that, In step S3, a load prediction deviation ΔL is inputted to the fuzzy-PID controller t and the temperature rise ΔT of each permanent magnet synchronous motor, a dynamic output power adjustment factor α is outputted, and a reference power P is generated base base = α · P rated ; P = Pnom rated P = Pnom P = Pnom The PID controller calculates an error between the reference power P base and the actual power P actual and calculates: e(t) = P base - P actual ; e(t) represents the instantaneous deviation between the reference power and the actual output power. A control signal is generated by defuzzifying multiple fuzzy control rules and using the centroid method to adjust the output power of the permanent magnet drive unit in real time. The target power calculation formula is as follows: where Kp, Ki, and Kd are proportional, integral, and derivative gain coefficients, respectively. p , Kp, Ki, and Kd are proportional, integral, and derivative gain coefficients, respectively. i , Kp, Ki, and Kd are proportional, integral, and derivative gain coefficients, respectively. d , K 6. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 5, characterized in that, Load prediction deviation AL t = L predicted - L actual , domain [-100%, +100%]; domain of the temperature rise AT [0°C, 80°C].
7. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 5, characterized in that, Under conditions of sudden load increase, the intelligent timing prediction unit predicts that the load will continue to increase in the future, and the adaptive power allocation unit increases the power adjustment factor α until the permanent magnet synchronous motor outputs maximum power; under conditions of sudden load decrease, the intelligent timing prediction unit predicts that the load will continue to decrease in the future, and decreases the power adjustment factor α to reduce the motor power output.
8. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 7, characterized in that, In step S3, the fuzzy control rules include: If the load prediction deviation is "positive" and the motor temperature rise is "low", then the power adjustment factor is "large increase". If the load prediction deviation is "positive" and the motor temperature rise is "medium", then the power adjustment factor is "medium increase". If the load prediction deviation is "positive" and the motor temperature rise is "high", then the power adjustment factor is "small increase". If the load prediction deviation is "moderate" and the motor temperature rise is "low", then the power adjustment factor is "medium increase". If the load prediction deviation is "positive" and the motor temperature rise is "medium", then the power adjustment factor is "small increase". If the load prediction deviation is "neutral" and the motor temperature rise is "high temperature", then the power adjustment factor is "maintain". If the load prediction deviation is "zero" and the motor temperature rise is "low", then the power adjustment factor is "small increase". If the load prediction deviation is "zero" and the motor temperature rise is "normal", then the power adjustment factor is "maintained". If the load prediction deviation is "zero" and the motor temperature rise is "high", then the power adjustment factor is "small decrease". If the load prediction deviation is "negative medium" and the motor temperature rise is "low temperature", then the power adjustment factor is "small decrease". If the load prediction deviation is "negative medium" and the motor temperature rise is "medium temperature", then the power adjustment factor is "medium reduction". If the load prediction deviation is "negative medium" and the motor temperature rise is "high temperature", then the power adjustment factor is "large reduction". If the load prediction deviation is "negative large" and the motor temperature rise is "low low", then the power adjustment factor is "medium reduction". If the load prediction deviation is "negative large" and the motor temperature rise is "medium temperature", then the power adjustment factor is "large decrease". If the load prediction deviation is "large negative" and the motor temperature rise is "high temperature", then the power adjustment factor is "large decrease".
9. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 5, characterized in that, In step S4, the target power instruction P target converted into a direct-axis voltage V d , a quadrature-axis voltage V q The adjustment amount is generated by the SVPWM module to generate a PWM signal to drive the inverter, control the motor output torque T e and the rotational speed ω.
10. The adaptive cooperative control method for multiple permanent magnet synchronous motors in a scraper conveyor according to claim 9, characterized in that, An abnormality processing mechanism is also included, which dynamically adjusts the abnormal triggering threshold based on sliding window statistics, wherein the abnormal threshold of temperature rise ΔT is determined by the mean value plus three times the standard deviation within the last one hour; the abnormal threshold of comprehensive load index L total is determined by the mean value plus two times the standard deviation within the last one hour; when the sensor fails, the system switches to a digital twin virtual sensor mode, and the missing sensor data is compensated by the predicted value of the long short-term memory network combined with the physical model.
