A precise parking method for electric locomotive in coal mine

Abstract

The invention discloses a precise parking method for an electric locomotive in an underground coal mine. A model of the braking torque of the electric locomotive is established and combined with a double closed-loop vector control model. The model of the braking torque of the electric locomotive and the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive collected by the double closed-loop vector control model are used as sample sets to train a single hidden layer feedforward neural network model. The trained single hidden layer feedforward neural network model is introduced into a double closed-loop vector regulation system of a permanent magnet synchronous motor of the electric locomotive. The double closed-loop vector regulation system controls the permanent magnet synchronous motor of the electric locomotive to generate a braking torque, so as to achieve precise parking of the electric locomotive, reduce parking errors, and improve production efficiency.

Description

Technical Field

[0001] The invention relates to a precise parking method for an electric locomotive, in particular to a precise parking method for an electric locomotive in an underground coal mine. Background Art

[0002] Electric locomotives are not only an important part of the auxiliary transportation vehicle system of coal mines, but also an important means to assist the normal production of mining equipment, reduce maintenance time, labor hours, and improve production efficiency. The use of electric locomotives in auxiliary transportation can effectively improve production efficiency.

[0003] As one of the key technologies of vehicle-track coordination, the precise parking technology of electric locomotives can realize the precise parking of electric locomotives autonomously. For electric locomotives, which are rail vehicles, precise parking is of great significance: the precise parking of electric locomotives in coal mines is a requirement for intelligent and unmanned transportation in coal mines; in order to ensure the safety of intelligent auxiliary transportation systems in coal mines, electric locomotives need to realize precise parking functions.

[0004] The precision parking methods in the prior art are mostly applied to ground vehicles, whose working conditions are less complex than those in coal mines, and the parking precision requirements are lower. In addition, the parking methods in the prior art rely on the driver's operation. For vehicles with large mass such as electric locomotives, the precise parking process will be a task with high technical requirements for the driver. The precise parking process is difficult, which will increase the driver's fatigue and driving danger. Inaccurate parking of electric locomotives will increase the difficulty of loading and unloading goods. Summary of the invention

[0005] The purpose of the present invention is to provide a method for accurately parking an electric locomotive in an underground coal mine.

[0006] Technical solution: The present invention provides a precise parking method for an electric locomotive in a coal mine. The electric locomotive includes a permanent magnet synchronous motor. When the electric locomotive receives a parking instruction, its braking process includes the following steps:

[0007] S1. Establish a model of the braking torque of the electric locomotive according to the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive;

[0008] S2. Establish a double closed-loop vector control model, combined with the model of the braking torque of the electric locomotive, and obtain the calculated value of the braking torque of the electric locomotive according to the dead weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive. The double closed-loop vector control model controls the permanent magnet synchronous motor of the electric locomotive to generate the braking torque, and adds the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive as samples to the sample set;

[0009] S3. Based on the whole sample set obtained in S2, a single hidden layer feedforward neural network model is trained, wherein the real-time load of the electric locomotive, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the real-time track slope are input, and the calculated value of the braking torque of the electric locomotive is output, and a single hidden layer feedforward neural network model that outputs the actual braking torque of the electric locomotive is obtained after training;

[0010] S4. Importing the trained single hidden layer feedforward neural network model into the dual closed-loop vector control system of the permanent magnet synchronous motor of the electric locomotive;

[0011] S5. The double closed-loop vector control system controls the permanent magnet synchronous motor of the electric locomotive to generate actual braking torque, and the actual braking torque is used to realize the parking of the electric locomotive.

[0012] As a preferred technical solution of the present invention: the establishment of the model of the real-time braking torque of the electric locomotive in step S1 includes the following steps:

[0013] S11. Establish the real-time ground braking force F and real-time inertial resistance W based on the whole electric locomotive a , basic resistance W0, real-time ramp resistance W i The equilibrium equation is: i When braking uphill, use -, and when braking downhill, use +;

[0014] F=W a -W0±W i (1)

[0015] S12. Establish real-time inertial resistance W a , basic resistance W0, real-time ramp resistance W i The model includes equations (2), (3), and (4);

[0016] W a =1000(P+Q)(1+γ)a0 (2)

[0017] W0=1000(P+Q)gω (3)

[0018] W i =1000(P+Q)g i (4)

[0019] Where P is the real-time load of the electric locomotive, Q is the deadweight of the electric locomotive, γ is the inertia coefficient, a0 is the real-time braking deceleration of the electric locomotive, ω is the running resistance coefficient of the electric locomotive, i is the track slope, and g is the acceleration of gravity;

[0020] S13. Arrange equations (1), (2), (3), and (4) to obtain a calculation formula for the real-time ground braking force F of the electric locomotive when the electric locomotive brakes:

[0021] F=1000(P+Q)[(1+γ)a0-ωg±ig] (5)

[0022] In the formula, the real-time track slope i is + when braking uphill, and - when braking downhill;

[0023] S14. According to the torque formula T = FL, where T is the braking torque of the electric locomotive and L is the length of the lever arm, convert equation (5) into:

[0024]

[0025] S15. According to the formula Where S is the real-time distance between the electric locomotive and the destination, v is the real-time speed of the electric locomotive, and equation (6) is converted into:

[0026]

[0027] According to formula (7), a model is established with the weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the braking torque of the electric locomotive as variables.

[0028] The dual closed-loop vector control model in step S2 is a vector control model based on a dq rotating coordinate system and combined with a PID control system. The PID control system includes a speed PI controller, a current PI controller 1, a current PI controller 2, a coordinate system transformation module I_PARK, a coordinate system transformation module CLARK, a coordinate system transformation module PARK, a space vector modulation module, a three-phase inverter, and a voltage-current converter. The control steps of the dual closed-loop vector control model are as follows:

[0029] S21: According to the formula v=2πrn, where r is the rotor radius of the permanent magnet synchronous motor of the electric locomotive, and n is the real-time rotor speed of the permanent magnet synchronous motor of the electric locomotive, the formula (7) is converted into:

[0030]

[0031] S22: Input the given speed N and the real-time speed n of the permanent magnet synchronous motor rotor into the speed PI controller to obtain the stator current I along the q axis after being adjusted by the speed PI controller. q and the stator current I along the d axis d ;

[0032] S23: I q The real-time current i of the stator along the q axis qInput current PI controller 1 to obtain the stator voltage value u along the q axis after being adjusted by current PI controller 1 q ;

[0033] Will I d The real-time current i of the stator along the d axis d Input current PI controller 2 to obtain the stator voltage value u along the d axis after being adjusted by current PI controller 2 d ;

[0034] S24: will u q 、u d Input the coordinate transformation module I_PARK respectively, and the coordinate transformation module I_PARK outputs u q Voltage value u in αβ coordinate system α 、u d Voltage value u in αβ coordinate system β ;

[0035] S25: Will u α 、u β An input space vector modulation module is input, and the space vector modulation module outputs a trigger signal to a three-phase inverter;

[0036] S26: The three-phase inverter receives the trigger signal and outputs a three-phase voltage u a 、u b 、u c , the three-phase voltage controls the permanent magnet synchronous motor to produce braking torque;

[0037] S27: three-phase voltage u output by the three-phase inverter a 、u b 、u c Input voltage and current conversion module, voltage and current conversion module output three-phase current i a 、i b 、i c , three-phase current i a 、i b 、i c Input coordinate system transformation module CLARK, coordinate system transformation module CLARK outputs the current i of the stator along the α axis in the αβ coordinate system α and the current γ along the β axis β ,i α and i β Coordinate system transformation module PARK, the coordinate system transformation module PARK outputs the current i of the stator along the d-axis in the dq coordinate system d and the current i along the q axis q ,i q Input current PI controller 1, i d The input current PI controller 2 makes the PID control system form a closed loop.

[0038] As a preferred technical solution of the present invention: the precise parking method for an electric locomotive in an underground coal mine also includes step S6: during the braking process of the electric locomotive, when the real-time distance S between the electric locomotive and the end point is less than the preset value S0, if the real-time speed v of the electric locomotive is greater than the preset value v0, the electric locomotive starts mechanical braking.

[0039] As a preferred technical solution of the present invention: the permanent magnet synchronous motor of the electric locomotive is a surface permanent magnet synchronous motor.

[0040] As a preferred technical solution of the present invention: the single hidden layer feedforward neural network model is an extreme learning machine.

[0041] As a preferred technical solution of the present invention: if the calculated values ​​of the electric locomotive's own weight, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the braking torque of the electric locomotive obtained in steps S1 and S2 are the same as the samples already existing in the sample set, they will not be added to the sample set.

[0042] A system based on a precise parking method for electric locomotives in coal mines, comprising a communication module, a braking torque model module, a double closed-loop vector control model module, a correction module, a single hidden layer feedforward neural network model module, and a feedback control module;

[0043] A communication module, used to obtain a stop instruction of the electric locomotive;

[0044] The braking torque model module is used to obtain the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive, and establish a model of the braking torque of the electric locomotive;

[0045] The double closed-loop vector control model module is used to obtain the calculated value of the braking torque of the electric locomotive according to the dead weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive. The double closed-loop vector control model module controls the permanent magnet synchronous motor of the electric locomotive to generate braking torque;

[0046] A correction module is used to perform error analysis and correction on the braking torque model module and the double closed-loop vector control model module, and to construct a sample set with the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive as samples;

[0047] A single hidden layer feedforward neural network model module is used to train the single hidden layer feedforward neural network model based on the sample set of the correction module to obtain a single hidden layer feedforward neural network model that outputs the actual braking torque of the electric locomotive;

[0048] The feedback control module is used to import the trained single hidden layer feedforward neural network model into the double closed-loop vector regulation system of the permanent magnet synchronous motor of the electric locomotive. The double closed-loop vector regulation system controls the permanent magnet synchronous motor of the electric locomotive to generate actual braking torque to realize the parking of the electric locomotive.

[0049] An electronic device comprises a storage device, one or more processors, and the storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement a precise parking method based on an underground electric locomotive in a coal mine as described in any one of claims 1-7.

[0050] A computer-readable storage medium storing a computer program therein, characterized in that when the computer program is executed by a processor, a precise parking method based on an underground electric locomotive in a coal mine is implemented as described in any one of claims 1-7.

[0051] Beneficial effects: Compared with the prior art, the advantages of the present invention include: the electric locomotive can autonomously identify the working conditions underground in the coal mine when receiving a stop command, achieve precise parking under complex working conditions, reduce parking errors, and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of precise parking of electric locomotive of the present invention;

[0053] Figure 2 It is a force diagram of the electric locomotive in the present invention;

[0054] Figure 3 It is a schematic diagram of double closed-loop vector control of the permanent magnet synchronous motor in the present invention;

[0055] Figure 4 It is a schematic diagram of the algorithm structure of a single hidden layer feedforward neural network model in the present invention. DETAILED DESCRIPTION

[0056] The present invention will be further described below in conjunction with the accompanying drawings.

[0057] refer to Figure 1 The embodiment of the present invention provides a method for accurately parking an electric locomotive in an underground coal mine. The electric locomotive includes a permanent magnet synchronous motor. When the electric locomotive receives a parking instruction, its braking process includes the following steps:

[0058] S1. Establish a model of the braking torque of the electric locomotive according to the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive;

[0059] S2. Establish a double closed-loop vector control model, combined with the model of the braking torque of the electric locomotive, and obtain the calculated value of the braking torque of the electric locomotive according to the dead weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive. The double closed-loop vector control model controls the permanent magnet synchronous motor of the electric locomotive to generate the braking torque, and adds the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive as samples to the sample set;

[0060] S3. Based on the whole sample set obtained in S2, a single hidden layer feedforward neural network model is trained, wherein the real-time load of the electric locomotive, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the real-time track slope are input, and the calculated value of the braking torque of the electric locomotive is output, and a single hidden layer feedforward neural network model that outputs the actual braking torque of the electric locomotive is obtained after training;

[0061] S4. Importing the trained single hidden layer feedforward neural network model into the dual closed-loop vector control system of the permanent magnet synchronous motor of the electric locomotive;

[0062] S5. The double closed-loop vector control system controls the permanent magnet synchronous motor of the electric locomotive to generate actual braking torque, and the actual braking torque is used to realize the parking of the electric locomotive.

[0063] Reference Figure 2 In this embodiment, the establishment of the model of the real-time braking torque of the electric locomotive in step S1 includes the following steps:

[0064] S11. Perform force analysis on the electric locomotive during braking. The ground braking forces on the two pairs of wheels of the electric locomotive are F1 and F2 respectively. Analyze the electric locomotive as a whole. The ground braking force on the electric locomotive is F=F1+F2. Establish the real-time ground braking force F and real-time inertial resistance W based on the electric locomotive as a whole. a , basic resistance W0, real-time ramp resistance W i The equilibrium equation is: i When braking uphill, use -, and when braking downhill, use +;

[0065] F=W a -W0±W i (1)

[0066] When the electric locomotive brakes, its motion state changes to deceleration, but due to the inertia of the electric locomotive, it will continue to move forward. The inertial force at this time is thrust, and the inertial force at this time is inertial resistance W. a, its unit is N; the basic resistance W0 is the friction between the axle and the journal of the electric locomotive and the mine car, the friction between the wheel and the rail, the impact and vibration resistance caused by the uneven track and the wheel tread scratches, and the air resistance, its unit is N; the slope resistance W i This is the component force generated by the locomotive's own weight when going uphill or downhill. Since the track slope is relatively small, sinθ1≈tanθ1=i, and its unit is N.

[0067] S12. Establish real-time inertial resistance W a , basic resistance W0, real-time ramp resistance W i The model includes equations (2), (3), and (4);

[0068] W a =1000(P+Q)(1+γ)a0 (2)

[0069] W0=1000(P+Q)gω (3)

[0070] W i =1000(P+Q)gi (4)

[0071] Where P is the real-time load of the electric locomotive, in t; Q is the deadweight of the electric locomotive, in t; γ is the inertia coefficient, ranging from 0.05 to 0.1; a0 is the real-time braking deceleration of the electric locomotive, in m / s 2 ; ω is the running resistance coefficient of the electric locomotive. For the same mine car model, the more mine cars the electric locomotive pulls, the greater the running resistance coefficient; i is the track slope, and g is the acceleration of gravity;

[0072] S13. Arrange equations (1), (2), (3), and (4) to obtain a calculation formula for the real-time ground braking force F of the electric locomotive when the electric locomotive brakes:

[0073] F=1000(P+Q)[(1+γ)a0-ωg±ig] (5)

[0074] In the formula, the real-time track slope i is + when braking uphill, and - when braking downhill;

[0075] S14. According to the torque formula T = FL, where T is the braking torque of the electric locomotive and L is the length of the lever arm, convert equation (5) into:

[0076]

[0077] S15. According to the formula Where S is the real-time distance between the electric locomotive and the destination, v is the real-time speed of the electric locomotive, and equation (6) is converted into:

[0078]

[0079] According to formula (7), a model is established with the weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the braking torque of the electric locomotive as variables.

[0080] The dual closed-loop vector control model in step S2 is established in the dq rotating coordinate system, which is a rotating coordinate system that rotates synchronously with the air gap magnetic field of the motor and can be regarded as a rotating coordinate system placed on the motor rotor. d and q The equation is:

[0081] Ψ d =L d i d +Ψ f

[0082] Ψ q =L q i q

[0083] Voltage u d and u q The equation is:

[0084]

[0085]

[0086] Electromagnetic torque T e The balanced equation is:

[0087]

[0088] Where Ψ d is the stator d-axis flux, which includes the flux generated by the stator d-axis current and the flux generated by the permanent magnet f ; q is the stator q-axis flux; L d and L q are the inductance of the stator winding along the d and q axes respectively; u d and u q are the voltages of the stator along the d and q axes respectively; i d and i q are the stator currents along the d and q axes respectively, ω is the rotor angular velocity; R s is the phase resistance of the stator winding; p is the number of motor pole pairs; T e is the electromagnetic torque of the motor.

[0089] The double closed-loop vector control model is combined with the PID control system. When the armature current reaches the set maximum reverse current, under the action of the current PI regulator of the PID control system, the motor terminal voltage and the armature current establish a new balance, and the motor enters a stable braking state. In this state, the motor flows through the braking current. The motor decelerates under the action of the set limited braking torque and friction resistance torque, and the back electromotive force decreases. The current PI regulator of the PID control system also reduces the output voltage synchronously to maintain the balance relationship of the phasor equation under the system set current. At this time, the motor runs in the generator state. This process is the main stage of the braking process. During this process, the three-phase inverter of the PID control system converts the mechanical energy stored in the rotor and the load into electrical energy and feeds it back to the DC side. The motion equation of the motor is:

[0090]

[0091] refer to Figure 3 The control model of the dual closed-loop vector control model combined with the PID control system is a speed and current dual closed-loop. The PID control system includes a speed PI controller, a current PI controller 1, a current PI controller 2, a coordinate system transformation module I_PARK, a coordinate system transformation module CLARK, a coordinate system transformation module PARK, a space vector modulation module, a three-phase inverter, and a voltage-current converter. The control steps of the dual closed-loop vector control model are as follows:

[0092] S21: According to the formula v=2πrn, where r is the rotor radius of the permanent magnet synchronous motor of the electric locomotive, and n is the real-time rotor speed of the permanent magnet synchronous motor of the electric locomotive, the formula (7) is converted into:

[0093]

[0094] S22: Input the given speed N and the real-time speed n of the permanent magnet synchronous motor rotor into the speed PI controller to obtain the stator current I along the q axis after being adjusted by the speed PI controller. q and the stator current I along the d axis d ;

[0095] S23: I q The real-time current i of the stator along the q axis q Input current PI controller 1 to obtain the stator voltage value u along the q axis after being adjusted by current PI controller 1 q ;

[0096] Will I d The real-time current i of the stator along the d axis d Input current PI controller 2 to obtain the stator voltage value u along the d axis after being adjusted by current PI controller 2 d ;

[0097] S24: will u q 、u d Input the coordinate transformation module I_PARK respectively, and the coordinate transformation module I_PARK outputs u q Voltage value u in αβ coordinate system α 、u d Voltage value u in αβ coordinate system β ;

[0098] S25: Will u α 、u β An input space vector modulation module is input, and the space vector modulation module outputs a trigger signal to a three-phase inverter;

[0099] S26: The three-phase inverter receives the trigger signal and outputs a three-phase voltage u a 、u b 、u c , the three-phase voltage controls the permanent magnet synchronous motor to produce braking torque;

[0100] S27: three-phase voltage u output by the three-phase inverter a 、u b 、u c Input voltage and current conversion module, voltage and current conversion module output three-phase current i a 、i b 、i c , three-phase current i a 、i b 、i c Input coordinate system transformation module CLARK, coordinate system transformation module CLARK outputs the current i of the stator along the α axis in the αβ coordinate system α and the current i along the β axis β ,i α and i β Input coordinate system transformation module PARK, coordinate system transformation module PARK outputs the current i of the stator along the d-axis in the dq coordinate system d and the current i along the q axis q ,i q Input current PI controller 1, i d The input current PI controller 2 makes the PID control system form a closed loop.

[0101] In this embodiment, the precise parking method for an electric locomotive in an underground coal mine also includes step S6: during the braking process of the electric locomotive, when the real-time distance S between the electric locomotive and the end point is less than a preset value S0, if the real-time speed v of the electric locomotive is greater than the preset value v0, the electric locomotive starts mechanical braking to achieve "secondary braking" of the electric locomotive, thereby achieving precise parking of the electric locomotive.

[0102] In this embodiment, the permanent magnet synchronous motor of the electric locomotive is a surface permanent magnet synchronous motor, which is divided into two types: surface permanent magnet synchronous motor and built-in permanent magnet synchronous motor. d =L q , built-in permanent magnet synchronous motor L d ≠L q In the surface permanent magnet synchronous motor, due to L d =L q , electromagnetic torque T e for:

[0103] T e =pΨ f i q

[0104] Surface permanent magnet synchronous motor adopts i d =0 control mode, i d =0 is a simple and easy control method. When the torque is constant, the required customized current is the minimum, which can greatly reduce copper loss and improve efficiency; electromagnetic torque T e It is proportional to the rotor flux and the q-axis component of the stator current. The parameters can be decoupled from each other, so as long as i is maintained during operation d =0, the electromagnetic torque is only controlled by the q-axis component of the stator current, so that the vector control of the permanent magnet synchronous motor obtains the same control performance as that of the DC motor.

[0105] In this embodiment, the single hidden layer feedforward neural network model is an extreme learning machine. Figure 4 The network model of the extreme learning machine is divided into three layers, namely the input layer, hidden layer and output layer. The input layer realizes the function of accepting input variables from the external environment; the hidden layer has activation functions mainly used to realize functions such as calculation and recognition; the output layer is used to output results. Different from the traditional BP neural network training method, the extreme learning machine abandons the back propagation algorithm based on gradient information, adopts random input layer weights and biases, introduces the output layer weight norm as a regular term in the loss function corresponding to the training error, and calculates the output layer weight according to the Moore-Penrose generalized inverse matrix theory. Therefore, the extreme learning machine has fewer training parameters, faster learning speed, and stronger generalization ability for new data.

[0106] In this embodiment, if the calculated values ​​of the electric locomotive's own weight, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the braking torque of the electric locomotive obtained in steps S1 and S2 are the same as the samples already existing in the sample set, they will not be added to the sample set to avoid the same samples occupying the sample set space, so as to improve the training speed of the single hidden layer feedforward neural network model.

[0107] A system based on a precise parking method for electric locomotives in coal mines, comprising a communication module, a braking torque model module, a double closed-loop vector control model module, a correction module, a single hidden layer feedforward neural network model module, and a feedback control module;

[0108] A communication module, used to obtain a stop instruction of the electric locomotive;

[0109] The braking torque model module is used to obtain the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive, and establish a model of the braking torque of the electric locomotive;

[0110] The double closed-loop vector control model module is used to obtain the calculated value of the braking torque of the electric locomotive according to the dead weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive. The double closed-loop vector control model module controls the permanent magnet synchronous motor of the electric locomotive to generate braking torque;

[0111] A correction module is used to perform error analysis and correction on the braking torque model module and the double closed-loop vector control model module, and to construct a sample set with the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive as samples;

[0112] A single hidden layer feedforward neural network model module is used to train the single hidden layer feedforward neural network model based on the sample set of the correction module to obtain a single hidden layer feedforward neural network model that outputs the actual braking torque of the electric locomotive;

[0113] The feedback control module is used to import the trained single hidden layer feedforward neural network model into the double closed-loop vector regulation system of the permanent magnet synchronous motor of the electric locomotive. The double closed-loop vector regulation system controls the permanent magnet synchronous motor of the electric locomotive to generate actual braking torque to realize the parking of the electric locomotive.

[0114] An electronic device includes a storage device, one or more processors, and the storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement a precise parking method based on an electric locomotive in an underground coal mine.

[0115] A computer-readable storage medium stores a computer program therein, wherein when the computer program is executed by a processor, a precise parking method for an electric locomotive in an underground coal mine is implemented.

[0116] The invention provides a precise parking method for an electric locomotive in an underground coal mine. The model of the braking torque of the electric locomotive is combined with a double closed-loop vector control model. The real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive are used as sample sets to train a single hidden layer feedforward neural network model. The trained single hidden layer feedforward neural network model is introduced into the double closed-loop vector regulation system of the permanent magnet synchronous motor of the electric locomotive to achieve precise parking of the electric locomotive, reduce parking errors, and improve production efficiency.

[0117] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A precise parking method for an electric locomotive in a coal mine, the electric locomotive comprising a permanent magnet synchronous motor, characterized in that: When the electric locomotive receives a stop command, its braking process includes the following steps: S1. Establish a model of the braking torque of the electric locomotive according to the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive; The establishment of the model of the electric locomotive braking torque in S1 includes the following steps: S11. Establish the real-time ground braking force F and real-time inertial resistance W based on the whole electric locomotive a , basic resistance W0, real-time ramp resistance W i The equilibrium equation is: i When braking uphill, use -, and when braking downhill, use +; F=W a -W0±W i (1) S12. Establish real-time inertial resistance W a , basic resistance W0, real-time ramp resistance W i The model includes equations (2), (3), and (4); W a =1000(P+Q)(1+γ)a0 (2) W0=1000(P+Q)gω (3) W i =1000(P+Q)gi (4) Where P is the real-time load of the electric locomotive, Q is the deadweight of the electric locomotive, γ is the inertia coefficient, a0 is the real-time braking deceleration of the electric locomotive, ω is the running resistance coefficient of the electric locomotive, i is the track slope, and g is the acceleration of gravity; S13. Arrange equations (1), (2), (3), and (4) to obtain a calculation formula for the real-time ground braking force F of the electric locomotive when the electric locomotive brakes: F=1000(P+Q)[(1+γ)a0-ωg±ig] (5) In the formula, the real-time track slope i is + when braking uphill, and - when braking downhill; S14. According to the torque formula T = FL, where T is the braking torque of the electric locomotive and L is the length of the lever arm, convert equation (5) into: S15. According to the formula Where S is the real-time distance between the electric locomotive and the destination, v is the real-time speed of the electric locomotive, and equation (6) is converted into: According to formula (7), a model is established with the weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the braking torque of the electric locomotive as variables; S2. Establish a double closed-loop vector control model, combined with the model of the braking torque of the electric locomotive, and obtain the calculated value of the braking torque of the electric locomotive according to the dead weight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive. The double closed-loop vector control model controls the permanent magnet synchronous motor of the electric locomotive to generate the braking torque, and adds the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive as samples to the sample set; The dual closed-loop vector control model in S2 is a vector control model based on the dq rotating coordinate system and combined with the PID control system. The PID control system includes a speed PI controller, a current PI controller 1, a current PI controller 2, a coordinate system transformation module I_PARK, a coordinate system transformation module CLARK, a coordinate system transformation module PARK, a space vector modulation module, a three-phase inverter, and a voltage-current converter. The control steps of the dual closed-loop vector control model are as follows: S21: According to the formula v=2πrn, where r is the rotor radius of the permanent magnet synchronous motor of the electric locomotive, and n is the real-time rotor speed of the permanent magnet synchronous motor of the electric locomotive, the formula (7) is converted into: S22: Input the given speed N and the real-time speed n of the permanent magnet synchronous motor rotor into the speed PI controller to obtain the stator current I along the q axis after being adjusted by the speed PI controller. q and the stator current I along the d axis d ; S23: I q The real-time current i of the stator along the q axis q Input current PI controller 1 to obtain the stator voltage value u along the q axis after being adjusted by current PI controller 1 q ; Will I d The real-time current i of the stator along the d axis d Input current PI controller 2 to obtain the stator voltage value u along the d axis after being adjusted by current PI controller 2 d ; S24: will u q 、u d Input the coordinate transformation module I_PARK respectively, and the coordinate transformation module I_PARK outputs u q Voltage value u in αβ coordinate system α 、u d Voltage value u in αβ coordinate system β ; S25: Will u α 、u β An input space vector modulation module is input, and the space vector modulation module outputs a trigger signal to a three-phase inverter; S26: The three-phase inverter receives the trigger signal and outputs a three-phase voltage u a 、u b 、u c , the three-phase voltage controls the permanent magnet synchronous motor to produce braking torque; S27: three-phase voltage u output by the three-phase inverter a 、u b 、u c Input voltage and current conversion module, voltage and current conversion module output three-phase current i a 、i b 、i c , three-phase current i a 、i b 、i c Input coordinate system transformation module CLARK, coordinate system transformation module CLARK outputs the current i of the stator along the α axis in the αβ coordinate system α and the current i along the β axis β ,i α and i β Input coordinate system transformation module PARK, coordinate system transformation module PARK outputs the stator current i along the d-axis in the dq coordinate system d and the current i along the q axis q ,i q Input current PI controller 1, i d Input current PI controller 2 to make the PID control system form a closed loop; S3. Based on the whole sample set obtained in S2, a single hidden layer feedforward neural network model is trained, wherein the real-time load of the electric locomotive, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the real-time track slope are input, and the calculated value of the braking torque of the electric locomotive is output, and a single hidden layer feedforward neural network model that outputs the actual braking torque of the electric locomotive is obtained after training; S4. Importing the trained single hidden layer feedforward neural network model into the dual closed-loop vector control system of the permanent magnet synchronous motor of the electric locomotive; S5. The double closed-loop vector control system controls the permanent magnet synchronous motor of the electric locomotive to generate actual braking torque, and the actual braking torque is used to realize the parking of the electric locomotive.

2. The precise parking method for electric locomotive in coal mine according to claim 1, characterized in that: The method further includes step S6: during the braking process of the electric locomotive, when the real-time distance S between the electric locomotive and the end point is less than a preset value S0, if the real-time speed v of the electric locomotive is greater than a preset value v0, the electric locomotive starts mechanical braking.

3. The precise parking method for electric locomotive in coal mine according to claim 1, characterized in that: The permanent magnet synchronous motor of the electric locomotive is a surface permanent magnet synchronous motor.

4. The precise parking method for electric locomotive in coal mine according to claim 1, characterized in that: The single hidden layer feedforward neural network model is an extreme learning machine.

5. The precise parking method for electric locomotive in coal mine according to claim 1, characterized in that: If the calculated values ​​of the electric locomotive's own weight, the electric locomotive's real-time load, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the braking torque of the electric locomotive obtained in steps S1 and S2 are the same as the samples already existing in the sample set, they will not be added to the sample set.

6. A system based on a precise parking method for electric locomotives in coal mines, characterized in that: It includes a communication module, a braking torque model module, a double closed-loop vector control model module, a correction module, a single hidden layer feedforward neural network model module, and a feedback control module, and realizes a precise parking method based on an underground electric locomotive in a coal mine as described in any one of claims 1 to 5; A communication module, used to obtain a stop instruction of the electric locomotive; The braking torque model module is used to obtain the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive, and establish a model of the braking torque of the electric locomotive; the double closed-loop vector control model module is used to obtain the calculated value of the braking torque of the electric locomotive according to the deadweight of the electric locomotive, the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, and the real-time speed of the electric locomotive. The double closed-loop vector control model module controls the permanent magnet synchronous motor of the electric locomotive to generate braking torque; A correction module is used to perform error analysis and correction on the braking torque model module and the double closed-loop vector control model module, and to construct a sample set with the real-time load of the electric locomotive, the real-time track slope, the real-time distance between the electric locomotive and the end point, the real-time speed of the electric locomotive, and the calculated value of the braking torque of the electric locomotive as samples; A single hidden layer feedforward neural network model module is used to train the single hidden layer feedforward neural network model based on the sample set of the correction module to obtain a single hidden layer feedforward neural network model that outputs the actual braking torque of the electric locomotive; The feedback control module is used to import the trained single hidden layer feedforward neural network model into the double closed-loop vector regulation system of the permanent magnet synchronous motor of the electric locomotive. The double closed-loop vector regulation system controls the permanent magnet synchronous motor of the electric locomotive to generate actual braking torque to realize the parking of the electric locomotive.

7. An electronic device, characterized in that: It includes a storage device, one or more processors, and the storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement a precise parking method based on an underground electric locomotive in a coal mine as described in any one of claims 1-5.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, a precise parking method for an electric locomotive in an underground coal mine as described in any one of claims 1 to 5 is implemented.

Images