Amphibian humanoid welding robot chassis control method based on magnetic wheel driving

Abstract

The invention provides an amphibian human welding robot chassis control method based on magnetic wheel driving. A dead-beat fast calculation three-vector model is adopted for conducting predictive control on a motor; the input of the dead-beat fast calculation three-vector model is a target current; a current parameter Pr (k) = [R Ls psi f] T in the dead-beat fast calculation three-vector model is obtained through real-time estimation by adopting a forgetting factor recursive least square method parameter identification algorithm; wherein R is the stator resistance of the motor; ls is motor stator inductance; psi f is motor rotor flux linkage; the forgetting factor recursive least square method parameter identification algorithm is realized based on a q-axis discretization mathematical model and a least square method recursive formula with forgetting factors. According to the method, the current harmonic distortion rate can be effectively reduced, the torque stability can be improved, the influence of temperature and other external environment changes on motor parameters can be restrained, accurate parameter identification is achieved, control can be better conducted through a model prediction control algorithm, the robot chassis can move more stably, and the welding stability is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of humanoid welding, and more specifically, to a chassis control method for an amphibious humanoid welding robot based on magnetic wheel drive. Background Art

[0002] Welding technology, as the tailor in the field of industrial manufacturing, can ensure strong industrial production capacity and is widely used in the manufacturing industry. In recent years, the demand for ocean resource development has been continuously increasing, and the demand for underwater pipeline welding has also been growing. Currently, the most commonly used method for underwater pipeline welding is wet welding. However, wet welding requires welders to operate underwater, and manual welding has problems such as low efficiency and unstable quality, which can no longer meet the requirements of modern industrial production. Moreover, being able to perform welding operations both on land and underwater is also a problem that needs to be solved.

[0003] To solve the above existing problems, such as the Chinese patent "An Underwater Welding Robot and Its Operating Process" (publication number: CN114669933A), when this device works, it can adjust the posture of the welding torch according to the welding shape and the spatial position of the welding. However, this device moves through a propulsion device. Due to the underwater undercurrent, it cannot reach the welding position stably, and it cannot ensure the accuracy and stability of welding. Since it works through a thruster, it cannot perform welding operations on land. Another example is the Chinese patent "An Air Chamber Type Local Dry Underwater Welding Robot" (publication number: CN112743192A). When this device works, it adopts the air chamber type local dry underwater welding technology, which can ensure the welding effect. However, this device has a large structure and moves through a triangular crawler wheel structure. It can work both underwater and on land, but it cannot move and operate on a negatively inclined magnetic surface and in a narrow space. At the same time, the crawler wheel structure is also easily affected by the underwater undercurrent fluctuation, resulting in unstable welding.

[0004] Therefore, it is necessary to design a chassis control method for an amphibious humanoid welding robot based on magnetic wheel drive, which can not only perform welding operations on the magnetic surface on land, but also perform welding operations on the magnetic surface underwater. The high control performance of the motor algorithm can ensure welding stability and improve welding quality. Summary of the Invention

[0005] To overcome the disadvantages and deficiencies in the prior art, the purpose of the present invention is to provide a chassis control method for an amphibious humanoid welding robot based on magnetic wheel drive; this method can effectively reduce the current harmonic distortion rate, improve the torque stability, and can suppress the influence of external environment changes such as temperature on the motor parameters, realize accurate parameter identification, and the model predictive control algorithm can better perform control, making the movement of the robot chassis more stable and ensuring welding stability.

[0006] To achieve the above object, the present invention is realized by the following technical solutions: A chassis control method for an amphibious humanoid welding robot based on magnetic wheel drive, which is used to control the magnetic wheels driven by motors in the chassis of the amphibious humanoid welding robot;

[0007] Predictive control is performed on the motor using a deadbeat speed calculation three-vector model; the input of the deadbeat speed calculation three-vector model is the target current I*; the current parameter Pr(k) in the deadbeat speed calculation three-vector model = [R L s ψ f T is obtained by real-time estimation using the forgetting factor recursive least squares parameter identification algorithm; where, R is the stator resistance of the motor; L s is the stator inductance of the motor; ψ f is the rotor flux linkage of the motor;

[0008] The forgetting factor recursive least squares parameter identification algorithm is implemented based on the q-axis discretized mathematical model and the least squares recursive formula with a forgetting factor;

[0009] The q-axis discretized mathematical model is:

[0010] where, y(k) = I q (k);

[0011] is the system input matrix;

[0012] η(k) = [β q1 β q2 β q3 β q4 T is the parameter matrix to be identified;

[0013] where, I q (k - 1) is the q-axis current at the previous moment; I d (k) is the d-axis current at the current moment; u q (k) is the q-axis voltage at the current moment; ω e (k) is the electrical angular velocity of the rotor at the current moment; β q1 、β q2 、β q3 、β q4 are coefficients that vary with time respectively;

[0014] The least squares recursive formula with a forgetting factor is:

[0015]

[0016] where, is the estimated value at the current moment, ​​is the estimated value at the previous moment; P(k) is the covariance matrix at the current moment, and P(k - 1) is the covariance matrix at the previous moment; L(k) is the gain matrix; E(k) is the identity matrix;

[0017] The forgetting factor recursive least squares parameter identification algorithm refers to:

[0018] Measure y(k); and calculate the estimated value at the current moment covariance matrix P(k) and gain matrix L(k);

[0019] When the convergence value of the parameter matrix η(k) to be identified is calculated then calculate the motor parameter Pr(k) to be identified = [R L s ψ f T :

[0020]

[0021] Calculate the relative error err of the motor identification parameters according to the identification parameter Pr(k - 1) at the previous moment and the identification parameter Pr(k) at the current moment, and judge whether the parameters are completely convergent according to the relative error err;

[0022] If the parameters have converged completely, substitute the identification parameter Pr(k) at the current moment into the deadbeat speed calculation three-vector model for predictive control; if the parameters have not converged, skip the current cycle and substitute the identification parameter of the previous cycle into the deadbeat speed calculation three-vector model for predictive control.

[0023] Preferably, in the q-axis discretized mathematical model, the coefficients β q1 β q2 β q3 β q4 and the d-axis current L d q-axis current L q stator resistance R, motor rotor flux linkage ψ f and the control algorithm period T have the following mathematical relationship:

[0024]

[0025] Preferably, for the forgetting factor recursive least squares parameter identification algorithm, when k = 1, first initialize the estimated value Initialize the covariance value P(0) = αE(k); then start the forgetting factor recursive least squares parameter identification algorithm.

[0026] The relative error err is:

[0027] ​

[0028] If err < threshold, it means the parameters have converged completely; otherwise, it means the parameters have not converged.

[0029] Preferably, the input of the deadbeat fast calculation three-vector model is the target current I*.

[0030] The deadbeat fast calculation three-vector model selects an optimal vector u from six effective voltage vectors Ui, i = [1, 2, 3, 4, 5, 6] according to the target current I*. p_1 and a sub-optimal vector u p_2 , and calculates the action time of the optimal vector u p_1 and the sub-optimal vector u p_2 ; According to the optimal vector u p_1 and the sub-optimal vector u p_2 and the action time, act on the inverter to make the inverter generate a corresponding voltage vector to act on the motor.

[0031] Preferably, in the deadbeat fast calculation three-vector model, selecting an optimal vector u p_1 and a sub-optimal vector u p_2 from six effective voltage vectors Ui, i = [1, 2, 3, 4, 5, 6] means:

[0032] First, according to the deadbeat control principle of d-q axis current, calculate the optimal voltage vector u p_dq (k + 1) at the next moment from the full voltage vector range:

[0033] u p_dq (k + 1) = AI dq (k) + BI dq * + C;

[0034]

[0035] Among them, u p_d (k + 1), u p_q (k + 1) are the global optimal voltage components of the d-axis and q-axis at the next moment respectively; I d (k), I q (k) represent the d-axis and q-axis currents collected at the current moment; I dq * represents the target current; A, B, and C represent coefficients respectively;

[0036] After that, transform the global optimal voltage vector u p_dq (k + 1) into the value u p_αβ (k + 1) in the two-phase stationary coordinate system; judge the sector where the value u p_αβ (k + 1) is located; according to the value u p_αβ(k + 1) selects the optimal vector u for the sector where it is located p_1 and the sub - optimal vector u p_2 .

[0037] Preferably, substituting the estimated motor parameters into the dead - beat speed - calculation three - vector model means: substituting the estimated motor parameters into the calculation formulas of coefficients A, B, and C; the calculation formulas of parameters A, B, and C are respectively:

[0038]

[0039] where T is the control period.

[0040] Preferably, the target current I* includes the target current d - axis component I d *, and the target current q - axis component I q *; it is set that the target current d - axis component I d * = 0;

[0041] The method for obtaining the target current q - axis component I q * is: inputting the target speed Speed * and the actual speed Speed after encoder acquisition and processing into a speed PI controller, and outputting to obtain the target current q - axis component I q *.

[0042] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0043] 1. The present invention adopts a new forgetting - factor recursive least - squares parameter identification - speed - calculation three - vector model predictive current control algorithm to control the current (torque) of the chassis permanent - magnet synchronous motor. Compared with the current PI closed - loop in the existing traditional vector control algorithm FOC, this method has better dynamic response performance, can effectively respond to the changes in current (torque), can better adapt to the pulse fluctuations generated during the welding process, ensure the stability of the chassis movement during welding, and does not require repeated adjustment of the parameters in the controller, which can achieve the purpose of simplifying the controller; compared with the traditional model predictive current control algorithm, this algorithm does not need to calculate the cost function to predict the voltage vector required at the next moment, but generates the voltage vector required at the next moment through dead - beat control of d - q axis current, which can reduce the prediction times and thus reduce the operation time and relieve the calculation burden of the control system;

[0044] 2. The algorithm adopted by the present invention combines the forgetting - factor recursive least - squares parameter identification algorithm; compared with the traditional offline parameter identification FOC algorithm, this algorithm performs online identification of motor parameters, can identify motor parameters in real - time, has strong parameter robustness, and can effectively suppress the influence of motor parameter changes on the model predictive control algorithm. Description of the Drawings

[0045] Figure 1 is a schematic structural diagram of the amphibious anthropomorphic welding robot of the present invention;

[0046] Figure 2 is a schematic layout diagram of the chassis watertight control box of the amphibious anthropomorphic welding robot of the present invention;

[0047] Figure 3 is a block diagram of the chassis control method of the amphibious anthropomorphic welding robot based on magnetic wheel drive of the present invention;

[0048] Figure 4 is a parameter identification flowchart of the chassis control method of the amphibious anthropomorphic welding robot based on magnetic wheel drive of the present invention;

[0049] Figure 5 is a performance comparison chart of the chassis control method of the amphibious anthropomorphic welding robot based on magnetic wheel drive of the present invention and traditional algorithms. Detailed implementation manners

[0050] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0051] Embodiment

[0052] A chassis control method of an amphibious anthropomorphic welding robot based on magnetic wheel drive in this embodiment; this method is used to control the magnetic wheels driven by motors in the chassis of the amphibious anthropomorphic welding robot.

[0053] The structure of the amphibious anthropomorphic welding robot is as Figure 1 and Figure 2 shown, and it includes a chassis 1, a six-degree-of-freedom robotic arm, a robot head 13, a body 14 connecting various components of the robot, an underwater searchlight 5, and an underwater camera 6.

[0054] The chassis includes a chassis transmission structure, a bracket for carrying the chassis transmission structure, a platform for connecting the upper robotic arm, magnetic wheels 4, and a chassis watertight control box 3. Among them, the chassis transmission structure includes a watertight high-torque permanent magnet synchronous motor and a harmonic reducer. The upper computer outputs control instructions to the motor driver, and the motor driver controls the motor to rotate and transmits the power to the magnetic wheels through the harmonic reducer to achieve the movement of the chassis. The chassis watertight control box 3 is mainly used to carry a brushless motor drive board, a robot main controller 9, a switch 11, a 48V DC power module 7, a 48V to 24V power converter 8, a wire feeder motor driver 10, and a wire splitter. The chassis watertight control box 3 realizes the connection between the upper computer and electrical equipment and the connection between the electrical equipment and the chassis motor through watertight terminals, preventing water from entering and damaging and short-circuiting electrical equipment such as the controller.

[0055] The brushless motor drive board includes a core circuit based on the STM32F405RGT6 main control chip, a drive circuit based on the DRV8301 drive chip, a three-phase full-bridge inverter circuit based on the parallel connection of Sic-MOSFETs, and a sampling circuit composed of a high-precision sampling resistor and an operational amplifier.

[0056] The robot main controller 9 includes connections to the chassis motor driver, the robotic arm motor driver, the wire feeder motor driver, the welding power supply signal reception and transmission, and the industrial computer. The robot main controller 9 is mainly responsible for sending control signals for the chassis motor, the six-joint robotic arm motor, the wire feeder motor, the welding power supply analog signal, and the underwater searchlight 5 and the underwater camera 6; the wire feeder motor driver 10 is mainly responsible for supplying driving energy to the wire feeder motor and is controlled by the control signal sent by the robot main controller; the switch 11 is mainly responsible for the communication between the chassis motor, the robotic arm motor and the upper computer.

[0057] According to the power requirements of the chassis, including the chassis moving speed, the chassis turning speed, the turning radius, etc., through the analysis and calculation of the chassis dynamics model, the target speeds of the four power motors can be obtained respectively, and then the target speeds are input into the chassis motor control algorithm to realize the open-loop control of the robot chassis movement. The speed closed-loop control of the chassis power motor of the amphibious humanoid welding robot based on the forgetting factor recursive least squares-fast calculation three-vector model predictive current control algorithm includes two closed-loop modules, an inverter module, a sampling module, a feedback module, a coordinate transformation module and a parameter identification module. Among them, the two closed-loop modules are the outer speed closed-loop and the inner current closed-loop respectively. The speed closed-loop is controlled by the traditional PI algorithm, and the inner current loop is controlled by the new parameter identification-three-vector model predictive current control algorithm.

[0058] Specifically, the present invention uses the deadbeat fast calculation three-vector model for predictive control, as Figure 3 shown, the input of the deadbeat fast calculation three-vector model is the target current I*; the current parameter Pr(k) = [R L s ψ f T in the deadbeat fast calculation three-vector model is obtained by real-time estimation using the forgetting factor recursive least squares parameter identification algorithm; where R is the stator resistance of the motor; L s is the stator inductance of the motor; ψ f is the rotor flux linkage of the motor; the estimated motor parameters are substituted into the deadbeat fast calculation three-vector model for prediction; the motor control of the magnetic wheel is carried out according to the prediction value obtained from the deadbeat fast calculation three-vector model.

[0059] The target current I* includes the target current d-axis component I d *, the target current q-axis component I q *. The target current q-axis component Iq The acquisition method of * is as follows: input the target speed Speed * and the actual speed Speed after encoder acquisition and processing into the speed PI controller, and output the target current q-axis component I q *. The present invention uses a control strategy with the target current d-axis component I d * = 0. Such a control strategy can achieve static decoupling of the d-q axis currents of the motor, that is, it can independently control the d-axis current and q-axis current of the motor, and can achieve better control of the motor speed and torque.

[0060] The forgetting factor recursive least squares parameter identification algorithm is implemented based on the q-axis discretized mathematical model and the recursive formula of the least squares method with a forgetting factor.

[0061] The construction process of the q-axis discretized mathematical model is as follows: First, identify the three parameters of the motor through the forgetting factor recursive least squares parameter identification algorithm.

[0062] Motor mathematical model

[0063]

[0064] Among them, R is the stator resistance; L d , L q are the stator inductances of the direct and quadrature axes of the motor; I d , I q , u d , u q are the direct and quadrature axis currents and voltages of the motor respectively; ψ f is the rotor flux linkage of the motor; ω e is the electrical angular velocity of the rotor of the motor.

[0065] Discretize the above differential model through forward Euler discretization and separate the parameters and variables. The discretized result is in the following form:

[0066]

[0067] Among them, I d (k), I q (k) are the d-axis and q-axis currents at the current moment respectively; I d (k - 1), I q (k - 1) are the d-axis and q-axis currents at the previous moment respectively; u d (k), u q (k) are the d-axis and q-axis voltages at the current moment respectively; ω e (k) is the electrical angular velocity of the rotor at the current moment; β d1 , β d2 , β d3 , β d4and β q1 、β q2 、β q3 、β q4 are coefficients that vary with time, and their mathematical relationships with L d 、L q 、R、ψ f and the control algorithm period T are as follows:

[0068]

[0069] The motor parameters to be identified are L d 、L q 、R、ψ f , so the q-axis discretized mathematical model is selected as the basic model for parameter estimation. Simplifying its model gives:

[0070]

[0071] where y(k) = I q (k);

[0072] is the system input matrix;

[0073] η(k) = [β q1 β q2 β q3 β q4 T is the parameter matrix to be identified;

[0074] where I q (k - 1) is the q-axis current at the previous moment; I d (k) is the d-axis current at the current moment; u q (k) is the q-axis voltage at the current moment; ω e (k) is the electrical angular velocity of the rotor at the current moment; β q1 、β q2 、β q3 、β q4 are coefficients that vary with time respectively;

[0075] Based on the above recursive parameter identification model, the recursive formula of the least squares method with a forgetting factor is established as:

[0076]

[0077] where, is the estimated value at the current moment, is the estimated value at the previous moment; P(k) is the covariance matrix at the current moment, P(k - 1) is the covariance matrix at the previous moment; L(k) is the gain matrix; E(k) is the identity matrix; the initial estimation value ​The initial value of covariance P(0) = αE(k), where α is generally a relatively large number between 1000 and 10000; λ is the forgetting factor, generally between 0.97 and 1.

[0078] The parameter identification algorithm of the forgetting factor recursive least squares method, as Figure 4 shown, means:

[0079] Measure y(k); and calculate the estimated value at the current moment of the covariance matrix P(k) and the gain matrix L(k);

[0080] When the convergence value of the parameter matrix η(k) to be identified is calculated After that, calculate the motor parameter Pr(k) to be identified = [R L s ψ f T :

[0081]

[0082] Calculate the relative error err of the motor identification parameters, and judge whether the parameters are completely convergent according to the relative error err:

[0083]

[0084] where Pr(k) represents the identification parameter at the current moment; Pr(k - 1) represents the identification parameter at the previous moment;

[0085] If err < threshold (e.g., 0.01%), it means that the identification error in the algorithm is small and the parameters have converged completely. Substitute the identification parameter Pr(k) at the current moment into the deadbeat speed calculation three-vector model for predictive control;

[0086] If err ≥ threshold, it means that the identification error in the algorithm is large and the parameters have not converged. Then skip the current cycle, do not substitute the identification parameter with large identification error into the controller, and substitute the identification parameter in the previous cycle into the deadbeat speed calculation three-vector model for predictive control.

[0087] Bringing the three motor parameters obtained by real-time estimation of the forgetting factor recursive least squares method parameter identification algorithm into the deadbeat speed calculation three-vector model predictive current control algorithm can effectively identify the motor parameters affected by environmental factors such as temperature, thereby reducing the impact of parameter changes on the model predictive control algorithm.

[0088] Secondly, by analyzing the motor parameters identified by the above algorithm and substituting them into the deadbeat speed calculation three-vector model predictive current control algorithm to replace the motor parameters of the original model, the accurate identification of the motor model and the control of the motor state can be realized.

[0089] ​The deadbeat rapid calculation three-vector model selects an optimal vector u from six effective voltage vectors Ui, i = [1, 2, 3, 4, 5, 6] according to the target current I*. p_1 and a sub-optimal vector u p_2 , and calculates the action times of the optimal vector u p_1 and the sub-optimal vector u p_2 ; According to the optimal vector u p_1 and the sub-optimal vector u p_2 and the action times, act on the inverter to make the inverter generate corresponding voltage vectors to act on the motor.

[0090] Specifically, in the deadbeat rapid calculation three-vector model, the method for obtaining the six effective voltage vectors Ui is as follows:

[0091] Calculate the d-axis voltage value u d [i] and the q-axis voltage value u q [i] of the six effective voltage vectors Ui in the switching state:

[0092]

[0093] Among them, U dc is the motor bus voltage; θ e is the motor electrical angle, θ e = pθ, p is the number of pole pairs of the motor, and θ is the detected angle; S_abc[i][j] is the j-th bit of the switching state value S_abc[i]. The switching state value S_abc[i] is shown in Table 1:

[0094] Table 1 Values of the switching state value S_abc[i]

[0095] Voltage vector S_abc Voltage vector S_abc U1 001 U4 100 U2 010 U5 101 U3 011 U6 110

[0096] Selecting an optimal vector and a sub-optimal vector from the six effective voltage vectors Ui means:

[0097] First, according to the deadbeat control principle of d-q axis current, calculate the global optimal voltage vector u p_dq (k + 1) at the next moment from the full voltage vector range:

[0098] u p_dq (k + 1) = AI dq (k) + BI dq * + C;

[0099]

[0100] Among them, u p_d (k + 1), u p_q(k + 1) are the global optimal voltage components on the d-axis and q-axis at the next moment; I d (k), I q (k) represent the d-axis and q-axis currents collected at the current moment; I dq * represents the target current; A, B, and C represent coefficients respectively; A, B, and C are respectively:

[0101]

[0102] Among them, R is the stator resistance; L s is the stator inductance; ω e is the electrical angular velocity of the motor; ψ f is the magnetic flux linkage of the permanent magnet; T s is the control period;

[0103] Substitute the identified motor parameters calculated above into the above formula, and the correction values of these three matrices can be obtained for subsequent model predictive control.

[0104] After that, transform the global optimal voltage vector u p_dq (k + 1) into the value u p_αβ (k + 1) in the two-phase stationary coordinate system; judge the sector where the value u p_αβ (k + 1) is located; the judgment of the sector where the value u p_αβ (k + 1) is located is realized by the sector judgment method of the SVPWM algorithm;

[0105] According to the sector where the value u p_αβ (k + 1) is located, select the optimal vector u p_1 and the sub-optimal vector u p_2 . The optimal vector and sub-optimal vector can be selected by looking up the corresponding table of sectors and optimal vectors and sub-optimal vectors. The corresponding table of sectors and optimal vectors and sub-optimal vectors is shown in Table 2;

[0106] Table 2 Corresponding table of sectors and optimal vectors and sub-optimal vectors

[0107] <![CDATA[u p_αβ (sector where (k + 1) is located)]]> <![CDATA[u p_1 ,u p_2 > <![CDATA[u p_αβ (sector where (k + 1) is located)]]> <![CDATA[u p_1 ,u p_2 > Ⅰ U4, U6 Ⅳ U3, U1 Ⅱ U6, U2 Ⅴ U1, U5 Ⅲ U2, U3 Ⅵ U5 U4

[0108] The calculation method of the action time of the optimal vector u p_1 and the sub-optimal vector u p_2 is:

[0109] Based on the d-q axis current deadbeat control, the output voltage is composed of the optimal vector u p_1 , the sub-optimal vector u p_2 and the zero vector u 0 . Therefore, set the control period as T s : T s = t p1 + tp2 +t 0 ; where t p1 is the action time of the optimal vector u p_1 ; t p2 is the action time of the sub-optimal vector u p_2 ; t 0 is the action time of the zero vector u 0 ;

[0110] After setting the optimal vector u p_1 , sub-optimal vector u p_2 and zero vector u 0 , the generated d-axis current change rates are s p1_d , s p2_d and s 0_d , and the generated q-axis current change rates are s p1_q , s p2_q and s 0_q ; The expression of each current change rate is as follows:

[0111]

[0112] where u p1_d and u p1_q are the d-axis and q-axis voltage components of the optimal vector u p_1 ; u p2_d and u p2_q are the d-axis and q-axis voltage components of the sub-optimal vector u p_2 ; I d , I q are the d-axis and q-axis components of the detected motor current respectively;

[0113] The action time is allocated according to the d-q axis current deadbeat principle, and the expression is as follows:

[0114] I d (k + 1) = s p1_d t p1 + s p2_d t p2 + s 0_d t 0 = I d *

[0115] I q (k + 1) = s p1_q t p1 + s p2_q t p2 + s 0_q t 0 = I q *

[0116] Through the above formula transformation, the optimal vector u can be obtained p_1 , the sub-optimal vector u p_2 and the zero vector u 0 The calculation formula for the action time is as follows:

[0117]

[0118] where I d *, I q * respectively represent the d-axis and q-axis components of the required current; K is a coefficient, and the expression is as follows:

[0119] K = s p0_q s p2_d + s p1_q s p0_d + s p2_q s p1_d

[0120] - s p1_q s p2_d - s p2_q s p0_d - s p0_q s p1_d .

[0121] After calculating the action time of the optimal vector u p_1 , the sub-optimal vector u p_2 and the zero vector u 0 , it is also necessary to judge whether t p1 + t p2 + t 0 > T s : If so, the value of the control period T s is equal to the switching period of the inverter, and the action time is reconstructed; the reconstruction formula is:

[0122]

[0123] To verify the technical effect of the present invention, the method of the present invention is simulated on the Matlab / Simulink platform. The parameters of the motor are shown in Table 3, and the parameters of the speed loop PID are shown in Table 4.

[0124] The simulation conditions are as follows: the switching frequency is 10 kHz, that is, the sampling time of the simulation actual system is 1e-5; the simulation step size is 1e-6, that is, the calculation accuracy of the simulation platform, indicating that sampling is performed every ten steps; the forgetting factor λ = 0.996.

[0125] Table 3 Motor parameter table

[0126] Table 4 PID parameter table

[0127] Parameter P I D Value 0.75 0.0001 0

[0128] The performance of DBQ3V-MPCC and the method of the present invention (FFRLS-DBQ3V-MPCC) when the three parameters of the motor are 1.5 times the normal value and the performance of the DBQ3V-MPCC algorithm when the parameters are at the normal value were compared by simulation. The motor parameters were set to 1.5 times the normal value, the rotational speed was 1000 rpm, and the load was 1.0 N·m.

[0129] From Figure 5 It can be seen that when the parameters are 1.5 times the normal value, compared with DBQ3V-MPCC, the total harmonic distortion rate (THD) of the current of FFRLS-DBQ3V-MPCC is relatively reduced by 0.3%, and the torque ripple rate (Tr) is relatively reduced by 0.11%; however, compared with DBQ3V-MPCC under normal parameters, the THD of FFRLS-DBQ3V-MPCC is relatively increased by 0.4%, and the Tr is relatively increased by 0.12%.

[0130] The results show that the performance of the FFRLS-DBQ3V-MPCC algorithm when the parameters are abnormal is better than that of the DBQ3V-MPCC algorithm without parameter identification, but slightly worse than that of the DBQ3V-MPCC algorithm under normal parameters. Therefore, when the parameters are abnormal, the FFRLS algorithm can accurately identify the parameters and apply them to DBQ3V-MPCC, reducing the total harmonic distortion rate of the three-phase current, thereby reducing the torque ripple.

[0131] This novel parameter identification method of forgetting factor recursive least squares - fast calculation three-vector model predictive current control method, compared with the current PI closed-loop control in the traditional voltage vector control algorithm FOC, has the advantages of being able to effectively reduce the total harmonic distortion rate of the current, improve the torque stability, and being able to effectively reduce the influence of parameter changes on the controller, improving the parameter robustness of the controller. In addition, the method does not require repeated adjustment of the controller parameters, nor does it require SVPWM modulation of the generated voltage, and it can conveniently constrain the state variables in the system, making the control simpler and faster.

[0132] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited by the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principle of the present invention shall be equivalent replacement methods and are all included in the protection scope of the present invention.

Claims

1. A method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive, characterized in that: Used to control the motor-driven magnetic wheels in the chassis of an amphibious humanoid welding robot; The motor is predictively controlled by using a three-vector model without a difference. The input of the three-vector model without a difference is the target current I*. The current parameter Pr(k) in the three-vector model without a difference is [RL s ψ f ] T The parameter identification algorithm of the forgetting factor recursive least squares method is used to estimate in real time; where R is the stator resistance of the motor; L s is the stator inductance of the motor; f is the motor rotor flux; The forgetting factor recursive least squares parameter identification algorithm is implemented based on the q-axis discretization mathematical model and the least squares recursive formula with forgetting factor; The q-axis discretization mathematical model is: Where y(k) = I q (k); Input matrix for the system; η(k)=[β q1 β q2 β q3 β q4 ] T is the parameter matrix to be identified; Among them, I q (k-1) is the q-axis current at the previous moment; I d (k) is the d-axis current at the current moment; u q (k) is the q-axis voltage at the current moment; ω e (k) is the electrical angular velocity of the rotor at the current moment; β q1 , β q2 , β q3 , β q4 are the coefficients that vary with time; The least squares recursive formula with forgetting factor is: in, is the estimated value at the current moment, is the estimated value of the previous moment; P(k) is the covariance matrix of the current moment, P(k-1) is the covariance matrix of the previous moment; L(k) is the gain matrix; E(k) is the unit matrix; The forgetting factor recursive least squares parameter identification algorithm refers to: Measure y(k); and calculate the estimated value at the current moment Covariance matrix P(k) and gain matrix L(k); When the convergence value of the parameter matrix η(k) to be identified is calculated Then calculate the motor parameter to be identified Pr(k) = [RL s ψ f ] T : The relative error err of the motor identification parameters is calculated based on the identification parameters Pr(k-1) at the previous moment and the identification parameters Pr(k) at the current moment, and whether the parameters are fully converged is determined based on the relative error err; If the parameters have fully converged, the current identification parameters Pr(k) are substituted into the three-vector model with zero-delay fast calculation for predictive control; if the parameters have not converged, the current cycle is skipped and the identification parameters of the previous cycle are substituted into the three-vector model with zero-delay fast calculation for predictive control.

2. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 1 is characterized in that: In the q-axis discretization mathematical model, the coefficient β q1 , β q2 , β q3 , β q4 and d-axis current L d , q-axis current L q , stator resistance R, motor rotor flux ψ f The mathematical relationship between and the control algorithm period T is:

3. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 1 is characterized in that: The forgetting factor recursive least squares parameter identification algorithm, when k = 1, first initializes the estimated value Initialize the covariance value P(0)=αE(k); then start the forgetting factor recursive least squares parameter identification algorithm.

4. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 1 is characterized in that: The relative error err is: If err < threshold, it means the parameters have fully converged; otherwise, it means the parameters have not converged.

5. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 1, characterized in that: The three-vector model without differential fast calculation selects an optimal vector u from six effective voltage vectors Ui, i = [1, 2, 3, 4, 5, 6] according to the target current I*. p_1 and a suboptimal vector u p_2 , and calculate the optimal vector u p_1 and the suboptimal vector u p_2 The action time; according to the optimal vector u p_1 and the suboptimal vector u p_2 and action time, acting on the inverter, causing the inverter to generate a corresponding voltage vector acting on the motor.

6. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 5 is characterized in that: In the three-vector model without differential calculation, an optimal vector u is selected from six effective voltage vectors Ui, i = [1, 2, 3, 4, 5, 6] p_1 and a suboptimal vector u p_2 , means: First, according to the dq axis current deadbeat control principle, the optimal voltage vector u at the next moment is calculated from the full voltage vector range. p_dq (k+1): u p_dq (k+1)=AI dq (k)+BI dq * +C; Among them, u p_d (k+1),u p_q (k+1) are the global optimal voltage components of the d-axis and q-axis at the next moment; I d (k) I q (k) represents the d-axis and q-axis currents collected at the current moment; Idq* represents the target current; A, B, and C represent coefficients respectively; Afterwards, the global optimal voltage vector u p_dq (k+1) is transformed into the value u in the two-phase stationary coordinate system p_αβ (k+1); judgment value u p_αβ (k+1) sector; according to the value u p_αβ The sector where (k+1) is located selects the optimal vector u p_1 and the suboptimal vector u p_2 .

7. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 6 is characterized in that: Substituting the estimated motor parameters into the three-vector model of zero-delay speed calculation means: substituting the estimated motor parameters into the calculation formulas of coefficients A, B, and C; the calculation formulas of parameters A, B, and C are respectively: Wherein, T is the control period.

8. The method for controlling the chassis of an amphibious humanoid welding robot based on magnetic wheel drive according to claim 1, characterized in that: The target current I* includes a target current d-axis component I d *、Target current q-axis component I q *; Set the target current d-axis component I d * = 0; The target current q-axis component I q * The method to obtain is: set the target speed Speed * The actual speed Speed ​​collected and processed by the encoder is input into the speed PI controller, and the output is the target current q-axis component I q *.

Images