A hot-press welding power supply temperature control method based on fuzzy active disturbance rejection

Abstract

The application discloses a hot press welding power supply temperature control method based on fuzzy self-disturbance resistance, which comprises the following steps: applying heat to a hot press welding head, measuring a temperature rising curve and a temperature falling curve, and feeding obtained data into a system identification tool to identify a mathematical model of a temperature sensor; designing an extended Kalman filter to compensate for a measured temperature, taking optimal estimated temperature and a control amount output by a control law as inputs of an extended observer, and estimating system state amounts and total disturbance of the system; comparing the temperature estimated by the extended observer with a target temperature to obtain an error signal and an error differential signal of a current temperature control system; feeding the error signal and the error change rate into a fuzzy controller to operate, obtaining an increment of control law parameter adjustment, feeding the error signal and the system state amounts estimated by the extended observer into the control law to operate on the control amount; and substituting the total disturbance of the system estimated by the extended observer into the control amount to obtain a control amount output of the system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of temperature control of hot bar soldering power supply, and particularly relates to a temperature control method of hot bar soldering power supply based on fuzzy active disturbance rejection. BACKGROUND

[0002] In recent years, hot bar soldering is widely applied to the connection of FPC (flexible circuit board) TO PCB, HSC (zebra paper) TO FPC, HSC TO LCD, TAB TO PCB and other products due to the advantages of short heating time, small heat affected zone, simultaneous multi-point welding and easy automation. The temperature control precision in the hot bar soldering process directly affects the quality of the welded product. Excessive temperature can cause the mechanical properties, creep strength and toughness of the welded joint to decrease, and low temperature can cause virtual welding and insufficient strength of the welded joint. In order to ensure the stability and consistency of product quality, accurate temperature control in the welding process plays a very important role. At present, PID algorithm is mostly used in industrial control, which has the advantages of simple structure, convenient adjustment and reliable work, and is more suitable for simple linear system control.

[0003] The hot bar soldering temperature control loop mixes time delay and nonlinear link, and has certain temperature control effect by using PID algorithm control, but it is difficult to realize rapid and accurate control. When there is large interference and delay in the system, the single PID algorithm cannot meet the control requirements of the system, which can cause large fluctuations in the control temperature and is not conducive to the formation of the welded joint. A high-precision constant temperature controller and method based on fuzzy adaptive PID control (CN201510456159.9) is the fuzzy PID temperature control algorithm in the industry at present, which can improve the temperature control precision by automatically adjusting the PID parameters through fuzzy principle. However, when there is large disturbance in the input system, the PID parameter adjustment is limited, and the error caused by the disturbance cannot be quickly and accurately eliminated. SUMMARY

[0004] In view of the technical problems existing in the prior art, the present application aims to: based on the problem that the fuzzy PID control algorithm is used in the temperature control in the background art, the response speed and precision are not high, and it is difficult to deal with a large number of nonlinearities and disturbances in the process. The present application proposes a hot bar soldering power supply temperature control method based on fuzzy active disturbance rejection, which combines fuzzy control algorithm and active disturbance rejection control algorithm, avoids the dependence of the controller on the system model, has good control effect and robustness, the controller structure is simple, the parameters are easy to adjust, and it is easy to realize. In the actual control process, the total disturbance composed of the nonlinearities, system disturbances and unmodeled parts in the system is estimated by using the extended observer, and the control law is designed to eliminate the total disturbance of the system. The fuzzy controller is used for online adjustment of the control law to further improve the control performance and anti-interference ability.

[0005] The present application is implemented by at least one of the following technologies.

[0006] A hot press welding power supply temperature control method based on fuzzy active disturbance rejection, comprising the following steps:

[0007] Step A, after continuously applying constant heat input to the hot press welding head, stop the heat input, measure the temperature rise and fall curve, and send the obtained data to the Matlab system identification tool for system identification to obtain the mathematical model of the temperature sensor;

[0008] Step B, according to the obtained temperature sensor mathematical model, design an extended Kalman filter to eliminate the time lag effect and measurement noise of the temperature sensor temperature measurement, and compensate the measured temperature;

[0009] Step C, send the optimal estimated temperature obtained by the extended Kalman filter and the control amount output by the control law to the extended observer of the fuzzy active disturbance rejection controller, for estimating the system state quantity and the total disturbance of the system;

[0010] Step D, compare the temperature estimated by the extended observer with the target temperature to obtain the error signal and the error change rate of the current temperature control system;

[0011] Step E, send the error signal and the error change rate to the fuzzy controller for operation to obtain the increment of the control law parameter adjustment, and send the error signal and the system state quantity estimated by the extended observer to the control law for operation on the control amount;

[0012] Step F, substitute the total disturbance of the system estimated by the extended observer into the control amount to eliminate the total disturbance of the system, thereby obtaining the system control amount, which controls the heat output of the hot press welding power supply power module, so that the temperature of the hot press welding head is kept within the target range.

[0013] Further, the extended observer of the fuzzy active disturbance rejection controller is a three-order linear extended state observer, and its state equation is:

[0014]

[0015] In the formula, t is time; y(t) is the optimal estimated temperature output by the extended Kalman filter; z1(t) is the estimation of y(t), z2(t) is the estimation of the differential of y(t), and z3(t) is the estimation of the total disturbance of the system; and are the differentials of z1(t), z2(t) and z3(t) respectively; e(t) is the observation error of y(t); β1, β2 and β3 are error gain parameters of the respective state equations, and L=[β1β2β3] TThe observer gain vector is formed; U(t) is the final control output of the system; b is the system control coefficient.

[0016] The observer gain vector is determined by configuring the observer poles, specifically by configuring the observer poles as s = -ω0, β1 = 3ω0. β3=ω0 3 , where ω0 represents the system bandwidth.

[0017] Furthermore, the system equations of the extended Kalman filter are as follows:

[0018]

[0019] In the formula, y(k) is the output vector; w(k) and v(k) are assumed to be uncorrelated Gaussian white noise, which are process noise and measurement noise, respectively; A is the state matrix; C is the constant output matrix; x(k) is the state vector; k is the k-th sampling time of the system;

[0020] The system matrix A, state vector x(k), output vector Q, and covariance matrix R of the extended Kalman filter are as follows:

[0021]

[0022] x(k)=[T m (k)T a (k)] T ,y(k)=T m (k)

[0023] Q = cov(w) = E{ww T}

[0024] R = cov(v) = E{vv T}

[0025] In the formula, b = 1 - a, τ is the time constant of the thermocouple sensor, τ0 is the sampling time; C is the constant output matrix; T m The temperature measured by the temperature sensor; T a y(k) represents the welding temperature; y(k) represents the output vector of the extended Kalman filter; k represents the kth sampling time of the system; w and v represent uncorrelated Gaussian white noise, which are process noise and measurement noise, respectively.

[0026] Furthermore, the inputs to the fuzzy controller are the error signal e1 and the error rate of change e2, and the output is the increment Δk of the proportional coefficient of the control law. p and the increment of the differential coefficient Δk d ;

[0027] The fuzzy inference rules of the fuzzy controller adjust the control law proportional coefficient Δk according to the size of the error signal e1 and the error change rate e2 p and the differential coefficient Δk d .

[0028] Further, the rules for adjusting the control law proportional coefficient Δk p and the differential coefficient Δk d of the fuzzy controller are: when the error signal e1 is small, Δk p is increased to improve the anti-interference performance of the system; when the error signal e1 is large, Δk d is increased to improve the system response; when the error change rate e2 is too large and the differential saturation causes over-regulation, Δk d should be reduced; according to the input error signal e1 and the error change rate e2, Δk p and Δk d are used to speed up the response of the system.

[0029] Further, the membership function of the fuzzy controller is a Gaussian membership function.

[0030] Further, the partitioning rules of the input and output universes of the fuzzy controller are: the error signal and the error change rate are greater than 20℃, which is too large, and less than 20℃, which is too small, the input fuzzy universe is set to [-20, 20], and the output fuzzy universe is set to [-10, 10].

[0031] Further, the control law is implemented using linear PD control, i.e.

[0032] u0 = k p1 (r-z1)-k d1 z2

[0033] where u0 is the preliminary calculation result of the control law, r is the target temperature, z1 is the estimate of the output y of the extended Kalman filter, z2(t) is the estimate of the differential of y, k p1 is the proportional coefficient, k d1 is the differential coefficient, the initial values of k p1 and k d1 are determined by the bandwidth method, k d1 = 2ξω c , k p1 = ω c 2 where ξ is the damping ratio and ω c is the natural resonant frequency.

[0034] Further, the control law coefficients are adjusted by the fuzzy controller output and described as:

[0035]

[0036] wherein: k p is the adjusted proportional coefficient; k d is the adjusted differential coefficient; k p1 is the initial proportional coefficient; k d1 is the initial differential coefficient; Ak p is the proportional coefficient adjustment amount; Ak d is the differential coefficient adjustment amount.

[0037] Further, the system disturbance elimination estimates the total system disturbance The system control amount u is compared with the control law calculation output u0 to obtain the final output of the system control amount u:

[0038]

[0039] wherein b0 is the estimation of the control amount coefficient b;

[0040] The final output of the control law u is described as:

[0041]

[0042] wherein: k p is the adjusted proportional coefficient; k d is the adjusted differential coefficient; e1 is the error signal; z2 is the estimation of the differential of the extended Kalman filter output y by the extended observer; z3 is the estimation of the total system disturbance by the extended observer; b0 is the estimation of the system control amount coefficient.

[0043] The system for implementing the temperature control method of the hot-press welding power supply based on fuzzy active disturbance rejection control comprises a fuzzy active disturbance rejection controller, an extended Kalman filter, a temperature sensor, and a power module.

[0044] The power module provides energy output for the hot-press welding power supply under the action of the system control amount.

[0045] The temperature sensor measures the temperature output of the hot-press welding power supply in real time and serves as the feedback data for the closed-loop control.

[0046] The extended Kalman filter is used to eliminate the noise and time lag of the temperature sampled by the temperature sensor, so that the optimal estimated temperature is as close as possible to the actual temperature.

[0047] The fuzzy active disturbance rejection controller comprises:

[0048] The extended observer estimates the system state quantity and the total system disturbance, and improves the anti-interference ability of the hot-press welding power supply system.

[0049] Fuzzy controller: according to the temperature error signal and error rate, the control law coefficient is adjusted online, which further enhances the disturbance adaptability of the hot press welding power supply system.

[0050] Control law module: the two parameters adjusted online are used to calculate the output control amount by using the linear PD formula.

[0051] Compared with the prior art, the present application has the following beneficial effects:

[0052] The present application combines fuzzy control and active disturbance rejection control, estimates and eliminates the total disturbance of the system using an extended observer when the system is disturbed, and adjusts the control law coefficient using a fuzzy controller to improve the temperature response speed and control accuracy of the hot press welding power supply in each stage of temperature control.

[0053] The fuzzy active disturbance rejection control method of the present application collects the real-time temperature output of the hot press welding power supply, transmits it to the extended Kalman filter for processing, and obtains the final control amount output of the system through the operation of the extended observer, the fuzzy controller and the control law module, which acts on the power output module, thereby realizing the control effect of high precision, fast speed and strong anti-interference performance, and greatly improving the welding stability and consistency of the hot press welding power supply. BRIEF DESCRIPTION OF DRAWINGS

[0054] Figure 1 The control structure block diagram of the hot press welding power supply temperature control method based on fuzzy active disturbance rejection in the embodiment of the present application is shown in the figure.

[0055] Figure 2 The temperature sensing curve provided in the embodiment of the present application is shown in the figure.

[0056] Figure 3 The membership degree of the input signal of the fuzzy controller provided in the embodiment of the present application is shown in the figure.

[0057] Figure 4 The membership degree of the output signal of the fuzzy controller provided in the embodiment of the present application is shown in the figure.

[0058] Figure 5 The temperature control effect comparison chart of the controller of the present application and other controllers is shown in the figure. DETAILED DESCRIPTION

[0059] In order to enable the personnel in the technical field to better understand the present application scheme, the present application will be further described in detail in combination with the drawings and specific embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by the person skilled in the art without creative labor belong to the scope of protection of the present application.

[0060] As Figure 1 shown, the temperature control system of the hot press welding power supply based on fuzzy active disturbance rejection control in the embodiment includes a fuzzy active disturbance rejection controller, an extended Kalman filter, a temperature sensor and a power module. The fuzzy active disturbance rejection controller includes an extended observer, a fuzzy controller and a control law module, so that the system can respond quickly and accurately under large disturbance; the extended Kalman filter is designed based on the transfer function of the temperature sensor, which effectively eliminates the noise and time lag in the temperature measurement of the temperature sensor; the temperature sensor is used to measure the heat output of the hot press welding power supply in real time, and is used as the parameter feedback of the temperature closed-loop control; the power module is controlled by the output of the total control amount of the fuzzy active disturbance rejection controller, so that the heat output of the hot press welding power supply is regulated, and the output temperature of the hot press welding power supply is stabilized at the set temperature value.

[0061] The temperature control method of the hot press welding power supply based on fuzzy active disturbance rejection control in the embodiment includes the following steps:

[0062] Step A, a constant heat input is applied to the hot press welding head, and after a period of time, the heat input is stopped, the temperature rising and falling curves are measured, the obtained data is sent to the Matlab system identification tool, and the mathematical model of the temperature sensor is obtained through system identification. In the embodiment, a K-type thermocouple is selected as the temperature sensor.

[0063] As a specific embodiment, a constant control amount is applied to the power module, so that the hot press welding head is heated under constant heat input, and after 400ms, the heat input is stopped. The temperature data measured by the thermocouple is recorded at a time interval of 0.001s sampling time, and the thermocouple temperature measurement curve as shown in Figure 2 is obtained, and sent to the system identification tool in Matlab, and the transfer function G(s) of the thermocouple is obtained as:

[0064]

[0065] In the formula, s is a complex variable in the transfer function.

[0066] Step B, according to the obtained mathematical model of the thermocouple, an extended Kalman filter is designed to eliminate the time lag effect and measurement noise of the thermocouple temperature measurement, and to compensate the measured temperature, including the following steps:

[0067] Step B-1, according to the mathematical model of the thermocouple obtained in step A, the change of the thermocouple temperature can be described by the following difference equation:

[0068] T m (k)=aT m (k-1)+bT a (k-1) (2)

[0069] where, b = 1 - a; τ is the time constant of the thermocouple sensor, which is 0.1 in this embodiment, and τ0is the system sampling period; T a is the temperature of the weld structure; T m is the temperature measured by the thermocouple; and k is the system's every sampling period.

[0070] Step B-2, linearize the thermocouple model to obtain the extended Kalman filter system equation

[0071]

[0072] where w(k) and v(k) are assumed to be mutually independent Gaussian white noise, which is process noise and measurement noise; A is the state matrix; C is the constant output matrix; x(k) is the state vector; and y(k) is the output vector.

[0073] Then the system matrix A, the state vector x(k), and the output vector Q and the covariance matrix R are as follows:

[0074]

[0075] x(k) = [T m (k) T a (k)] T , y(k) = T m (k) (4)

[0076] Q = cov(w) = E{ww T} (5)

[0077] R = cov(v) = E{vv T} (6)

[0078] where, b = 1 - a, τ is the time constant of the thermocouple sensor, and τ0is the sampling time; C is the constant output matrix; T m is the temperature measured by the temperature sensor; T a is the temperature of the weld structure; y(k) is the output vector; k is the system's kth sampling time; and w and v are mutually independent Gaussian white noise, which is process noise and measurement noise.

[0079] Step B-3, use the extended Kalman filter to obtain the optimal estimated temperature of the current control period, and the iteration process based on the extended Kalman filter is as follows:

[0080] (1) Calculate the prior estimated temperature and the corresponding error

[0081]

[0082]

[0083] where: is the prediction function; A k-1 is the state transition matrix; L k-1 is the noise input matrix; Q k-1 is the process noise covariance matrix; is the transpose of L k-1 .

[0084] (2) Compute the Kalman gain

[0085]

[0086] where: M k is the measurement noise input matrix; R k is the measurement noise covariance matrix; is the transpose of M k .

[0087] (3) Compute the current optimal estimate from the measurements

[0088]

[0089] where: y k represents the extended Kalman filter system equation output vector.

[0090] (4) Compute the error in the current optimal estimate

[0091] P k|k = [I - K k C] P k|k-1 (11)

[0092] where: I is the identity matrix, of the same dimension as the state vector x(k).

[0093] Discrete quantities in the process A k-1 , L k-1 , M k can be obtained using forward Euler integration:

[0094]

[0095] L k-1 = M k = T e I (13)

[0096]

[0097] where, P k|k-1are the priori estimated value and the error priori estimated value of the state variable at time k, respectively, and the calculation result in the process; P k|k 、P k-1|k-1 are the posteriori estimated value and the error posteriori estimated value of the state variable at time k and k-1, respectively, and are the optimal estimated results of the state and the error; K k is the filter gain matrix; T e is the system control period, and g is the transfer function.

[0098] The optimal estimated temperature obtained after the above process eliminates the measurement noise and time lag, and improves the reliability of the measured temperature. The optimal estimated temperature is taken as the feedback variable of the hot press bonding temperature control system, and is used for the operation in the subsequent steps.

[0099] Step C, taking the real-time temperature obtained after the processing of the extended Kalman filter and the control amount output by the control law as the input of the extended observer, the state variable of the system and the total disturbance of the system are estimated, including the following steps:

[0100] Step C-1, designing the extended observer according to the hot press bonding temperature control system, constructing a three-order linear extended state observer, which is described as:

[0101]

[0102] In the formula, y is the optimal estimated temperature output by the extended Kalman filter; z1 is the estimation of y; z2 is the estimation of the differential of y; z3 is the estimation of the total disturbance of the system; and are the differentials of z1(t), z2(t) and z3(t), respectively; e(t) is the observation error; β1, β2 and β3 are error gain parameters of the state equation, respectively, and L = [β1β2β3] T is the observer gain vector; and U(t) is the final control amount output of the system, and b is the system control amount coefficient.

[0103] In this embodiment, the observer gain vector is determined by configuring the observer pole as s = -ω0, β1 = 3ω0, β3 = ω0 3 , wherein ω0 represents the system bandwidth.

[0104] Step C-2, inputting the optimal estimated temperature of the extended Kalman filter and the system control amount output into the extended observer to obtain the temperature estimation value and the total disturbance estimation of the system, which are used for error signal calculation and total disturbance elimination of the system, respectively.

[0105] Step D, comparing the temperature estimated by the extended observer with the target temperature to obtain the error signal and the error change rate of the current temperature control system, including the following steps:

[0106] Step D-1, determine the temperature error signal e1(t) of the temperature control system as:

[0107] e1(t) = R(t) - z1(t) (16)

[0108] wherein R(t) is the target temperature, and z1(t) is the estimated temperature obtained in step C.

[0109] Step D-2, determine the temperature error change rate e2(k) as:

[0110] e2(k) = (e1(k) - e1(k-1)) / h (17)

[0111] wherein h is the system sampling interval, as a specific embodiment, h is 0.001; and e1(k) is the temperature error signal.

[0112] Step E, send the error signal into the fuzzy controller for operation to obtain the increment of the control law parameter adjustment, and send the error signal into the control law for operation on the control amount.

[0113] The design of the fuzzy controller in the step E is described as follows:

[0114] S1, determine the temperature error e1 and the temperature error change rate e2 and the output signal Δk p and Δk d In the embodiment, the temperature error and the temperature error change rate are selected as the input of the fuzzy controller, the control law proportion coefficient increment Δk p and the differential coefficient increment Δk p are selected as the output of the fuzzy controller.

[0115] S2, fuzzy domain division, it is stipulated that the temperature error and the temperature error change rate are greater than 20℃ and are too large, and are less than 20℃ and are too small, the input fuzzy domain is set as [-20, 20], and the output fuzzy domain is set as [-10, 10].

[0116] S3, fuzzy subset division, the fuzzy subsets are divided into seven language variable levels as {negative big (NB), negative medium (NM), negative small (NS), zero (ZO), positive small (PS), positive medium (PM), and positive big (PB)}.

[0117] S4, membership function selection, the membership functions of the input and output variables of the fuzzy controller are the membership functions combined by the Gaussian type and the triangular type. The membership of the fuzzy input and output rules of the fuzzy controller can be obtained as Figure 3 and Figure 4 .

[0118] S5, fuzzy inference rule determination, according to engineering debugging experience and Mamdani theory, when the error e1 is small, k should be appropriately increased p to improve the anti-interference performance of the system; when the error e1 is large, k should be appropriately increased d to improve the system response; when the error change rate e2 is too large and the differential saturation causes over-regulation, k should be appropriately reduced d ; when the input error e1 and the error change rate e2 are moderate, appropriate k p and k d should be used to speed up the response of the system. Therefore, the fuzzy control rule table 1 and table 2 are formulated.

[0119] Table 1 Δk p fuzzy rule

[0120]

[0121]

[0122] Table 2 Δk d fuzzy rule

[0123]

[0124] S6, de-fuzzification rule selection, the gravity method is used to obtain the specific output of Δk p and Δk d .

[0125] The design of the temperature control system control law in step E is described as follows:

[0126] S61, the control law is realized by linear PD control, that is,

[0127]

[0128] wherein u0 is the preliminary calculation result of the control law; r is the target temperature; z1 is the estimation of the extended Kalman filter output y, and z2(t) is the estimation of the differential of y; k p1 is the proportional coefficient; k d1 is the differential coefficient; the initial values of k p1 and k d1 can be determined by the bandwidth method, k d1 = 2ξω c , k p1 = ω c 2 , wherein ξ is the damping ratio, ω c is the natural resonant frequency;

[0129] S62, system disturbance elimination, compare the system total disturbance estimated by the extended observer in step C with the control law calculation output to obtain the final output of the system control quantity:

[0130]

[0131] Wherein, b0 is the estimation of the control quantity coefficient b.

[0132] The proportional coefficient k of the control law in step E p And the differential coefficient k d After online adjustment, it can be described as:

[0133]

[0134] Step F, substitute the system total disturbance estimated by the extended observer into the control quantity to eliminate the system total disturbance, thereby obtaining the control quantity output u of the system:

[0135]

[0136] Wherein, k p Is the adjusted proportional coefficient; k d Is the adjusted differential coefficient; e1 is the error signal; z2 is the estimation of the differential of the extended Kalman filter output y by the extended observer; z3 is the estimation of the system total disturbance by the extended observer; b0 is the estimation of the system control quantity coefficient.

[0137] The system control quantity controls the heat output of the hot press welding power supply power module, so that the temperature of the hot press welding head is kept within the target range, so as to realize accurate control of the temperature output of the hot press welding power supply.

[0138] In this embodiment, the controller parameter setting is shown in Table 3

[0139] Table 3 Controller parameter setting in this embodiment

[0140]

[0141] Through the above steps, the temperature of the hot press welding power supply is controlled, the real-time temperature output of the hot press welding power supply is collected and transmitted to the extended Kalman filter for processing, and the final control quantity output of the system is obtained through the operation of the extended observer, the fuzzy controller and the control law module, and is applied to the power output module, so as to realize the control effect of high precision, fast speed and strong anti-interference performance. It is verified by experiment that the welding stability and consistency of the hot press welding power supply are greatly improved. Figure 5 The temperature response curves of the implementation examples and other controllers are shown, and it can be concluded that the fuzzy active disturbance rejection controller proposed in this embodiment has better set value tracking ability and robustness.

[0142] Obviously, the above embodiments are merely example for clearly illustrating but not limitation to the embodiments. Based on the above description, other different forms of changes or variations can be made by those skilled in the art. Here, all the embodiments are not required to be enumerated. The obvious changes or variations derived from the above description are still within the protection scope of the present application.

Claims

1. A temperature control method for hot press welding power supply based on fuzzy active disturbance rejection, characterized in that, The method comprises the following steps: Step A, after continuously applying constant heat input to the hot press welding head, stop the heat input, measure the temperature rising and falling curves, and input the obtained data into a Matlab system identification tool to obtain a mathematical model of the temperature sensor through system identification; Step B, according to the obtained mathematical model of the temperature sensor, design an extended Kalman filter to eliminate the time lag effect and measurement noise of the temperature sensor, and compensate the measured temperature; Step C, input the optimal estimated temperature obtained through the extended Kalman filter and the control amount output by the control law into an extended observer of the fuzzy active disturbance rejection controller, to estimate the system state quantity and the total disturbance of the system; Step D, compare the temperature estimated by the extended observer with the target temperature to obtain the error signal and the error change rate of the current temperature control system; Step E, input the error signal and the error change rate into a fuzzy controller to obtain the increment of the control law parameter adjustment, and input the error signal and the system state quantity estimated by the extended observer into the control law to calculate the control amount; Step F, input the total disturbance of the system estimated by the extended observer into the control amount to eliminate the total disturbance of the system, so as to obtain the system control amount, and the system control amount controls the heat output of the hot press welding power module, so that the temperature of the hot press welding head is kept in the target range.

2. The FBD-IF thermal compression bonding power supply temperature control method according to claim 1, wherein The extended observer of the fuzzy active disturbance rejection controller is a third-order linear extended state observer, and the state equation thereof is: where t is time; y(t) is the optimal estimated temperature from the extended Kalman filter; z1(t) is the estimate of y(t), z2(t) is the estimate of the derivative of y(t), and z3(t) is the estimate of the total disturbance to the system; and are the derivatives of z1(t), z2(t), and z3(t), respectively; e(t) is the observation error of y(t); β1, β2, and β3 are the error gain parameters for each state equation, respectively, and L = [β1β2β3] T is the observer gain vector; and U(t) is the final control output of the system; and b is the control coefficient of the system. The observer gain vector is determined by configuring the observer pole, in particular by configuring the observer pole to s = -ω0, β1= 3ω0, β3= ω0 3 where ω0denotes the system bandwidth.

3. The fuzzy active disturbance rejection thermal press welding power supply temperature control method of claim 1, wherein, The system equation of the extended Kalman filter is: In the formula, y(k) is an output vector; w(k) and v(k) are assumed to be mutually independent Gaussian white noises, which are process noise and measurement noise; A is a state matrix; C is a constant output matrix; x(k) is a state vector; and k is the kth sampling time of the system; The system matrix A, the state vector x(k), the output vector Q and the covariance matrix R of the extended Kalman filter are as follows: x(k) = [T m (k)T a (k)] T y(k) = T m (k) Q = cov(w) = E{ww T} R = cov(v) = E{vv T} wherein b = 1 - a, τ is the time constant of the thermocouple sensor, τ0is the sampling time; C is the constant output matrix; T m is the temperature measured by the temperature sensor; T a is the temperature of the welded structure; y(k) is the extended Kalman filter output vector; k is the kth sampling time of the system; w and v are mutually independent Gaussian white noises, which are process noise and measurement noise.

4. The fuzzy active disturbance rejection thermal press welding power supply temperature control method of claim 1, wherein, The input of the fuzzy controller is error signal e1 and error change rate e2, and the output is control law proportional coefficient increment Δk p and differential coefficient increment Δk d ; The fuzzy inference rules of the fuzzy controller adjust the control law proportional coefficient Δk according to the magnitudes of the error signal e1 and the error rate of change e2 p and the differential coefficient Δk d .

5. The FBD thermal compression bonding power supply temperature control method of claim 4, wherein, The adjustment control law proportional coefficient Δk of the fuzzy controller p and the differential coefficient Δk d The rule is: when the error signal e1 is small, increase Δk p to improve the anti-interference performance of the system; when the error signal e1 is large, increase Δk d to improve the system response; when the error change rate e2 is too large and the differential saturation causes over-regulation, Δk d should be reduced; according to the input error signal e1 and the error change rate e2, Δk p and Δk d are used to accelerate the response of the system.

6. The FBD-TC method of claim 4, wherein, The membership function of the fuzzy controller is a Gaussian membership function.

7. The FBDT power supply temperature control method of claim 4, wherein, The partition rule of the input and output universe of the fuzzy controller is that the error signal and the error change rate are greater than 20 DEG C, which is too large, and less than 20 DEG C, which is too small, the input fuzzy universe is set to [-20, 20], and the output fuzzy universe is set to [-10, 10].

8. The fuzzy active disturbance rejection thermocompression bonding power supply temperature control method of claim 1, wherein, The control law adopts linear PD control, that is: u0 = k p1 (r - z1) - k d1 z2 where u0 is the preliminary result of the control law calculation; r is the target temperature; z1 is an estimate of the extended Kalman filter output y, and z2(t) is an estimate of the derivative of y; k p1 is a proportional coefficient; k d1 is a derivative coefficient; k p1 and the initial values of k d1 are determined by the bandwidth method, k d1 = 2ξω c , k p1 = ω c 2 where ξ is the damping ratio and ω c is the natural resonant frequency.

9. The fuzzy active disturbance rejection thermal press welding power supply temperature control method of claim 8, wherein, The control law coefficient is adjusted by the fuzzy controller and described as: where: k p is the adjusted proportional coefficient; k d is the adjusted derivative coefficient; k p1 is the initial proportional coefficient; k d1 is the initial derivative coefficient; Δk p is the proportional coefficient adjustment amount; Δk d is the derivative coefficient adjustment amount.

10. The fuzzy active disturbance rejection thermocompression bonding power supply temperature control method according to any one of claims 1-9, wherein, System disturbance cancellation by estimating the total system disturbance with an extended observer The final system control output u is compared with the control law calculation output u0: In the formula, b0 is the estimation of the control amount coefficient b; The final output u of the control law is described as: where k p is the adjusted proportional coefficient; k d is the adjusted derivative coefficient; e1 is the error signal; z2 is the extended observer's estimate of the derivative of the extended Kalman filter output y; z3 is the extended observer's estimate of the total disturbance to the system; and b0 is the estimate of the coefficient of the system control variable.

Images