Fuzzy PI controller design method based on system response characteristics

Abstract

The invention discloses a fuzzy torque PI controller design method based on system response characteristics, and the method comprises the steps: firstly carrying out the calculation analysis and visual simulation of a PI controller and a transfer function of a closed-loop control system under the control of the PI controller, and obtaining the parameter ranges of Kp and Ki, which meet the performance requirements, and the corresponding response characteristics and trend trends of the parameter ranges; secondly, establishing a fuzzy domain of a fuzzification function of fuzzy control judgment parameters and a corresponding subsection membership function; establishing a fuzzy domain of the defuzzification function and a corresponding subsection membership function; respectively setting fuzzy rules only between the error value of the controlled target and the Kp and Ki, and fuzzy rules between the error change rate of the controlled target and the Kp and Ki under cross judgment of the error change rate of the controlled target and the error; acquiring a fuzzy control judgment parameter in each control period; then the PI controller calculates to obtain Kp (t + 1) and Ki (t + 1) of the PI controller in the next control period; and finally, the PI controller receives the Kp (t + 1) and the Ki (t + 1) and updates the Kp (t + 1) and the Ki (t + 1) to existing parameters, and the updated parameters are used for controlling the controlled target in the next period.

Description

Technical Field

[0001] The invention belongs to the field of motor control, and in particular relates to a fuzzy PI controller design method based on system response characteristics. Background Art

[0002] Asynchronous motors are widely used in industrial transmission, rail transit, and electric vehicles due to their simple structure, high efficiency, strong overload capacity, strong robustness, low cost, mature manufacturing technology, and easy maintenance.

[0003] Existing motor control methods such as vector control and direct torque control mostly use traditional PI controllers to control speed, torque, current and other parameters. Traditional PI control uses fixed control parameters. As the motor runs longer, the internal temperature of the motor rises, causing the motor parameters such as the resistance and inductance of the motor stator and rotor to change, resulting in a decrease in the control effect of the original control parameters. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the present invention provides a fuzzy torque PI controller design method based on system response characteristics. The fuzzy torque PI controller design method detects the system operation status and updates the K value of the torque controller in real time. p , K i Parameters are used to address the problem of decreased control effect caused by changes in physical parameters when the motor is running for a long time.

[0005] The technical solution of the present invention to solve the above technical problems is:

[0006] A fuzzy torque PI controller design method based on system response characteristics includes the following steps:

[0007] S1: By calculating and analyzing the transfer function of the PI controller and the closed-loop control system under its control and visually simulating it, we can obtain the K that meets the performance requirements. p , K i The parameter range and its corresponding response characteristics and trends;

[0008] S2: Establish the fuzzy domain of the fuzzy function of the fuzzy control judgment parameter and the corresponding segmented membership function. p The system response, the controlled quantity error and K i The fuzzy domain of the defuzzification function and the corresponding piecewise membership function are established by the mapping relationship between the system response and the overshoot;

[0009] S3: According to the control requirements of the controlled system, K p , K i For the purpose of correcting the error and reducing the error change rate, the controlled error value and K are respectively setp , K i The fuzzy rules between the two factors of the rate of change and error value of the controlled target are used as the fuzzy rules between the judgment conditions and the parameters Kp and Ki;

[0010] S4: In each control cycle, the system obtains the fuzzy control judgment parameters, namely the controlled target error and the controlled target error change rate, through sensors or function calculation;

[0011] S5: The PI controller calculates the corresponding membership of the current fuzzy control judgment parameter according to the membership function in step S2, and calculates the K of the PI controller in the next control cycle based on the fuzzy rule in step S3 and the fuzzy reasoning method. p(t+1) , K i(t+1) ;

[0012] S6: PI controller receives the control parameter K for the next control cycle p(t+1) , K i(t+1) And update to the existing parameters, and use the updated parameters to control the controlled target in the next cycle.

[0013] Preferably, the specific process of step S1 is:

[0014] S11: Determine the physical equations between the controlled parameters of the selected control system before and after the PI controller at the position to be designed;

[0015] S12: Arrange the obtained physical equation into a differential equation form, and obtain the transfer function between the controlled parameter and the output parameter through Laplace transformation;

[0016] S13: combining the transfer function with the transfer function of the PI controller to obtain an open-loop transfer function, adding a feedback loop to form a closed-loop control system, and obtaining its closed-loop transfer function;

[0017] S14: Through theoretical calculation or simulation, the K in the PI controller is p , K i Perform response speed and overshoot analysis and determine K p , K i The interval that meets the control requirements and its internal trend, as well as the range of the controlled target error change rate when the system is in a relatively steady state and in a dynamic state.

[0018] Preferably, the specific process of step S2 is:

[0019] S21: A fuzzy domain is established according to the fluctuation range of the controlled target error, and the fuzzy domain of the controlled target error is divided into five fuzzy subsets of "NL", "NS", "Z", "PS" and "PL" according to the positive and negative rated fluctuation ranges and the fluctuation ranges in steady state, where Z is [-rated value*10%, rated value*10%]; NS is [0, rated value*40%]; PS is [-rated value*40%,0]; NL is [-rated value, -rated value*40%); PL is (rated value*40%, rated value], where the rated value is the rated value of the motor torque; the corresponding fuzzy membership function is designed according to the relationship between the control response requirement and the controlled target error;

[0020] S22: A fuzzy domain is established according to the fluctuation range of the controlled target error change rate, and the fuzzy domain of the controlled target error change rate is segmented into three fuzzy subsets of "N", "Z" and "P" according to the positive and negative rated fluctuation range and the fluctuation range under steady state, wherein N represents that the change rate is negative and the system is in the process of tending to steady state; Z represents that the change rate fluctuates within the interval belonging to the steady state of the system; P represents that the change rate is positive and the system is in the process of tending to steady state; the corresponding fuzzy membership function is designed according to the relationship between the control response demand and the controlled target error change rate;

[0021] S23: According to K p The response characteristics and available range of K are used to establish the fuzzy domain, and p According to the mapping relationship between the system response and the controlled target error, the fuzzy domain is divided into three fuzzy subsets: "SS", "SP", and "PP" according to the torque fluctuation intensity range, where SS is [0, 30%]; SP is (30%, 70%); PP is [70%, 100%]; according to the control system state's response requirements and K p According to the trend of response intensity, the defuzzified membership function corresponding to each fuzzy subset is established;

[0022] S24: According to K i The response characteristics and available range of K are used to establish the fuzzy domain, and i The mapping relationship between parameters and system response and overshoot is as follows: the fuzzy domain is divided into three subsets according to the torque overshoot range: “MAX”, “MED”, and “MIN”, where MIN is [0, 30%]; MED is (30%, 70%); MIN is [70%, 100%]; according to the control system state’s response requirements and K i According to the trend of response intensity, the defuzzified membership function corresponding to each fuzzy subset is established.

[0023] Preferably, the specific process of step S3 is:

[0024] S31: Based on the fuzzy subsets established in steps S21 to S24, a single-factor fuzzy rule for the controlled target error is established, specifically:

[0025] When the error is in the fuzzy subsets "NL", "PL", select K p Fuzzy subset "PP" and K i fuzzy subset "MAX";

[0026] When the error is in the fuzzy subsets "NS" and "PS", select K p Fuzzy subset "SP" and K i fuzzy subset "MED";

[0027] When the error is in the fuzzy subset "Z", select K p Fuzzy subsets "SS" and K i fuzzy subset "MIN";

[0028] S32: Based on the fuzzy subsets established in steps S21 to S24, establish fuzzy rules under the combined influence of the controlled target error and the error change rate, specifically:

[0029] First, the error change rate is used as a preliminary judgment of the system state, and then the current state is determined by the error;

[0030] (1) When the fuzzy subset where the error change rate is located is "Z", K i Select "MIN" for all fuzzy subsets of parameters and continue to judge the error:

[0031] If the error is in the fuzzy subsets "NL", "PL", select K p fuzzy subset "PP";

[0032] If the error is in the fuzzy subset "NS", "PS", then select K p fuzzy subset "SP";

[0033] If the error is in the fuzzy subset "Z", then select K p fuzzy subset "SS";

[0034] (2) When the error change rate is "N", "P", K p , K i The parameters need to be adjusted according to the range of the error:

[0035] If the error is in the fuzzy subsets "NL", "PL", select K p Fuzzy subset "PP" and K i fuzzy subset "MAX";

[0036] If the error is in the fuzzy subset "NS", "PS", then select K p Fuzzy subset "SP" and K i fuzzy subset "MED" or "MIN";

[0037] If the error is in the fuzzy subset "Z", then select K p Fuzzy subsets "SS" and K i fuzzy subset "MIN";

[0038] S33: Arrange the fuzzy rules obtained in step S31 and step S32 and enter them into the fuzzy rule base.

[0039] Preferably, the specific process of step S4 is:

[0040] S41: using a sensor or function estimation method to obtain a true value of the controlled target, and subtracting it from the expected value to obtain an error value;

[0041] S42: Derivative processing is performed on the error value obtained in step S41, and the difference is taken from the error value of the previous control cycle and divided by the control cycle to obtain the error change rate.

[0042] Preferably, the specific process of step S5 is:

[0043] S51: The controlled target error and error change rate obtained in step S41 and step S42 are respectively obtained through the fuzzification function in step S21 and step S22 to obtain the corresponding membership degree;

[0044] S52: Based on the membership degree obtained in step S51, the K in step S23 and step S24 are calculated after fuzzy reasoning. p , K i Fill in the defuzzification function;

[0045] S53: According to the selected defuzzification method and weight, calculate the K for the next control cycle p(t+1) , K i(t+1) .

[0046] Preferably, the specific process of step S6 is:

[0047] S61: At the beginning of each control cycle, the PI controller clears the K used in the previous control cycle. p(t) , K i(t) ;

[0048] S62: At the beginning of each control cycle, the PI controller converts the K obtained in step S53 into p(t+1) , K i(t+1) Update to the PI controller and keep the control parameters in this cycle;

[0049] S63: The PI controller participates in the system control and uses the updated PI parameters to control the controlled quantity of the system.

[0050] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0051] 1. The fuzzy torque PI controller design method based on system response characteristics detects and calculates the difference between the current torque and the target value and the rate of change of the difference as fuzzy input, and calculates the corresponding K according to the fuzzy rules. p , K i Decentralize to the PI controller to complete the parameter update of the PI controller; by detecting the system operation status, update the K of the PI controller in real time p , K i This is to address the problem of decreased control effect caused by changes in physical parameters when the motor is running for a long time.

[0052] 2. The fuzzy PI controller design method based on system response characteristics of the present invention can update the parameters of the PI controller by detecting the system response and performing the designed fuzzy control calculation processing without relying on motor parameter detection, so as to cope with the problem of decreased dynamic performance of the traditional fixed-parameter PI controller when the motor parameters change, and improve the system's anti-interference ability and overall stability.

[0053] 3. The fuzzy PI controller design method based on system response characteristics of the present invention aims at the system instability phenomenon caused by the change of system parameters. It uses the error and error change rate of the control parameters as the input of the fuzzy controller, avoids the introduction of influencing factors such as sensors or complex numerical estimation, and reduces the cost and calculation complexity.

[0054] 4. The fuzzy PI controller design method based on system response characteristics of the present invention uses fuzzy control combined with PI controller to update Kp and Ki parameters in real time according to the system status, which can make the system respond more timely and have better robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of the fuzzy PI controller design method based on system response characteristics of the present invention.

[0056] Figure 2 Schematic diagram of the fuzzy function of the controlled target error.

[0057] Figure 3 It is a schematic diagram of the fuzzy function of the error change rate of the controlled target.

[0058] Figure 4 K p Schematic diagram of the parameter defuzzification function.

[0059] Figure 5 K i Schematic diagram of the parameter defuzzification function. DETAILED DESCRIPTION

[0060] The present invention is further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0061] See also Figure 1 The fuzzy PI controller design method based on system response characteristics of the present invention comprises the following steps:

[0062] S1. By calculating and analyzing the transfer function of the PI controller and the closed-loop control system under its control and visually simulating it, we can obtain the K that meets the performance requirements. p , K i The parameter range and its corresponding response characteristics and trends;

[0063] The specific steps include:

[0064] S11, determining the physical equations between the controlled parameters before and after the PI controller of the selected control system at the position to be designed;

[0065] S12, arranging the physical equation into a differential equation form, and obtaining the transfer function between the controlled parameter and the output parameter through Laplace transformation;

[0066] S13, combining the transfer function in S12 with the transfer function of the PI controller to obtain an open-loop transfer function, adding a feedback loop to form a closed-loop control system, and obtaining its closed-loop transfer function;

[0067] S14. Through theoretical calculation or simulation, the parameter K of the PI controller is p , K i Analyze the response characteristics (response speed, overshoot) and determine K p , K i The range that meets the control requirements and its internal trend, as well as the range of the controlled target error change rate when the system is in a relatively steady state and in a dynamic state;

[0068] Taking this design as an example, we get the following K p , K i Response characteristics of parameters and stability judgment interval of controlled target error change rate:

[0069] Table 1: K p Parameter response capability and fluctuation relationship table

[0070]

[0071] Table 2: K iParameter response capability and fluctuation relationship table

[0072]

[0073]

[0074] Table 3: Stability judgment interval of controlled target error change rate

[0075] Interval Steady State [-6000,6000] Non-steady state [-25000,-6000]U[6000,25000]

[0076] Note: The steady state is taken as 1.2 times of the common steady state error change rate, and the non-steady state interval boundary is taken as 1.2 times of the extreme value of the common error change rate.

[0077] Table 4: Fluctuation range of controlled target error

[0078] Interval Maximum fluctuation [-13,13] Steady-state fluctuations [-1,1]

[0079] Note: The maximum fluctuation is calculated based on the maximum system capacity, which is the rated value here; the steady-state fluctuation takes a margin of 1.5 times the general steady-state.

[0080] S2. Establish the fuzzy domain of the fuzzy function of the fuzzy control judgment parameter (controlled target error and error change rate) and the corresponding segmented membership function, and p Parameters and system response and controlled quantity error, K i The correlation between the parameters and the system response and overshoot is used to establish the fuzzy domain of the defuzzification function and the corresponding piecewise membership function;

[0081] The specific process includes the following:

[0082] S21. According to the fluctuation range of the controlled target error, a fuzzy domain is established, and according to the positive and negative rated fluctuation range and the fluctuation range under steady state, the fuzzy domain of the controlled target error is divided into five fuzzy subsets of "NL", "NS", "Z", "PS" and "PL", which represent the positive and negative error (N-negative\P-positive) and the degree (L-large / S-small) and very small (Z-zero); according to the relationship between the control response demand and the controlled target error, the corresponding fuzzy membership function is designed;

[0083] Among them, Z is [-rated value*10%, rated value*10%]; NS is [0, rated value*40%]; PS is [-rated value*40%,0]; NL is [-rated value,-rated value*40%); PL is (rated value*40%, rated value), where the rated value is the motor torque rating;

[0084] like Figure 2 As shown, taking this design as an example, the design method is as follows:

[0085] According to the controlled target error fluctuation range in Table 4 in step S14, design each fuzzy subset range and the corresponding fuzzy membership function:

[0086] "NL", "PL": The extreme value of the error is only reached when the system is started, and a faster response is required when the system is started to reduce the error quickly. Therefore, the Gaussian function is used in the design, and the membership obtained when the error is large is higher. In this embodiment, if the input error is -13, the point corresponding to the membership function value of the fuzzy subset "NL" is 1; due to the symmetry of the system, the same is true for "PL";

[0087] "NS", "PS": The error is in the process of decreasing from a larger error value to a smaller error value. In the face of the above situation, a relatively gentle control parameter is required, and the smaller the error value, the smoother the control process needs to be. Therefore, a triangular function is used in the design, and the center of gravity is biased toward the direction of the error of 0. This design can make the membership function value larger when the error is smaller, and obtain a more stable control parameter; in this embodiment, the horizontal coordinates of the three vertices of the triangle are taken as [-6, -1, 0], half of the maximum error is taken as the starting point, and the center of gravity is designed to tend to the direction of the error of 0;

[0088] "Z": This interval is axisymmetric and is 1.5 times the peak value of the error in the system steady-state fluctuation; the center of gravity of the triangular function is taken at the horizontal coordinate 0, which means that the smaller the error, the larger the membership function value.

[0089] S22. A fuzzy domain is established according to the fluctuation range of the error change rate of the controlled target, and the fuzzy domain of the error change rate is segmented into three fuzzy subsets of "N", "Z" and "P" according to the positive and negative rated fluctuation range and the fluctuation range under steady state, which respectively represent:

[0090] N: The rate of change is negative and the system is in the process of moving toward a steady state;

[0091] Z: The rate of change fluctuates within the range belonging to the steady state of the system;

[0092] P: The rate of change is positive and the system is in the process of moving toward a steady state;

[0093] The corresponding fuzzy membership function is designed according to the relationship between the control response requirement and the error change rate of the controlled target.

[0094] Take this design as an example:

[0095] like Figure 3 As shown, according to the error change rate stability judgment interval in Table 3 in step S14, the range of each fuzzy subset and the corresponding fuzzy membership function are designed:

[0096] “Z” is the fluctuation range of the error change rate when the system is in steady state. A triangular function is used to make it gain more weight when the error change rate is close to 0, that is, closer to stability;

[0097] "N" and "P" are the fluctuation ranges of the error change rate when the system is dynamic. The membership function uses a triangular function, and the vertex abscissa is set to 1.2 times the peak value of the steady-state change rate to obtain more weight when the system change rate is the largest;

[0098] S23, according to the K in the PI controller in step S14 p The response characteristics and available range of K are used to establish the fuzzy domain, and p The mapping relationship between parameters, system response and controlled target error is shown in Table 1. The fuzzy domain is divided into three fuzzy subsets according to the torque fluctuation intensity range and mapped to the corresponding K p Parameter range: "SS", "SP", "PP", respectively represent:

[0099] SS: The response is relatively gentle; SS is [0, 30%];

[0100] SP: moderate response intensity; SP is (30%, 70%);

[0101] PP: Strong response; PP is [70%, 100%];

[0102] According to the control system state's response requirements and K in step S14 p The response intensity trend of the parameters is used to establish the defuzzification membership function corresponding to each fuzzy subset;

[0103] Take this design as an example:

[0104] like Figure 4 As shown, according to K p The parameter response capability and fluctuation relationship table, design the range of each fuzzy subset and the corresponding defuzzification membership function:

[0105] “SS”: The horizontal coordinates of the three vertices of the triangle [0.05, 0.2, 0.28], the response capacity within this interval is expressed in K p =0.2 is the minimum value, so the vertex is set to 0.2 to obtain more membership filling;

[0106] "SP": the horizontal coordinates of the three vertices of the triangle [0.005, 0.01, 0.2]. The responsiveness in this range decreases from left to right, so the center of gravity is placed on the left to obtain stronger responsiveness;

[0107] “PP”: The horizontal coordinates of the three vertices of the triangle [0.28, 0.28, 0.4]. The response capacity in this interval varies with K pThe increase is significant, so the center of gravity is shifted to 0.28.

[0108] S24, according to the controller parameter K in step S14 i The response characteristics and available range of K are used to establish the fuzzy domain, and i The mapping relationship between parameters and system response and overshoot. According to the torque overshoot range in Table 2, the fuzzy domain is divided into three subsets and mapped to the corresponding K i Parameter range: "MAX", "MED", "MIN", respectively represent:

[0109] MAX: large overshoot and strong response; MIN is [0, 30%];

[0110] MED: Overshoot and response intensity are moderate; MED is (30%, 70%);

[0111] MIN: minimum overshoot and weakest response; MIN is [70%, 100%];

[0112] According to the control system state's response requirements and K in step S14 i The response intensity trend of the parameter is used to establish the defuzzification membership function corresponding to each fuzzy subset; for example:

[0113] like Figure 5 As shown, according to K i The parameter response capability and fluctuation relationship table, design the range of each fuzzy subset and the corresponding defuzzification membership function:

[0114] "MIN": The horizontal coordinates of the three vertices of the triangle [6,7,8], the responsiveness of this interval is expressed in K i =7 is the minimum value, so the vertex is set to 7 to obtain more membership filling;

[0115] “MED”: The horizontal coordinates of the three vertices of the triangle are [2,7,7]. The responsiveness in this interval decreases from right to left, so the center of gravity is placed on the right to make the responsiveness gradually weaken;

[0116] "MAX": The horizontal coordinates of the three vertices of the triangle are [0, 0.01, 2]. The response ability in this range is relatively strong. i =0.01 is the weakest and gradually becomes stronger on both sides, so the center of gravity is shifted toward 0.01 to obtain a comprehensive stronger response capability.

[0117] S3. According to the control requirements of the controlled system, K p , K i The purpose of the parameters is to correct the error and reduce the error change rate. The controlled error value and K p , K iThe fuzzy rules between parameters, and the error change rate of the controlled target and the error cross judgment with K p , K i The fuzzy rules between the parameters, that is, to set up the error value and K when only considering the error value of the controlled target p , K i The fuzzy rules between parameters, and the error value and error change rate when the controlled target error value and error change rate are considered comprehensively with K p , K i Fuzzy rules between parameters;

[0118] The specific process includes the following:

[0119] S31, based on the fuzzy subsets established in step S21 to step S24, a single factor fuzzy rule for the controlled target error is established, wherein the design method is as follows:

[0120] When the error is large, that is, the error is in the fuzzy subsets "NL" and "PL", the system needs a larger correction strength, so the K with a stronger response is selected. p Fuzzy subset "PP" and K i fuzzy subset "MAX";

[0121] When the error is small, that is, the error is in the fuzzy subsets "NS" and "PS", the system needs a moderate correction strength, so K with a medium response strength is selected. p Fuzzy subset "SP" and K i fuzzy subset "MED";

[0122] When the error is within the steady-state fluctuation range, that is, the error is within the fuzzy subset "Z", the system needs to maintain a steady state and does not require much correction, so the K with the smallest response intensity is selected. p Fuzzy subsets "SS" and K i fuzzy subset "MIN";

[0123] Table 5: Single factor fuzzy rule table

[0124]

[0125] S32, based on the fuzzy subsets established in steps S21-S24, establish fuzzy rules under the combined influence of the controlled target error and the error change rate, wherein the design method is as follows:

[0126] First, the error change rate is used as a preliminary judgment of the system state, and then the current state is determined by the error;

[0127] When the error change rate is in the fuzzy subset "Z", the system is likely to be in a steady state, so it is necessary to suppress K i Overshoot and oscillation caused by parameters, K iSelect "MIN" for all parameter fuzzy subsets. Continue to judge the error:

[0128] If the error is in the fuzzy subsets "NL" and "PL", the system is in the initial stage of transient change. Therefore, K with a stronger response is selected. p fuzzy subset "PP";

[0129] If the error is in the fuzzy subsets "NS" and "PS", the system is in the middle of transient change. Therefore, K with medium response intensity is selected. p fuzzy subset "SP";

[0130] If the error is in the fuzzy subset "Z", the system is in a steady state. To maintain the steady state, no more corrections are needed, so the K with the smallest response intensity is selected. p fuzzy subset "SS";

[0131] When the error change rate is "N" or "P", the system is likely to be in a transient state. p , K i The parameters need to be adjusted according to the range of the error:

[0132] If the error is in the fuzzy subsets "NL" and "PL", the system is in the initial stage of transient change, and the controlled target is far from the target value. Therefore, K with stronger response strength is selected. p Fuzzy subset "PP" and K i fuzzy subset "MAX";

[0133] If the error is in the fuzzy subsets "NS" and "PS", the system is in the middle of change, the controlled target is close to the target value, but it needs to slow down to prevent overshoot. Therefore, a K with medium response strength is selected. p Fuzzy subset "SP" and K i fuzzy subset "MED" or "MIN";

[0134] If the error is in the fuzzy subset "Z", the system is in the final stage of transient change or the stage of approaching the target, and the controlled target is very close to the target value. Therefore, the K with the smallest response intensity is selected. p Fuzzy subsets "SS" and K i fuzzy subset "MIN";

[0135] Table 6: Fuzzy rules table under the influence of controlled target error and error change rate

[0136]

[0137] S33. Arrange the fuzzy rules obtained in S31 and S32 and enter them into the fuzzy rule base.

[0138] S4. In each control cycle, the system obtains the fuzzy control judgment parameters, namely the controlled target error and the controlled target error change rate, through sensors or function calculation.

[0139] The specific process includes the following

[0140] S41, using a sensor or function estimation method to obtain the true value of the controlled target, and subtracting it from the expected value to obtain an error value;

[0141] S42, performing derivative processing on the error value obtained in S41, subtracting it from the error value of the previous control cycle and dividing it by the control cycle to obtain the error change rate.

[0142] S5, the fuzzy controller calculates the corresponding membership of the current judgment parameter according to the membership function in S2, and calculates the PI controller parameter K of the next control cycle based on the fuzzy rule in S3 and the fuzzy reasoning method. p(t+1) , K i(t+1) .

[0143] The specific process includes the following:

[0144] S51, the controlled target error and error change rate obtained in step S41 and step S42 are respectively obtained through the fuzzification function in step S21 and step S22 to obtain the corresponding membership degree;

[0145] S52, according to the membership degree obtained in step S51, after fuzzy reasoning, the K in step S23 and step S24 are p , K i Fill in the defuzzification function;

[0146] S53, according to the selected defuzzification method and weight, calculate the K for the next control cycle p , K i Parameter K p(t+1) , K i(t+1) .

[0147] S6. The PI controller receives the control parameter K of the next control cycle obtained in step S4. p(t+1) , K i(t+1) And update to the existing parameters, and use the updated parameters to control the controlled target in the next cycle.

[0148] The specific process includes the following:

[0149] S61, PI controller clears the K used in the previous control cycle at the beginning of each control cycle. p(t) , K i(t) parameter;

[0150] S62, PI controller, at the beginning of each control cycle, converts the K obtained in S53 into p(t+1) , K i(t+1) Update to the PI controller and keep the control parameters in this cycle;

[0151] S63, the PI controller participates in the system control and uses the updated PI parameters to control the system controlled quantity.

[0152] In this embodiment, steps S1 to S3 are the preliminary design phase of the fuzzy PI controller. After the design is completed, the control system repeats steps S4 to S6 to complete the closed-loop operation.

[0153] The present invention utilizes K in PI controller p , K i Starting from the effect of the size of parameters individually and in combination on the response performance of the control system, a variable parameter PI controller based on a fuzzy controller is designed according to the error between the controlled parameter and the target value and the error change rate.

[0154] The present invention utilizes K p , K i The effect of parameters on system response and the K of PI controllers available when output constraints are present p , K i Based on the changing trend of the influence of parameters on the system within the parameter range, the fuzzy rules in the fuzzy controller are designed to improve the responsiveness and anti-interference ability of the PI controller.

[0155] The above is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited to the above content. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A fuzzy torque PI controller design method based on system response characteristics, characterized in that: The following steps are involved: S1: By calculating and analyzing the transfer function of the PI controller and the closed-loop control system under its control and visually simulating it, we can obtain the K that meets the performance requirements. p , K i The parameter range and its corresponding response characteristics and trends; S2: Establish the fuzzy domain of the fuzzy function of the fuzzy control judgment parameter and the corresponding segmented membership function. p The system response, the controlled quantity error and K i The fuzzy domain of the defuzzification function and the corresponding piecewise membership function are established by the mapping relationship between the system response and the overshoot; S3: According to the control requirements of the controlled system, K p , K i For the purpose of correcting the error and reducing the error change rate, the controlled error value and K p , K i The fuzzy rules between the two factors of the rate of change and error value of the controlled target are used as the fuzzy rules between the judgment conditions and the parameters Kp and Ki; S4: In each control cycle, the system obtains the fuzzy control judgment parameters, namely the controlled target error and the controlled target error change rate, through sensors or function calculation; S5: The PI controller calculates the corresponding membership of the current fuzzy control judgment parameter according to the membership function in step S2, and calculates the K of the PI controller in the next control cycle based on the fuzzy rule in step S3 and the fuzzy reasoning method. p(t+1) , K i(t+1) ; S6: PI controller receives the control parameter K for the next control cycle p(t+1) , K i(t+1) And update to the existing parameters, and use the updated parameters to control the controlled target in the next cycle.

2. The fuzzy torque PI controller design method based on system response characteristics according to claim 1 is characterized in that: The specific process of step S1 is as follows: S11: Determine the physical equations between the controlled parameters of the selected control system before and after the PI controller at the position to be designed; S12: Arrange the obtained physical equation into a differential equation form, and obtain the transfer function between the controlled parameter and the output parameter through Laplace transformation; S13: combining the transfer function with the transfer function of the PI controller to obtain an open-loop transfer function, adding a feedback loop to form a closed-loop control system, and obtaining its closed-loop transfer function; S14: Through theoretical calculation or simulation, the K in the PI controller is p , K i Perform response speed and overshoot analysis and determine K p , K i The interval that meets the control requirements and its internal trend, as well as the range of the controlled target error change rate when the system is in a relatively steady state and in a dynamic state.

3. The fuzzy torque PI controller design method based on system response characteristics according to claim 2 is characterized in that: The specific process of step S2 is: S21: A fuzzy domain is established according to the fluctuation range of the controlled target error, and the fuzzy domain of the controlled target error is divided into five fuzzy subsets of "NL", "NS", "Z", "PS" and "PL" according to the positive and negative rated fluctuation ranges and the fluctuation ranges in steady state, where Z is [-rated value*10%, rated value*10%]; NS is [0, rated value*40%]; PS is [-rated value*40%,0]; NL is [-rated value, -rated value*40%); PL is (rated value*40%, rated value], where the rated value is the motor torque rated value; the corresponding fuzzy membership function is designed according to the relationship between the control response requirement and the controlled target error; S22: A fuzzy domain is established according to the fluctuation range of the controlled target error change rate, and the fuzzy domain of the controlled target error change rate is segmented into three fuzzy subsets of "N", "Z" and "P" according to the positive and negative rated fluctuation range and the fluctuation range under steady state, wherein N represents that the change rate is negative and the system is in the process of tending to steady state; Z represents that the change rate fluctuates within the interval belonging to the steady state of the system; P represents that the change rate is positive and the system is in the process of tending to steady state; the corresponding fuzzy membership function is designed according to the relationship between the control response demand and the controlled target error change rate; S23: According to K p The response characteristics and available range of K are used to establish the fuzzy domain, and p According to the mapping relationship between the system response and the controlled target error, the fuzzy domain is divided into three fuzzy subsets according to the torque fluctuation intensity range: "SS", "SP", and "PP", where SS is [0, 30%]; SP is (30%, 70%); PP is [70%, 100%]; according to the control system state's response requirements and K p According to the trend of response intensity, the defuzzified membership function corresponding to each fuzzy subset is established; S24: According to K i The response characteristics and available range of K are used to establish the fuzzy domain, and i The mapping relationship between parameters and system response and overshoot is as follows: the fuzzy domain is divided into three subsets according to the torque overshoot range: "MAX", "MED", and "MIN", where MIN is [0, 30%]; MED is (30%, 70%); MIN is [70%, 100%]; according to the control system state's response requirements and K i According to the trend of response intensity, the defuzzified membership function corresponding to each fuzzy subset is established.

4. The fuzzy torque PI controller design method based on system response characteristics according to claim 3 is characterized in that: The specific process of step S3 is: S31: Based on the fuzzy subsets established in steps S21 to S24, a single-factor fuzzy rule for the controlled target error is established, specifically: When the error is in the fuzzy subsets "NL" and "PL", select K p Fuzzy subset "PP" and K i Fuzzy subset "MAX"; When the error is in the fuzzy subsets "NS" and "PS", select K p Fuzzy subset "SP" and K i Fuzzy subset "MED"; When the error is in the fuzzy subset "Z", select K p Fuzzy subsets "SS" and K i Fuzzy subset "MIN"; S32: Based on the fuzzy subsets established in steps S21 to S24, establish fuzzy rules under the combined influence of the controlled target error and the error change rate, specifically: First, the error change rate is used as a preliminary judgment of the system state, and then the current state is determined by the error; (1) When the fuzzy subset where the error change rate is located is "Z", K i Select "MIN" for all fuzzy subsets of parameters and continue to judge the error: If the error is in the fuzzy subsets "NL" and "PL", select K p Fuzzy subset "PP"; If the error is in the fuzzy subsets "NS" and "PS", select K p fuzzy subset "SP"; If the error is in the fuzzy subset "Z", then select K p Fuzzy subset "SS"; (2) When the error change rate is "N", "P", K p , K i The parameters need to be adjusted according to the range of the error: If the error is in the fuzzy subsets "NL" and "PL", select K p Fuzzy subset "PP" and K i Fuzzy subset "MAX"; If the error is in the fuzzy subsets "NS" and "PS", select K p Fuzzy subset "SP" and K i Fuzzy subset "MED" or "MIN"; If the error is in the fuzzy subset "Z", then select K p Fuzzy subsets "SS" and K i Fuzzy subset "MIN"; S33: Arrange the fuzzy rules obtained in step S31 and step S32 and enter them into the fuzzy rule base.

5. The fuzzy torque PI controller design method based on system response characteristics according to claim 4 is characterized in that: The specific process of step S4 is as follows: S41: using a sensor or function estimation method to obtain a true value of the controlled target, and subtracting it from the expected value to obtain an error value; S42: Derivative processing is performed on the error value obtained in step S41, and the difference is taken from the error value of the previous control cycle and divided by the control cycle to obtain the error change rate.

6. The fuzzy torque PI controller design method based on system response characteristics according to claim 4 is characterized in that: The specific process of step S5 is as follows: S51: The controlled target error and error change rate obtained in step S41 and step S42 are respectively obtained through the fuzzification function in step S21 and step S22 to obtain the corresponding membership degree; S52: Based on the membership degree obtained in step S51, the K in step S23 and step S24 are calculated after fuzzy reasoning. p , K i Fill in the defuzzification function; S53: According to the selected defuzzification method and weight, calculate the K for the next control cycle p(t+1) , K i(t+1) .

7. The fuzzy torque PI controller design method based on system response characteristics according to claim 6 is characterized in that: The specific process of step S6 is as follows: S61: At the beginning of each control cycle, the PI controller clears the K used in the previous control cycle. p(t) , K i(t) ; S62: At the beginning of each control cycle, the PI controller converts the K obtained in step S53 into p(t+1) , K i(t+1) Update to the PI controller and keep the control parameters in this cycle; S63: The PI controller participates in the system control and uses the updated PI parameters to control the controlled quantity of the system.

Images