JP2011138200A - Adaptive controller - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an adaptive controller capable of coping promptly with changes in load and changes in characteristics of a control object and achieving more excellent responsive characteristic. <P>SOLUTION: The adaptive controller includes: a loop gain adjuster 7, an adaptive controlling device 5, and a PI controller 3. In a variable gain calculator 11 of the adaptive controlling device 5, which is provided with a multiplier 16 that squares control deviation e<SB>a</SB>(t), and a primary delay element 17 that sets output from the multiplier 16 as input, when the variation inclination of the control deviation e<SB>a</SB>(t) is negative, a time constant of a primary delay element 17 is set as a larger value than a time constant of the primary delay element 17 when the variation inclination of the control deviation e<SB>a</SB>(t) is constant or positive. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an adaptive control apparatus based on a simple adaptive control system, and more particularly to an adaptive control apparatus that can immediately respond to a load change or a characteristic change of a control target and can realize a better response characteristic.

  In recent years, a control method has been proposed in which adaptive control for estimating (part of) parameters of a control target while stabilizing the control system is applied to a control target of a plant or servo system (for example, Japanese Patent Laid-Open No. 8-254136). No. publication). In particular, Non-Patent Document 1 and Non-Patent Document 2 are proposed as processes that are controlled, and a simplified control system based on the simple adaptive control system.

  This conventional simple adaptive control method (for example, Non-Patent Document 1) is a simplified configuration in which the variable gain is determined only by the minimum required control deviation, and is related to the control deviation. Due to the variable gain based on the square integral value of the obtained value, when the peak value of the control deviation is large, the large gain state is maintained and active control is performed. As a result, the peak value of the control deviation is reduced, and at the same time, the gain is returned to the original small value to stabilize the steady state.

  In the techniques disclosed in Non-Patent Document 1 and Non-Patent Document 2, in a transient state immediately after a disturbance or a transient state immediately after a target value is changed, a gain is obtained by a variable gain based on a square integral value of a control deviation. Is increased to suppress the control deviation, and when the steady state is reached, the gain is returned to the original small value to ensure a stable state. That is, in the variable gain calculation unit, the control deviation is squared, and the value obtained through the first-order lag element is used as the variable gain, but the time constant in the first-order lag element is the total time of the process to be controlled. The value is based on an estimated value of a constant (including a dead time value). As a result, a transient response characteristic with excellent followability is obtained, and as soon as the control deviation converges to zero, the variable gain also converges to the minimum set value.

JP-A-8-254136

Toshikatsu Fujiwara et al., High Functional Control System “SNAC” and its Application to Thermal Power Plant Steam Temperature Control, Mitsubishi Heavy Industries Technical Report, Vol 31, No. 6 (1994.11), pp. 388-391 Toshikatsu Fujiwara, Simple Adaptive Control (SAC) and its Application to Boiler Plant Steam Temperature Control, 34th Annual Conference of the Society of Instrument and Control Engineers SICE '95 in Sapporo Proceedings, 307 C-3 (1995), pp. 867-868

  By the way, in process control, even after the control amount converges to a set value and the control deviation converges to zero, a load change or a characteristic change of a process to be controlled often occurs. In response to such a load change or process characteristic change, there is a situation that it takes time to respond from a variable gain converged to a minimum set value.

  The present invention has been made in view of such circumstances, and is to provide an adaptive control device that can immediately respond to a load change or a characteristic change of a control target and can realize a better response characteristic. Objective.

In order to solve the above problems, the present invention employs the following means.
An adaptive control apparatus according to the present invention includes a loop gain adjusting means for adjusting a steady gain of the closed loop control system to 1 by using a deviation between an output of a control target and a target value as input, and an output of the closed loop control system. Feedback means for returning to the output side of the loop gain adjusting means through the phase delay element as an input, and calculating a variable gain using the control deviation between the output of the phase delay element and the output of the loop gain adjusting means as an input, A closed loop control system comprising: a variable gain calculating means for compensating a gain of the closed loop control system; a PI control means for performing PI control with an output of the loop gain adjusting means as an input; an output of the closed loop control system and the PI control Adding means for adding the output of the means and giving the control object as an operation amount, the adaptive control device comprising the variable gain The calculating means includes a calculating means for squaring the control deviation, and a first-order lag element having the output of the calculating means as an input, and the time constant of the first-order lag element according to the fluctuation slope of the control deviation. It is characterized by variable setting.

  According to the present invention, since the time constant of the first-order lag element is variably set according to the fluctuation slope of the control deviation in the variable gain calculating means, the time constant of the first-order lag element is set to a larger value and the minimum setting of the variable gain is made. It is possible to delay the convergence to a value and to respond immediately to a load change or a characteristic change of a control target that can occur after the control deviation converges to zero, and to realize a better response characteristic.

  The present invention also provides coefficient means for multiplying the deviation between the output of the control target and the target value by a predetermined coefficient, a phase delay element that receives the output of the coefficient means, and the output of the closed loop control system as input. Feedback means for returning via an element, variable gain calculation means for calculating a variable gain by inputting a control deviation between the output of the phase delay element and the output of the phase delay element, and compensating the gain of the closed loop control system, A closed loop control system comprising: a PID control means for performing PID control with the deviation as an input; an addition means for adding the output of the closed loop control system and the output of the PID control means as an operation amount to the control object; The variable gain calculation means includes calculation means for squaring the control deviation, and a first-order lag element having the output of the calculation means as input. The provided, characterized by variably set according to the time constant of the primary delay element to change the inclination of the control deviation.

  According to the present invention, since the time constant of the first-order lag element is variably set according to the fluctuation slope of the control deviation in the variable gain calculating means, the time constant of the first-order lag element is set to a larger value and the minimum setting of the variable gain is made. It is possible to delay the convergence to a value and to respond immediately to a load change or a characteristic change of a control target that can occur after the control deviation converges to zero, and to realize a better response characteristic.

  In the adaptive control apparatus according to the present invention, the variable gain calculation means may use a time constant of a first-order lag element when the fluctuation slope of the control deviation is negative, to make the fluctuation slope of the control deviation constant or positive. It is characterized in that it is set to a value larger than the time constant of the first-order lag element at.

  According to the present invention, in the variable gain calculation means, the time constant of the first-order lag element when the fluctuation slope of the control deviation is negative is the time constant of the first-order lag element when the fluctuation slope of the control deviation is constant or positive. Since it is set to a larger value, the convergence time to the minimum setting value of the variable gain is lengthened, and it is possible to respond quickly to load changes and control target characteristic changes that may occur after zero convergence of the control deviation. Response characteristics can be realized.

  According to the present invention, in the variable gain calculation means for compensating the gain of the closed loop control system, the time constant of the first-order lag element is variably set according to the fluctuation slope of the control deviation. Since the convergence to the minimum setting value of the variable gain can be delayed as a larger value, it is possible to respond immediately to load changes and control target characteristic changes that may occur after the control deviation has converged to zero, realizing a superior response characteristic There is an effect that can be done.

It is a block diagram of the adaptive control apparatus which concerns on 1st Embodiment of this invention, (a) is a schematic block diagram of the whole adaptive control apparatus, (b) is a schematic block diagram of an adaptive controller. It is a detailed block diagram of the various functional blocks of the adaptive controller of 1st Embodiment of this invention, (a) is a block diagram of the circuit which produces | generates the time constant of the primary delay element in a variable gain calculating part, (b). Is a detailed configuration diagram of a variable gain calculation unit, and (c) is a detailed configuration diagram of a phase delay element. It is explanatory drawing which illustrates the step response of the plant output of 1st Embodiment of this invention. It is explanatory drawing which illustrates the step response of the plant input of 1st Embodiment of this invention. It is explanatory drawing which illustrates the step response of the variable gain of the variable gain calculating part of 1st Embodiment of this invention. It is explanatory drawing which illustrates the disturbance response of 1st Embodiment of this invention, (a) illustrates plant output y (t), (b) illustrates a disturbance input, respectively. It is explanatory drawing which illustrates the disturbance response of 1st Embodiment of this invention, (a) illustrates the variable gain K (t) of a variable gain calculating part, (b) illustrates the plant input u (t), respectively. . It is a block diagram of the adaptive control apparatus which concerns on 2nd Embodiment of this invention, (a) is a schematic block diagram of the whole adaptive control apparatus, (b) is a schematic block diagram of an adaptive controller. It is a detailed block diagram of the various functional blocks of the adaptive controller of 2nd Embodiment of this invention, (a) is a block diagram of the circuit which produces | generates the time constant of the primary delay element in a variable gain calculating part, (b). Is a detailed configuration diagram of a variable gain calculation unit, and (c) is a detailed configuration diagram of a phase delay element.

[First Embodiment]
Hereinafter, an adaptive control apparatus according to a first embodiment of the present invention will be described.
FIG. 1 is a block diagram of an adaptive control apparatus according to the first embodiment of the present invention. FIG. 1A shows a schematic block diagram of the entire adaptive control apparatus, and FIG. 1B shows a schematic block diagram of the adaptive controller 5.

  In FIG. 1A, the adaptive control apparatus of the present embodiment is roughly configured to include a PI controller 3 and a closed loop control system for a plant 1 that is a control target. The PI controller 3 generates a control amount based on proportional control (P control) and integral control (I control) for the plant 1. Further, the output y (t) of the plant 1 is fed back, and the deviation e (t) between the target value r (t) and the output y (t) of the plant 1 is obtained by the subtracter 8.

In FIG. 1A, the closed loop control system includes a loop gain adjuster 7 and an adaptive controller 5. The loop gain adjuster 7 receives the deviation e (t) between the output y (t) of the plant 1 and the target value r (t) as input and adjusts the steady gain of the closed loop control system to be 1.
The loop gain adjuster 7 is realized by a coefficient unit. The coefficient α is such that the output e ^* (t) of the coefficient unit is steady when the manipulated variable u (t) for the plant 1 is increased by a value of 1. It is set so as to have a value of approximately −1 in the state.

Further, as shown in FIG. 1B, the closed loop control system adaptive controller 5 sends the output u _a (t) of the closed loop control system to the output side of the loop gain adjuster 7 via the phase delay element 13. The variable gain K (t) is calculated by using the feedback portion to be returned, and the control deviation e _a (t) between the output y _f (t) of the phase delay element 13 and the output e ^* (t) of the loop gain adjuster 7 as input. And a variable gain calculator 11 for compensating for the gain of the closed loop control system.

Reference numeral 20 is a subtractor for obtaining _a control deviation e _a (t) between the output y _f (t) of the phase delay element 13 and the output e ^* (t) of the loop gain adjuster 7. This is a multiplier that multiplies the control deviation e _a (t) by the variable gain K (t) that is the output of the variable gain calculator 11. Accordingly, if the transfer characteristic of the phase delay element 13 is F ₁ (s), the transfer characteristic Gc () between the output e ^* (t) of the loop gain adjuster 7 and the output u _a (t) of the closed loop control system. s) is represented by Gc (s) = K (s) / (1 + K (s) · F ₁ (s)). Here, s is a Laplace operator.

Various characteristics can be considered as the transfer characteristic F ₁ (s) of the phase delay element 13, but here, in order to obtain the effect of the second derivative, it is assumed that it is given by the following expression.

F ₁ (s) = 1 / 2K _max (1 / (1 + τs) + 1 / (1 + τs) ² )) (1)

Here, K _max is the maximum value of the steady gain between the output e ^* (t) of the loop gain adjuster 7 and the output u _a (t) of the closed loop control system.

  By substituting the equation (1) into the equation of the transfer characteristic Gc (s) and changing K (s) → ∞, the relationship of the following equation is obtained.

Gc (s) = K _max (1 + τs) ² /(1+0.5τs) (2)

Here, the values of K _max and τ are specified values, and the adaptive controller 5 does not exceed K _max even if the value of the variable gain K (s) becomes infinite. Similarly, if K (s) → K _min (K _min << K _max ), the relationship of Gc (s) ≈K _min is obtained, and the adaptive controller 5 has a small value of the variable gain K (s). As a result, the proportional operation with the gain K _min is performed. The value of K _min is a specified value.

Next, in FIG. 1B, the variable gain calculation unit 11 includes a coefficient unit 15, a multiplier 16, a first-order lag element 17, a coefficient unit 18, and an adder 19. That is, the variable gain K (t) is obtained by multiplying the control deviation e _a (t) by the coefficient γ by the coefficient unit 15, squares it by the multiplier 16, and passes this through the first-order lag element 17 of the time constant μ. It is calculated by adding the coefficient β / K _max of the coefficient unit 18 to the obtained value. The calculation of the variable gain K (t) in the variable gain calculator 11 is expressed as follows.

K (t) = (μ / (μs + 1)) · (γe _a (t)) ² + β / K _max (3)

Here, when d / dt | e _a (t) | <0, μ = μ ₁₁
When d / dt | e _a (t) | ≧ 0 μ = μ ₁₂ (μ ₁₂ <μ ₁₁ )

As shown in the equation (3), in the adaptive control device of this embodiment, the time constant μ of the first-order lag element 17 in the variable gain calculation unit 11 is variably set according to the fluctuation slope of the control deviation e _a (t). There is a feature in the point. That is, when the fluctuation slope d / dt | e _a (t) | of the control deviation e _a (t) is negative, the time constant of the first-order lag element 17 is μ = μ _11, and this is the control deviation e _a (t ) Fluctuation slope d / dt | e _a (t) | is constant or positive and is set to a value larger than the time constant μ ₁₂ of the first-order lag element 17.

FIG. 2A is a configuration diagram of a circuit that generates the time constant μ (1 / μ) of the first-order lag element 17 in the variable gain calculation unit 11. Time constant 31 constant _{mu 11,} the time constant _{mu 12} constants 32, selector switch 33, a differentiator 35, the switching signal generator 36, a configuration in which a constant 38 of the divider 37 and "1". That is, the differentiator 35 obtains the fluctuation slope d / dt | e _a (t) | of the control deviation e _a (t), and the switching signal generator 36 determines negative, positive, or zero to generate a switching signal, Based on the switching signal, the output of either the constant unit 31 having the time constant μ ₁₁ or the constant unit 32 having the time constant μ ₁₂ is selected. As will be described later, since 1 / μ is used for the first-order lag element 17, 1 / μ is obtained by the divider 37.

  FIG. 2B shows a detailed configuration diagram of the variable gain calculation unit 11. In FIG. 2B, the first-order lag element 17 includes an integrator 21, a coefficient unit 22 having a gain (coefficient) 1 / μ, and a subtracter 23. In FIG. 2B, an adder 24 is provided between the variable gain calculator 11 and the adder 19, and a signal (variable gain) Ke (t) of another control system is added. Here, the other control system is another control system in the case where the control target is composed of two or more variable systems, and such a configuration makes it possible to apply to a multivariable control system.

The coefficient unit 15 having the gain γ desirably obtains the gain (coefficient) γ as follows. That is, the reciprocal of the maximum deviation estimated value e ^* (t) _max of the output e ^* (t) of the loop gain adjuster 7 was obtained via the first-order lag element (1 / (1 + σs)) having the time constant σ. Value. Here, the time constant σ is approximately equal to the value of mu _12.

FIG. 2C shows a detailed configuration diagram of the phase delay element 13 described above. The phase delay element 13 includes a coefficient unit 41, a first primary delay element 42, a second primary delay element 43, and an adder 44. The output u _a (t) of the multiplier 10 is multiplied by 1 / 2K _max by the coefficient unit 41, the output is input to the first primary delay element 42, and the output is further input to the second primary delay element 43. input. Then, the outputs of the first primary delay element 42 and the second primary delay element 43 are added by the adder 44, and the addition result becomes the output y _f (t) of the phase delay element 13. Note that the time constant tau in the first primary delay element 43 and the second primary delay element 44, Ru can be specified from the outside (e.g. is set to τ = 0.2μ _12).

  Next, simulation results of step response and disturbance response by the adaptive control apparatus of this embodiment are illustrated in FIGS. 3 to 6 to verify the effect of the present invention. Here, FIG. 3 is an explanatory diagram illustrating the step response of the plant output y (t) of this embodiment, FIG. 4 is an explanatory diagram illustrating the step response of the plant input u (t), and FIG. It is explanatory drawing which illustrates the step response of the variable gain K (t) of the variable gain calculating part 11. 6 and 7 are explanatory diagrams illustrating the disturbance response of this embodiment. FIG. 6 (a) shows the plant output y (t), FIG. 6 (b) shows the disturbance input, and FIG. 7 (a). Indicates the variable gain K (t) of the variable gain calculator 11, and FIG. 7 (b) indicates the plant input u (t).

First, as shown in FIG. 3, the step response is performed for a step input in which the target value r (t) is set to “1” at a certain timing. In the conventional example, the time constant mu variable gain calculation unit 11 in the first-order lag element 17, is constant at mu = mu ₁₂ regardless of variations inclination of the control deviation e _{a (t).} On the other hand, in the present invention, the fluctuation slope of the control deviation e _a (t) is negative until the plant output y (t) first reaches the target value r (t), and the time constant μ is set to a larger μ = μ ₁₁ , and thereafter, until the peak point of the overshoot is reached, the fluctuation slope of the control deviation e _a (t) is positive, and the time constant μ is smaller μ = μ Is set to ₁₂ , and the time constant of the first-order lag element is switched at the point where the target value r (t) is reached and the peak point of overshoot or undershoot.

  As shown in FIG. 5, the variable gain K (t) of the variable gain calculation unit 11 is set to be larger than the conventional value, and gradually decreases as the deviation e (t) converges (compared to the conventional value). Converging to the set value. As a result, the step response of the plant output y (t) and the plant input u (t) can be quickly responded by the change in the target value r (t), the overshoot is suppressed, and the settling time is shortened. As a result, it was confirmed that step response characteristics superior to those of the prior art can be realized.

Further, as shown in FIG. 6B, the disturbance response was applied to a disturbance input having a vibration property that swings in a minus direction at a certain timing and then swings in a plus direction while converging. In the conventional example, the time constant mu variable gain calculation unit 11 in the first-order lag element 17, is constant at mu = mu ₁₂ regardless of variations inclination of the control deviation e _{a (t).} On the other hand, in the present invention, the fluctuation of the control deviation e _a (t) occurs until the plant output y (t) (see FIG. 6A) swings in the negative direction and reaches the peak point of the undershoot. slope becomes positive, the time constant mu is set to a smaller mu = mu _12, thereafter, until it reaches the target value r (t), the variation slope of the control deviation e _{a (t)} is negative Thus, the time constant μ is set to a larger value μ = μ ₁₁ , and thereafter, until the peak point of the overshoot is reached, the fluctuation slope of the control deviation e _a (t) becomes positive, and the time constant μ It is set to a smaller mu = mu _12, ... and, in the boundary of the peak point of the destination point and overshoot or undershoot of the target value r (t), so that the time constant of the primary delay element is switched .

  As shown in FIG. 7A, the variable gain K (t) of the variable gain calculation unit 11 is set to be larger than that of the prior art, and with the convergence of the deviation e (t) (compared to the prior art). Slowly converges to the minimum setting value. As a result, the step response of the plant output y (t) and the plant input u (t) can be quickly responded to a change in disturbance, and undershoot and overshoot are suppressed, and the settling time is shortened. Thereby, it has been confirmed that disturbance response characteristics superior to those of the prior art can be realized.

As described above, the adaptive control apparatus according to the present embodiment receives the deviation e (t) between the output y (t) of the plant 1 and the target value r (t) as an input, and the steady-state gain of the closed loop control system is 1. A loop gain adjuster 7 that adjusts so as to be, a feedback unit that takes the output of the closed loop control system as an input and returns it to the output side of the loop gain adjuster 7 via the phase delay element 13, and an output y _{f of the} phase delay element 13 A variable gain calculation for calculating a variable gain K (t) using _a control deviation e _a (t) between (t) and the output e ^* (t) of the loop gain adjuster 7 as input and compensating for the gain of the closed loop control system. A closed loop control system including the unit 11, a PI controller 3 that performs PI control using the output of the loop gain adjuster 7 as an input, an output u _a (t) of the closed loop control system, and an output of the PI controller 3. Add to plastic An adder 9 to provide a manipulated variable u (t) to sheet 1, a adaptive controller having a variable gain calculator 11, a multiplier 16 which squares the control deviation e _a a _(t), a multiplier 16 A first-order lag element 17 that receives the output of the first-order lag element 17, and the time constant of the first-order lag element 17 is variably set according to the fluctuation slope of the control deviation e _a (t).

As described above, in the adaptive control apparatus of the present embodiment, the variable gain calculation unit 11 variably sets the time constant of the first-order lag element 17 according to the fluctuation slope of the control deviation e _a (t). The time constant of 17 is set to a larger value, the convergence of the variable gain K (t) to the minimum set value is delayed, and the load change that may occur after the control deviation e _a (t) converges to zero and the characteristic change of the control target are also detected. It is possible to respond immediately and realize a better response characteristic.

In the adaptive control apparatus of the present embodiment, the variable gain calculation unit 11 uses the time constant of the first-order lag element 17 when the fluctuation slope of the control deviation e _a (t) is negative as the fluctuation of the control deviation e _a (t). Since it is set to a value larger than the time constant of the first-order lag element 17 when the slope is constant or positive, the convergence time of the variable gain K (t) to the minimum set value is lengthened, and the control deviation e _a ( It is possible to immediately respond to a load change that may occur after the zero convergence of t) and a characteristic change of the controlled object, and to realize a more excellent response characteristic.

[Second Embodiment]
Next, FIG. 8 is a block diagram of an adaptive control apparatus according to the second embodiment of the present invention. FIG. 8A shows a schematic block diagram of the entire adaptive control apparatus, and FIG. 8B shows a schematic block diagram of the adaptive controller 105.

  In FIG. 8A, the adaptive control apparatus of the present embodiment is roughly configured to include a PID controller 103 and a closed loop control system for a plant 101 that is a control target. The PID controller 103 generates a control amount based on proportional control (P control), integral control (I control), and differential control (D control) for the plant 101. Further, the output y (t) of the plant 101 is fed back, and the deviation e (t) between the target value r (t) and the output y (t) of the plant 101 is obtained by the subtractor 108.

The closed loop control system includes an adaptive controller 105. As shown in FIG. 8B, the adaptive controller 105 of the closed loop control system includes a coefficient unit 107 that multiplies the deviation e (t) between the target value r (t) and the output y (t) of the plant 101 by k. A phase advance element 112, 114 and 119 that receives the output of the coefficient unit 107, _a feedback unit that returns the output u _a (t) of the closed loop control system via the phase delay element 113, and The variable gain K (t) is calculated using the control deviation e _a (t) between the output y _f (t) and the output e ^* (t) of the phase advance elements 112, 114 and 119, and the gain of the closed loop control system is calculated. And a variable gain calculator 111 for compensating for the above.

The phase advance element includes a coefficient multiplier 112 having a gain λ, a differentiator 114 having a time constant τ, and an adder 119, and has a transfer characteristic (λ + τs). Reference numeral 120 is a subtractor for obtaining _a control deviation e _a (t) between the output y _f (t) of the phase delay element 113 and the output e ^* (t) of the phase advance element. Reference numeral 110 is a control deviation. This is _a multiplier that multiplies e _a (t) by the variable gain K (t) that is the output of the variable gain calculator 111. Therefore, if the transfer characteristic of the phase delay element 13 is F ₁ (s), the transfer characteristic Gc (s) between the deviation e (t) and the output u _a (t) of the closed loop control system is Gc (s ) = K · (λ + τs) · K (s) / (1 + K (s) · F ₁ (s)).

Various characteristics are conceivable as the transfer characteristic F ₂ (s) of the phase delay element 113. Here, in order to obtain the effect of the second order differentiation, it is assumed that it is given by the following expression.

F ₁₂ (s) = 1/2 (1 / (1 + τs) + 1 / (1 + τs) ² )) (4)

  By substituting the equation (4) into the equation for the transfer characteristic Gc (s) and changing K (s) → ∞, the relationship of the following equation is obtained.

Gc (s) = k · (λ + τs) · (1 + τs) ² /(1+0.5τs) (5)

  Here, the values of k, λ and τ are specified values, and the adaptive controller 105 does not exceed kλ even when the value of the variable gain K (s) becomes infinite. Similarly, if K (s) → smaller, the relationship Gc (s) ≈k · (λ + τs) · K (s) is obtained, and the adaptive controller 105 has a smaller variable gain K (s). As a result, the first-order differential operation is performed.

Next, in FIG. 8B, the variable gain calculation unit 111 includes a coefficient unit 115, a multiplier 116, and a first-order lag element 117. That is, the variable gain K (t) is obtained by multiplying the control deviation e _a (t) by the coefficient γ by the coefficient unit 115, squares it by the multiplier 116, and passes this through the first-order lag element 117 of the time constant μ. It is calculated as the obtained value. The calculation of the variable gain K (t) in the variable gain calculator 111 is expressed as follows.

K (t) = (μ / (μs + 1)) · (γe _a (t)) ² (6)

Here, when d / dt | e _a (t) | <0, μ = μ ₂₁
When d / dt | e _a (t) | ≧ 0 μ = μ ₂₂ (μ ₂₂ <μ ₂₁ )

As shown in the equation (6), in the adaptive control device of the present embodiment, the time constant μ of the first-order lag element 117 in the variable gain calculation unit 111 is variably set according to the fluctuation slope of the control deviation e _a (t). There is a feature in the point. That is, when the fluctuation slope d / dt | e _a (t) | of the control deviation e _a (t) is negative, the time constant of the first-order lag element 117 is set to μ = μ ₂₁ , which is defined as the control deviation e _a (t ) Fluctuation slope d / dt | e _a (t) | is constant or positive and is set to a value larger than the time constant μ ₂₂ of the first-order lag element 117.

FIG. 9A shows a configuration diagram of a circuit that generates the time constant μ (1 / μ) of the first-order lag element 117 in the variable gain calculation unit 111. Constant 132 constant 131, the time constant _{mu 22} time constant _{mu 21,} the changeover switch 133, a differentiator 135, the switching signal generator 136, a configuration in which a divider 137 and the constant 138 "1". That is, the differentiator 135 obtains the fluctuation slope d / dt | e _a (t) | of the control deviation e _a (t), and the switching signal generator 136 determines negative, positive, or zero to generate a switching signal, Based on the switching signal, the output of either the constant unit 131 having the time constant μ ₂₁ or the constant unit 132 having the time constant μ ₂₂ is selected. As will be described later, since 1 / μ is used for the first-order delay element 117, 1 / μ is obtained by the divider 137.

  FIG. 9B shows a detailed configuration diagram of the variable gain calculation unit 111. In FIG. 9B, the first-order lag element 117 includes an integrator 121, a coefficient unit 122 having a gain (coefficient) 1 / μ, and a subtractor 123. In FIG. 9B, an adder 124 is provided at the subsequent stage of the variable gain calculation unit 111 to add a signal (variable gain) Ke (t) of another control system. Here, the other control system is another control system in the case where the control target is composed of two or more variable systems, and such a configuration makes it possible to apply to a multivariable control system.

The coefficient unit 115 having the gain γ desirably obtains the gain (coefficient) γ as follows. That is, the reciprocal of the maximum deviation estimated value e ^* (t) _max of the output e ^* (t) of the phase advance elements 112, 114 and 119 is passed through the first-order lag element (1 / (1 + σs)) having the time constant σ. This is the value obtained. Here, the time constant σ is approximately equal to the value of mu _22.

FIG. 9C shows a detailed configuration diagram of the phase delay element 113 described above. The phase delay element 13 includes a coefficient unit 141, a first primary delay element 142, a second primary delay element 143, and an adder 144. The output u _a (t) of the multiplier 110 is multiplied by 1/2 by the coefficient unit 141, the output is input to the first first-order lag element 142, and the output is input to the second first-order lag element 143. To do. Then, the outputs of the first primary delay element 142 and the second primary delay element 143 are added by the adder 144, and the addition result becomes the output y _f (t) of the phase delay element 113. Note that the time constant tau in the first primary delay element 143 and the second primary delay element 144, Ru can be specified from the outside (e.g. is set to τ = 0.0833μ _22).

As described above, the adaptive control apparatus of the present embodiment includes the coefficient multiplier 107 that multiplies the deviation e (t) between the target value r (t) and the output y (t) of the plant 101 by k, Phase advance elements 112, 114, and 119 that receive outputs as inputs, a feedback unit that receives the output of the closed loop control system and returns it via the phase delay element 113, and outputs y _f (t) of the phase delay elements 113 and the phase advance A variable gain calculation unit 111 for calculating a variable gain K (t) by taking _a control deviation e _a (t) from the output e ^* (t) of the elements 112, 114 and 119 as an input and compensating for the gain of the closed loop control system; , A PID controller 103 that performs PID control with the deviation e (t) as an input, and adds the output of the closed loop control system and the output of the PID controller 103 to the plant 101. A adaptive controller having an adder 109 to provide as u (t), a variable gain calculation unit 111, a multiplier 116 for squaring the control deviation _e a a (t), and inputs the output of the multiplier 116 And a time constant of the first-order lag element 117 is variably set according to the fluctuation slope of the control deviation e _a (t).

As described above, in the adaptive control apparatus of this embodiment, the time constant of the first-order lag element 117 is variably set according to the fluctuation slope of the control deviation e _a (t) in the variable gain calculation unit 111. The time constant of 117 is set to a larger value to delay the convergence of the variable gain K (t) to the minimum set value, and also to a load change or a control target characteristic change that may occur after the control deviation e _a (t) converges to zero. It is possible to respond immediately and realize a better response characteristic.

In the adaptive control apparatus of the present embodiment, the variable gain calculation unit 111 uses the time constant of the first-order lag element 117 when the fluctuation slope of the control deviation e _a (t) is negative as the fluctuation of the control deviation e _a (t). Since it is set to a value larger than the time constant of the first-order lag element 117 when the slope is constant or positive, the convergence time of the variable gain K (t) to the minimum set value is lengthened, and the control deviation e _a ( It is possible to immediately respond to a load change that may occur after the zero convergence of t) and a characteristic change of the controlled object, and to realize a more excellent response characteristic.

1,101 plant (control target)
3 PI controller 5, 105 Adaptive controller 7 Loop gain adjuster 8, 20, 23, 108, 120, 123 Subtractor 9, 19, 24, 109, 119, 124 Adder 10, 16, 110, 116 Multiplier 11, 111 Variable gain calculation units 13, 113 Phase delay elements 15, 18, 22, 41, 107, 112, 115, 122, 141 Coefficient units 17, 117 Primary delay elements 21, 121 Integrators 31, 32, 38, 131, 132, 138 Constant unit 33, 133 Changeover switch 35, 114, 135 Differentiator 36, 136 Switching signal generator 37, 137 Divider 42, 142 First primary delay element 43, 143 Second primary delay Element 103 PID controller

Claims (3)

Loop gain adjusting means for adjusting the steady-state gain of the closed-loop control system to be 1 with the deviation between the output of the control target and the target value as input, and the output of the closed-loop control system as an input via the phase delay element A feedback means for returning to the output side of the loop gain adjusting means, and a variable for calculating a variable gain with a control deviation between the output of the phase delay element and the output of the loop gain adjusting means as an input, and a variable for compensating the gain of the closed loop control system A closed loop control system comprising a gain calculation means;
PI control means for performing PI control using the output of the loop gain adjusting means as an input;
An adding control unit that adds the output of the closed-loop control system and the output of the PI control unit and gives the output to the control target as an operation amount;
The variable gain calculating means includes a calculating means for squaring the control deviation, and a first-order lag element having the output of the calculating means as an input, and the time constant of the first-order lag element is set as a fluctuation slope of the control deviation. Adaptive control device that variably sets according to Coefficient means for multiplying the deviation between the output of the control target and the target value by a predetermined coefficient, a phase delay element that receives the output of the coefficient means, and feedback that inputs the output of the closed loop control system and returns it via the phase delay element A closed loop control system comprising: a means, and a variable gain calculating means for calculating a variable gain using a control deviation between the output of the phase delay element and the output of the phase delay element as an input, and compensating the gain of the closed loop control system; ,
PID control means for performing PID control with the deviation as an input;
An adding control unit that adds the output of the closed-loop control system and the output of the PID control unit and gives the output to the control target as an operation amount;
The variable gain calculating means includes a calculating means for squaring the control deviation, and a first-order lag element having the output of the calculating means as an input, and the time constant of the first-order lag element is set as a fluctuation slope of the control deviation. Adaptive control device that variably sets according to   The variable gain calculating means has a value larger than the time constant of the first-order lag element when the fluctuation slope of the control deviation is constant or positive as the time constant of the first-order lag element when the fluctuation slope of the control deviation is negative. The adaptive control device according to claim 1 or 2, wherein

Images