JPS6343523Y2 - - Google Patents

Description

[Detailed description of the invention] <Industrial application field> The invention is designed to enable switching between two types of control algorithms: the deviation differential type PID algorithm (hereinafter simply referred to as the PID algorithm) and the I-PD algorithm. This invention relates to an algorithm-switching analog controller.

<Background technology and its problems> In recent years, various algorithms have been tried and proposed as controllers, and the most suitable one has been selected depending on the object to be controlled. I'm getting used to it.

For example, as a controller control algorithm that has been widely known for a long time, the above deviation differential form is used.
There is a PID algorithm. This PID algorithm uses P (proportional term), I (integral term), D
(differential term) and is expressed by the following relational expression.

MV=100/PB(SV-PV) (1+1/T _I S+T _D S/1+αT _D S) ...(1) Here, MV is the controller's operation signal output, PB is the proportional band (%), and SV is the target value setting signal input, PV
is the control measurement signal input, T _I is the integration time, S is the Laplace operator, T _D is the differential time, and is the reciprocal of the differential gain of α.

This PID algorithm is excellent in feedpack controlling the controlled object so that the controlled variable matches the target value, but when the target value setting signal is changed, the controller's operation signal output changes suddenly. , the control response to changes in the setting signal and the control response to disturbances are different, and the optimal values of the PID parameters for each are different, so that P, I, D
When each parameter is set, the control response to disturbance deteriorates, and when trying to improve the followability to changes in the setting signal, the amount of overshoot increases.

In contrast, recent DDC (Direct, Digital,
Control) The I-PD algorithm is a control algorithm used in controllers. This I-PD algorithm is expressed by the following relational expression.

MV=100/PB {(SV-PV)1/ _TIS -PV(1+ _TDS )}...(2) Note that each symbol here is the same as in the case of the above-mentioned PID algorithm.

According to this I-PD algorithm, steady position deviation is corrected by integral calculation and (proportional + differential) feed hack compensation is performed on the controlled object.
There is no sudden change in the controller's operation signal output in response to a setting change, the quick response of the control is improved regardless of the amount of overshoot, the response to a setting change is the same as the response to a disturbance, and a set of
PID parameters have the advantage of being able to perform optimal control for both, but on the other hand, a tuning method for each parameter in the field has not yet been established, and the limit of control responsiveness is lower than PID, and cascade control The disadvantages were that it was unsuitable for use as a slave-side controller or a controller that could be combined with a program setting device, and that the sensor noise characteristics were slightly worse than PID.

Therefore, in DDC controllers with built-in microcontrollers, each calculation of P, I, and D can be easily and freely realized by programming on software, so in some cases, these PID algorithms and I-PD A controller having two control algorithms as software has been proposed.

However, in analog controllers, the control algorithm must be configured as hardware using an analog computing unit, so if the algorithm becomes complex, the circuit becomes complex and requires high-performance parts. The number of parts increases. For this reason, various problems arise, such as not being able to downsize the device, increasing costs, and deteriorating reliability.

Therefore, if you try to implement an analog controller with two types of algorithms, the PID algorithm and the I-PD algorithm, it will be difficult in terms of cost, product dimensions, etc. compared to a DDC controller with similar functions. So far,
This kind of analog algorithm switching controller has not been proposed.

However, if the above-mentioned drawbacks can be overcome, the analog type will not have any adverse effects on sampling operation control, which is inevitable in principle with DDC controllers. Better control can be expected than with a controller.

<Purpose of the invention> The present invention was made in view of the above circumstances, and its purpose is to improve the circuit configuration and combinations of circuits in analog controllers. The purpose of the present invention is to propose a controller that has two types of control algorithms, the PID algorithm and the I-PD algorithm, and allows switching between the two types of control algorithms, the PID algorithm and the I-PD algorithm, with only a small number of parts.

<Summary of the invention> The algorithm switching analog controller of the invention is outlined as shown in the block diagram of FIG.

This controller uses PID algorithm and I-PD
In order to coexist with the algorithm and to enable switching between them, first, the PID algorithm side is made non-interfering in its circuit configuration, and each element of proportional term 1, integral term 2, and differential term 3 is placed in parallel. It has been placed. Further, a proportional amplifier 4 is provided after each of the proportional, integral, and differential terms 1, 2, and 3. The reason for this is as follows.

That is, the gain of a proportional amplifier is proportional bandwidth (PB)
It generally ranges from 0.1 to several 100 times. However, the output of an analog computing unit saturates at a certain value, so proportional amplification is
The point where the term, I term, and D term are added, i.e. PID
(or I-PD) If performed before the point where the algorithm is formed, the P term may become saturated depending on its gain. When the P term saturates, the control algorithm becomes an algorithm without the P term.

However, in the PID algorithm, the input of the P term is the deviation, and if the deviation is larger than the proportional band width, the above situation may occur, but the saturated output of the P term in this case is in the direction of reducing the control deviation. , In addition, if the deviation is at least smaller than the proportional band width, the P term will not be saturated, so each operation of P, I, and D will operate according to the principle. Therefore, in the case of the PID algorithm, the proportional amplifier can be placed either immediately after the deviation amplifier or after the P, I, and D terms have been added. When the width is narrower than 100%, the input level of the integrator can be increased and the residual deviation of the controller can be reduced, so a proportional amplifier is provided immediately after the deviation amplifier.

On the other hand, in the I-PD algorithm, as the control operation progresses and approaches a stable state, the PV value, which is the input to the P term, approaches the SV value, and as in the PID algorithm, the input to the P term becomes smaller. Therefore, if the gain of the P term is large up to the point where the P term, I term, and D term are added, the P term may remain saturated no matter how long (in this case, P
Since the saturated output of the term is absorbed by the I term), control is performed only by the I term and the D term. I-
In the PD algorithm, the quick response of control characteristics is mainly due to feedback compensation by the P term, so in that case, the fact that the P term does not work means that
This means that the control characteristics will be significantly deteriorated.
Therefore, in the I-PD algorithm, when the saturation level of the output of the arithmetic unit is limited, as in the case of an analog arithmetic unit, the proportional amplifier has a P term, an I term, and a D term.
This is because it must be provided after the terms are added.

Even with this I-PD algorithm, for example,
In the calculation of a DDC type controller, it is easy to increase the number of digits (number of bytes) of the calculation using software, and by doing so, the proportional amplification can be
It is possible to perform the calculations before each of the calculations for the term and D term, and it may be possible to reduce the residual deviation of the controller with calculations using fewer bytes. So, each term 1,
There is a configuration in which two proportional amplifiers 4, 4 are provided in front of the amplifiers 2, 3, and their gains are changed in conjunction with each other.

However, in an analog controller that performs analog calculations, the input signal level of the analog calculation unit
Since the output saturation level is limited, it is almost impossible to perform proportional amplification first. Furthermore, even if the calculation is performed by limiting the proportional bandwidth and input signal level to a very narrow range,
As shown in Figure 2, it is necessary to provide two proportional amplifiers for the two signals of SV value and PV value, and to change their gains in conjunction, which complicates the circuit and increases cost. Therefore, as mentioned above, it is better to provide it after each term 1, 2, and 3. Furthermore, in order to realize and switch between both the PID algorithm and the I-PD algorithm, first, the input of the integral term 2 is the control deviation (SV-PV) output from the deviation amplifier 5.
An interlocking algorithm switching switch 6 is provided in the proportional term 1 and differential term 3, and this switch 6 switches either the control deviation (SV-PV) or the measurement signal PV according to the algorithm. It is designed to be switched.

That is, when this switch 6 is connected to the contact side, a PID algorithm is configured, and its operation signal MV is determined by the following equation, which is the same as the above-mentioned relational equation (1).

MV=100/PB(SV-PV) (1+1/T _I S+T _D S/1+αT _D S) ...(3) Also, when switch 6 is connected to the opposite contact, the I-PD algorithm is configured. ,
The operation signal MV is obtained by the following equation, which is similar to the above-mentioned relational equation (2).

MV=100/PB {(SV-PV)1/T _I S -PV(1+T _D S/1+αT _D S)} ...(4) Here, the differential term has a first-order lag 1/(1+αT _D S) ) to achieve incomplete differential operation and to prevent the negative effects of high frequency noise.

<Example> Next, an example of the controller of the present invention will be described in more detail with reference to FIG.

In the figure, 11 is a deviation amplifier, 12 is an integrator, 13 is a proportional amplifier, 14 is a (proportional + differential) device including a proportional circuit, a differential circuit, and an amplifier, and 15 is a 2
This is an algorithm switching switch consisting of two interlocking switches SW ₁ and SW ₂ .

In this controller, a setting signal E _S and a measurement signal E _P feed-hacked from the controlled object are input to the deviation amplifier 11 . This amplifier 11 outputs the deviation E _E of each of these signals E _S and _EP . If R ₁ = R ₂ = R ₃ = R ₄ , this deviation E _E becomes E _E = E _S −E _P (5). The output of this deviation E _E is input to an integrator 12, integrated, and output. Integral output at this time
If E _I is the variable resistor RV ₁ = (m-1) R ₅ , then E _I = R ₅ /R ₅ + (m-1) R ₅ -E _E /C ₁ R ₆ S = -E _E /m・R ₆ C ₁・S ...(6) It is obtained as follows. Here, m is a value that changes with the rotation angle of RV ₁ (m≧1), and S is a Laplace operator.

The above (proportional + differentiator) 14 receives the measurement signal E _P and the above deviation E _E as input signals by switching the algorithm switching switch 15 as described later.
input as E _C , output E _PD for this signal E _C
is, E _PD = R ₇ /R ₇ +R ₈・(1+R ₇ +Rs/R ₇・C ₂ R
V ₂ S / 1 + C ₂ RV ₂ S)・R ₉ + R ₁₀ /R ₉ E _C , where, if R ₇ = R ₉ and R ₈ = R ₁₀ , then E _PD = 1 + R ₇ + Rs /R ₇・C ₂ RV ₂ S/1+C ₂ RV ₂ S・E _C ……(
7) is obtained as.

In this (proportional + differentiator) device 14, the interlocking switch SW ₁ of the algorithm switching switch 15,
When SW ₂ is connected to the contact side, the measurement signal E _P is directly input, and its output E _PD is sent to the proportional amplifier 1 above.
It is input to the negative terminal side of 3. In this proportional amplifier 13, if RV ₃ ≪ R ₁₅ , its output (operation signal) _EM is: _EM = -RV ₃ /nRV ₃ · R ₁₅ + (R ₁₃ R ₁₄ ) / (R _{13 R14}
)・{(R ₁₃ R ₁₅ )/R ₁₄ + (R ₁₃ R ₁₅ )E _PD + (R ₁₄ R
₁₅ )/R ₁₃ + (R ₁₄ R ₁₅ )E _I }, and if we further set R ₁₄ = R ₁₅ = 2R ₁₃ , then E _M = -1/n (E _PD + 2E _I ) ...is output as (8). However, n is a value that changes depending on the rotation angle of RV ₃ (0≦n≦1).

In addition, in the above-mentioned case where the interlocking switches SW ₁ and SW ₂ are connected to the contact side of the algorithm switching switch 15, E _C = E _P , so from the above equation (7), the (proportional + differential) The output E _PD of 14 is also E _PD =1+R ₇ +Rs/R ₇・C ₂ RV ₂ S/1+C ₂ RV ₂ S・E _P ……(
9) can be rewritten as

As mentioned above, algorithm switching switch 1
When the interlocking switches SW ₁ and SW ₂ of No. 5 are connected to the contact side, the operation signal E _M is also E _M = - from the above equations (5), (6), (8), and (9). 1/n{1+R ₇ +Rs/R ₇・C ₂ RV ₂ S/1+
C ₂ RV ₂ S・E _P −2/m・R ₆ C ₁・S( _ES −E _P )} =1/n{1/m・R ₆・C ₁ /2・S( _ES −E _P
)−1+R ₇ +Rs/R ₇・C ₂ RV ₂ S/1+C ₂ RV ₂ SE _P }……(10)
It is expressed as

On the other hand, when the interlocking switches SW ₁ and SW ₂ are connected to the contact side of the algorithm switching switch 15, the deviation E _E from the deviation amplifier 11 is input to the input side of the (proportional + differentiator) unit 14. ,
The output E _PD is input to the positive terminal side of the proportional amplifier 13. The operating signal E _M of the proportional amplifier 13 at this time is, if RV ₃ ≪ R ₁₅ , E _M = RV ₃ /nRV ₃・R ₁₃ +R ₁₅ /R ₁₃・(R ₁₂ /R ₁₁ +
R ₁₂ E _PD − _{R 15} / R ₁₃ + R ₁₅ E _I ), where, as mentioned above, R ₁₅ = 2R ₁₃ and R ₁₁ = 2R ₁₂ , then E _M = 1 /n(E _PD −2E _I ) ...(11) is output.

In addition, in the above case where the interlocking switches SW ₁ and SW ₂ are connected to the contact side of the algorithm changeover switch 15, E _C = E _E = E _S - E _P , so the above equation (6), From (7) and (11), the operation signal E _M of the proportional amplifier 13 is: E _M = 1/n・{1+R ₇ +R ₈ /R ₇・C ₂ RV ₂ S/
1+C ₂ RV ₂ S・( _ES −E _P )+2/m・R ₆ C ₁・S( _ES −E _P )
} =1/n(E _S −E _P ) {1+R ₇ +R ₈ /R ₇・C ₂ RV
₂ S/1+C ₂ RV ₂ S+1/m・R ₆ C ₁ /2・S}...(12)

In the above formulas (10) and (12), 1/n=100/PB: Proportional gain (PB=Proportional band width [%]
) R ₇ +R ₈ /R ₇ =1/α: Differential gain R ₇ +R ₈ /R ₇・C ₂ RV ₂ = T _D : Differential time m・R ₆ C ₁ /2=T _I : Integral time, _Equations (10) _{and ( 12} ) are _each
+αT _D S+T _D S/1+αT _D S)…(13) E _M =100/PB(E _S −E _P )(1/1+αT _D S+1
/T _I S+T _D S/1+αT _D S) (14) It can be seen that the I-PD algorithm is realized by equation (13) and the PID algorithm is realized by equation (14). In addition, the above formulas (13) and (14)
The proportional term is 1/1+αT _D S in , and this is because a high frequency limiting filter is included in the proportional term.

That is, with this controller, when the interlocking switches _SW1 , _SW2 of the algorithm changeover switch 15 are connected to the contact side, the I-PD algorithm is realized, and when they are connected to the contact side, the PID algorithm is realized.

As described above, according to the controller of this embodiment, only four ICs (Q to Q ₄ ) forming the OP amplifier are required, and only various resistors and capacitors are required. Therefore, it can be said that the number of parts is approximately the same or less than that of this type of conventional non-interference type PID algorithm controller. Furthermore, when switching between the PID algorithm and the I-PD algorithm, it is only necessary to insert an algorithm switching switch 15 into the circuit.
As mentioned above, there is almost no increase in the number of parts, and the switching can be performed extremely easily.
Furthermore, when the I-PD algorithm is selected by switching the interlocking switches SW ₁ and SW ₂ of the algorithm switching switch 15 to the contact side, the proportional term is added to the integral term as described above. The proportional amplifier is placed after the integral term and the proportional term are added in order to prevent the control operation from becoming saturated before reaching the integral term and resulting in only an integral operation. Then, in front of the proportional amplifier, (proportional + differential) is applied so that the proportional term does not saturate over the entire input range of the measurement signal input PV.
The proportional circuit of the amplifier 14 and the gain of the amplifier _Q3 are set. That is, in this embodiment, the gain of (proportional circuit + amplifier Q ₃ ) for the proportional term is set to 1. Furthermore, in order to be able to move the operation signal output MV from 0 to 100% with the output of the integrator 12, regardless of the value of the output of the (proportional + differentiator) 14, a proportional term and a differential term are added. The outputs are added in advance by the amplifier Q ₃ of the proportional + differentiator 14, the maximum value of the output is regulated by the saturation level of the amplifier Q ₃ , and the maximum value of the output of one integral term is also added by the integrator 12. amplifier
It is regulated at the saturation level of Q ₂ , and the resistance R ₁₃ and
The gain of the proportional amplifier 13 for each of the integrator 12 and (proportional+differentiator) 14 is optimized by changing the value of _R14 . In this embodiment, the gain of the integrator 12 versus the gain of the (proportional+differentiator) device 14 is set to 2:1.

<Effects of the invention> As is clear from the above description of the present embodiment, the invention has a small number of parts, is advantageous in terms of cost, and can switch between the PID algorithm and the I-PD algorithm, allowing for a wide range of control. It is possible to provide an algorithm-switching analog controller that can extremely smoothly respond to the target and has good handling properties.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an outline of the controller of the present invention, FIG. 2 is a block diagram of an example of a conventional controller, and FIG. 3 is a circuit diagram showing an embodiment of the controller of the present invention. In the figure, 11... deviation amplifier, 12... integrator,
13... Proportional amplifier, 14... (Proportional + Differential)
15...Algorithm switching switch.

Claims (1)

[Scope of utility model registration request] A deviation amplifier that compares the measurement signal and the setting signal and outputs a deviation signal, an integrator that integrates the deviation signal, and a (proportional + differential) device that outputs the proportional and differential components of the deviation signal or measurement signal. , has a proportional amplifier that adds and amplifies the output of the integrator and the output of the (proportional + differentiator) at a certain ratio to obtain an operation signal output, and adds the output of the integrator and the output of the (proportional + differentiator) to a deviation signal. It is equipped with an algorithm changeover switch that switches between the control algorithm and the measurement signal.
An algorithm switching analog controller characterized by being able to select either a PID algorithm or an I-PD algorithm.