JPS6327721B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an improvement in a multivariable process control device used, for example, to control the liquid level of a concentrator. Conventionally, multivariable process control devices using the classic P.I.D control method (so-called P.I.D. controller) have the following problems. The input and output signals of a multivariable process must be allocated in combinations of one input and one output control system. There is no one-to-one correspondence between the inputs and outputs of a multivariable process, and if there is mutual interference, it is necessary to determine parameters by considering the whole at the same time, which is complicated. In a multivariable process, parameters can only be selected for a one-input, one-output system, so the range of stability is narrow overall, and it is difficult to evaluate optimality because decisions need to be made through trial and error. difficult. The above-mentioned problems are particularly noticeable in the case of the evaporator liquid level control system. Hereinafter, a process control device using a P.I.D controller in the evaporator liquid level control system will be described with reference to FIG. That is, this process control device controls the supply liquid flow rate by adjusting the supply liquid flow rate control valve 1.
At the same time, the liquid level in the evaporator 2 is detected by a liquid level detector 3, and this detected value is compared with a set value l by a deviation calculator 4 to obtain a deviation. Then, after performing P-I-D calculation on this deviation using the P-I-D controller 5 to obtain a manipulated variable, the heated steam flow rate control valve 6 is operated using this manipulated variable to achieve appropriate heating. The liquid level in the evaporator 2 is controlled by supplying the vapor flow rate to the evaporator 2. In the figure, 7 is a liquid supply flow rate setting device, 8 is an evaporated steam flow rate discharge pipe, and 9 is a concentrated liquid discharge valve. However, in this process control device, since both the feed liquid flow rate and the heated steam flow rate are the manipulated variables for the evaporator liquid level, the P/I/D controller 5 cannot control the two quantities at the same time, resulting in unstable stability. Process control is not possible.
Another problem is that while the evaporator liquid level is controlled to be constant, the flow rate of liquid supplied must be arbitrarily set as the throughput of the concentrator. Furthermore, the liquid supply flow rate control valve 1 is semi-fixed and controlled at a constant value by the setting device 7, and the control of the evaporator liquid level must rely solely on the operation of the heated steam flow rate. This has a significant impact on the responsiveness and stability of liquid level control when a disturbance to the liquid level occurs, such as when a part of the concentrated liquid is discharged. FIG. 2 shows the disturbance characteristics of the evaporator liquid level when the concentrator process is simulated by a computer and only the heated steam flow rate is manipulated. However, this is an example when the liquid supply flow rate fluctuates by 0.1l/min. On the other hand, according to the optimal regulator method, which incorporates a method of modern control theory, the system is as follows (1)
By solving this equation using the Ritsukachi equation (2), the optimal control evaluation function can be determined as an optimal solution using equation (3). x=Ax+Bv …(1) v=R ^-1 B ^T π _x …(2) J=∫ ^∞ ₀ (xPx ^T +vQv ^T )dt …(3) However, A and B are matrices representing the system, v is the input, x is the output, R is the positive definite matrix of Ritsukachi equation,
π is the solution to Ritsukachi's equation, and PQ is a positive definite matrix. However, this optimal regulator method has many problems to be fully applied due to the following problems. First, it is difficult to completely represent a system using an equation of state. The values of the weighting coefficient matrices P and Q in the evaluation function, which determine the weight of the transient response and stability of the system when minimizing the evaluation function by solving the Ritsukachi equation, must be determined by human trial and error. It's not easy. If the optimal regulator method is applied completely to a multivariate rate process, the order of the state equation becomes high, making it difficult to accurately obtain a solution to the Ritsukachi equation by calculation. Even if a solution is obtained, the parameters of the state equation are often internal states of the system (such as enthalpy) that cannot be measured in the actual plant, and if the order of the solution to the Ritsukachi equation becomes high, the calculation unit that realizes it is required. There are hardware problems such as difficulty in manufacturing. The above problems also apply to the evaporator liquid level control system. In this case, in order to minimize the deviation between the liquid level, the feed liquid flow rate, and the set value in the process control device expressed by equation (1), it is necessary to newly supply liquid to the feed liquid flow inflow pipe 11 as shown in Fig. 3. A flow rate detector 12 is provided, the deviation between the output of each of the detectors 3 and 12 and the set value m·l is determined, and these deviations and the integrators 13 and 14 are calculated.
4 is inputted into the multivariable control unit 15, where the manipulated variables of the supply liquid flow rate and the heated steam flow rate are determined, and then the respective manipulated variables are added to the control valves 1 and 6 to control the control valves 1 and 6.
This is the next input system. In the figure, 16 is a deviation calculator. Therefore, when replacing the conventional classical control system as described above, the control parameters are at least 4: P (proportional), I (integral), D (differential), and D ² (secondary derivative).
is required. As a result, calculations to find the optimal solution become difficult. FIG. 4 is a transient response characteristic diagram obtained under the same conditions as the classical control means of FIG. As described above, conventional multivariable process control equipment has the disadvantage that it cannot handle two or more manipulated variables at the same time when using the classic P-I-D control method, and if the optimal regulator method is used as is, the system This has the disadvantage that the order (dimension) becomes high, making it difficult to calculate the regulator. The present invention has been made in view of the above-mentioned circumstances, and it combines classical P-I-D control and modern control, making it possible to handle two or more manipulated variables at the same time, and increasing the order of the system. It is an object of the present invention to provide a multivariable process control device that facilitates the design of a regulator by lowering the temperature, and further improves transient response characteristics and stability of liquid level during disturbances. Hereinafter, one embodiment of the present invention will be described with reference to FIGS. 5 to 7. Furthermore, Fig. 5 is a principle block diagram of the device of the present invention applied to a two-input, one-output multivariable process, Fig. 6 is a block diagram showing a specific example of application to an evaporator liquid level control system, and Fig. 6 is a signal processing configuration diagram embodying a part of FIG. 6; FIG. First of all,
As shown in FIG. 5, the device of the present invention includes an optimal control system 21 based on current control theory, and
The optimal control system 21 comprises a process 211 and a state variable feedback type controller 212 to form a one-loop feedback control system.
In this configuration, a manipulated variable C is determined from the deviation between the controlled variable b outputted from the controller 1 and the set value l, and is inputted to the process 211 in consideration of the state variables. On the other hand, the classical controller 22 forms a feedback loop that encompasses the optimal control system 21, and controls the controlled object m using the control signal obtained by this controller 22 to provide another manipulated variable a of the process 211. This is the configuration that obtains the following. At this time, the classical controller 22 calculates an arithmetic expression using the root locus method (or pole placement method) so that the placement of poles and zeros of the transfer function is optimal as a total feedback. Next, FIG. 6 shows a specific example applied to an evaporator liquid level control system. This control system adjusts the supply liquid flow rate m1 with the supply liquid flow rate control valve 31, and then the evaporator 32.
supply to. The fluid in the evaporator 32 is heated and evaporated and concentrated by the heated steam m2 introduced into the heating pipe 34 via the heated steam flow rate control valve 33. This heated steam is outputted from a heated steam discharge passage 35, while the concentrated liquid is discharged through a concentrated liquid discharge valve 36. 37 is a liquid level detector, and the fluid level signal in the evaporator 32 outputted from this detector 37 is sent to the deviation calculator 38, where it is compared with the liquid level setting value l and the deviation signal is determined. can get. Reference numeral 39 denotes a state variable feedback type controller which inputs a deviation signal and obtains a signal one order higher than the order of the system (in this case, 1 input and 1 output) from an integrator 40 to form a state variable type feedback calculation formula. The manipulated variable is determined by
3 to control the heating steam flow rate.
A controller 41 corresponds to the classical controller 22, and after performing calculations using the output of the state variable feedback type controller 39 and the deviation signal, the output of the controller 41 is transferred to the deviation calculator. 42 and compares it with the liquid supply flow rate set value h to find a deviation, and this deviation signal is supplied to the liquid supply flow rate control valve 31 for use in controlling the liquid supply flow rate. In addition, the seventh
This figure is a diagram showing the arithmetic expressions and signal processing configuration performed by the controller 39, integrator 40, and controller 41 shown in FIG. 6. In FIG. 7, reference numeral 51 is a section for obtaining a state feedback output by the optimal regulator method using a state variable feedback type controller, and corresponds to the optimal control system 21 in FIG. Reference numerals 52 and 53 receive the output and deviation signal of the state variable feedback type controller, and perform stability analysis by the root locus method using a classical controller, and correspond to the classical controller 22 in FIG. 53 means a state variable composition matrix. Next, the operation of the device configured as described above will be explained. First, the evaporator 32 obtained by the liquid level detector 37
The liquid level signal is input to the deviation calculator 38, where it is compared with the liquid level set value l to obtain a liquid level deviation signal, and this signal is sent as it is and via the integrator 40 to the state variable feedback type controller 39. input. Here, the controller 39 minimizes the liquid level deviation signal and deviation integral signal of the evaporator 32 as shown in FIG.
Calculation is performed using the formula Kp ₁ (1fT ₁・1/S+T _D1・S) to find the manipulated variable. Here, the state variable feedback type controller 39
Explain the effect of In other words, as mentioned above, the evaporator liquid level loop is described by an equation of state with heated steam as input and liquid level as output, and the value that minimizes the evaluation function is given by the solution to Ritsukachi's equation. .
Specifically, the arithmetic expression of the controller 39 is determined from the relationship between the deviation of the evaporator liquid level and the manipulated variable of the heated steam flow rate that minimizes the deviation integral signal. At this time, in order to convert the detected state into an actually measured quantity (for example, flow rate, liquid level), the two inputs of the feed liquid flow rate and the heated steam flow rate are synthesized using a state variable synthesis matrix 53 shown in FIG. . Also, the controller 41
Another feature is that a differential operation with a negative coefficient (T _D2 ) is inserted as a characteristic. Controller 39
The evaporator level control system using the feed liquid flow rate as a disturbance and the heated steam flow rate as the control input (operated variable) is optimally determined, but in reality, simply manipulating the heated steam flow rate does not provide sufficient transient response. It is difficult to control the liquid level with good performance. Therefore, a one-input, one-output system (input: liquid supply flow rate, output: evaporator liquid level) is considered, in which the range including the controller 39 is controlled, and this control is executed by the controller 41. Note that the arithmetic expression of the controller 41 is 1 input, 1
Considering the poles and zeros of the output system, to eliminate the existence of poles on the right plane and move the poles to the stable left plane , use not only p operation but also P/D operation with negative coefficients and place the zero on the right plane. It can be seen that the convergence of the system is better when Coefficient of D action
T _D2 is selected to be an appropriate negative value when varied from -∞ to +∞, and the arithmetic expression of the controller 41 is determined by determining the gain ( p constant) from the root locus diagram. Since this controller 41 does not include an integral operation, the output is zero when the liquid level deviation input is zero and stable, and the deviation output signal from the liquid supply flow rate is only the liquid supply flow rate, so in a steady state, the liquid supply flow rate is It is also a system that can be configured as desired. If this control method is used, a control system design that can operate both the heating steam flow rate and the feed liquid flow rate simultaneously during transient changes such as disturbances will be achieved, resulting in a good design. Control results can be expected.
Note that FIG. 8 is a diagram showing control results under the same disturbance conditions as FIGS. 2 and 4. Therefore, since the present invention is configured as described above,
It is possible to configure a control system with two inputs and one output, which cannot be achieved with the classical control method, and better control results can be obtained. Furthermore, by effectively applying modern control theory and keeping it to a minimum, the order of the system can be reduced and the design of the control system can be made easier. In addition, by not relying solely on the optimal regulator for control results, but also using the pole placement method, we were able to construct a control system that can be achieved with the current hardware. In particular, in the evaporator liquid level control system, it not only improves the convergence (transient response) and stability of liquid level deviation, but also allows the supply liquid flow rate to be arbitrarily set during steady state. In order to determine the evaporation throughput, it is possible to provide a multivariable process control device that can realize a control system that is excellent in terms of operation and operation.

[Brief explanation of the drawing]

Figure 1 is a configuration diagram of a conventional device using a classic P-I-D controller, Figure 2 is a transient response characteristic diagram of the device shown in Figure 1 during disturbances, and Figure 3 is one of modern control theories. FIG. 4 is a diagram of the transient response characteristics of the device shown in FIG. 3 during disturbances, and FIGS. 5 to 8 show the multivariable process control device according to the present invention. These figures are shown to explain one embodiment, and FIG. 5 is a diagram showing the principle configuration, FIG. 6 is a configuration diagram showing a specific example applied to an evaporator liquid level control system, and FIG. 7 is a diagram showing the configuration. 6 is a diagram showing a part of the arithmetic expressions, and FIG. 8 is a transient response characteristic diagram at the time of disturbance in the apparatus of FIG. 6. 31... Liquid supply flow rate control valve, 32... Evaporator, 3
3... Heating steam flow rate control valve, 37... Liquid level detector, 38... Deviation calculator, 39... State variable feedback type controller, 40... Integrator, 41...
Controller, 42...deviation calculator.

Claims (1)

[Scope of Claims] 1. In a multivariable process control device that controls a process by determining two manipulated variables from one controlled variable, a deviation signal is obtained from the controlled variable detection signal and a preset setting signal. an arithmetic unit, an integrator that integrates the deviation signal of the deviation arithmetic unit, and an integrator that receives the deviation integral signal of the integrator and the deviation signal, and uses an optimal regulator method to minimize the deviation integral signal and the deviation signal. A state variable feedback type controller for calculating one manipulated variable, the output of this state variable feedback type controller, and the deviation signal of the deviation calculation section are used to perform stability analysis using the root locus method to calculate the other manipulated variable. A multivariable process control device characterized by being equipped with the desired classical controller.