JP2000222005A - Controller for process - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a controller with high accuracy because of operating a predicted value in different predicted time in accordance with a manipulated variable for a process in which plural manipulated variables exist to a single controlled variable and deciding a manipulated variable on the basis of respective estimated values. SOLUTION: A steam temperature estimation controller 380 contains controllers 320a and 320b which respectively output a manipulated variable for the same controlled variable, an estimating part 330 calculating an estimated value, a switching part 339 for feedbacking either an estimated value Ys being its output or a plant output value (controlled variable: steam temperature) Y to the controller 320 and a parameter adjusting part 340 for adjusting parameters for various models included in the part 330, and the part 330 operates and outputs two estimated values Ys(t+τa) and Ys(t+τb) whose estimation times are different to different manipulated variables (fuel flow rate and desuperheater spray water flow rate).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for predicting a future state of a process, determining an operation amount based on the prediction result, and controlling the process.

[0002]

2. Description of the Related Art As a response characteristic of a process, there is one having a large response time delay. For example, in a thermal power plant,
The time from the change in the fuel input to the change in the main steam temperature is long, and the time constant is several minutes to 20 minutes.

Therefore, there is a problem that the control deviation (main steam temperature deviation) becomes large in the ordinary feedback control.

On the other hand, there is a prediction control method for predicting a future process state in advance and determining an operation amount based on the predicted value. As the predictive control method, for example, there is a conventional technique disclosed in the following document.

(1) Japanese Patent Application Laid-Open No. 9-274507 (2) Japanese Patent Application Laid-Open No. 6-257702 (3) Japanese Patent Application Laid-Open No. 10-116105 A method is described in which a process model is composed of a generalized model and a dead time model, input variables are estimated by a state observer, and future values of process quantities are calculated.

The prior art (2) includes a model reference type adaptive control method using a process and a distributed PID (proportional, integral, differential).
A method of configuring an adaptive control system by adding in parallel to a control target configured by a control device is described.

[0007] The above prior art (3) has a problem in controlling a plurality of control systems including a generalized predictive control system while switching the control systems.
A method is described in which a control command value is switched smoothly without a large deviation.

By applying such a predictive control technique, since the manipulated variable is determined in advance by predicting the state of the future process, the control deviation can be reduced.

[0009]

However, the desired prediction time differs depending on the characteristics of the process to be applied, the type of process amount to be predicted, and the type of operation amount for controlling the process amount. Therefore, when determining a plurality of operation amounts based on a predicted value of one process amount, the effect of the prediction control may not be able to be maximized.

However, the above prior arts (1) to (3) do not disclose a method of determining a predicted time according to a relationship with an operation amount.

For example, in a thermal power plant, the manipulated variables for controlling the steam temperature include a fuel flow rate and a spray water flow rate of a steam temperature reducer. Response time is shorter. Therefore, if the predicted time of the steam temperature, which is the control amount, is set according to, for example, the fuel, the spray water flow rate cannot be determined effectively.

An object of the present invention is to provide a process control apparatus and method capable of accurately controlling a controlled object having a plurality of operation amounts for the same control amount by predictive control.

[0013]

In order to achieve the above object, the present invention provides a process amount (controlled amount) to be controlled.
And a manipulated variable calculating means for determining a process quantity (operated variable) to be operated to control the controlled variable based on the predicted value output from the predicting module. In the control device or the control method of the process, when there are a plurality of manipulated variables for a single controlled variable, the predicting unit determines whether the same controlled variable includes a predicted time and a predicted time interval. At least one of the plurality of different predicted values is calculated and output, the manipulated variable calculation stage,
It is characterized in that the plurality of operation amounts are respectively determined using the plurality of predicted values.

The predicting means simulates, for example, an input or output characteristic equivalent to the operation amount calculating means for determining an operation amount having a smaller time constant with respect to the controlled amount among a plurality of operation amounts. It is preferable to include at least a controller model.

The predicting means may be, for example, a process model which simulates a relationship between at least one of a control signal or an operation amount relating to a predetermined process and a process amount which is a controlled amount, for example, equivalent to the operation amount calculating means. Alternatively, a configuration including a controller model simulating the input / output characteristics may be adopted.

Further, in the above-mentioned present invention, a configuration may be further provided that includes predicted time determining means for respectively determining a plurality of predicted times for a corresponding operation amount based on a predefined control performance evaluation value.

Further, in the present invention described above, it is more preferable that the apparatus further comprises parameter setting means for setting parameter values of the controller model and the process model based on a control performance evaluation value defined in advance.

Further, in order to achieve the above object, the present invention provides a prediction means for predicting a future value of a steam temperature of a thermal power plant, and the thermal power generation based on a steam temperature predicted value which is an output value of the prediction means. A thermal power plant having one or more operation amount calculating means for determining at least one of a fuel flow rate, a steam temperature desuperheater spray flow rate, a recirculation gas flow rate, and a gas distribution damper opening degree of the plant as an operation amount. In the control device or method, for the steam temperature at the same position, calculate and output a plurality of different predicted values at least one of the predicted time and the predicted time interval,
The operation amount of two or more of the fuel flow rate, the steam temperature desuperheater spray flow rate, the recirculation gas flow rate, and the gas distribution damper opening degree is determined using the plurality of output predicted values. And

Further, in the control apparatus or method for a thermal power plant according to the present invention, at least the estimated value of the combustion gas temperature and the measured steam temperature are used to determine the steam temperature after the time when the steam temperature is measured. When estimating the heat transfer tube temperature of the heat exchanger and estimating the steam temperature after a lapse of a predetermined time, the estimated steam temperature and the estimated heat transfer tube temperature are used, and the combustion gas is used. The temperature may be calculated using at least the estimated value of the heat transfer tube temperature before the time and the measured value of the steam temperature after the time.

Further, the control device or method according to the present invention can be applied to a waste incineration process in the same manner as a thermal power plant. However, in this case, the controlled amount is at least one of a CO (carbon monoxide) concentration, a phenol chloride concentration, a nitrogen oxide concentration, and a dioxin concentration in the exhaust gas related to the waste incineration process. The operation amount is one of the following: the amount of waste input into the incinerator, the transport speed in the incinerator, the air flow rate for drying, the air flow rate for combustion, the air temperature, and the chemical input rate for detoxifying the combustion gas. At least.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described.

In this embodiment, an example is described in which the present invention is applied to steam temperature prediction control of a thermal power plant. Hereinafter, description will be made with reference to the drawings.

The target coal-fired thermal power plant 10
FIG. 2 shows an example of the basic configuration of 0.

In the coal-fired thermal power plant 100 of the present embodiment, the fuel and the air are
And the water is circulated by the water supply pump 140 and evaporated by the furnace water wall 152. Further superheater 1
The superheated steam that has been heated in 54 is guided to the high-pressure turbine 130 via the turbine control valve 121 to drive the high-pressure turbine 130. The steam that has passed through the high-pressure turbine is heated again by the reheater 156 and enters the low-pressure turbine 120. Electric power is generated in the generator 110 by the rotation of the high-pressure turbine 130 and the low-pressure turbine 120.

Hereinafter, the steam at the inlet of the high-pressure turbine 130 is referred to as main steam, and the steam at the inlet of the low-pressure turbine 120 is referred to as reheat steam. In order to control the temperatures of the main steam and the reheat steam, the superheater 154a and the superheater cooler 154a are provided at the inlets of the reheater 156.
And a reheater desuperheater 156a. Temperature reducers 154a, 156
Low-temperature water after passing through the water supply pump 140 is guided to a, and is injected into high-temperature steam through the desuperheater spray water amount control valves 154b and 156b, respectively.

The reheater desuperheater 156a is an emergency and operates only when the reheat steam temperature exceeds a set value.

The boiler 150 is provided with a gas recirculation fan 142 for recirculating the combustion gas, which also serves as a means for adjusting the steam temperature.

Further, depending on the plant, a gas distribution damper may be provided between the superheater and the reheater to divide the gas flow path by dividing the gas flow path into two and adjust the distribution of each gas flow rate. . In this case, the gas distribution damper also serves as a means for adjusting the steam temperature.

In the thermal power plant, in addition to the above components, there are also devices such as a condenser 125 for cooling the steam after driving the turbine with cooling water 126 and a combustion exhaust gas treatment device 170. The gas that has passed through the exhaust gas treatment device 170 is released from the chimney 175 to the atmosphere.

The operation states of the plant 100 are as follows: generator output measuring device 111, main steam temperature (superheater outlet steam temperature) measuring device 122, superheater inlet steam temperature measuring device 127, main steam pressure measuring device 123, reheated steam (Reheater outlet steam temperature) Measured by a data measuring device such as a temperature measuring device 124 and a reheater inlet steam temperature measuring device 124a, and transmitted to the operation control device 300. The plant is provided with other devices for measuring various process quantities, and the operation control device 300 also captures the measured values. Here, the detailed description thereof is omitted.

The operation control device 300 grasps the operation state of the plant based on these process data, and controls the fuel flow control valve 162,
Devices such as an air flow control valve 161, a turbine control valve 121, and a water supply pump 140 are controlled.

In a thermal power plant, there is a control variable having a relatively long response time, such as a steam temperature, and this control is generally considered to be difficult. Therefore, a predictive control technique which is a feature of the present invention is applied to control of the main steam temperature and the like.

The purpose of predictive control is to predict the future behavior of a process value having a large time constant and a slow response,
An object of the present invention is to improve control accuracy by determining an operation amount in advance.

The functional configuration of the operation control device 300 will be described. As shown in FIG. 3, it is composed of a master control section 370 and a sub-loop control section 390 based thereon. The master controller 370 includes a normal controller 375 and a steam temperature prediction controller 380 to which the present invention is applied.
And divided into The master control unit 370 creates various operation amount command signals based on the load command signal, and adds the correction based on measured values such as the steam temperature, the steam pressure, and the gas O ₂ concentration to determine the operation amount.

The sub-loop control unit 390 includes a turbine control controller 391 and a feedwater pump controller 39
2. Fuel flow control valve controller 393, pushing fan controller 394, induction fan controller 395,
There is a spray flow controller 396 and a gas recirculation flow controller 397.

These controllers are mutually connected to the signal transmission network 400, and can exchange signals. The output from each controller of the sub-loop control unit 390 is sent to each actuator 101 of the plant 100 to operate the device.

The operation control device 300 has a hardware configuration shown in FIG. It is connected to a signal transmission network 400 via an external input interface 301 and an output interface 302. The arithmetic processing unit 304 calculates and generates various command signals while storing the received signal in the storage device 303 as necessary.
The command signal is sent to the control target via the output interface.

The external input interface 301 is connected to an external input device 900 including a keyboard 930 and a mouse 940, and a data storage device 500. An image display device 910 and a magnetic disk device 950 are connected to the output interface 302, and function as an interface with an operator.

Next, the steam temperature prediction controller 380 to which the present invention is applied will be described with reference to FIG. The steam temperature predicting controller 380 includes a predicting unit 330 and a switching unit 339 for feeding back one of the predicted value Ys and the plant output value (control amount: steam temperature) Y to the controller 320. Parameter adjustment unit 34
0 is provided. When the prediction control is not performed, the switching unit 339 selects the steam temperature Y, and outputs the current measured value of the steam temperature for both the fuel flow rate and the spray water flow rate.

The manipulated variables for controlling the main steam temperature include a manipulated variable command value Ua for the fuel flow rate and a manipulated variable command value Ub for the desuperheater spray water flow rate. The manipulated variable command value Ua generates a manipulated variable reference value in the unit master 315 based on the load command, and that value and the measured value or the predicted value Ys of the control variable Y
(Proportional / integral) controller 3 based on the deviation between the target and the target value Rs
It is calculated as the sum with the output value of 20a.

The manipulated variable command value Ub is calculated as the output value of the P (proportional) controller 320b based on the deviation between the measured value or the predicted value Ys of the control amount Y and the target value Rs.

A feature of the present invention is that two predicted values Ys (t +) having different prediction times for different manipulated variables (fuel flow rate, desuperheater spray water flow rate) with the same process quantity to be predicted.
τa) and Ys (t + τb) are calculated and output.

The fuel flow rate, which is one of the manipulated variables, takes a relatively long time, from several minutes to about 20 minutes, after operating the fuel flow rate to exert its influence on the steam temperature. This is due to the presence of factors such as a fuel combustion process, a heat transfer process from combustion gas to steam, and a gas or steam flow delay. On the other hand, since the desuperheater spray water, which is the other operation amount, acts at the inlet of the superheater 154, the main steam temperature, which is the steam temperature at the outlet position of the superheater 154, is within several minutes. The effect appears.

For this reason, for example, when selecting an effective predicted time for the fuel flow rate, the predicted time is too long for the desuperheater spray water flow rate. In this case, not only the effect of the predictive control on the spray water flow rate is not sufficiently exerted, but also the control performance of the steam temperature is reduced.

According to the present invention, a prediction value having a different prediction time is output according to the temporal characteristic of the operation amount to be determined, so that the effect of the prediction control is maximized even for a plurality of operation amounts. be able to.

In this example, although two predicted values are output,
In the present invention, the number of predicted values to be output is not limited.

The function of the prediction unit 330 in FIG. 1 will be described.

First, the target value of the main steam temperature before the target time (predicted time) is calculated. The main steam temperature target value may be constant, but may change with a change in load.

When the load change plan is known, a future load change is calculated from the steam temperature target value Rs (Equation 1) which is a function of the load, based on the load change plan.

[0050]

(Equation 1)

If the future load change is unknown, the target steam temperature Rs (k) at the current sampling step k, the target value Rs (k-1) one sampling step before, and the sampling time interval Δt seconds are used. , And a target value n seconds ahead is calculated by the following equation.

[0052]

(Equation 2)

In the present embodiment, the change in the target value is obtained by approximating it with a linear expression, but the present invention is not limited to this method. Further, the target value may be the value R (k) in the current sampling step k.

The future steam temperature target value thus obtained is predicted how many times the steam temperature will be C at the same time in the future.

Next, a method of determining the manipulated variable will be described. The control system employs PI (proportional / integral) or P (proportional) control. The specific calculation method will be described using PI control as an example.

[0056]

(Equation 3)

[0057]

(Equation 4)

Here, U is the aforementioned manipulated variable command value, ΔU is a change in the manipulated variable command value, Ys is a predicted steam temperature value, Kc is a proportional gain, and Tc is an integration time.

When the manipulated variable command values Ua and Ub are determined by Equations 3 and 4, the opening of the fuel flow control valve 162 and the spray water flow control valve opening 154b of the plant 100 are operated from these values.

As shown in FIG. 5, for example, the prediction unit 330 of this example includes a furnace water wall model 331, a superheater desuperheater (spray) model 332, a superheater model 333, and a load command (plan). A control target model including function generators 331a to 331e for setting various process quantities based on the values, controller models 334 and 335 equivalent to the functions of the PI controller 320a and the P controller model 320b in FIG. It is composed of

Therefore, the prediction unit 330 of this embodiment is a model that calculates a predicted value in consideration of the influence of the control system of the plant.

The furnace water wall model 331, the superheater desuperheater model 332, and the superheater model 333 are modeled based on the energy conservation equation.

The model equation of the furnace water wall model 331 is shown below. In the furnace water wall, the amount of heat transfer is expressed as the sum of radiant heat transfer and convective heat transfer

[0064]

(Equation 5)

[0065]

(Equation 6)

[0066]

(Equation 7)

[0067]

(Equation 8)

[0068]

(Equation 9)

Here, V is volume [m ³ ], γ is specific weight [kg /
m ³ ], H is enthalpy [J / kg], F is flow rate [kg / s], A is heat transfer area [m ² ], α is heat transfer coefficient [J / (m ² · s · K)], β is a coefficient related to the radiation heat transfer coefficient and the effective radiant heat transfer area ^{[J / (s · K 4} )], θ
Is temperature [° C], M is weight [kg], and C is specific heat [J / (kg · K)]. Also, suffix s is steam, m is heat transfer tube (metal), f is fuel, a is air, grf is recirculated gas, ms is steam from heat transfer tube, gm is heat transfer tube from combustion gas, in is inlet position, o represents each exit position. The numbers 1 in the subscript indicate the furnace water wall, 2 indicates the superheater desuperheater section, and 3 indicates the superheater section.

The model formula of the superheater desuperheater model 332 is shown by the following formula.

[0071]

(Equation 10)

[0072]

[Equation 11]

Here, the subscript s2o is the outlet of the desuperheater, 2in
Represents spray water.

In the superheater model 333, the characteristics of the heat exchanger are modeled based on the energy conservation equation. The model formula is shown.

[0075]

(Equation 12)

[0076]

(Equation 13)

[0077]

[Equation 14]

[0078]

(Equation 15)

The superheater model 333 constitutes a state observer based on Equations 14 and 15, and has a function of estimating the superheater inlet gas temperature using this.

The gas temperature estimating function will be specifically described.

The enthalpy and the temperature have the following relationship:
It is possible to convert values to each other. Equations 16 and 17 are relational expressions for steam, but the same relation holds for gas.

[0082]

(Equation 16)

[0083]

[Equation 17]

Here, C _Ps is constant pressure specific heat [J / (kg · K)], H _sBo
Is the reference enthalpy [J / kg].

By substituting Equations 16 and 17 into Equations 14 and 15, the following equation is obtained.

[0086]

(Equation 18)

[0087]

[Equation 19]

Here, when Equations 18 and 19 are arranged, the following equation is obtained.

[0089]

(Equation 20)

[0090]

(Equation 21)

[0091]

(Equation 22)

[0092]

(Equation 23)

[0093]

(Equation 24)

[0094]

(Equation 25)

[0095]

(Equation 26)

[0096]

[Equation 27]

[0097]

[Equation 28]

[0098]

(Equation 29)

[0099]

[Equation 30]

[0100]

(Equation 31)

Equations (20) and (21) are discretized in time and are expressed in a matrix as follows.

[0102]

(Equation 32)

[0103]

[Equation 33]

[0104]

(Equation 34)

[0105]

(Equation 35)

[0106]

[Equation 36]

[0107]

(37)

Equation 32 is a state equation, and an output equation is represented by the following equation.

[0109]

(38)

Here, Z _M is an observation vector (x1: corresponding to the outlet steam temperature), C _M is an observation matrix, and V _M is an observation noise vector.

Next, the steam temperature is predicted based on the physical equation model.

Of the state X _M (k), x ₁ (superheater outlet steam temperature) can be measured, but x ₂ (superheater heat transfer tube temperature) is difficult to measure. X
Estimate _M (k).

State observer 31 using Kalman filter
4 will be described.

Assuming that the current sampling step is k,
(K-1) The state value calculated from the state equation (Equation 32) using various values in the sampling step is a superscript P
Is expressed by adding a superscript SP to the maximum likelihood estimation value obtained by forming a Kalman filter.

X _M ^SP (k) is obtained by the following equation.

[0116]

[Equation 39]

[0117]

(Equation 40)

[0118]

[Equation 41]

[0119]

(Equation 42)

Here, the superscript T indicates a transposed matrix, and the superscript -1 indicates an inverse matrix.

Using this state observer, the superheater outlet steam temperature can be calculated. As the previous value of the superheater outlet steam temperature, the measured value is used only for the first time of the calculation, but the second and subsequent times are calculated using the calculated value. Controller model 3 based on the deviation between the calculated value of the superheater outlet steam temperature and the target value
At 34 and 335, the manipulated variables (fuel flow rate,
Spray water flow rate) is calculated, the fuel flow rate of the furnace water wall model 331 and the spray water flow rate of the desuperheater model 332 are changed, and the superheater outlet steam temperature is calculated again. This operation is repeated up to the number of times corresponding to the predetermined prediction time to calculate the predicted value of the steam temperature.

The superheater inlet gas temperature θ _g3in is required on the right side of Expression _15, but it is difficult to directly measure a high gas temperature. Although the gas temperature can be calculated by the furnace water wall model 331, the gas temperature can be more accurately determined using the above-described state observer.

If the current sampling step is k,
Using the values in step (k-1), the current superheater outlet steam temperature is estimated by Expression 36. In order to predict the future steam temperature, it is assumed that the steam temperature at the current step k is correctly determined from the information obtained at the time (k-1).

[0124] When the estimated to have gas temperature _θ g3in is not appropriate, it is cause for an error in the X _M ^SP (k). Therefore, in the present invention, the superheater outlet steam temperature x ₁ in step k at present.
The measured value of (k) (= θ _s3o (k)) and the maximum likelihood estimate x ₂ ^SP (k-1) of the heat transfer tube temperature in step (k-1) (= θ _m3 ^SP (k-1))
_{Is used} to derive the following equation from the relation of the equation of state (Equation 32) to calculate the gas temperature θ _g3in .

[0125]

[Equation 43]

Here, θ _s3o (k) is the measured value of the outlet steam temperature in step k.

The gas temperature can be determined such that the estimated value of the superheater outlet steam temperature in step k now matches the measured value. Outlet gas temperature θ of furnace water wall model
Based on the deviation between the calculated gas temperature theta _G3in in _g1o the formula (43), and sequentially modify the beta _gm1 of the furnace waterwall model correction section 336. According to this method, an appropriate gas temperature can be determined successively, and the characteristics of the furnace water wall model 331 can be corrected sequentially, so that the model error in the current step is reduced. In addition, since the model error is small, the prediction accuracy is improved.

In this example, the superheater inlet steam temperature θ _s3in
Uses the value in the sampling step (k-1), but the measured value in step k may be used.

The specific method of the prediction method has been described above.
The present invention is not limited to the prediction method shown in the present embodiment. The furnace water wall model 331, the superheater desuperheater model 332, and the superheater model 333 may be other than the physical model that simulates the physical causal relationship as long as the model simulates the characteristics of the target device. For example, a regression model using a statistical method or a learning model using a neural network may be used. At this time, a correlation model that outputs a deviation between the target value and the measured value of the steam temperature may be used.

In the present invention, different predicted values with different predicted times are calculated for different manipulated variables for controlling the same controlled process amount. The parameter adjustment unit 340 determines the predicted time τa for the fuel flow rate and the predicted time τb for the spray flow rate as shown in FIG. 6, for example, and also determines the proportional gain and the integration time of the controllers 320a and 320b.

As shown in FIG. 7, for example, the parameter adjustment unit 340 includes a plant model 355, a prediction function model 350 simulating the prediction unit 330, and a prediction function model 35.
0 and a model adjustment unit 342 that adjusts a control target model parameter in the plant model 355;
(Proportional gain Kc1, integration time Tc), the parameter of controller 320b (proportional gain Kc2), and parameter optimizing unit 348 for calculating predicted time intervals τa, τb.

The plant model 355 includes the controller 320a
And 340b equivalent to controller models 346a, 346b
And a model 344 of the control target 100, and the prediction function model 350 includes the same model as the plant model 355.

The controller model 346a is PI (proportional / integral)
It is defined by the transfer function of the following equation representing the controller.

[0134]

[Equation 44]

Here, s is a Laplace operator.

Similarly, the controller model 346b has P (proportional)
The controller is defined by the following equation.

[0137]

[Equation 45]

The control object model 344 includes a model 344a simulating the relationship between the fuel flow rate and the steam temperature, and a model 344b simulating the relationship between the temperature reduction spray water flow rate and the steam temperature.

The two models 34 of the control target model 344
4a and 344b are simulated with first order delay and dead time, respectively.
It is defined by the following transfer function.

[0140]

[Equation 46]

[0141]

[Equation 47]

Here, Kp1 and Kp2 are gains, Tp1 and Tp2 are time constants, and Lp1 and Lp2 are model parameters corresponding to dead time.

Since the prediction function model 350 has exactly the same model as the plant model 355, the output value of the plant model 355 is predicted by performing simulation calculation up to the previous time step. The predicted value is output to the plant model 355, the deviation from the target value is calculated, and the output of the plant model 355 is calculated via the controller models 346a and 346b.

That is, the case where the prediction control is performed on the actual control target 100 is simulated. The parameter adjustment unit 340 calculates the prediction times τa and τb and the proportional gain, which is a controller parameter, so that the control target model 344 becomes the most desirable state by the simulation calculation of the prediction control.
Kc1, Kc2 and integration time Tc are determined.

A specific method for determining the predicted time, controller parameters, and the like will be described. First, it is necessary to match the response characteristics of the control target model 344 with the characteristics of the control target 100.
In a thermal power plant, test operation is performed at the time of new plant construction or at the end of periodic inspections to understand plant characteristics.

As one of them, a step response test is performed for various manipulated variables. The model adjustment unit 342 fetches the data of the step response test, and
Model parameters Kp1, Kp2, Tp1, T of 44 and 344 '
Adjust p2, Lp1, Lp2.

An operator intervenes in these adjustments, and the adjustment is performed offline using the test operation data. FIG. 8 shows a man-machine interface for adjusting model parameters.
FIG. 8 is displayed on the image display device 910 and the external input device 9 is displayed.
Operation and data input can be performed using 00.

As an example, the adjustment process of the control target model 344a will be described. The control target model 344b can be adjusted in the same procedure.

In the graph display area 911, a response waveform 915 of the step response at the time of the trial operation of the plant is displayed in a graph. Auxiliary lines 912, 914, 916 are displayed to determine model parameters Kp1, Tp1, Lp1 from this waveform. The auxiliary line 914 indicates the steady-state output level A at the start of the step response test, and can be moved up and down on the screen using the mouse pointer cursor 918. The output level is shown as a percentage with respect to the rated level.

Further, in the numerical value display area 920, in addition to the output level A, various numerical values are displayed.
It can be changed directly by input from 30. By clicking the increase / decrease pointer 922 with a mouse, the displayed value can be increased or decreased at minute intervals. Auxiliary line 914 and numerical display area 920
The value inside changes in conjunction with it.

With these methods, the operator determines the value of the output level A. Similarly, the output level B in the steady state at the end of the step response test indicated by the auxiliary line 912 is determined.

Further, the tangent line 916 at the inflection point at the time of the rise of the response waveform is determined. The tangent line 916 is the output level A
And B are determined by determining the points C and D shown in FIG.

When the points A to D are determined from the above, the model parameters Kp1, Tp1, and Lp1 are determined as follows.

Change width K2 of manipulated variable during step response test
The gain Kp1 is

[0155]

[Equation 48]

[0156]

[Equation 49]

It is assumed that:

Step response test start (operation amount change start)
From time Q, the dead time Lp1 is

[0159]

[Equation 50]

It is assumed that

Further, the time constant Tp1 is expressed by a line segment E on the tangent line 916.
From time P when the length of P becomes 0.632K1,

[0162]

(Equation 51)

It is assumed that

FIG. 9 summarizes the processing procedure of the model adjustment unit 342 as described above. That is, in this processing procedure, the target value R and the controlled variable Y1 are read as step response test data (step 5001), and the output level is
Points C and D for specifying the shapes of A and B and the close battle 916
Are determined (steps 5002 and 5003), and these points A
From the values of ~ D, the parameters Kp1, Tp1, Lp of the plant model
1 is determined (step 5004).

Next, a method of determining the predicted times τa, τb and the controller parameters Kc1, Kc2, Tc will be described.

The parameter optimizing unit 348 sets parameters τa, τb, Kc1, Kc2, and Tc for minimizing the control performance evaluation value J.
Find the combination of Here, the control performance evaluation value J is defined by the following equation.

[0167]

(Equation 52)

Here, Rm (t) is a model target value, Ym (t) is an output value of the controlled object model 344, Uam (t) and Ubm (t) are output values of the controller models 346a and 346b, respectively. This corresponds to the operation amount for the target model 344. p and q are weighting factors.

The value of J can be approximately calculated by simulation. The values of the first, second, and third terms in the right-hand side of Equation 52 are calculated in each calculation time step, and the sum of them is calculated. That is, if the calculation time interval is Δt, it is expressed by the following equation.

[0170]

(Equation 53)

The controller parameters Kc1, Kc2, and Tc define upper and lower limits, respectively, as possible values of the controller. The prediction times τa and τb are respectively 0 to Tp1 × αa and Tp
Up to 2 × αb. αa and αb are coefficients and can be set arbitrarily.

Within this range, the simulation calculation is performed from t = 0 to T by sequentially changing the parameters at a predetermined step size, and the evaluation value J is calculated each time. Finally the evaluation value J
The combination of the parameters τa, τb, Kc1, Kc2, and Tc that minimizes is defined as the optimum value.

Since the control system may not be stable depending on the combination of τa, τb, Kc1, Kc2, and Tc, it is necessary to confirm the stability using Hurwitz's stability judgment method before calculating the evaluation value J. Is desirable. For an unstable combination, the parameter value is changed without calculating the evaluation value J, and the optimum value search is advanced.

The method for obtaining the optimum combination of the parameters τs, Kc, and Tc is not limited to the above method, and an optimization method such as a hill-climbing method or a genetic algorithm may be used.

Even when the model target value Rm (t) changes, the optimum value of the parameter can be calculated similarly.

At the time of starting the thermal power plant, there is a heating process in which the steam temperature target value is increased according to a predetermined time schedule. In the heating process, it is important not only to reduce the deviation from the target value but also to match the heating rate with the target heating rate in consideration of the influence on the equipment material.

In this case, the evaluation value J (Equation 52, right-hand side {｛})
It is desirable to add the following item of temperature rise rate to (in).

[0178]

(Equation 54)

Here, r is a weighting factor like p and q.

The controller model 346, the control object model 344, and the prediction function model 350 are represented by the following equations (44) and (45), respectively.
Although defined in 46, 47, these modelings are not limited to this method. For example, the prediction unit 330 described above
It may be configured with the exact same model as.

The evaluation value J is not limited to the expression 52 or 53, but can be set arbitrarily. Further, a method of calculating an evaluation value for a closed loop system with feedback is disclosed in, for example, the following document.

"Robust IP with Generalized ISE as Evaluation Criteria"
D Controller Minimax Optimization Design ", Toru Kawabe,
Toru Katayama, Transactions of the Society of Instrument and Control Engineers, Vol.32, No.8, pp1
226-1233 (1996).

With this method, regardless of the simulation
In some cases, the evaluation value from t = 0 to the evaluation time ∞ can be calculated.

The parameter optimizing unit 348 described above
10 is shown in FIG.

In this processing procedure, step 110 in FIG.
As shown in 1-1104, parameters τa, τb, Kc
The values of 1, Kc2, and Tc are each changed within a predetermined range at a predetermined interval, and when the evaluation of all the combinations is completed, the optimal value search ends.

As described above, an optimization method such as a hill-climbing method may be introduced for the search for the optimum parameter value.

As a method for determining the predicted time used in the present invention, other methods such as the method described in Japanese Patent Application Laid-Open No. 9-14612 may be used.

Next, when the target value changes over time,
A method for setting the target value Rm (t) will be described.

FIG. 11 shows a man-machine interface for supporting the setting of the model target value Rm (t). In the graph area 911, the value 93 of the target value Rm with respect to the time axis (horizontal axis)
A graph of 0 is displayed.

By clicking and dragging a point on the line of the Rm value 930 on the graph with the mouse, the graph waveform shape can be changed. The coordinates of the polygonal line at that time are displayed in the coordinate display area 932 as numerical values. If all the coordinate points cannot be displayed at once in the coordinate display area 932, a scroll bar 934 is displayed, and the user can scroll up and down by operating the mouse.

The evaluation time T required for calculating the evaluation value J in Expression 47 is input numerically from the keyboard in the evaluation time input field 936. After inputting the target value Rm and the evaluation time T, the setting is completed by clicking the setting button 938 with a mouse.

FIG. 12 shows a simulation example of the control result when the prediction time intervals τa, τb and the controller parameters Kc1, Kc2, Tc are determined as described above. This example is a response when the target value is changed stepwise.

According to the above configuration, it can be seen that the follow-up to the target value is faster, the amount of overshoot is reduced, and the control performance is improved as compared with the case where the predictive control is not performed.

Further, according to the above configuration of the present invention,
The prediction time at which the control performance is most improved can be determined according to the operation amount, and the optimal value of the controller parameter when the prediction control is performed can be determined, so that the effect of the prediction control method is maximized. be able to.

Further, according to the above configuration of the present invention,
Conventionally, these controller parameters are adjusted by trial and error, but there is also an effect that the adjustment time can be reduced.

The effect of the simulation is displayed on the screen as shown in FIG. 13 so that the user can confirm the effect.

For example, in a graph area 911, a response waveform based on a simulation result is displayed. In the parameter display field 924, the parameters τa, τb,
Displays the values of Kc1, Kc2, and Tc. Also, the switch button 92
6 is displayed in the graph area 911 each time the mouse is clicked on. The model output value Ym (t) and the manipulated variable Uam
(t) and then to Ubm (t).

Further, a function may be provided for switching between the model output value Ym (t) or the manipulated variables Uam (t) and Ubm (t) by using the switch button 928 or displaying the graphs side by side. good.

In the above, the method of performing the off-line adjustment using the parameter adjustment unit 340 at the time of the trial operation of the plant or at the time of the periodic inspection has been described. The present invention controls the operation of the plant as described below. You can also use it online.

For example, the parameter adjusting unit 340 is periodically activated at predetermined time intervals during the operation of the plant. For example, it starts every five minutes, and the parameter adjustment unit 348 determines the predicted times τa and τb and the controller parameters Kc1, Kc2, and Tc. The processing procedure of the parameter adjustment unit 348 is the same as that of FIG.

In this way, an appropriate prediction time interval τs and controller parameters Kc and Tc can be determined at any time according to the operation state of the plant 100 (load zone, load change conditions), so that control performance is further improved. Can be.

Further, before determining the predicted time interval τs and the controller parameters Kc and Tc, the model adjustment section 342 changes the control target model 344 as needed so that the characteristics of the control target model 344 match the characteristics of the control target 100. It can also be adjusted. If the control target model is a regression equation model or a neural network model, the configuration may be such that the correlation equation is recreated with the latest (immediately before) data, or the model characteristics are sequentially adjusted by re-learning with the neural network.

According to the above configuration, the characteristics of the controlled object model 344 can always be matched to the characteristics near the operating point (operating state) of the plant, so that the effect of predictive control is further improved. In addition, since the control target model 344 does not need to simulate a wide range of operating conditions of the plant, it can be approximated by a relatively simple model, and has an effect of reducing the calculation load.

Next, a second embodiment of the present invention will be described.

The control device of this embodiment uses the steam temperature prediction controller 380 having the configuration shown in FIG. 14 in the first embodiment.

The difference from the steam temperature prediction controller 380 in the first embodiment is that the predicted value Ys and the target value
Instead of feeding back the deviation from Rs to the PI controller 320a and the P controller 320b, the plant output value (measured steam temperature) is fed back, and from the predicted value Ys, the manipulated variable is output by the P controllers 320c and 320d. This is a configuration in which correction is made to. In this case, plant model 3
55 has the same configuration.

Further, the switching unit 339 has a function of switching between outputting the correction value based on the predicted value as it is, or outputting 0 and not using prediction control substantially.

In the present embodiment, the P controllers 320c and 320d are used as the correction amount calculation units for the operation amounts based on the predicted values, but other methods may be used.

According to the configuration of the present embodiment, similarly to the first embodiment, in a process in which a plurality of manipulated variables exist for the same control variable, each process corresponds to each of the plurality of manipulated variables. It is possible to calculate prediction values at different prediction times and determine the operation amount based on each prediction value.

Next, a third embodiment of the present invention will be described.

In this embodiment, the present invention is applied to an operation control device of a waste incineration plant. The outline of the waste incineration plant will be described with reference to FIG.

FIG. 15 shows an example of the configuration of a waste incineration plant 1000 having a stoker type incinerator.

[0213] The waste (garbage) discharged from ordinary households is
Collected by the garbage truck 700. The collected waste is temporarily stored in a waste pit 614.

[0214] The waste in the waste pit 614 is
At 2, it is carried to the incinerator hopper 610. On the way, there is a weight measuring device 201 for measuring the weight of the refuse. Hopper 61
After being put into 0, it is carried on the stoker by the extruder 630. The stoker is divided into a former drying stoker 626 and a latter burning stoker. The drying stoker mainly evaporates the moisture of the input waste. The combustion stoker burns dry refuse. The ash after burning is ash pit 638
Once deposited on the land, it is landfilled.

The drying state of the refuse is mainly controlled by the transfer speed of the drying stoker, and the amount of combustion is mainly controlled by the transfer speed of the combustion stoker.

Dry stoker 626 and combustion stoker 62
The drying air and the combustion air are blown into the portion 8 from below. The air is sent in by a fan 620, and an air heater 622 for heat exchange (not shown) with a part of the combustion gas to raise the temperature is installed in the middle of the air.

Dry stoker 626 and combustion stoker 628
The amount of air to be supplied is controlled by the opening of the air damper 624 that adjusts the amount of distribution to the fans 620 and respective stokers.

After the dust and harmful substances have been removed to some extent by the exhaust gas treatment device 632, the combustion gas
Is discharged from

A part of the combustion gas is sent to the gas analyzer 115 before being discharged from the chimney, and the CO (carbon monoxide) concentration, O ₂ (oxygen) concentration and NOx (nitrogen oxide) concentration are adjusted as needed. Measured online. In addition, dioxin concentrations are regularly measured off-line.

[0220] In addition, two places above the stokers 626 and 628, one place at the entrance of the exhaust gas treatment device 632, and a chimney 634.
Gas temperature sensors 640, 642, 64
4, 646 are installed to measure the combustion gas temperature.

It is important to control the emission of harmful substances in facilities that incinerate waste. Particularly, it is known that dioxins are thermally decomposed by keeping the inside of a furnace at a high temperature. Maintaining a high temperature state basically means maintaining a complete combustion state. However, in the incineration of waste, the composition and water content of the waste to be introduced into the incinerator are not constant, and it is difficult to stably burn the waste.

Thus, the predictive control method of the present invention is applied for the purpose of suppressing the emission of dioxins.

It is desirable that the concentration of dioxins can be measured on-line, but if it cannot be measured, process data serving as an index of the concentration of dioxins is substituted. As one of them, there are CO concentration and chlorinated phenol, and these values can be measured in a relatively short time, and a correlation with the dioxin concentration is recognized to some extent.

Therefore, the process amount to be predicted is set as the CO concentration in the exhaust gas. This value is measured online by the gas analyzer 115. In this example, the CO concentration in the exhaust gas is used, but any other substance that can estimate the concentration of dioxins such as phenol chloride and can be measured in a relatively short time may be used.

The process amount (operation amount) operated based on the predicted value is the transfer speed of the combustion stoker and the amount of waste input (speed).

The operation control device 300 of this embodiment is similar to the case of the thermal power plant described in the first embodiment, in that a control system for determining the operation amount of each actuator disposed in the incineration plant 1000 is provided. In the control system, the process amount to be controlled (CO concentration in this example)
Of the prediction controller 380 are provided.

The prediction controller 380 is, for example, as shown in FIG.
As shown in (1), an operation amount 1 calculation unit 321a and an operation amount 2 calculation unit 321b for calculating an operation amount for changing the control process amount (CO concentration), and prediction values of different prediction times corresponding to the respective operation amounts , A parameter adjustment unit 340 that adjusts parameters of various models such as a controller model included in the prediction unit 330, and a switching unit 3
32 and P controllers 321c and 321d.

In the prediction controller 380 of the present embodiment, the operation stoichiometry calculation units 321a and 321b calculate the combustion stoker transfer speed and the amount of waste input (speed) based on the weight of the waste measured by the weight measuring device 201. Furthermore, the future CO concentration is predicted by the predictive control method of the present invention, and the P controller 321c and 321d correct the transfer speed of the combustion stoker and the operation amount with respect to the amount of waste input.

[0229] The input waste is first dried stoker 626.
After drying above, it moves to the combustion stoker 628. For this reason, the time until the change in the amount of waste input affects the combustion state is relatively long. In contrast, changes in the combustion stoker transfer rate immediately affect the current combustion state. Therefore, when performing prediction control for both operation amounts, it is effective to use a prediction time that is appropriate for each.

Next, the prediction section 330 of this embodiment will be described.

The CO concentration in the exhaust gas was modeled by the following equation as a function of the amount of waste supplied (speed), the transfer rate of the combustion stoker, the flow rate of the dry air, the flow rate of the combustion air, and the measured value of the CO concentration (up to the previous sampling). I have.

[0232]

[Equation 55]

Here, CO is the CO concentration in the exhaust gas, Fw is the amount of waste supplied (speed), CSV is the transfer rate of the combustion stoker, Fda is the flow rate of the dry air, Fca is the flow rate of the combustion air, and k is the current sampling step. Represent.

At the time of prediction, while the values of Fw, CSV, Fda, and Fca are held at the values of the current step, the value of CO on the right side of Equation 55 is sequentially replaced with the calculated value on the left side to calculate the future CO concentration.

FIG. 17 shows the configuration of the prediction unit 330. number
55 CO correlation model 360, models 322a and 322b of the operation amount calculation units 321a and 321b, and a P controller 321
and 321d models 322c and 322d.

In the present embodiment, a parameter adjusting unit 340 is provided as in the first embodiment, and the estimated time corresponding to the combustion stoker transfer speed and the amount of waste input (speed), and the P controller 321c And 321d are determined.

In this embodiment, the combustion stoker transfer speed and the amount of waste are selected as the manipulated variables for the predictive control. However, the dry stoker transfer speed, the dry air amount, the combustion air amount, and the like may be used. It may be a combination.

Further, since dioxins are compounds containing Cl (chlorine), there is a method of suppressing the emission of dioxins by fixing and removing chlorine in exhaust gas as another stable chlorine compound. In this case, for example, a calcium compound such as limestone or slaked lime can be used as a dioxin inhibitor. At that time, the chlorine content is fixed as CaCl ₂ (calcium chloride) by the reaction shown in the following equation.

[0239]

[Equation 56]

Therefore, the input amount of a dioxin inhibitor such as a calcium compound may be used as the operation amount. In addition, not only the concentration of dioxins as a control amount,
(Nitrogen oxide) concentration or NOx concentration may be predicted.

[0241]

According to the present invention, for a process in which a plurality of manipulated variables exist for a single controlled variable, predicted values at different prediction times corresponding to each manipulated variable are calculated. Since the operation amount is determined based on the predicted value, a control device with high control accuracy can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a prediction controller according to a first embodiment.

FIG. 2 is an explanatory diagram showing a configuration of a thermal power plant.

FIG. 3 is an explanatory diagram showing a relationship between a configuration of a control device and a plant.

FIG. 4 is a block diagram showing a hardware configuration of the control device.

FIG. 5 is a block diagram illustrating a configuration of a prediction unit.

FIG. 6 is an explanatory diagram of a prediction time interval.

FIG. 7 is a block diagram illustrating a configuration of a parameter adjustment unit.

FIG. 8 is an explanatory diagram showing a parameter adjustment support screen of a control target model.

FIG. 9 is a flowchart illustrating a processing procedure of a model adjustment unit.

FIG. 10 is a flowchart illustrating a processing procedure of a parameter optimization unit.

FIG. 11 is an explanatory diagram showing a target value setting screen.

FIG. 12 is an explanatory diagram showing the effect of predictive control.

FIG. 13 is an explanatory diagram showing an adjustment effect confirmation screen.

FIG. 14 is a block diagram illustrating a configuration of a prediction controller according to the second embodiment.

FIG. 15 is a block diagram showing a configuration of a waste incineration plant.

FIG. 16 is a block diagram illustrating a configuration of a prediction controller according to a third embodiment.

FIG. 17 is a block diagram illustrating a configuration of a prediction unit according to the third embodiment.

[Explanation of symbols]

100: thermal power plant, 110: generator, 111:
Generator output measuring device, 120 ... low / medium pressure turbine, 12
DESCRIPTION OF SYMBOLS 1 ... Turbine control valve, 122 ... Superheater outlet steam temperature measuring device, 123 ... Main steam pressure measuring device, 124 ... Reheat steam temperature measuring device, 125 ... Condenser, 126 ... Cooling water, 130 ... High pressure turbine, 140 ... water supply pump, 150 ... boiler, 1
52 ... furnace water wall, 154 ... superheater, 156 ... reheater, 1
Reference numeral 60: burner, 161: air flow control valve, 162: fuel flow control valve, 170: exhaust gas treatment device, 175: chimney, 300: operation control device, 320a, b: controller, 33
0: prediction unit, 331a: furnace water wall model, 331b: spray model, 331c: superheater model, 339: switching unit,
340: Parameter adjustment unit, 342: Model adjustment unit, 3
44: Control target model, 346: Controller model, 348
... parameter optimizing unit, 350 ... prediction function model.

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Masahide Nomura 7-1-1, Omikacho, Hitachi City, Ibaraki Prefecture Inside the Hitachi Research Laboratory, Hitachi, Ltd. (72) Inventor Toshihiko Harashima 5-chome Omikacho, Hitachi City, Ibaraki Prefecture No. 1 Inside Hitachi Ltd. Omika Plant (72) Inventor Toru Kimura 5-2-1 Omikacho, Hitachi City, Ibaraki Prefecture Inside Hitachi Ltd. Omika Plant (72) Inventor Shinya Kikuchi Omika Hitachi City, Ibaraki Prefecture 5-2-1, Machi F-term (reference) at Hitachi, Ltd. Omika Plant 3L021 BA08 DA05 DA25 DA26 DA27 DA28 EA01 FA08 FA12 FA13 FA15 5H004 GA10 GB04 GB20 HA01 HA04 HB01 HB03 HB04 JB23 KA01 KA65 KB02 KB03 KB04 KB19 KB33 KB38 KB38 KB39 KC09 KC10 KC12 KC26 KC28 KC44 KC45 KC48 KC50 KD46 KD67 KD70 LA01 LA03 LA12 LA18 MA27 MA50

Claims (15)

[Claims] 1. A process amount to be controlled (a controlled amount)
And a manipulated variable calculating means for determining a process quantity (operated variable) to be operated to control the controlled variable based on the predicted value output from the predicting module. In the control device, there is a plurality of manipulated variables for one controlled variable, and the prediction means differs in at least one of a predicted time and a predicted time interval for the same controlled variable. A process control device, wherein a plurality of predicted values are calculated and output, and the manipulated variable calculation stage determines each of the plurality of manipulated variables using the plurality of predicted values. 2. The control apparatus for a process according to claim 1, wherein the predicting unit includes a process model simulating the process to be controlled and a controller model simulating the manipulated variable calculating unit. A control device for a process, comprising: 3. The control device for a process according to claim 1, wherein the predicting means determines an operation amount having a smaller time constant with respect to the control target amount among the plurality of operation amounts. A process control device comprising at least a controller model simulating a partial configuration of the operation amount calculation means provided. 4. The process control device according to claim 1, wherein an evaluation value for evaluating a predefined control performance is calculated, and the plurality of predicted values are calculated using the evaluation value. A control device for a process, further comprising a predicted time determining means for determining a predicted time or a predicted time interval for each. 5. The control device for a process according to claim 2, wherein an evaluation value for evaluating a control performance defined in advance is calculated, and a parameter value of the controller model is set using the evaluation value. A predictive control device for a process, further comprising parameter setting means for performing the following. 6. The control device for a process according to claim 5, wherein the manipulated variable calculating means includes at least one of a proportional controller and an integral controller, and the parameter setting means sets the manipulated variable. The process control device according to claim 1, wherein the parameter value corresponds to a proportional gain of the proportional controller or an integration time of the integral controller. 7. A predicting means for predicting a future value of steam temperature of a thermal power plant, and a fuel flow rate and a steam temperature reducer spray of the thermal power plant based on a steam temperature predicted value which is an output value of the predicting means. A control device for a thermal power plant having one or more operation amount calculating means, wherein at least one of the flow rate, the recirculation gas flow rate, and the gas distribution damper opening degree is determined as an operation amount, For the steam temperature at the location
Calculating and outputting a plurality of predicted values at least one of which is different from a predicted time and a predicted time interval, wherein the manipulated variable calculating means uses the output plurality of predicted values to calculate the fuel flow rate and the steam temperature desuperheater. Spray flow rate,
A controller that determines two or more operation amounts of the recirculation gas flow rate and the gas distribution damper opening. 8. The control device for a thermal power plant according to claim 7, wherein the predicting means simulates a relationship between at least one of a control signal relating to operation of the thermal power plant and the manipulated variable and the steam temperature. A process model comprising at least one of a heat exchanger model and the steam temperature desuperheater model; and a manipulated variable calculation model equivalent to the manipulated variable calculation means or simulating the characteristics thereof. A control device that performs a calculation by combining a model and the manipulated variable calculation model to calculate a steam temperature at a specific time. 9. The control device for a thermal power plant according to claim 7, wherein the predicting means is configured to determine the fuel flow rate, the steam temperature desuperheater spray flow rate, the recirculation gas flow rate, and the gas distribution damper opening degree. And a controller model simulating an input / output characteristic equivalent to a part of the operation amount calculating means for determining an operation amount having a small time constant with respect to the steam temperature. Control device. 10. The control device for a thermal power plant according to claim 7, wherein an evaluation value for evaluating a control performance defined in advance is calculated, and the plurality of prediction values are calculated using the evaluation value. A control device, further comprising prediction time determination means for determining one of a prediction time and a prediction time interval for each value. 11. The control device for a thermal power plant according to claim 7, wherein the manipulated variable calculating means includes at least one of a proportional controller and an integral controller. It further comprises controller parameter setting means for calculating an evaluation value for evaluating a predefined control performance, and calculating a proportional gain of the proportional controller or an integration time of the integral controller using the evaluation value. Characteristic control device. 12. The control device for a process according to claim 1, wherein the controlled amount is a concentration of CO (carbon monoxide), a concentration of phenol chloride, a concentration of CO in exhaust gas related to a waste incineration process. Nitrogen oxide concentration, and at least one of dioxin concentration, the manipulated variable is the amount of waste input to the incinerator, transport speed in the incinerator, drying air flow rate, combustion air flow rate, A control device for a process, comprising at least one of an air temperature and a charge amount of a chemical for detoxifying a combustion gas. 13. A future value of a process amount (controlled amount) to be controlled is predicted, and a process amount (operated amount) operated to control the controlled amount is determined based on the predicted value. In the method of controlling a process, a plurality of manipulated variables exist for one controlled variable, and for a same controlled variable, at least one of a predicted time and a predicted time interval is different. A method for controlling a process, comprising: calculating and outputting a predicted value; and determining each of the plurality of manipulated variables using the plurality of output predicted values. 14. A predicting means for predicting a future value of a steam temperature of a thermal power plant, and a fuel flow rate of the thermal power plant based on a steam temperature predicted value which is an output value of the predicting means.
A method for controlling a thermal power plant having one or more operation amount calculating means, wherein at least one of a steam temperature desuperheater spray flow rate, a recirculation gas flow rate, and a gas distribution damper opening is determined as an operation amount, For a steam temperature at the same position, at least one of a predicted time and a predicted time interval is calculated and output a plurality of different predicted values, and the fuel flow rate and the steam temperature are calculated using the output plurality of predicted values. A control method comprising determining at least two operation amounts among a desuperheater spray flow rate, the recirculation gas flow rate, and the gas distribution damper opening. 15. The method for controlling a thermal power plant according to claim 14, wherein at least a combustion gas temperature estimated value and a steam temperature measurement value are used, and the steam temperature and the heat after the time when the steam temperature is measured are determined. A method of using the estimated steam temperature and the estimated heat transfer tube temperature when estimating the heat transfer tube temperature of the exchanger and estimating the steam temperature after the lapse of a predetermined time, comprising: Is calculated using at least the estimated value of the heat transfer tube temperature before the time and the measured value of the steam temperature after the time.