JPH08339204A - Autonomous adaptive optimization control system for thermal power station - Google Patents

Abstract

PURPOSE: To automatically produce a starting schedule which can minimize the starting time while satisfying plural operation limit conditions at a time by increasing the adoption ratio of the starting schedule that is acquired by an adaptive knowledge acquisition means. CONSTITUTION: An actual starting schedule is produced from a 1st starting schedule which is produced based on the schedule correction value calculated by a schedule correction value calculation means 600 and a 2nd starting schedule which is produced based on the schedule correction value acquired by an adaptive knowledge acquisition means 800. Then an autonomy management means 200 increases the adoption ratio of the 2nd starting schedule by increasing the autonomous coefficients that prescribe the adoption ratios of the 1st and 2nd starting schedules every time the start is repeated. Therefore, the rate of the 2nd starting schedule is increased and it is possible to produce a more suitable starting schedule in response to the change of the plant operation condition as the plant starting frequency is increased.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control system for a thermal power plant, and in particular, to create a start-up schedule suitable for starting up the plant in a short time while satisfying various operational restriction conditions. Control system for a thermal power plant in Japan.

[0002]

2. Description of the Related Art When a thermal power plant is to be started up rapidly, many problems are operating restriction conditions, nonlinear characteristics between input and output, and the presence of time delay. That is, the startup schedule to be created needs to satisfy not only the operation amount such as the fuel flow rate and the water supply flow rate, but also many process state values such as thermal stress generated in equipment and NOx discharged from the boiler to satisfy the operation restriction condition. . Further, since there is a large non-linearity and a time delay between the turbine acceleration pattern and the load increase pattern, that is, the operation amount that defines the startup schedule, that is, the turbine acceleration rate and the load increase rate, and the above process state value, there is a simple delay during startup. With feedback control, it was impossible to realize rapid start-up while satisfying the operation restriction conditions with high accuracy. For this reason, in the first conventional method, as described in Japanese Patent Laid-Open No. 63-94009, a dynamic system model of the entire system of the plant is built into the control system, and this is done before the actual plant startup. The start-up schedule was optimized by predicting the start-up characteristics by repeating the start-up simulation using. That is, the above non-linearity was modeled as table information or a physical formula and used to create the activation schedule. Also, as a second conventional method, as described in Japanese Patent Laid-Open No. 3-164804, the starting characteristic is evaluated based on the actual driving result, and the amount of modification of the starting schedule is calculated from this evaluation result by fuzzy inference. Having a means for calculating, learning the correction amount obtained at each startup to the neural network, and correcting the schedule at the previous startup with the value output from the neural network when the driving target is input to the neural network. Was used to optimize the startup schedule. In other words, the control system does not have a dynamic characteristic model of the entire plant system, and the start schedule is created by utilizing the qualitative knowledge of the expert and the learning ability of the neural network by fuzzy reasoning.

[0003]

However, according to the above-mentioned prior art, the following problems arise.

(1) According to the first conventional method, the entire system dynamic characteristic model used for predicting the starting characteristic takes into consideration operating conditions such as the temperature, humidity, water temperature surrounding the plant and the temperature condition of the plant equipment before starting. It is necessary to make it something that has been done, it takes a lot of time to make it, and it has to be made for each target plant. In addition, there are many uncertainties such as heat transfer in the boiler and denitration reaction in the denitration equipment, which limits the accuracy of dynamic characteristics prediction. Also, in the second conventional method, it is difficult to construct a knowledge base for fuzzy reasoning in consideration of the fluctuations of the operating conditions and the uncertain factors because of the complicated causal relationship, and it takes a lot of time. Not only does this require reliability of the knowledge base.

(2) According to the first conventional method, in order to obtain the optimum start schedule within the practical time by using the whole system dynamic characteristic model of the plant, a high processing capacity and a large memory capacity are provided. You need a computer system.

(3) If the operating conditions such as the initial state of the plant before starting and the temperature are different, it is difficult to utilize the starting schedule created in the past, and the optimum solution must be obtained each time. That is, since the control system does not have the adaptability to the change in the above operating conditions, it is necessary to fully operate the high performance computer each time it is started. Further, in the second conventional method, as described above, it is difficult to construct the knowledge base for fuzzy inference in consideration of the fluctuation of the operating conditions and the uncertain factors. Similarly, it has no ability to adapt to changes in operating conditions.

The present invention has been made in view of the above problems, and its purpose is to (a) determine the operation limiting factor without using a complex large-scale dynamic system model of a power plant. Be able to monitor and evaluate. In other words, use the minimum necessary dynamic characteristic model to monitor the driving restriction factors, (b)
It is possible to optimize the startup schedule in the process of repeating actual startup without obtaining the optimum startup schedule from the first startup. (C) Even if the operating conditions such as the plant initial state and temperature are different, the startup schedule created in the past can be It can be used adaptively, (d) Furthermore, it is possible to add an appropriate adaptability to changes in the operating conditions described above,
It is to provide an autonomous adaptive optimization control system for a thermal power plant that is capable of

[0008]

A thermal power plant autonomous adaptive optimization control system according to the present invention for achieving the above object is a boiler for generating steam, and is driven by the steam generated by the boiler. In the control system of the thermal power plant having the steam turbine
It is characterized by comprising the following means (i) to (vii).

(I) Basic schedule creating means for creating a start-up schedule from ignition of the boiler to completion of start-up of the plant based on a temperature state before starting of the thermal power plant.

(Ii) Margin value evaluating means for evaluating a margin value with respect to the operation restriction condition of a predetermined operation restriction factor in the starting process of the thermal power plant.

(Iii) Schedule modification amount calculation means for calculating the modification amount of the activation schedule using fuzzy inference based on the margin value obtained by the margin value evaluation means.

(Iv) When the past starting schedule and operating conditions of the thermal power plant are given as input values to the neural network, the schedule correction amount obtained by the schedule correction amount calculating means is obtained as an output value. An adaptive knowledge acquisition means for storing the input / output correspondence (this is called a learning sample).

(V) The input / output correspondence is learned by determining the internal connection state of the neural network in the adaptive knowledge acquisition means, and the average value of the time-series data and the average value thereof as an input signal. A learning management method that uses the deviation from the value.

(Vi) Created based on the first start-up schedule created based on the schedule modification amount obtained by the schedule modification amount calculation means and the schedule modification amount obtained by the adaptive knowledge acquisition means. An actual start-up schedule is created using the second start-up schedule, and the autonomous coefficient λ that defines the adoption ratio of the first and second start-up schedules when creating the actual start-up schedule is set to start. Autonomous management means for increasing the adoption rate of the second activation schedule by increasing the repetition rate.

(Vii) Schedule execution control means for controlling startup of the thermal power plant according to the startup schedule created by the autonomy management means.

A combined cycle power plant having a gas turbine, an exhaust heat recovery boiler for generating steam using exhaust gas of the gas turbine, and a steam turbine driven by the steam generated in the exhaust heat recovery boiler Even in the control system, the object of the present invention can be achieved by providing the above means (i) to (vii).

The autonomy management means outputs the activation schedule X ₁ determined by the modification amount obtained from the schedule modification amount calculation means as the first activation schedule, and when X ₁ is input to the neural network together with the operating conditions. The start-up schedule X ₂ obtained by adding the correction amount ΔX _OUT and X ₁ obtained from the section is used as a second start-up schedule, and the complement (1-λ) of the autonomous coefficient is added to the first start-up schedule X _1. Value obtained by multiplying (1-λ) X
₁ and the autonomous coefficient λ in the _second startup schedule X _2.
It is also possible to adopt the activation schedule X _R obtained by adding the value λ X ₂ obtained by multiplying by as the actual activation schedule and transfer it to the schedule execution control means.

Further, the autonomy managing means has an autonomic coefficient generating means for varying the autonomic coefficient λ, and λ = 0 at the time of initial activation, and is obtained by the number of activations i or the schedule correction amount calculating means. In accordance with the schedule modification amount, λ may be increased in the range of 0 ≦ λ ≦ 1 as the number of times of activation is increased.

The schedule modification amount calculation means stores a plurality of fuzzy rules for defining a schedule modification amount of the start schedule corresponding to the size of the margin value obtained by the margin value evaluation means from a knowledge base. The schedule modification amount may be calculated by incorporating a fuzzy rule.

The predetermined operation limiting factor may be at least one or more of stress in the steam header at the outlet of the superheater of the boiler, stress in the steam turbine, and NOx discharged from the boiler. In the cycle power plant, at least one of stress of the steam header of the superheater outlet of the exhaust heat recovery boiler, stress of the steam turbine, and NOx discharged from the exhaust heat recovery boiler may be used.

The start-up schedule in the combined cycle power plant comprises a start-up schedule of the gas turbine and a start-up schedule of the steam turbine, and the start-up schedule of the gas turbine is:
The startup parameter is configured by a predetermined parameter that defines at least one of a speed increase rate, a rated speed holding time, an initial load, an initial load holding time, a load rising rate, and a load holding time. The steam turbine start-up schedule is configured by predetermined parameters that define the start-up schedule, and the predetermined parameters include a high-pressure turbine bypass valve operating speed, an intermediate-pressure turbine bypass valve operating speed, a low-pressure turbine bypass valve operating speed, It may be at least one of a high pressure control valve operation speed, an intermediate pressure control valve operation speed, and a low pressure control valve operation speed.

The predetermined operation limiting factors are the superheater outlet steam header stress of the exhaust heat recovery boiler, the steam turbine stress, and NOx discharged from the exhaust heat recovery boiler, and the margin value evaluation means is For the superheater outlet steam header stress of the exhaust heat recovery boiler and the steam turbine stress, a minimum margin value for the operation limit value is obtained for each section obtained by dividing the starting process of the thermal power plant into a plurality of sections, and the exhaust For NOx discharged from the heat recovery boiler, an instantaneous value and a moving average value are obtained to obtain a minimum margin value for the operation limit value for each section obtained by dividing the starting process of the thermal power plant into a plurality of sections. May be

The schedule correction amount calculation means is a steam turbine main plan (STPS) based on the minimum margin value of the exhaust heat recovery boiler superheater outlet steam header stress and the minimum margin value of the steam turbine stress obtained from the margin value evaluation means. Inference for gas turbine wide area adjustment (GTGT), fuzzy inference for gas turbine main plan (GTPS) and fuzzy reasoning for steam turbine local adjustment (STLT) based on the minimum margin value of NOx discharged from the exhaust heat recovery boiler Reasoning is performed, the STPS targets a wider range of schedule parameters than the STLT, and the GT
The PS may make more precise correction as compared with the GTGT, and the GTGT may make more global correction as compared with the GTPS.

The schedule correction amount calculation means calculates the speedup rate by the priority value determination means from a plurality of correction amounts for the same schedule parameter obtained from the fuzzy inferences for STPS, STLT, GTPS, and GTGT. A low value may be selected for the parameters relating to the load change rate and the valve operating speed, and a high value may be selected for the parameters relating to the speed holding time and the load holding time.

The neural network used in the adaptive knowledge acquisition means has a multilayer structure composed of at least an input layer and an output layer by collecting a plurality of units having a signal conversion function, and has an output unit and a next layer of a front layer unit. The internal connection strength provided in the connection part with the input part of the unit can be set, the number of units corresponding to the startup schedule parameter and the operating condition is arranged in the input layer, and the startup schedule parameter is set in the output layer. A corresponding number of units may be arranged, and the signal input to the input layer may be preset by the learning management means, converted according to the internal connection strength, and output from the output layer.

The learning management means includes cumulative correction amount calculation means for cumulatively calculating a new correction amount obtained by the schedule correction amount calculation means each time the thermal power plant is repeatedly started, and the cumulative correction amount. Cumulative correction amount storage means for storing the cumulative correction amount obtained by the calculation means, schedule storage means for storing the start schedule applied to the actual start in the past, and operating conditions at the time of actual start in the past Of the past activation schedule and operating conditions from the operating condition storing means for storing the operation schedule storing means, the schedule storing means and the operating condition storing means, respectively, and input them to the input section of the neural network in the adaptive knowledge acquiring means. The output value obtained from the output unit when the cumulative correction amount is stored in the cumulative correction amount storage means. And learning means for determining the internal connection strength of the neural network to match the amount, may have a.

The operating conditions stored in the operating condition storage means are one of the thermal power plant shutdown period, plant equipment temperature, pressure, condenser cooling water temperature, atmospheric temperature, humidity, and plant operation restriction conditions. It may be at least one.

The operating conditions stored in the operating condition storage means may be separated into an average value which is statistically processed in advance for each individual operating condition and a deviation from the average value.

Even if the operation limiting factor is at least the difference in expansion between the rotor of the steam turbine and the casing, the stress generated in the drum of the exhaust heat recovery boiler, and the SO or CO discharged from the exhaust heat recovery boiler. Good.

The thermal power plant autonomous adaptive optimization control system having the above configuration may be used as a start control system for a gasification combined cycle power plant, an atmospheric pressure and a pressurized fluidized bed boiler power plant.

A series of means for creating a startup schedule in the thermal power plant autonomous adaptive optimization control system according to the present invention may be used as guidance presenting means for creating or modifying a startup schedule in the plant operation training simulator. The thermal power plant autonomous adaptive optimization control system according to the present invention may be provided with means for displaying the actual startup schedule.

[0032]

According to the present invention, based on the first activation schedule created based on the schedule modification amount obtained by the schedule modification amount calculation means, and the schedule modification amount obtained by the adaptive knowledge acquisition means. An actual startup schedule is created using the created second startup schedule, and an autonomous coefficient λ that defines the adoption ratio of the first and second startup schedules when creating the actual startup schedule is set. Since the autonomous management means is provided which increases the adoption ratio of the second startup schedule by increasing it as the startup is repeated, when the actual startup schedule is determined, the fuzzy inference is performed as the startup is repeated. To the conclusion from the neural network obtained by the adaptive knowledge acquisition means Lies will be intensified the degree, it is possible to create a suitable start-up schedule in response to changes in the operating conditions of the plant.

That is, the adaptive knowledge acquisition means has a neural network for storing the correspondence between the schedule and the operating conditions of the plant every time the plant is started and the correction amount obtained by the schedule correction amount determining means. Therefore, when the actual activation schedule is created, the first schedule created by the modification amount from the schedule modification amount determining means by the fuzzy inference and the first schedule created by the modification amount from the adaptive knowledge acquisition means. By increasing the ratio of the second schedule adopted (autonomous coefficient λ) as the start-up is repeated, the actual start-up schedule is the ratio of the second schedule created corresponding to the operating conditions of the plant. The higher the start-up frequency and the greater the number of times the plant is started, the better It is possible to create a rule.

The above-mentioned means (i) to (vi) of the present invention and the function of the knowledge base will be described below.

In the basic schedule creating means, paying attention to the stop period having a relatively high correlation with the temperature state of the plant,
From a plurality of schedules prepared in advance as table information for each stop period, a corresponding schedule is selected corresponding to the start time commanded by the central power feeding command center, and this is used as the basic schedule. This creates a schedule that, although not necessarily optimal, can safely start the plant. The operation restriction factor monitoring means monitors a process state in which direct measurement is possible by a signal obtained from a measuring instrument while the plant is started by the schedule execution control means, and measures a process state in which direct measurement is difficult. The behavior of operation limiting factors is estimated and monitored by inputting possible process states into the dynamic characteristic model.

The allowance value evaluation means evaluates the allowance value with respect to the limit value for the operation limit factor obtained by the operation limit factor monitoring means in the entire starting process.

The schedule correction amount calculation means calculates the correction amount of the schedule according to the size of the margin value obtained by the margin value evaluation means. At this time, it is calculated by fuzzy inference using a knowledge base described later.

The adaptive knowledge acquisition means acquires adaptive knowledge using a neural network.

Here, the adaptive knowledge is not only the schedule correction measure represented by the fuzzy rule, but also the knowledge having the adaptability capable of coping with changes in plant operating conditions. In order to acquire this knowledge, the neural network stores the correspondence relationship between the schedule and the plant operating conditions and the correction amount obtained by the schedule correction amount determining means each time the plant is started. Here, as the operating conditions, those separated into an average value and a deviation from the average value which are statistically processed in advance for each individual operating condition,
It is stored in association with the schedule correction amount.

In determining the actual activation schedule, the autonomy management means depends on the conclusion from the neural network having the ability to acquire the adaptive knowledge rather than the conclusion from the fuzzy inference as the activation is repeated. It works to strengthen the degree. That is, the adoption ratio (autonomous coefficient λ) of the schedule created by the modification amount from the schedule modification amount determining means by the fuzzy reasoning and the schedule created by the modification amount from the adaptive knowledge acquisition means is made variable, and the activation is repeated. By increasing the ratio of the latter in accordance with the above, the function of enhancing the adaptability and autonomy with respect to changes in the operating conditions is exerted.

The schedule execution control means controls the process states necessary for plant startup, such as turbine acceleration control, load increase control, steam pressure control, etc., in accordance with a startup schedule created by the autonomy management means described later.

The knowledge base is a set of a plurality of rules that define a schedule correction policy in a causal relationship with the margin value obtained by the margin value evaluation means, and each rule depends on the size of the margin value. It functions as a piece of knowledge that is a computer-processable expression of the expert's knowledge about the method of modifying the schedule.

Further, the learning management means causes the adaptive knowledge acquisition means to learn the correspondence. For this purpose, the correspondence between the schedule and plant operating conditions for each start of the plant and the correction amount obtained by the schedule correction amount determining means is sequentially presented as a learning sample to the neural network for learning. That is, the learning management means functions to determine the internal connection state of the neural network so that the latter is output when the former is input to the neural network.

[0044]

Embodiments of the present invention will be described below.

FIG. 1 shows the basic configuration of a thermal power plant autonomous adaptive optimization control system 1000 of the present invention. The present system operates upon receipt of a start time T _Ri (i indicates the next start, for example, i = 1 indicates the first start) which is a start command from the central power feeding command station 2000, and schedule execution control means in the system. 300 thermal power plant 30
00 is actually started.

This system 1000 includes a basic schedule creating means 100, a schedule modifying means 150, a schedule execution control means 300, and an operation restriction factor monitoring means 40.
0, margin value evaluation means 500, schedule correction amount calculation means 600, knowledge base 700, correction amount storage means 650,
It comprises adaptive knowledge acquisition means 800, learning management means 900, and autonomy management means 200.

In the basic schedule creating means 100, a corresponding schedule is selected from a plurality of schedules prepared in advance for each stop period as table information, corresponding to the start time T _Ri instructed by the central power feeding command station,
This is the basic schedule X _0i . Where X
Is a vector quantity consisting of a plurality of parameter values that define the schedule, and details will be described later. This creates a schedule that is not optimal, but safe and reliable for plant startup.

In the schedule correction means 150, the schedule correction amount calculation means 60 using fuzzy inference described later is used.
A new schedule X _i is obtained by using the correction amount ΔX _i-1 obtained from 0 and stored in the correction amount storage means 650 and the previous activation schedule X _Ri-1 stored in the learning management means 900 described later. create. Here, at the first startup (i =
In 1), ΔX ₀ = 0 and X ₀ = X ₀₁ .

The schedule execution control means 300 controls the process state associated with the start-up such as turbine speed, load, steam pressure in the thermal power plant 3000 according to the start-up schedule X _Ri created by the autonomous management means 200 described later. .

The operation restriction factor monitoring means 400 operates by inputting a measurable process state into the dynamic characteristic model for a process state that is difficult to directly measure while the plant is started by the schedule execution control means 300. Estimate and monitor the behavior of limiting factors. In addition, the process state where direct measurement is possible is monitored by the signal obtained from the measuring instrument. The operation restriction factors to be monitored here will be specifically described later.

The allowance value evaluation means 500 evaluates the allowance value for the operation restriction factor obtained by the operation restriction factor monitoring means 400 with respect to the restriction value in the entire starting process. The definition of this margin value will also be specifically described later.

In the schedule correction amount calculation means 600,
The correction amount ΔX _{i of the} schedule is calculated according to the size of the margin value obtained by the margin value evaluation means 500. At this time, it is calculated by fuzzy inference using the knowledge base 700 described below.

The knowledge base 700 is a set of a plurality of rules that define a schedule correction policy in a causal relationship with the margin value obtained by the margin value evaluation means 500, and each rule is dependent on the size of the margin value. It works as a piece of knowledge that is a computer-processable representation of the expert's knowledge about how to modify the schedule.

In the correction amount storage means 650, the correction amounts ΔX ₁ to
ΔX _i is stored in association with each past plant activation.

The adaptive knowledge acquisition means 800 acquires adaptive knowledge using a neural network. Here, the adaptive knowledge is not only the schedule correction measure represented by the fuzzy rule, but also the knowledge having the adaptability capable of coping with changes in plant operating conditions. In order to acquire this knowledge, the neural network has a schedule and plant operating conditions each time the plant is started, and a correction amount ΔX ₁ obtained by the schedule correction amount calculation means 600.
The correspondence with ~ ΔX _i is stored. Here, as the plant operation conditions, in the embodiment of the present invention, the stop period T _Si (which means the elapsed time between the previous stop time and the command start time T _Ri ) and the temperature T _Ai that most affect the start-up temperature state of the plant. , Humidity H _Ai and water temperature T _Wi are used. In addition, Z _i will be used hereinafter as a general term for these plant operating conditions. However, since the temperature T _Ai and the humidity H _Ai have large daily fluctuations, in order to improve the learning accuracy by the neural network, they are _{separated into} respective average values T _AAi , humidity H _AAi and deviations ΔT _Ai and ΔH _Ai from the average values. And handle. Note that the average value is a value estimated from past actual values of a plurality of days, but it will be properly defined later.

The learning management means 900 causes the adaptive knowledge acquisition means 800 to learn the correspondence. For that purpose, the correspondence relationship between the schedule X _Ri and the plant operating condition Z _{i for} each plant startup and the cumulative value of the correction amount ΔX _i obtained by the schedule correction amount calculation means 600 is sequentially presented as a learning sample, Let them learn. That is, when the learning management unit 900 inputs X _Rk (k = 1 to i−1) and Z _k (k = 1 to i−1) to the input layer of the neural network used as the adaptive knowledge acquisition unit 800. In addition, the connection state in the neural network is determined so that the output value ΔX _OUTk from the output layer _matches the cumulative correction amount ΔX _Qk (k = 1 to i−1) of ΔX _i . The cumulative correction amount ΔX _Qk will be specifically described later.

In determining the actual activation schedule, the autonomy management means 200 increases the degree of dependence on the conclusion ΔX _OUTi from the neural network rather than the conclusion ΔX _i from the fuzzy inference as the activation is repeated. Work. That is, the adoption ratio of the schedule created by the modification amount obtained from the schedule modification amount calculation means 600 based on the fuzzy inference and the schedule created by the modification amount obtained from the adaptive knowledge acquisition means 800 is represented by an autonomous coefficient λ. Is made variable, and the ratio of the latter is increased as the activation is repeated, thereby increasing the adaptability and autonomy to the change in the operating condition. The above-mentioned autonomous coefficient λ will be specifically described separately.

The outline of the present invention has been described above.
As a thermal power plant to which the present invention is applied, an embodiment will be described by taking a combined cycle power plant including a gas turbine, an exhaust heat recovery boiler, and a steam turbine as an example.

FIG. 2 shows a combined cycle power plant 300.
The relationship between the device configuration of 0 and the schedule execution control means 300 is shown. The schedule execution control means 300 further comprises a gas turbine control system 330 and a steam turbine control system 360. In the gas turbine equipment 3200,
The energy generated by combustion by injecting the fuel 22 into the combustor 23 and pressurizing the combustion air 24 by the compressor 20 is converted into mechanical energy by the gas turbine 26, and as a result, power generation connected to the common shaft 28 is performed. Machine 3
5 is converted into electric energy and a part of the electric power is used as the driving force of the compressor 20. At startup, the fuel control valve opening command 311 from the gas turbine control system 330
By operating the fuel control valve 21 to adjust the fuel flow rate, the gas turbine equipment 3200 and the steam turbine equipment 3300 connected to the common shaft 28 and the generator 35 are accelerated.

Exhaust gas 27 from the gas turbine 26
Is guided to the exhaust heat recovery boiler facility 3100, and the thermal energy of the exhaust gas 27 is recovered. At this time, the exhaust gas 27
As a result, the fluid in the various heat exchangers arranged in the flue 14 of the exhaust heat recovery boiler facility 3100 receives heat, evaporates, and is overheated. The exhaust heat recovery boiler equipment 3100 of this embodiment is composed of a steam system having three pressure levels, and the steam generated from each is a high pressure steam 341, an intermediate pressure steam 342, and a low pressure steam 343. In the steam turbine equipment 3300, the high-pressure turbine 31, the intermediate-pressure turbine 32, and the low-pressure turbine 33 are driven by the thermal energy of these steams, and play a role in power generation by the generator 35 connected to the common shaft 28.

At start-up, the exhaust heat recovery boiler equipment 3100
High pressure main steam 341, medium pressure main steam 342, generated from the above
By bypassing the low-pressure main steam 343 via the high-pressure bypass valve 334, the intermediate-pressure bypass valve 335, and the low-pressure bypass valve 336, the individual pressures are controlled to predetermined values, and the high-pressure control valve 331, the medium-pressure control valve 332, and the like. By operating the low pressure regulator valve 333, the high pressure turbine 31,
The output of the intermediate pressure turbine 32 and the low pressure turbine 33 is increased. Therefore, the flow rates of the high-pressure bypass steam 344, the medium-pressure bypass steam 345, and the low-pressure bypass steam 346 increase as the opening degrees of the respective bypass valves 334, 335, 336 increase, and the opening degrees of the regulator valves 331, 332, 333 increase. When is increased, the amount of steam flowing into the steam turbine increases, and the amount decreases. Opening degree command 32 to these bypass valves
4, 325, 326 and opening degree commands 321 and 3 to the regulator valve
22 and 323 are both steam turbine control system 36
Output from 0. Further, when the temperature deviation between the medium-pressure superheated steam 1 and the high-pressure turbine exhaust 2 is within a predetermined value during the startup of the plant, the medium-pressure stop valve 337 is opened. The operation signal 327 at this time is also instructed by the steam turbine control system 360. Condensed water 37 from the condenser 34 is supplied to the low-pressure water supply pump 16, the medium-pressure water supply pump 17, and the high-pressure water supply pump 18.
Thus, the flow rates of the low-pressure feed water 3, the medium-pressure feed water 4, and the high-pressure feed water 5 are controlled so that the water levels of the low-pressure drum 6, the medium-pressure drum 7, and the high-pressure drum 8 are maintained within predetermined values.

Here, the factors that limit the operation are the stress generated in the rotors of the high-pressure turbine 31 and the intermediate-pressure turbine 32, and the stress generated in the headers at the outlets of the high-pressure superheater 11 and the intermediate-pressure superheater 12. , NOx emissions from the flue outlet 15 of the exhaust heat recovery boiler facility 3100 to the atmosphere. Each of the stresses described above is a dynamic process with a large time delay, that is, heat transfer from the gas turbine exhaust gas 27 to the metal of the heat exchanger, heat transfer from the metal to the internal fluid, and heat transfer from the internal fluid to the metal of interest. Appears as a result. Further, the centrifugal force is added to the rotor, and the mechanical stress corresponding to the steam pressure is added to the header. Hereinafter, unless otherwise specified, these are collectively referred to as stress. The NOx emission amount from the boiler also largely depends on the NOx emission characteristic of the gas turbine itself and the temperature characteristic of the denitration device 13 installed in the flue. Further, the various operation limiting factors are the plant stop period T _Si , the temperature T _Ai , the humidity H _Ai , and the water temperature T _Wi which are the operating conditions Z _i.
Is greatly affected by. That is, the temperature and humidity affect the exhaust gas temperature of the gas turbine and the exhausted NOx, and as a result,
It affects the stress characteristics of boilers and steam turbines, and the amount of NOx discharged into the atmosphere after passing through the denitration device. Further, the water temperature affects the condensate performance of the steam turbine exhaust, and changes in the plant efficiency due to changes in plant efficiency have an effect. This affects both the stress and NOx characteristics as well as the effect of the stop time. Therefore, in order to accurately manage these operation limiting factors, the coordination and consistency of the above control operations are required. Therefore, as described above, in the thermal power plant autonomous adaptive optimization control system 1000 of the present invention, the operation restriction factor monitoring means 400 and the margin value evaluation means 500 for monitoring and evaluating these dynamic behaviors are provided.

FIG. 3 shows the relationship between the starting process and the starting schedule parameter of the combined cycle power plant to which the present invention is applied.

[0064] As shown in the figure, as the scheduling parameters of the gas turbine relationships, speed increasing ratio (DN), rated speed holding time (DT _NL), first load (L _I), first load retention time (DT _LI), The first load increase rate (DL ₁ ), the load holding time (DT _HL ), the second load increase rate (DL ₂ ), and the third load increase rate (DL ₃ ). The fuel control valve 2 which is the operating end has a start-up schedule defined by these as a control target.
The gas turbine is started by adjusting the opening degree of 1. Further, as the operating end relating to the steam turbine, as described above, the high pressure bypass valve (HPBV) 334, the intermediate pressure bypass valve (IPBV) 335, the low pressure bypass valve (LPBV) 3
36, high pressure control valve (HPCV) 331, medium pressure control valve (I
PCV) 332, low pressure control valve (LPCV) 333, intermediate pressure stop valve (ISHV) 337, and high pressure bypass valve operating speed (DA _HBV ), intermediate pressure bypass valve operating speed (DA _IBV ), low pressure bypass as schedule parameters. Valve operating speed (D
A _LBV ), low pressure bypass valve operation waiting time (DT _LBV ), high pressure control valve first operation speed (DA _HCV1 ), high pressure control valve second operation speed (DA _HCV2 ), low pressure control valve operation speed (DA _LCV ). Controlled. The control targets and operation timings other than these schedule parameters are as illustrated.
Here, T _MS as the steam condition is the high pressure main steam temperature, P _MS
Is the high pressure main steam pressure, P _CRP is the high pressure turbine exhaust pressure, P
_IS is the medium-pressure main steam pressure, P _LS is the low-pressure main steam pressure, and ΔT is the deviation between the medium-pressure superheater steam temperature and the high-pressure turbine exhaust temperature. Further, in the figure, the% display of each valve indicates the opening.

Next, the detailed configuration of the control system and each means will be described step by step with reference to FIG. 4 and the subsequent figures.

FIG. 4 shows a detailed configuration of a thermal power plant autonomous adaptive optimization control system for a combined cycle power plant which is an embodiment of the present invention.

The present system operates in response to the start time T _Ri (i indicates the next start, for example, i = 1 indicates the first start) which is the start command from the central power feeding command station 2000,
The combined cycle power plant 3000 is actually started by the schedule execution control means 300 in the system. The system 1000 includes a basic schedule creating means 100,
Schedule correction means 150, schedule execution control means 300, operation restriction factor monitoring means 400, margin value evaluation means 500, schedule correction amount calculation means 600, knowledge base 700, correction amount storage means 650, adaptive knowledge acquisition means 800, learning management means. 900 and an autonomy management means 200. Hereinafter, each of the above means will be described step by step.

FIG. 5 shows the basic schedule creating means 100.
Shows the basic schedule creation method in. here,
When the start time T _Ri is commanded from the central power feeding command station 2000, first, the stop period T _Si , which is the elapsed time from the previous plant stop time T _STPi ,

[0069]

## EQU1 ## T _Si = T _Ri −T _STPi (1) Next, a table corresponding to the suspension period T _Si is selected from a plurality of schedules prepared in advance for each suspension period as table information, and this is set as the basic schedule X _0i . Therefore, X _0i
Is a vector quantity consisting of multiple parameter values that define the schedule.

In the schedule correction means 150, the schedule correction amount calculation means 60 using fuzzy reasoning described later.
A new schedule X _i is obtained by using the correction amount ΔX _i-1 obtained from 0 and stored in the correction amount storage means 650 and the previous activation schedule X _Ri-1 stored in the learning management means 900 described later.

[0071]

[Formula 2] X _i = X _Ri-1 + ΔX _i-1 (2) Here, at the first startup (i = 1), Δ
X ₀ = 0 and X ₀ = X ₀₁ .

In the schedule execution control means 300, the combined cycle power plant 300 is _{operated in} accordance with the starting schedule X _Ri created by the autonomy management means 200 described later.
The control of the process state associated with startup such as turbine speed, load, steam pressure at 0 is carried out by the above-mentioned method.

In the operation restriction factor monitoring means 400, during the process in which the plant is started by the schedule execution control means 300, the steam turbine stress and the boiler stress, which are process conditions that are difficult to be directly measured, move in a measurable process condition. Estimate and monitor dynamic behavior by inputting into the characteristic model. Further, NOx emitted from the exhaust heat recovery boiler, which is a process state in which direct measurement is possible, is monitored by a signal obtained from a measuring instrument.

FIG. 6 shows a method of monitoring the steam turbine stress in the operation limiting factor monitoring means 400, which is generated on the rotor surface and the rotor bore of the high pressure turbine and the intermediate pressure turbine, which are points to be monitored as an operation limiting factor. It was constructed as a dynamic characteristic model for stress. That is, the steam condition (temperature, pressure) inside the turbine and the heat transfer coefficient on the rotor surface are estimated from the turbine inlet steam condition, speed and load, and the unsteady temperature distribution inside the rotor metal is calculated to determine the rotor surface and bore. In addition to calculating thermal stress, centrifugal stress was added to obtain total stress. The stress generated in the exhaust heat recovery boiler is also monitored by the same method. That is, the internal heat transfer coefficient of the high-pressure superheater outlet header and the medium-pressure superheater outlet header, which are the points to be monitored as an operation limiting factor, are estimated from the steam conditions (temperature, pressure) and the flow rate,
As in the case of the steam turbine, the unsteady temperature distribution inside the metal is calculated to calculate the thermal stress inside and outside the header metal, and the stress due to steam pressure is added to obtain the total stress. Next, the margin value evaluation means 500 for evaluating the margin value with respect to the limit value of the operation restriction factor obtained by the operation restriction factor monitoring means 400 will be described.

FIG. 7 shows a margin value evaluation method for the limiting value of the steam turbine stress. First, a time period (t ₁ ~
t ₆ ) is divided into a plurality of sections (in this example, the case of five divisions is shown), and the minimum stress margin value m (i) in the i-th section is obtained. Here, the stress margin value m is defined by the following equation, where the limit value is S _L and the value obtained by the operation limit factor monitoring means 400 is S.

[0076]

Equation 3] _{m = S L -S ... (3} ) In addition, the stress value is positive when the tensile stress, if negative means compressive stress. As described above, the stress points are actually the rotor surface and the bore of the high-pressure turbine and the rotor surface and the bore of the medium-pressure turbine, so there are four minimum stress margin values to be obtained for each section. These are described below.

M _HS (i): Minimum stress margin value of high pressure turbine rotor surface in section i m _HB (i): Minimum stress margin value of high pressure turbine rotor bore in section i m _IS (i): Medium pressure turbine rotor in section i Minimum surface stress margin m _IB (i): Medium pressure turbine rotor bore minimum stress margin in section i The exhaust heat boiler stress margin evaluation method is basically the same as the turbine stress margin evaluation method. . However, two points to be noted are the header outer surface at the high pressure superheater outlet and the medium pressure superheater outlet. This is because the stress generated on the outer surface of the header is larger than that on the inner surface.
Therefore, there are two minimum stress allowance values to be obtained for each section, and they are set as follows.

_{M HHD} (i): Minimum stress margin value of outer surface of high pressure superheater header in section i m _IHD (i): Minimum stress margin value of outer surface of medium pressure superheater header in section i FIGS. 8 (A) and 8 (B) , A margin value evaluation method for the limit value of the exhausted NOx characteristics is shown. Similarly to the case of the turbine stress, this method is also the time period (t _{1 to} t ₇ ) from the start of the gas turbine start to the elapse of a predetermined time after the completion of the plant start.
Is divided into a plurality of sections (in this example, the case of 6 divisions is shown),
Minimum NOx instantaneous value margin value m _PS (i) in the i-th section
And the minimum discharge NOx mean value margin value m _PA (i). Here, the exhausted NOx instantaneous value margin value m _PS and the exhausted NOx average value margin value m _PA are defined by the following equations, where the respective limit values are P _SL and P _AL and the measured values are P _S and P _A.

[0079]

(4) m _PS = P _SL −P _S (4)

[0080]

## _EQU00005 ## m _PA = P _AL −P _A (5) FIG. 9 shows the evaluation results obtained from the margin value evaluation means 500 m _HS , m _HB , m _IS , m _IB , m _HHD , and m _IHD. , M _PS , m
Upon receiving the _PA , an overall outline of the schedule correction amount calculation method in the schedule correction amount calculation means 600 will be shown. The schedule correction amount calculation means 600 calculates the schedule correction amount ΔX _i according to the size of the margin value obtained by the margin value evaluation means 500. The schedule correction amount calculating means 600 further includes steam turbine stress adjusting means 610,
Boiler stress adjusting means 620, exhaust NOx adjusting means 63
0, priority value determining means 640. The above-mentioned adjusting means are low-value selecting means 611, 621, 6 respectively.
31 and the low value selecting means 611 has margin values m _HS , m _HB ,
From among m _IS and m _IB , the low value selection means 621 is the margin value m.
_{From HHD} and m _IHD , the low value selecting means 631 selects the minimum margin values from the margin values m _PS and m _PA , and sets these as m _T , m _B and m _P , respectively. That is, m _T ,
m _B and m _P are defined by the following equations.

[0081]

M _T = Min (−m _HS , m _HB , −m _IS , m _IB ) ... (6)

[0082]

[ _Equation 7] m _B = Min (m _HHD , m _IHD ) (7)

[0083]

[Equation 8] m _P = Min (m _PS , m _PA ) ... (8) Next, fuzzy inference 612, using the above m _T , m _B , and m _P.
Schedule correction amounts ΔX _T (i), ΔX _B (i), and ΔX _P (i) are calculated by 613, 622, 623, 632, and 633. Here, the subscripts T, B, and P indicate that the correction amounts are determined according to the steam turbine stress, the boiler stress, and the margin value of exhausted NOx, respectively.

FIG. 10 shows the membership function used in the fuzzy rule at this time. The membership function shown in FIG. 10 (1) is for the steam turbine main plan (STPS) for adjusting the stress by correcting the steam system schedule from the turbine stress margin value m _T and the boiler stress margin value m _B. a correction factor k _S as a definition of the schedule correction coefficient k _G as for gas turbine wide adjustment for adjusting the stress by modifying the schedule for a gas turbine system (GTGT). Further, the membership function shown in FIG. 10 (2) is based on the emission NOx.
The correction coefficient k _G for the gas turbine main plan (GTPS) for adjusting the emission NOx by adjusting the schedule of the gas turbine system from the margin m _P and the emission NOx by adjusting the schedule of the steam system. The schedule correction coefficient k _S for the local adjustment (STLT) of the steam turbine for adjustment is specified. More specifically, FIG. 10 (1) (a) shows a membership function used in the conditional part of the fuzzy rule for stress evaluation, and shows four functions (NS, ZO, PS, PB).
(B) shows the membership function for determining the schedule correction coefficient used in the conclusion part, and the five functions (NB,
NS, ZO, PS, PB). Here, the meaning of each membership function is NB: Negative Big, N
S: Negative Small, ZO: Zero, PS: Positive Sma
ll, PB: Positive Big. FIG. 10 (2) (a) shows a membership function used in the conditional part of the fuzzy rule for evaluating emission NOx, and shows four functions (NS, ZO,
PS, PB), and (b) shows the membership function for determining the schedule modification coefficient used in the conclusion part, which is composed of five functions (NB, NS, ZO, PS, PB). Here, the meaning of each membership function is shown in FIG. 9 (1) above.
Is the same as in.

Next, a fuzzy rule using the above membership function will be described.

FIG. 11 shows a steam turbine main plan (STPS).
Shows the turbine stress adjustment rule used in. Here, 4
High pressure bypass valve operating speed (DA _HBV ) and medium pressure bypass valve operating speed (DA
_IBV ), high pressure control valve first operating speed (DA _HCV1 ), high pressure control valve second operating speed (DA _HCV2 ), turbines in the second, third, fourth and fifth sections among the five _target sections shown in FIG. A fuzzy rule for correction in relation to the stress margin value m _T is shown. That is, the margin value m in the second and third sections
A rule table defining the correction amounts of DA _HCV1 and DA _HCV2 in the relationship between _T (2) and m _T (3), and the margin values m _T (3) and m _T (4) in the third and fourth sections DA _{HCV1 in a} relationship
And the rule table that defines the amount of modification of DA _HCV2 , and
And a rule table defining correction amounts of DA _HBV , DA _IBV , DA _HCV1 and DA _HCV2 in the relation between the margin values m _T (4) and m _T (5) in the fifth section. However, the blank portion of the rule table means that the causal relationship between each stress margin value and the schedule parameter is small or almost nonexistent.

FIG. 12 shows the gas turbine wide area adjustment (GTG).
The turbine stress adjustment rule used in T) is shown. Here, as the six correction schedule parameters, the acceleration rate
(DN), initial load holding time (DT _LI ), first load rise rate (D
L ₁ ), load holding time (DT _HL ), second load increase rate (D
A fuzzy rule for correcting L ₂ ), the third load increase rate (DL ₃ ) in relation to the turbine stress margin value m _T in the five target sections shown in FIG. 7 is shown. That is, the first
When a rule table that defines DN, the correction amount of DT _LI and DL ₁ in relation allowance m _T (1) and m _T (2) in the second interval, margin value in the second and third sections m
_In the relationship between _T (2) and m _T (3), DT _LI , DL ₁ , DT _HL
And a rule table that defines the modification amount of DL ₂ , and
And a rule table defining the correction amounts of DL ₁ , DT _HL , DL ₂ and DL ₃ in the relation between the margin values m _T (3) and m _T (4) in the fourth and fifth intervals, It is composed of a rule table that defines the correction amounts of DT _HL , DL ₂ and DL ₃ in the relationship between the margin values m _T (4) and m _T (5). Also in this rule table, the blank portion means that the causal relationship between each stress margin value and the schedule parameter is small or almost nonexistent.

As described above, the fuzzy rules shown in FIGS. 11 and 12 are for STPS and GTG based on the boiler stress margin value.
It is also used for T.

FIG. 13 shows the gas turbine main plan (GTPS).
The emission NOx adjustment rule used in is shown below. Here, the speedup rate (DN) is set as eight correction schedule parameters,
Rated speed holding time (DT _NL), first load (L _I), first load retention time (DT _LI), the first load increase rate (DL _1), the load holding time
(DT _HL ), second load increase rate (DL ₂ ), third load increase rate (D
L ₃ ) is the emission NO in the six target sections shown in FIG.
A fuzzy rule for correction in relation to the x margin value m _P is shown.

That is, the margin value m in the first and second sections
A rule table defining the correction amounts of DN, DT _NL , L _I and DT _LI in the relationship between _P (1) and m _P (2), and the margin values m _P (2) and m in the second and third sections A rule table defining the correction amounts of DT _NL , L _I , DT _LI, and DL ₁ in the relationship of _P (3), and a margin value m in the third and fourth sections
_In the relationship between _P (3) and m _P (4), DT _LI , DL ₁ , D
A rule table defining correction amounts of T _HL , DL ₂ and DL ₃ , and margin values m _P (4) and m in the fourth and fifth sections
DL ₁ , DT _HL , DL ₂ and DL _{3 in} the relationship of _P (5)
It consists of a rule table that defines the modification amount of. Also in this rule table, the blank portion means that the causal relationship between each stress margin value and the schedule parameter is small or almost nonexistent.

FIG. 14 shows the steam turbine local adjustment (STL
The emission NOx adjustment rule used in T) is shown. here,
High pressure bypass valve operating speed (DA _HBV ) and medium pressure bypass valve operating speed (DA
_IBV ), the high pressure control valve first operating speed (DA _HCV1 ) in relation to the emission NOx margin value m _{P in the first, second} and third sections among the six sections of interest shown in FIG. Show the rules. That is, in the relationship between the margin values m _P (1) and m _P (2) in the first and second sections, DA _HBV , DA
A rule table defining the correction amounts of _IBV and DA _HCV1 , and margin values m _P (2) and m _P (3) in the second and third sections
In the above relation, it comprises a rule table defining the correction amounts of DA _HBV , DA _IBV and DA _HCV1 .

Compared with STLT, STPS described above targets a wider range of schedule parameters, GTPS performs more precise correction than GTGT, and GTGT has a broader perspective than GTPS. It works to make various corrections.

FIG. 15 shows a schedule modification amount calculation method by fuzzy inference. This figure shows calculation of the steam system schedule correction coefficient k _S as an example. In this example, for a certain schedule parameter, the membership function and membership value as conclusions from the four rules are (NS, 0.6), (ZO, 0.8),
The case where (PS, 0.4) and (PB, 0.2) are obtained is shown. The comprehensive evaluation value is defined by the weight W (i) of the trapezoidal portion determined by each membership value and the center of gravity k _SG of the correction coefficient k _S (i). That is, W (1) = 0.168, W (2) = 0.09
6, W (3) = 0.096, W (4) = 0.09, and k
_S (1) =-0.2, k _S (2) = 0, k _S (3) = 0.15, k _S
Since (4) = 0.35, k _SG is calculated as follows.

[0094]

[Equation 9]

Therefore, the modification amount k _SG for this activation schedule parameter is 0.0273. Using this result, the schedule correction amount ΔX _Ti is calculated according to the following equation.

[0096]

ΔX _T = k _SG (X _Max −X _Min ) (10) Here, X _Max and X _Min represent upper and lower limit values of X.

According to the method described above, the ST
Plural correction amounts are calculated for the same schedule parameter by fuzzy inference of PS, GTGT, STPS, STLT. Therefore, in order to surely keep all the operation limiting factors within the limit values, the priority value determining means 64 shown in FIG.
At 0, the final correction amount ΔX (i) is determined. That is, the rate of speed increase,
The low value on the safe side is selected by the low value selection means LVG in the figure for the parameters relating to the load change rate and the valve operation speed.
The high value selecting means HVG selects a high value on the safe side for the parameters relating to the speed holding time and the load holding time.

In the correction amount storage means 650, the correction amounts ΔX ₁ to
ΔX _i is stored in association with each past plant activation.

In the learning management means 900, the schedule X _Rk (k = 1 to i-) for each plant start is applied to the adaptive knowledge acquisition means 800 composed of a neural network.
1) and plant operation conditions Z _k (k = 1 to i-1)
And the corresponding relationship between the cumulative correction amount ΔX _Qk (k = 1 to i−1), which is the cumulative value of the schedule correction amount ΔX _i , are sequentially presented as learning samples and learned. However, in order to prevent the learning samples from increasing endlessly as the plant accumulates start-up records, old samples are excluded and only r samples (for example, r = 15) from the newest one are stored for learning. It is a matter of course to set it, and the following description will also adopt this measure unless otherwise noted.

FIG. 16 shows a learning sample presented to the neural network. If the start is completed up to the nth time, the cumulative correction amount ΔX as the teacher data
_Qi is a value defined by the following equation.

[0101]

[Equation 11]

That is, ΔX _Qi is obtained by cumulatively calculating a new correction amount obtained each time the activation is repeated by the cumulative correction amount calculating means 910 according to the above equation (11). The cumulative correction amount ΔX _Qi is stored in the cumulative correction amount storage means 920.

The learning management means 900 uses X _Rk (k = 1 to i−1) and Z _k (k = _k ) in the input layer of the neural network.
1 to i-1), the output value ΔX from the output layer
_OUTk (k = 1 to i−1) is the cumulative correction amount ΔX _Qk (k = 1 to i
It functions to determine the connection state in the neural network so as to match -1).

Next, the operating condition storage means 940 will be specifically described. In the embodiment of the present invention, as described above as the operating conditions, the average temperature T _AAi , the average humidity H _AAi , the temperature deviation ΔT _Ai , the humidity deviation ΔH _Ai , and the water temperature T as the meteorological data outside the plant shutdown period T _{Si. Use Wi} . here,
The reason for separating the average value and the deviation value about the temperature and humidity is that both of them have a large daily fluctuation range compared to the water temperature.
This is because high learning accuracy cannot be expected if the neural network is used as it is as a learning sample. Therefore, both parties decided to estimate the average value by statistical processing and determine the deviation between this estimated value and the actual value.

Hereinafter, the method of calculating the average value will be described with reference to the average temperature T.
The case of _AAi will be described as an example.

First, the estimation formula is defined by the following formula.

[0107]

T _AA (x) = Ax ² + Bx + C (12) where A, B, and C are m days in the past (for example, m = 30)
Temperature T _A (x) at the reference time (for example, 6 am) in
It is obtained by the least squares method using (x = 1 to m). In the above equation, the value T _AA (i) when x = i is the estimated average temperature T _AAi of the day.

In order to obtain the above A, B and C by the method of least squares, the error function J is first defined by the following equation.

[0109]

(Equation 13)

In order to minimize the error function J, A, B, and C satisfying the following equation should be obtained.

[0111]

∂J / ∂A = 0, ∂J / ∂A = 0, ∂J / ∂A = 0 (14) Therefore, A, B, and C are solutions of the following equation.

[0112]

[Expression 15] | X ₄ X ₃ X ₂ || A | | Y ₃ | | X ₃ X ₂ X ₁ || B | = | Y ₂ | ... (15) | X ₂ X ₁ m || C | | Y ₁ | where X ₁ = Σx, X ₂ = Σx ² , X ₃ = Σx ³ , X ₄ = Σ
^{_{x 4, Y 1 = ΣT A}} (x), Y 2 = ΣxT A (x), Y 3 = Σx 2 T
_{It is A} (x).

The temperature deviation ΔT _Ai is defined by the following equation.

[0114]

ΔT _Ai = T _Ai −T _AAi (16) The average humidity and H _AAi and the humidity deviation ΔH _Ai are also defined in the same manner as the case of the above temperature.

The operating condition as a learning sample described above is a method of preparing r operating conditions obtained daily from the newest one in order, but it is very difficult to let the neural network learn many samples. It takes a lot of learning time, and there is a limit naturally. Moreover, if the number of samples is reduced, it is not possible to uniformly learn a wide range of operating conditions. In order to solve this problem, the present invention considers the distribution of temperature and humidity, which are the meteorological conditions, and stores and manages learning samples that are not biased.

FIG. 17 shows a specific storage management method. In this method, when it is assumed that the temperature deviation ΔT _A and the humidity deviation ΔH _A show a normal distribution, their median values ΔT A
_AA and ΔH _AA are found, and the plane is divided into five regions on a plane composed of ΔT _A and ΔH _A with the origin as the origin, and the number of samples contained in each region is made equal.
Here, the fifth region is a region defined by multiplying the respective standard deviations S _T and S _H obtained from the distributions of ΔT _A and ΔH _A , and the magnification at that time is the number of populations N _{When the} number is _{0, the} value includes N _0/5 . It is sufficient to use the data for several months as the population necessary for determining the scaling factor.

It is determined which of the five areas defined by the above method the meteorological data obtained every day belongs to, and management is performed so that an equal number of samples are stored in each area. For example, assuming that the total number of samples to be stored is 15, 5 will be stored for each area. Therefore, the oldest sample may be deleted from the newly added area. Needless to say, along with this addition and deletion, other operating conditions are also added and deleted as learning samples. In this way, the samples can always be made to be evenly distributed over the entire area, and old samples are excluded from the samples in order for each area. Therefore, the neural network can adapt efficiently and accurately with a short learning time. Acquire knowledge.

The median values ΔT _AA and Δ obtained by the above method
Since H _AA is essentially close to zero, the mean values T _AAi and H _AAi described above are used as the origins, and the deviation from this is taken into _account to add and delete learning samples for the same five regions as described above. Of course, the same effect can be expected even if it is used.

Next, the neural network and its learning method will be described.

This neural network has an input layer,
It is composed of a three-layer structure consisting of an intermediate layer and an output layer. Regarding this basic structure, the MIT Press and Neurocomputing Foundations of
Research, 1988, pages 318-362 (The M
IT Press, Neurocomputing Foundations of research,
1988, pp318-362). Further, in this paper, when a certain input signal pattern is given to the input layer, the output layer pattern is matched with the desired signal pattern, that is, the teacher data. A learning algorithm (referred to as backpropagation) that modifies the connection strength of the input section to each unit is shown. Also in the embodiment of the present invention,
The structure of the neural network and the learning algorithm itself use those shown in the above paper.

FIG. 18 shows the relationship between the structure of the neural network and the input / output variables. Input variables are the activation schedule parameter X _IN and the operating condition Z. As X _IN , the learning sample X _Rk read from the schedule storage means 950 that stores the past activation schedule at the time of learning
(k = 1 to i−1), the schedule parameter X _i is input when the activation schedule is created. As Z, the learning sample Z _k (k = 1) read from the operation condition storage means 940 that stores past operation conditions at the time of learning.
~ I-1), Z _i is input when the activation schedule is created. When X _i and Z _i are input to the neural network, ΔX _OUTi determined by the internal connection strength determined by learning is output as an output variable. At the time of learning, when the above X _Rk and Z _k are input, ΔX _OUTk (k
= 1 to i−1) is output, ΔX _OUTk (k = 1 to i−
Learning is performed by modifying the internal connection strength of the neural network so that 1) matches ΔX _Qk (k = 1 to i−1).

Next, the autonomy management means 200 will be specifically described. When determining the actual activation schedule X _Ri , the autonomy management means 200 functions to strengthen the degree of dependence on the conclusion ΔX _OUTi from the neural network rather than the conclusion ΔX _i from the fuzzy inference as the activation is repeated. do. That is, the schedule created by the correction amount ΔX _i obtained from the schedule correction amount calculation means 600 by the fuzzy inference and the adaptive knowledge acquisition means 8
The adaptability and autonomy to the change in the operating conditions are made variable by changing the autonomous factor λ, which is the adoption ratio of the schedule created by the correction amount ΔX _OUTi obtained from 00, and expanding the latter ratio as the activation is repeated. Function to increase This autonomous coefficient λ is the autonomous coefficient generating means 210.
Is determined. Therefore, the activation schedule X _Ri defined by the following equation is adopted for the actual activation.

[0123]

X _Ri = λ (X _i + ΔX _OUTi ) + (1−λ) X _i (17) Here, if the autonomous coefficient λ is 0 ≦ λ ≦ 1, then X _Ri when λ = 0. = X _i and λ = 1,

[0124]

(18) X _Ri = X _i + ΔX _OUTi (18)

That is, the former is a starting schedule in which only the conclusions from the fuzzy inference are adopted and the output from the neural network is not used. The latter is a startup schedule in which the basic schedule is corrected by the output from the neural network, and as described above, the adaptability to changes in operating conditions is fully utilized. Therefore, as the value of λ increases, the control system becomes more autonomous. However, if the autonomic coefficient λ is increased before the neural network gains sufficient start-up experience (before providing sufficient adaptability by learning), the optimumness and high reliability of the start-up schedule created cannot be expected. Therefore, how to select this autonomous coefficient λ is important. The following method is applied to the autonomy management means 200 in the present embodiment.

FIG. 19 shows an autonomous management system for determining the autonomous coefficient λ in the autonomous management means 200. In this embodiment, the three methods can be applied depending on how to determine the autonomous coefficient λ as shown in the figure. The first method is the leap-growth type autonomy management method shown in (1) of this figure, in which λ = 0 is set until the number of activations i reaches a predetermined value i _A , and the neural network continuously learns and adapts. Is set higher and λ = 1 with i ≧ i _A , so that the autonomy is increased at once. The second method is a linear continuous growth Autonomous management scheme shown in the figure (2), i <the i _B lambda =
0, the neural network enhances the adaptive ability to some extent by learning, and in i _B ≤i≤i _C , the autonomous ability is gradually increased while increasing the adaptive ability, and i ≥i
_This is a method of setting λ = 1 in _C. The third method is the nonlinear continuous growth type autonomy management method shown in (3) of this figure.
λ is a schedule correction amount calculation means 6 by fuzzy inference
This is a function of the correction amount ΔX _i obtained from 00. This method is a method in which λ is increased in accordance with a decrease in the correction amount ΔX _i . The function that defines λ is specifically the following expression.

[0127]

[Formula 19] λ = 1−‖ΔX _i ‖ / ‖ΔX ₁ ‖ (19) Alternatively, the same function can be achieved by the following equation.

[0128]

[Formula 20] λ = 1-M _AX (| ΔX _i | / (X _MAX −X _MIN )) / M _AX (| ΔX ₁ | / (X _MAX −X _MIN )) (20) FIG. 7 shows the effect of improving the optimal value convergence by the autonomous adaptive optimization control system of the invention. Fig. 1 (1) shows the case without the autonomous control, that is, the case where the activation schedule is optimized only by the fuzzy inference, and (2) shows the case with the autonomous adaptive control, that is, the combination of the fuzzy inference and the neural network. This is a case where the system has adaptability and the startup schedule is optimized. All of them show how the initial schedule converges toward the optimum value as the plant is repeatedly started. However, in the case of (1), since there is no adaptability to changes in operating conditions as described above, there is a non-convergence region, and convergence beyond the convergence limit cannot be expected.
On the other hand, in the case of (2), since the neural network has the adaptive ability to acquire the schedule correction method including the causality of the operating conditions and the activation schedule, the neural network can converge to an optimum value without limit. This is evaluated from another angle as follows.

FIG. 21 shows the effect of improving the starting characteristic of the plant by the autonomous adaptive optimization control system of the present invention, which is shown in comparison with the case where the autonomous adaptive optimization control is not used. As described above, the non-convergence region exists when the autonomous adaptive optimization control is not used, and thus the created activation schedule has a range. Therefore, the stress and the exhausted NOx, which are operation limiting factors, also have a range. Therefore,
In order to satisfy all the limiting conditions, the schedule to be created needs to have a sufficient margin in advance, and the start-up time becomes longer accordingly. On the other hand, in the case of the autonomous adaptive optimization control, it is possible to flexibly adapt to changes in operating conditions as described above, so that an optimal schedule can always be created. Therefore, it is possible to start in the shortest necessary time by the start schedule that can satisfy the operation restriction condition with high accuracy.

Although the embodiments of the present invention described above have been specifically described for the combined cycle power generation plant, the present invention is applicable to other plants such as a normal power generation plant including a boiler, a steam turbine, and a generator. Of course, it can be applied to a coal gasification power generation plant and an atmospheric pressure or pressurized fluidized bed boiler power generation plant. It is also clear that the fuel used is not limited to coal, petroleum, LNG and the like.

Further, in the embodiment of the present invention, the stress of the steam turbine and the exhaust heat recovery boiler header and the exhausted NOx are treated as the operation limiting factors, but depending on the characteristics of the applied plant,
It is also possible to adopt a method that takes into consideration other factors, for example, the difference in expansion between the rotor of the steam turbine and the casing, the stress of other parts such as the drum of the exhaust heat recovery boiler, and the exhaust SOx and CO. Further, it is clear that even if the stress is not necessarily estimated as a dynamic characteristic model, indirect limit value management such as the rate of change or the range of change of the steam temperature or the metal temperature can be performed without changing the essence of the present invention. Is.

Further, in the embodiment of the present invention, the plant stop period, air temperature, humidity, and water temperature are adopted as the operating conditions, but in addition to these, the plant starting characteristics such as atmospheric pressure, wind speed, solar radiation, equipment temperature, and pressure are used. It is possible to add the various state quantities to the essence of the present invention without changing the basic principle. Further, when the plant is operated while changing the operation restriction condition at each startup, the operation restriction condition can be handled as a part of the operation condition without changing the basic principle of the present invention.

Further, in the embodiment of the present invention, the average temperature ΔT _AAi and the average humidity ΔH _AAi are estimated by the least square approximation as the operating condition which is the input information to the neural network, and these are used. It is obvious that the function used does not necessarily have to be a quadratic function as in the embodiment, and other nonlinear functions or linear functions can be used without changing the essence of the present invention.
Further, as the method of estimating the average value, it is possible to use a moving average value, although the accuracy is somewhat lowered, but this does not change the essence of the present invention. Further, a method of using the annual average value at the estimation target day and the target time without sequentially estimating the average value can be implemented without changing the essence of the present invention.

Further, in the embodiment of the present invention, the starting schedule parameters to be corrected are eight for the gas turbine-related and four for the steam turbine-related, but they are not necessarily limited to these, and are shown in FIG. It is obvious that the present invention can be implemented without changing the basic principle as long as it is a parameter that defines the plant starting pattern such as the other parameters shown, the opening / closing timing condition of the regulator valve, the set value for steam pressure control, and the like. In the embodiment of the present invention, the start start time commanded by the central power feeding command station is exactly followed,
In addition, it is possible to create a start-up schedule of the shortest time, but instead of the start-up completion time instructed by the central power supply command center, the reference time is shifted even if it is gas turbine ignition time, load combination time, target load arrival time, etc. It is obvious that the present invention can be implemented without changing the principle of the present invention.

[0135]

The first effect of the present invention is that, in a start control system for a thermal power plant, a plurality of operation restrictions such as equipment life standards and equipment protection standards such as NOx emissions, which are not possible with the conventional methods, and environmental regulation values. It is to enable automatic creation and execution of a startup schedule that minimizes the startup time while satisfying the conditions at the same time. As a result, the burden on the operator is significantly reduced, and the energy loss accompanying the shortened start-up time can be reduced, so that the plant operation cost can be significantly reduced.

The second effect of the present invention is that in a start control system for a thermal power plant, a central power supply command station is provided while simultaneously satisfying a plurality of operation restriction conditions such as equipment protection standards such as equipment life and NOx emissions and environmental regulation values. It is possible to complete the startup at the time specified by. This enables stable and accurate power supply to the power system, which requires frequent start and stop of the plant due to fluctuations in power demand.

The third effect of the present invention is to recreate the start schedule even when the operation limit value is changed or the start command time is changed from the central power supply command station in the start control system of the thermal power plant. And this is possible. This enables flexible and safe plant operation and power system operation.

The fourth effect of the present invention is that, in the startup control system for a thermal power plant, since it has the ability to adapt to changes in operating conditions, a large-scale complex dynamic characteristic model for predicting the startup characteristics of the plant is constructed. , It does not have to be built into the startup control system. As a result, the system design and construction time can be significantly reduced, the computer load can be reduced, and the startup schedule can be created faster.

The fifth effect of the present invention is to realize a start schedule optimization function having adaptability to fluctuations in operating conditions in a start control system for a thermal power plant, by utilizing the features of fuzzy inference and neural networks. It is possible to simplify the knowledge base used in fuzzy inference. This makes it possible to significantly reduce the time required to build the knowledge base.

A sixth effect of the present invention is that, in the starting control system of a thermal power plant, when the temperature, humidity, etc. are input as input information of the neural network, they are separated into an average value and a deviation, and a change region thereof. It is possible to increase the scale of the neural network and shorten the learning time by adding and deleting the learning samples that are not even in order to prepare them. As a result, it becomes possible to add adaptability to operating conditions that change daily with a small number of learning samples.

[Brief description of drawings]

FIG. 1 is a basic configuration diagram of a thermal power plant autonomous adaptive optimization control system of the present invention.

FIG. 2 is a diagram showing the relationship between the equipment configuration of the combined cycle power plant to which the present invention is applied and the startup schedule execution control.

FIG. 3 is a diagram showing a relationship between a startup process and a startup schedule parameter of a combined cycle power plant to which the present invention is applied.

FIG. 4 is a diagram showing a detailed configuration of a thermal power plant autonomous adaptive optimization control system for a combined cycle power plant that is an embodiment of the present invention.

FIG. 5 is a diagram showing a basic schedule creation method in the embodiment of the present invention.

FIG. 6 is a diagram showing a method of monitoring steam turbine stress as an operation limiting factor of the combined cycle power plant according to the embodiment of the present invention.

FIG. 7 is a diagram showing a steam turbine stress margin value evaluation method as an operation limiting factor of the combined cycle power plant according to the embodiment of the present invention.

8 (A) and 8 (B) are emission N as an operation limiting factor of the combined cycle power plant according to the embodiment of the present invention.
The figure which shows the margin evaluation method of Ox.

FIG. 9 is a diagram showing an overall outline of a schedule modification amount calculation method in consideration of all operation limiting factors in the embodiment of the present invention.

10 (1) and 10 (2) are views showing membership functions used in the fuzzy rule in the embodiment of the present invention.

FIG. 11 is a diagram showing a steam turbine stress adjustment rule used for a steam turbine main plan (STPS) by the schedule correction amount calculation means in the embodiment of the present invention.

FIG. 12 is a diagram showing a steam turbine stress adjustment rule used for gas turbine wide area adjustment (GTGT) in the schedule correction amount calculation means in the embodiment of the present invention.

FIG. 13 is a diagram showing an emission NOx adjustment rule used for a gas turbine main plan (GTPS) by the schedule correction amount calculation means in the embodiment of the present invention.

FIG. 14 is a diagram showing an emission NOx adjustment rule used for steam turbine local adjustment (STLT) in the schedule correction amount calculation means in the embodiment of the present invention.

FIG. 15 is a diagram showing a schedule modification amount calculation method by fuzzy inference according to an embodiment of the present invention.

FIG. 16 is a diagram showing a learning sample for a neural network according to the embodiment of the present invention.

FIG. 17 is a diagram showing a method of storing meteorological conditions as a learning sample in a neural network according to the embodiment of the present invention.

FIG. 18 is a diagram showing a structure of a neural network and definitions of input / output variables in the embodiment of the present invention.

19 (1), (2) and (3) are diagrams showing an autonomy management system in the embodiment of the present invention.

20 (1) and 20 (2) are diagrams showing the effect of improving the optimal value convergence by the autonomous adaptive control system of the present invention.

FIG. 21 is a diagram showing the effect of improving the starting characteristics of the combined cycle power plant by the autonomous adaptive optimization control system of the present invention.

[Explanation of symbols]

100 ... Basic schedule creation means, 150 ... Schedule correction means, 200 ... Autonomy management means, 300 ... Schedule execution control means, 400 ... Operation restriction factor monitoring means, 500 ... Margin value evaluation means, 600 ... Schedule correction amount calculation means, 650 ... Correction amount storage means, 700 ... Knowledge base, 800 ... Adaptive knowledge acquisition means, 900 ... Learning management means, 1000 ... Autonomous adaptive optimization control system for thermal power plant, 2000 ... Central power supply command station.

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location F02C 9/00 F02C 9/00 F22B 35/14 7526-3L F22B 35/14 Z G06F 15/18 550 G06F 15/18 550Z (72) Inventor Takao Akiyama 3-1-1 1-1 Saiwaicho, Hitachi, Ibaraki Stock company Hitachi Ltd. Hitachi factory

Claims (14)

[Claims] 1. A control system for a thermal power plant having a boiler for generating steam and a steam turbine driven by the steam generated by the boiler, from ignition of the boiler to completion of start-up of the plant. A starting schedule of a basic schedule creating means for creating the starting schedule of the thermal power plant based on the temperature state before the start of the thermal power plant; A value evaluation means, a knowledge base for storing a plurality of fuzzy rules defining a schedule modification amount of the startup schedule corresponding to the size of the allowance value obtained by the allowance value evaluation means, and a knowledge base stored in the knowledge base. A schedule for calculating the schedule modification amount by fuzzy inference using a fuzzy rule. When the Joule correction amount calculation means and the past start schedule and operating conditions of the thermal power plant are given as input values to the neural network, the schedule correction amount obtained by the schedule correction amount calculation means is obtained as an output value. like,
The adaptive knowledge acquisition means for storing the input / output correspondence, the learning management means for learning the input / output correspondence by determining the internal connection state of the neural network in the adaptive knowledge acquisition means, and the schedule correction amount calculation means An actual startup schedule using a first startup schedule created based on the calculated schedule modification amount and a second startup schedule created based on the schedule modification amount obtained by the adaptive knowledge acquisition means. And when creating the actual start-up schedule,
And increasing the autonomous coefficient λ that defines the adoption ratio of the second startup schedule as the startup is repeated.
Of the start schedule of the thermal power plant according to the start schedule created by the autonomy managing means, and the schedule execution control means for controlling the start of the thermal power plant. An autonomous adaptive optimization control system for a thermal power plant, which uses an average value of time series data and a deviation from the average value as a signal. 2. A thermal power plant having a gas turbine, an exhaust heat recovery boiler for generating steam using the exhaust gas of the gas turbine, and a steam turbine driven by the steam generated in the exhaust heat recovery boiler. In the control system, the starting schedule from ignition of the gas turbine to completion of starting the plant, a basic schedule creating means for creating based on a temperature state before starting the thermal power plant, and starting the thermal power plant A margin value evaluation means for evaluating a margin value with respect to an operation restriction condition of a predetermined operation restriction factor in the process, and a schedule correction amount of the startup schedule corresponding to the size of the margin value obtained by the margin value evaluation means are defined. A knowledge base that stores a plurality of fuzzy rules, and fuzzy rules stored in the knowledge base When a schedule correction amount calculation means for calculating the schedule correction amount by fuzzy inference using, and a past start schedule and operating conditions of the thermal power plant are given as input values to a neural network, the schedule correction is performed as an output value. The adaptive knowledge acquisition means for storing the input / output correspondence so that the schedule modification amount obtained by the amount calculation means is obtained, and the input / output correspondence is determined by determining the internal connection state of the neural network in the adaptive knowledge acquisition means. Based on the learning management means for learning the relationship, the first start schedule created based on the schedule modification amount obtained by the schedule modification amount calculation means, and the schedule modification amount obtained by the adaptive knowledge acquisition means. With the created second activation schedule Together create the actual startup schedule had, the first in creating a power-on schedule of said actual
And increasing the autonomous coefficient λ that defines the adoption ratio of the second startup schedule as the startup is repeated.
Of the start schedule of the thermal power plant according to the start schedule created by the autonomy managing means, and the schedule execution control means for controlling the start of the thermal power plant. An autonomous adaptive optimization control system for a thermal power plant, which uses an average value of time series data and a deviation from the average value as a signal. 3. The thermal power generation according to claim 1 or 2, wherein the average value in the learning management means is obtained by a linear or non-linear function with respect to time approximated by least square approximation using time series data. Plant autonomous adaptive optimization control system. 4. The autonomous adaptive optimization control system for a thermal power plant according to claim 1 or 2, wherein the average value in the learning management means is a moving average value using time series data. 5. The number of learning samples included in a plurality of sample areas defined in advance according to the magnitude of the deviation from the average value of the input signal of the learning management means according to claim 1 or 2. , It is determined which of the plurality of sample areas the newly obtained data corresponds to, the new learning sample is added to the corresponding area, and the oldest sample is selected from the existing learning samples included in the corresponding area. A thermal power plant autonomous adaptive optimization control system characterized by exclusion. 6. The plurality of sample regions according to claim 5, wherein the distribution number of past time series data is checked in advance and the distribution number is defined to be uniform in each region. Thermal power plant autonomous adaptive optimization control system. 7. An autonomous adaptive optimization control system for a thermal power plant according to claim 1 or 2, wherein temperature or humidity is used as the time-series data in the learning management means. 8. The autonomous adaptive optimization control system for a thermal power plant according to claim 7, wherein the humidity is absolute humidity. 9. The autonomy managing means according to claim 1 or 2, wherein the starting schedule X ₁ determined by the correction amount obtained from the schedule correction amount calculating means is a first starting schedule, and X ₁ is the neural network. The activation schedule X ₂ obtained by adding the correction amount ΔX _OUT and X ₁ obtained from the output section when inputting together with the operating conditions into the second activation schedule is the first activation schedule.
Complement of the autonomous coefficient (1-λ) in the startup schedule X ₁ of
The over-obtained value _(1-λ) X 1 and the second start scheduled X ₂ have the obtained multiplied by the autonomous coefficient lambda of the value .LAMBDA.x ₂
The thermal power plant autonomous adaptive optimization control system is characterized in that the startup schedule X _R obtained by adding is adopted as an actual startup schedule and is transferred to the schedule execution control means. 10. The neural network used in the adaptive knowledge acquisition means according to claim 1, wherein a plurality of units having a signal conversion function are collected to form a multilayer structure including at least an input layer and an output layer, It is possible to set the internal connection strength provided at the connection between the output section of the front layer unit and the input section of the next layer unit, and the input layer is provided with a number of units corresponding to the startup schedule parameters and operating conditions, A number of units corresponding to the startup schedule parameter are arranged in the output layer, and the signal input to the input layer is preset by the learning management means, converted according to the internal connection strength, and output from the output layer. Autonomous adaptive optimization control system for thermal power plants. 11. The cumulative correction according to claim 1, wherein the learning management unit cumulatively calculates a new correction amount obtained by the schedule correction amount calculation unit every time the thermal power plant is repeatedly started. An amount calculation means, a cumulative correction amount storage means for storing the cumulative correction amount obtained by the cumulative correction amount calculation means, and a schedule storage means for storing the activation schedule applied to the actual activation in the past, The past operating schedule and operating conditions are respectively taken out from the operating condition storing means for storing the operating conditions at the time of actual start-up in the past, the schedule storing means and the operating condition storing means, and the adaptive knowledge acquiring means obtains them. The output value obtained from the output unit when input to the input unit of the neural network in Thermal power plant autonomous adaptive optimization control system characterized by having a learning means for determining the internal connection strength of the neural network to match the cumulative correction amount stored in Seiryo storage means. 12. The startup schedule according to claim 2, wherein the startup schedule comprises a startup schedule of the gas turbine and a startup schedule of the steam turbine, and the startup schedule of the gas turbine is a predetermined parameter defining the startup schedule. And the predetermined parameter is at least one of a speed-up rate, a rated speed holding time, an initial load, an initial load holding time, a load rising rate, and a load holding time, and the steam turbine starting schedule is It is configured by a predetermined parameter that defines the startup schedule, and the predetermined parameter includes a high-pressure turbine bypass valve operating speed, a medium-pressure turbine bypass valve operating speed, a low-pressure turbine bypass valve operating speed, a high-pressure control valve operating speed, and a medium-pressure control At least one of valve operating speed and low pressure control valve operating speed Thermal power plant autonomous adaptive optimization control system according to claim Rukoto. 13. The operating condition stored in the operating condition storage means according to claim 11, is a shutdown period of the thermal power plant, equipment temperature and pressure of the plant, condenser cooling water temperature, atmospheric temperature and humidity, A thermal power plant autonomous adaptive optimization control system, which is at least one of plant operation restriction conditions. 14. The autonomy managing means according to claim 1 or 2, wherein the autonomy managing means has an autonomic coefficient generating means for varying the autonomic coefficient λ, and λ = 0 at the first activation, and the number of activations i or In accordance with the schedule modification amount obtained by the schedule modification amount calculation means, λ is increased in the range of 0 ≦ λ ≦ 1 as the number of startups is increased, and the thermal power plant autonomous adaptive optimization control is characterized. system.