JPH0454808B2 - - Google Patents

Description

[Detailed description of the invention]

[Industrial Application Field] The present invention relates to a plant startup system, and particularly to a startup method suitable for observing operational constraints during the plant startup process and completing startup in a short time. [Prior Art] The conventional method for starting a thermal power plant is based on the initial amount of fuel input to the boiler, the time function of main steam temperature and pressure rise, the turbine In this method, the time function of speed increase and load increase is determined as a startup schedule, and this startup schedule is executed by a control system provided in each system of the plant. The most typical method is the Thermal Stress Influence Starting, Loading of Boilers and Turbines (Electrical World, Vo. 165, No. 6).
of Boilers, Turbines). This method uniquely determines the startup schedule based on the initial state of a limited portion of the plant. That is, the steam turbine speed increase rate, initial load amount, speed maintenance, steam turbine warm-up time by load maintenance, and load change rate are determined according to the initial values of boiler steam pressure, boiler outlet steam temperature, and steam turbine casing temperature. This is the way to do it. Since this method absorbs variations in the temperature rise characteristics of boiler-generated steam as a margin for the startup schedule, the created startup schedule tends to be longer than necessary. Further, as other conventional methods, USP3446224 and USP4228359 are known. These methods attempt to quickly start up a steam turbine while monitoring the thermal stress generated in the steam turbine online in real time, but, like the conventional methods described above, there is no mention of a method for starting the boiler. Japanese Patent Laid-Open No. 157402/1984 is known as a conventional method aimed at shortening the startup time of a boiler. This method aims to rapidly raise the temperature of the steam generated by the boiler while monitoring the thermal stress generated in the boiler online in real time. However, this method makes no mention of starting the turbine. [Problems to be Solved by the Invention] All conventional methods are rapid startup methods that focus on only one side of the boiler or steam turbine. However, even if such individual methods are combined, there is no guarantee that the start-up time of the entire plant will be the shortest. This is because the boiler and the steam turbine have extremely strong mutual interference, and individual optimization does not necessarily result in overall optimization. The present invention is characterized in that, while the conventional method described above targets either the boiler or the steam turbine, the startup time of the entire plant is optimized and minimized by comprehensively considering the startup characteristics of both. be. An object of the present invention is to provide an optimal plant startup system that can optimize and shorten the startup time of the entire plant by considering the interaction regarding the startup characteristics of the boiler and steam turbine, which was not considered in the conventional method. It is about providing. [Means for Solving the Problems] The above purpose is to provide a startup schedule creation means for setting control target values of operating parameters necessary for startup of a power generation plant and creating a startup schedule for the power generation plant, and a method for creating a startup schedule in accordance with the startup schedule. A plant start-up system comprising schedule execution means for actually starting up a power generation plant, wherein said start-up schedule creation means includes means for storing a dynamic characteristic model of a power generation plant, and a means for storing a power generation plant start-up schedule and executing said start schedule. process state calculation means for calculating boiler stress and turbine stress using a characteristic model; and means for optimizing a target operating state of the turbine when the plant is started according to the assumed start-up schedule in accordance with the turbine stress. This is achieved by an optimal start-up system for the plant. More specifically, it includes a startup schedule creation function for creating a temporal startup schedule regarding operations and control target value settings necessary for startup of a thermal power plant before starting the plant, and In a plant startup system that has a schedule execution function to actually start up the plant, there is a dynamic characteristics model that can simulate startup characteristics corresponding to an arbitrary startup schedule, and a dynamic characteristics model that can simulate startup characteristics that correspond to an arbitrary startup schedule. A dynamic characteristic prediction function that uses a model to predict the behavior of process states related to operational constraints among process conditions that fluctuate in response to an arbitrary startup schedule, and a process state predicted by the dynamic characteristic prediction function. A constraint judgment function to check whether the operational constraints are satisfied, and an optimal start-up in which the process state predicted by the dynamic characteristic prediction function satisfies the constraints throughout the entire start-up process and the start-up time is the shortest. An optimal schedule search function for determining the schedule,
It has an optimum schedule setting function for setting the optimum start schedule in the schedule execution function, and starts the thermal power plant in the shortest possible time. [Operation] A plant dynamic characteristics model is prepared to confirm in advance whether the process state values during the entire process of startup satisfy the operation limit conditions, and a kind of hill-climbing method using this model is used to find the optimal operation policy, that is, the shortest Determine the startup schedule. [Embodiment] FIG. 1 shows the basic functional configuration of the shortest startup method for a thermal power plant of the present invention. Broadly speaking, the functions are comprised of a startup schedule creation function 1000 and a schedule execution function 2000. The former creates an optimal startup schedule signal 101 that minimizes the plant startup time, and the latter changes the manipulated variable signal 201 for the plant from time to time in order to start the actual plant 3000 according to the optimal startup schedule. The startup schedule creation function 1000 further includes a schedule optimization function 1100 and a plant dynamic characteristic prediction function 1200. The former further includes an offline optimization function 1110 and an online optimization function 11.
20, the latter further comprising a plant dynamic characteristic model 1210, a boiler stress calculation function 1220, and a turbine stress calculation function 1230. The offline optimization function 1110 assumes the startup schedule 111 and sets it in the plant dynamic characteristic model 1210 to determine the startup characteristics (213, 211,
212), and the boiler stress 221 and turbine stress 231 are calculated by the boiler stress calculation function 1220 and the turbine stress calculation function 1230, respectively. On the other hand, the online optimization function 1120 sequentially optimizes the target operating state 121 of the turbine according to the turbine stress 231 when the plant dynamic characteristic model 1210 operates. The offline optimization function 1110 evaluates the behavior of process variables related to the operational limit conditions calculated as a result and the required startup time, assumes a new startup schedule 111, and converts this into the plant dynamic characteristic model 1.
Set to 210. By repeating such processing, an optimal startup schedule 101 is determined that allows startup to be completed in an optimal time without causing the process state to violate operational restrictions. The determined optimal startup schedule 101 is set in the schedule execution function 2000, and becomes a target value for actually starting the plant. Here, the plant dynamic characteristic model 121
0, the boiler stress calculation function 1220 and the turbine stress calculation function 1230 each have an initial value of 321,
322 and 323 are required, and these are process state values measured before startup. Next, parameters defining the startup schedule in the embodiment will be defined. Since the plant start-up time basically depends on the temperature characteristics of the plant,
Parameters that have a strong causal relationship with the temperature increase characteristics of the plant should be selected. Based on this basic concept, four factors are defined: igniter ignition interval (T _IG ), mill start interval (T _PLV ), main steam temperature rise rate (L _TMS ), and reheat steam temperature rise rate (T _RHH ). selected as a parameter. The igniter ignition interval (T _IG ) is the interval at which, when a boiler ignition command is generated, the igniters corresponding to each burner stage of the boiler are sequentially ignited at time intervals T _IG and the light oil burner is ignited. The mill start-up interval ((T _PLV ) is the time interval when the pulverized coal mills are started up in sequence after all the light oil burners have been ignited. In this case, the mills up to the second one are started according to the operational standards. When the third mill is started, the total flow rate of pulverized coal input to the boiler is 40% of the rated value.
(Each mill's share is 67%).
After that, when the turbine is started and the generator output reaches 40%, it shifts to normal load operation mode, and the mills are started according to the load request, and up to five mills are operated. As each mill is started up, the light oil burner is extinguished. The main steam temperature increase rate (L _TMS ) is a parameter that defines the rate of increase in main steam temperature in the normal load operation region (40 to 100% load), and is the main steam temperature target value calculated by the schedule execution function 2000. For T _MSSET , it has the function shown in the following equation. T _MSSET =M _IN [{T _MS40 +T _MSR −T _MS40 /L _TMS −40(L−4
0)}, T _MSR ] ...(1) Here, T _MS40 : Main steam temperature when 40% load is reached (℃) T _RHR : Main steam temperature rated value (℃) L: Load (%) i.e. , means to make the main steam temperature reach the rated value when the load reaches L _TMS (%). The reheat steam temperature increase rate (L _TRH) is a parameter that defines the rate of increase in the reheat steam temperature in the normal load operation region, similar to the main steam temperature, and the reheat steam temperature calculated by the schedule execution function 2000. Target value
It has the function shown in the following equation for T _RHSET . T _RHSET =M _IN [{T _RH40 +T _RHR −T _RH40 /L _TRH −40(L−4
0)}, T _RHR ] ...(2) Here, T _RH40 : Reheat steam temperature when 40% load is reached (℃) T _RHR : Reheat steam temperature rated value (℃) L: Load (% ) That is, it means that the reheat steam temperature reaches the rated value when the load reaches L _TRH (%). Next, the optimization algorithm in the offline optimization function 1110 will be described in detail. FIG. 2 shows the basic processing procedure of the startup schedule optimization algorithm to which the complex method, which is one of the nonlinear optimization methods, is applied to the present invention. Here, _the schedule parameters ^are : X = _[ X ⁽ 1), X (2), X (3 _{) ,} It is written as. Each processing function will be explained below. (1) Initialization 100 Figure 3 shows the symbols, values, units, and meanings of the constants and initial values used in the optimization algorithm. (2) Initial simplex formation 200 The processing procedure is shown in FIG. The design value X _D is set at the initial trial point X ₁ and the simulation is executed. Simulation in this case means operating the plant dynamic characteristic prediction function 1200 to predict the startup characteristics when the plant is started according to the startup schedule _XD . At this time, the operation limiting factor Y (N _MV ) for X ₁ is the implicit constraint Y _L
(N _MV ) (see FIG. 5), an initial simplex is formed in the vicinity of X ₁ according to the following equation. _{X J =} X _{1 + B J} If X _J does not satisfy the implicit constraints, modify the trial points using the procedure shown in Figure 7. (a) Random number generation 220 FIG. 6 shows the procedure for calculating pseudorandom numbers using M as a variable. This algorithm is a method that uses the fifth number appearing from the highest digit of the square root. (b) Initial simplex correction 240 Trial point X according to extension magnification correction coefficient D()
(I, J) is modified as shown below. X(I,J)=X(I,J)+(1-D())(X _MAX
() -M _MIN ()) ...(5) (c) Determination of extension magnification correction coefficient 260 As shown in Figure 8, the sensitivity to the operation limiting factor when changing the operating parameters is large, medium,
Different from small and zero. Therefore, depending on which operation limiting factor does not satisfy the implicit constraint condition (see Figure 5), it is better to modify the extension magnification of the trial point than to uniformly modify it. is considered to be higher. FIG. 9 shows an algorithm for determining the extension magnification correction coefficient based on the implicit constraint monitoring algorithm (see FIG. 10) in accordance with this idea. (3) Characteristic evaluation ranking 300 As shown in Figure 11A and B, this function selects the following three characteristic points from among the vertices of a simplex consisting of K (K = 8 in this example). This is an algorithm for making decisions. Best vertex (X _Q , T _X,Q ) Operation parameters ( _{X Q} ) and activation time _{( T} ) Operation parameters (X _S ) and activation time (T _X,S ) corresponding to the vertex with the longest activation among K vertices Second worst vertex (X _{S2 ,} T Operation parameters ( _{X S2} ) and startup time ( _T ) Find the geometric centroid coordinates X _G of a simplex consisting of vertices. (5) New trial point determination 500 As shown in FIG. 13, a new trial point is defined as X _K+1 with coordinates that satisfy the following equation. This point is on the straight line connecting the worst vertex X _S and the center of gravity X _G , and is at a distance R (X _G −X _S ) from the center of gravity. Here, R is the extension magnification described in (7). X _K+1 = X _{G +} R ₍ X _G −X _S ) ...(6) However, X _MIN The point retreats in the direction of the center of gravity, but instead of retreating indefinitely, as shown in Figure 14, when R-0.1R ₀ is reached, the retreat is stopped, the entire simplex is degenerated, and a new search direction is started. Find out. This degeneracy method will be explained in (9). (7) Extension magnification correction 700 If the implicit constraint condition is violated, the extension magnification R is corrected by the method shown in FIG. (8) New trial point extension 800 If the startup time T _{X,K+1 at the new trial point X K+} 1 is shorter than the previous minimum time _T Extend in the same direction to approach the optimal point. Let this re-extension point be _XE . (9) Trial point regression 900 If the startup time T _{X,K+1 at the new trial point X K} +1 is longer than the previous second longest time T _X,S2
There is a possibility that X _K+1 has jumped over the optimal point. Therefore, as shown in Fig. 17, in the case of T _{X,K+1 T}
), and if T _X,K+1 >T _X,S , it is greatly retreated (R = -0.5). At this time, set the trial point to X _C
shall be. (10) Simplex degeneracy 1300 If no characteristic improvement point is found on the straight line connecting the worst point X _S and the center of gravity X _G (if there is no X _C such that T _X,C < _T The best point is the size of _Q
By shrinking in the direction of , we find a new possibility of approaching the optimal point. In this case, as shown in FIG. 18, the degeneracy rate is initially set to 1/2, but if each vertex does not satisfy the constraint conditions, the degeneracy rate is set to 3/4. If the constraint condition is still not satisfied, the vertex is returned to its original position. Here, the explicit constraints are the upper and lower limits of the optimization parameters themselves, and
X _MAX and X _MIN . As shown in FIG. 19, the simulation is executed after confirming that all parameters satisfy the explicit constraints. (11) Worst point exclusion 1420, 1440, 146
0 As shown in Figures 20 A to C, if X _E , X _{K+1 or X C} is improved over X _S , exclude X _{S and} A new simplex is formed by adding . (12) Judgment of reaching the limit on the number of searches 1500 The number of searches is the number of simulations,
By limiting this, we prevent our algorithm from falling into an infinite loop. FIG. 21 shows the processing procedure for this purpose, and the meanings of the symbols are as follows. N _T : Total number of simulations N _AD : Number of times a simulation result was used as a vertex of a simplex N _NG : Number of times a simulation result was not used as a vertex of a simplex N _KAD : X _K+1 as a vertex of a simplex Number of times used N _EAD : Number of times X _E is used as a vertex of a simplex N _CAD : Number of times X _C is used as a vertex of a simplex N _SAD : Simulation result for simplex degeneracy is a vertex of a simplex N _KNG : Number of times X _K+1 was not used as a vertex of a simplex N _ENG : Number of times X _E was not used as a vertex of a simplex N _CNG : Number of times X _C was not used as a vertex of a simplex Number of times not used N _SNG : Number of times a simulation result for simplex degeneracy was not used as a vertex of a simplex (13) Simulation 1600 The basic procedure of simulation is shown in FIG. In the simulation, the plant startup process was divided into three phases: boiler startup, speed increase, and load increase. In the boiler startup phase, pressure increase control (this function is built into the plant dynamic characteristic model) is executed from igniter ignition to reach the startup set pressure (94.9ata for main steam, 8.16ata for reheat steam pressure). This shows the startup process. The speed-up phase indicates a startup process in which the speed is increased to the rated speed by the metal match control function including the speed-up control function, and the metal match condition for the high-pressure turbine is satisfied. The load increase phase indicates a startup process in which a load is added by the addition condition determination function and the load increase control function reaches the rated load (target load in actual operation). (14) Optimum point convergence determination 1700 The optimal point, that is, the shortest startup schedule is assumed to be X _Q that satisfies the following equation. T _X,S −T _X,Q /T _X,Q <ε ...(7) X _Q at this time is written as X _OPT . This completes the detailed explanation of the offline optimization function 1110. Next, online optimization function 11
20 will be explained in detail. In order to achieve rapid startup through online optimization, we focused on the following points. (1) Rapid pressure increase considering the drum steam temperature distribution rate An increase in main steam pressure means an increase in drum pressure, and an increase in drum pressure is reflected in an increase in the saturation temperature determined by the pressure. Additionally, changes in drum steam temperature create thermal stresses in the drum. In order to keep this drum thermal stress below a permissible value, it is necessary to keep the rate of change in steam temperature below a permissible steam temperature. In the present invention, in consideration of the nonlinearity of the relationship between pressure and saturation temperature, a method is adopted in which the target pressure is determined so that the maximum allowable rate of temperature change is always achieved. This minimizes the time required for boosting. (2) Increased ventilation by calculating optimal metal match conditions The plant to be controlled was set to medium-pressure startup (speed-up method using an intermediate-pressure turbine). For metal match conditions, it is necessary to consider both the high-pressure turbine and the intermediate-pressure turbine. In the present invention, when the reheated steam temperature reaches the ventilable temperature determined from the metal temperature of the intermediate pressure turbine, the intermediate pressure turbine is immediately vented to increase the speed. When the speed increase is completed and the main steam temperature reaches the ventilable temperature determined from the metal temperature of the high-pressure turbine, control immediately advances to the load increase phase.
This keeps the turbine ventilation waiting time to the necessary minimum. (3) Rapid speed increase taking into account the stress of the intermediate pressure turbine The stress (thermal stress + centrifugal stress) generated on the rotor surface and bore of the intermediate pressure turbine is kept below the allowable value, and
Moreover, by sequentially determining the maximum speed-up rate, speed-up is completed in the shortest possible time. (4) Early annexation based on determination of conditions for annexation Immediately after annexation, the boiler generated steam temperature rises rapidly.
If this phenomenon is not taken into consideration and the metal match condition is established for the high pressure turbine, excessive thermal stress will be generated in the rotor even though the load is maintained. This minimizes the waiting time for merging and shortens the startup time. (5) Rapid load increase taking into account the stress of high-pressure and intermediate-pressure turbines The stress (thermal stress + centrifugal stress) generated on the rotor surface and bore of high-pressure and intermediate-pressure turbines can be suppressed to below allowable values, and the maximum load increase rate can be reduced. By making sequential decisions, the load increase can be completed in the shortest possible time. The processing method of the online optimization function 1120 created based on the basic idea described above will be described in detail. (1) Pressure increase control When the plant starts up, thermal stress occurs in the boiler drum as the temperature of the internal fluid changes. In order to prevent excessive thermal stress from occurring at this time, the rate of change in internal fluid temperature must be kept below a permissible value.
Since the internal fluid temperature is regarded as the saturation temperature uniquely determined by the pressure at that time, the allowable temperature change rate can be expressed by the allowable pressure change rate. As shown in Fig. 23, the relationship between pressure P and saturation temperature _TSAT11
23 is nonlinear. Now, if we write the set α(P) of saturation temperature change rate at pressure P as α(P) = (dT _SAT /dt)p...(8) and let α _L be the saturation temperature change rate tolerance, then
Allowable pressure change rate β(P) at pressure P 1124
is expressed as β(P)= _Max (dP/dt|α(P)<α _L ) (9), and the characteristic curve shown in FIG. 24 is obtained. This characteristic indicates that the higher the pressure level, the greater the allowable pressure change rate. A control system block diagram 1125 shown in FIG. 25 takes advantage of this feature in boost control. (2) Metal match control 1610 The basic processing procedure of metal match control is shown in FIG. This plant uses an intermediate pressure turbine startup method, so the reheat steam temperature T _RH is T _RMCHN (the value obtained by converting the metal match condition limit temperature of the intermediate pressure turbine to the reheat steam temperature, hereinafter referred to as the intermediate pressure turbine If it is higher than Negative Max), it means that the metal match condition has been established, and it is possible to increase the speed using the intermediate pressure turbine. If it is low, wait for the temperature to rise. However, when the metal match condition is established, the main steam temperature T _MS is T _MMCHP
(This is the value obtained by converting the metal match condition upper limit temperature of the high-pressure turbine to the main steam temperature, and is hereinafter referred to as the Positive Max for the high-pressure turbine). Since it is impossible to increase the load due to ventilation, increasing the speed using the intermediate pressure turbine is no longer meaningful. In other words, the metal match has failed. Also, during acceleration
If T _MS > T _MMCHP , it is also a metal match failure. If the main steam temperature T _MS after completion of speed increase is lower than T _MMCHN (the value obtained by converting the metal match condition limit temperature of the high-pressure turbine to the main steam temperature, hereinafter referred to as the negative max value for the high-pressure turbine), the main steam temperature Waiting for the steam to heat up. Then _TMS >
If T _MMCHN and the metal match condition is established,
The process moves to a function that determines the conditions that allow the addition of the load increase phase. If the engine continues to wait for the metal match condition to be established after completing the speed increase, the simulation time will be limited (T _LIMIT ) and it will be treated as a startup failure. This saves simulation calculation time. Next, the procedure for calculating the metal match condition will be explained. Negative Max for medium pressure turbine
Value (T _RMCHN ) 1611 FIG. 27 shows the procedure for calculating the negative maximum value (T _RMCHN ) of the reheat steam temperature for the intermediate pressure turbine. In this example, the metal match lower limit temperature T _RSMIN of the steam temperature in the intermediate pressure turbine ball during ventilation is set to a value that is 50° C. lower than the ball temperature T _IBO .
The process shown in the figure is for calculating the reheating steam temperature T _RSMIN such that the steam temperature in the bowl becomes T _RSMIN . This process is carried out by the turbine stress calculation function 1.
230 (a method for calculating the steam temperature in the bowl from the reheat steam temperature),
The method is to calculate T _RMCHN from T _RSMIN using convergence calculation. Positive Max Value (T _MMCHP ) for High Pressure Turbine 1612 FIG. 28 shows a procedure for calculating the Positive Max value (T _MMCHP ) of main steam temperature for the high pressure turbine.
In this example, the methyl match upper limit temperature T _MSMAX of the steam temperature after the first stage of the high-pressure turbine during ventilation is assumed to be the rotor surface temperature (equal to the casing inner wall temperature).
The value is set to 50°C higher than the actual temperature. The process shown in the figure is for calculating the main steam temperature T _MMCHP such that the first stage post-stage steam temperature becomes T _MSMAX . Similar to the previous section, this process also uses the calculation routine included in the turbine stress calculation function 1230 (method for calculating the first stage post-steam temperature from the main steam temperature), and conversely calculates T _MMCHP from T _MSMAX using convergence calculation. We decided to do this. Negative Max Value (T _MMCHN ) for the High Pressure Turbine 1613 FIG. 29 shows a procedure for calculating the Negative Max Value (T _MMCHN ) of the main steam temperature for the high pressure turbine. This processing procedure is exactly the same as the previous section, and the main steam temperature corresponding to the metal match lower limit temperature T _MSMIN is
This is to find T _MMCHN . (3) Speed-up control 1640 FIG. 30 shows the processing procedure of speed-up control. The characteristics of this processing are as follows. Predict the stress generated in intermediate pressure turbines,
By successively determining the maximum speed increase rate at which this predicted value is less than or equal to the allowable value, the turbine can be started with a speed increase pattern that provides the shortest speed increase time. To improve the accuracy of stress prediction, use the plant model as is for prediction. In this method, it is assumed that the speed is increased at the maximum speed increase rate DN1 between the reference time T _IMEO and T _NVARY , and the speed is maintained thereafter, and the stress generated in the turbine is predicted until the time T _IMEO +T _NUP . As a result, if the predicted stress value is less than the allowable value at any time, the speed is actually increased (as a simulation of startup) at the speed increase rate DN1 until the time _TIMEO +T _NVARY . On the other hand, if the predicted value is greater than or equal to the allowable value, the acceleration rate DN2, which is one rank lower, is set in the model to predict the generated stress. Even in the case of DN3, if the predicted stress value does not fall below the allowable value, set DN4 to maintain the speed. In this way, the time T _IMEO + T _NVARY
When this time is reached, this time is set as the reference time _TIMEO again and the same process is performed. By repeating the above process, the speed increase control is completed when the rated speed is reached.
The process moves on to the next combinability condition determination process. (4) Judgment of merging possible conditions 1620 FIG. 31 shows a processing procedure for determining merging possible conditions. The contents of this process are roughly divided into the following two parts, as shown by the broken line in the figure. Predicting the state after application The part that uses the plant model to predict the stress that will occur after applying the initial load (L = 3%) and determines whether this will be below the allowable value in the entire prediction interval _TIL . . If it is below the allowable value, load addition is performed. Waiting for combination conditions If the prediction result in the previous section is negative (the predicted stress is greater than or equal to the allowable value), the control cycle T _CS (until the next state prediction is performed) is not performed and the no-load operation continues. time). During this time, the main steam temperature
If T _MS exceeds the upper limit temperature T _MMCHP of the high-pressure turbine metal match condition, the metal match has failed.
Startup fails. If the metal match does not fail, the process returns to () again in the next control cycle and the state after the merging is predicted. If the conditions for concurrent use are satisfied, the process moves to the next load increase control. (5) Load increase control 1630 FIG. 32 shows the processing procedure for load increase control.
This control method is basically the same as the acceleration control method. In this method, it is assumed that the load is increased at the maximum load change rate DL1 between the reference time T _IMEO and T _LVARY , and the load is maintained thereafter, and the stress generated in the turbine is predicted until the time T _IMEO +T _LUP . As a result, if the predicted stress value is less than the allowable value at any time, the load is actually increased (as a startup simulation) at the load change rate DL1 until time _TIMEO +T _LVARY . On the other hand, if the predicted value is greater than or equal to the allowable value, the load change rate DL2, which is one rank lower, is set in the model to predict the generated stress. Even in the case of DL3,
If the predicted stress value does not fall below the allowable value,
Set DL4 to enter load holding state. In this way, T _IMEO + T _NVARY whose time is the next control period
When this time is reached, this time is set as the reference time _TIMEO again and the same processing is performed. By repeating the above process, when the target load is reached, the startup is completed. As described above, the startup time T required from boiler ignition to reaching the target load is defined as _X. [Effects of the Invention] According to the present invention, the startup time, that is, the time required from boiler ignition to reaching the rated load, can be significantly shortened compared to the conventional method, and can be reduced by 40% in a plant that starts and stops every day. becomes possible. Since the startup time can be shortened, the load adjustment ability of the power system is improved, and the stability of power supply is improved. Additionally, in each power plant, there is an accompanying effect that the start-up loss can be reduced as the start-up time is shortened, and the burden on operators is reduced.

[Brief explanation of the drawing]

Figure 1 shows the basic functional block configuration diagram of the present invention.
Figure 2 shows the basic processing procedure of the startup schedule optimization algorithm to which the complex method, which is one of the nonlinear optimization methods, is applied to the present invention, Figure 3 shows the definition of constants and initialization values, and Figure 4 shows the Figure 2 shows the processing procedure for initial complex formation, Figure 5 shows the constraints, Figure 6 shows the process flow for calculating pseudo-random numbers, Figure 7 shows the correction procedure, and Figures 8 to 10 show the operating parameters. 11A and 11B are flowcharts explaining the characteristic evaluation ranking function, and FIG. 12 is a detailed flowchart of the center of gravity calculation 400 in FIG. 2. FIG. 13 is a detailed flowchart of the new trial point determination 500 in FIG. 2, and FIG.
0, FIGS. 15 to 17 are detailed flow diagrams of blocks 700, 800, and 900 in FIG. 2, and FIG.
19 is a detailed flow diagram of step 130 in FIG. 18, FIG. 20 is a detailed flow diagram of steps 1420, 1440, and 1460 in FIG. Detailed flow diagram of step 1500,
Figure 22 shows a detailed flowchart of the simulation of step 1600 in Figure 2, Figure 23 shows the relationship between pressure and saturation temperature, Figure 24 shows the relationship between pressure and allowable pressure change rate, and Figure 25 shows the pressure increase. 26 is a flowchart of the procedure for metal match control, FIG. 27 is a procedure for calculating reheat steam temperature for an intermediate pressure turbine, and FIG. 28 is a calculation means for a high pressure turbine similar to FIG. 27. Fig. 29 shows the procedure for calculating the main steam temperature for the high-pressure turbine, Fig. 30 shows the processing procedure for speed-up control, Fig. 31 shows the processing procedure for determining conditions that allow combination, and Fig. 32 shows the load-increase control procedure. The processing steps are shown below. 1000...Startup schedule creation function block, 2000...Schedule execution function block, 3000...Plant.

Claims (1)

[Scope of Claims] 1. Startup schedule creation means for setting control target values of operating parameters necessary for startup of the power generation plant and creating a startup schedule for the power generation plant, and actually starting up the power generation plant in accordance with the startup schedule. A plant start-up system comprising a schedule execution means, the start-up schedule creation means storing a dynamic characteristic model of a power generation plant, and a means for storing a power generation plant dynamic characteristic model and calculating boiler stress by assuming a power generation plant startup schedule and using the dynamic characteristic model. and a process state calculation means for calculating turbine stress; and means for optimizing a target operating state of the turbine when the plant is started according to the assumed startup schedule in accordance with the turbine stress. Optimal start-up system for plants.