JPS6053161B2 - Turbine minimum life consumption scheduled start-up method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for starting a steam turbine, in which the startup of the turbine is completed in a specified time, and the life span of metal is reduced due to the stress in the rotor that occurs as the speed increases and the load increases. Concerning the starting method that makes it possible to minimize the amount.

When starting a steam turbine and changing load, it is necessary to suppress fatigue, ie, life consumption, due to thermal stress generated in the thick portion of the turbine to within an allowable value.

Therefore, in order to achieve safe and rapid startup, it is important to accurately determine and control the generated thermal stress. on the other hand,
From the standpoint of economic operation and stability of the power system, it is also required that the turbines of each power plant belonging to the power system complete startup on time specified by an external source such as a central power dispatch center. However, the time it takes to start up a turbine and its lifetime consumption have contradictory characteristics. In other words, if the lifetime consumption is suppressed, the time required for starting will become longer, and if the lifetime consumption is increased, the required starting time will be shortened. For this reason, we use a startup method that knows the start-up completion time in advance and keeps the life consumption as low as possible with respect to the specified time (hereinafter referred to as the fixed time).
In other words, it is desired to establish a method for minimizing the life consumption, as well as setting a regular start-up completion time. When the turbine is started, stress generation points that should be noted are the rotor surface and bore (center hole) of the first stage rear labyrinth packing, which are exposed to high-temperature, high-velocity leaked steam from the high-pressure and intermediate-pressure turbines.

However, the rotor is a rotating body, and it is difficult to actually measure the stress or the temperature distribution that is the basis for stress calculation. In conventional startup methods, a startup schedule is determined according to the main steam conditions and turbine metal temperature immediately before ventilation to the turbine, and speed and load increases are controlled in accordance with this schedule. However, in this method, the startup time is likely to be longer than necessary because the startup schedule is created with a margin that takes into account that the main steam conditions may deviate from the scheduled values during the startup process. Another conventional method is to consider the inner wall metal temperature of the casing after the first stage actually measured during the startup process as the rotor surface temperature, or to calculate the rotor surface temperature from the actual measured values of steam temperature and pressure after the first stage. The method used was to calculate the rotor stress by calculating the rotor stress, and to adjust the speed increase rate and load change rate accordingly.

However, in the former case, it is difficult to understand the correlation between the casing and the rotor, and in the latter case, measurement accuracy becomes a problem during no-load operation with a small flow rate and during low-load operation. In this way, it cannot be said that the conventional methods utilize the allowable lifetime consumption effectively. Furthermore, it has not been possible to meet the specified start-up completion time and to minimize the lifetime consumption.
The present invention eliminates the drawbacks of the above conventional methods of turbine starting,
To provide a method for starting a turbine, which enables starting to be completed on time at a specified starting completion time and furthermore, minimizing the life consumption of a rotor.

In order to achieve the above object, the start-up method of the present invention incorporates an online turpin start-up control system and a plant dynamic characteristic model, and first operates both models before ventilating the turbine to achieve stress conflict corresponding to the allowable life consumption. 1. Measure the start-up completion time under one limit value.

If this predicted value is earlier than the designated time, the stress limit value is set to a lower value than the previous time and the startup completion time is predicted again.
In this way, the calculation is converged until the predicted time becomes close to the specified value, and the actual startup is started using the stress limit value at the time of convergence as the final limit value, and the acceleration rate and load change rate are sequentially adjusted during the startup process. It is characterized by optimization. Examples of the present invention will be described below.

FIG. 1 shows the relationship of input/output signals between a minimum life consumption scheduled start-up system 100, a control system interlocked therewith, and a plant when a digital computer is used as an embodiment of the present invention. This figure also clearly shows the first stage rear labyrinth packing part 1 of the high pressure turbine 200 and the first stage rear labyrinth packing part 2 of the intermediate pressure turbine 300. This is where high-temperature, high-pressure steam leaks at high speed from both turbines, so heat exchange between the steam and the metal is most intense. This heat exchange causes a temperature gradient in the radial direction of the rotor in this vicinity, and a large thermal stress is generated in the rotor surface and bore (center hole) 3. The basic function of the starting system 100 of the present invention is to observe the starting completion time TR specified from an external source such as a central power dispatch center, and to determine the stress limit value that minimizes the life consumption of the rotor before starting starting, and to set this limit. The turbine is actually started under the specified value, and the stress generated during the process of speed increase and load increase is below the limit value, and the governor 10 is set to the optimum speed increase rate 4 allowed under those conditions, or the optimum Load change rate 6 is calculated by ALR (AutoOmaticlOOdReg
By sequentially setting ulatOr)7 as a target value, scheduled start-up with minimum life consumption is realized.

Note that Al. For example, a detection signal 28 of the steam pressure PHL after the high-pressure first stage is fed back to R as a signal representative of the output.

An instantaneous target load 9 is applied to the governor 10 from the ALR 7.
A speed signal N is fed back to the governor 10. This governor ultimately gives a valve position command 13 to the actuator 12 for controlling the position of the regulating valve 11. Furthermore, the starting system 100 of the present invention predicts the stress that will occur after the combination, determines the possibility of load combination, and issues a combination permission command 15 to the synchronous combination function 14. In addition, to explain other parts in FIG. 1, 400 is a low-pressure turbine, 500 is a generator connected to the turbine, 16 is a disconnector that disconnects the generator from the power system, and 17 is a disconnector. Disconnector on/off status signal,
18 is a load request signal sent from the central power supply station, 20 is main steam guided to a high pressure turbine, 21 is reheat steam guided to an intermediate pressure turbine, 22 is a control valve position detection signal, 23
is the main steam pressure detection signal, 24 is the main steam temperature detection signal, 2
5 is a reheat steam temperature detection signal, 26 to 30 are high pressure first stage rear casing outer wall temperature, high pressure first stage casing inner wall temperature, high pressure first stage rear steam pressure, medium pressure steam room outer wall temperature,
The detection signal 31 of the medium-pressure steam indoor wall temperature is an instruction signal of the start-up completion time TR sent from the central power supply control center. Next, details of the minimum life consumption scheduled activation system 100 of the present invention will be explained.

FIG. 2 shows the processing procedure of this system 100, which will be explained in order below. When this system starts operating, the startup completion time setting function 181 (hereinafter, one function will be omitted for simplicity. Other functions such as calculation, prediction, and correction will also be omitted) is specified from the outside. The startup completion time TR is imported into the system. Next, the stress limit value initial value setting 181 sets the stress limit value σ given by the operator according to the startup mode to an initial value σ. into the system as . However, it is assumed that σLO is given to each of the four stress points. The convergence calculation number counter reset 183 is a function for resetting the counter content of the number of repetitions of convergence calculation to 640゛1, which will be described later. first,
The convergence calculation number counter 184 adds 1 to the content and stores that the current calculation is the first calculation. The plant state value input 185 inputs an initial value of the plant state necessary for the start-up completion time prediction 250 to predict the start-up completion time prediction value τ under the stress limit value σ. The startup completion time prediction 250 predicts τ, and the punctuality determination 187 determines whether this i falls within the range of tolerance ±Δt with respect to the specified time TR. If the error falls within the allowable error range, the turbine is actually started by the thermal response predictive turbine control system 50. If not, time comparison 188 specifies the specified time T.
Determine whether it is slower than R. If not, the stress limit value σL is corrected in order to predict the startup completion time when the stress limit value is further suppressed. In this case, if it is the first correction, the stress limit value is corrected in stress limit value correction 196, and if it is a second or subsequent correction, the stress limit value is corrected in stress limit value correction 197.
However, this modification is carried out using the same concept for the four points of interest. In order to prevent this convergence calculation from being repeated infinitely, a convergence calculation termination judgment 191 is provided, and the number of repetitions is 1.
. I decided to discontinue it. do. However, stress comparison 189 and time comparison 188 show that the stress limit value σL is equal to the initial value σLO.
If it is determined that startup before TR is impossible with a value lower than , a message output 198 is output to the effect that scheduled startup is impossible with the specified σLO. This place. In this case, the initial stress limit value σ, is determined by the operator. is corrected to a value larger than the original value, or the activation conditions are changed, such as by extending the specified time TR for activation completion. Therefore, the starting condition change 199 is generally performed by the operator. Note that the variable storage 193 is a function to store variables from the previous time, and the variable storage 194 is a function to store variables from the previous time. A method of correcting the stress limit value in the minimum life consumption scheduled start-up system 100 described above will be explained using FIGS. 3, 4, and 5.

Three sets of curves, a broken line, a dashed-dotted line, and a solid line, shown in FIG. 3 indicate the stress limit values σ1, σ2, and σ, the corresponding turbine startup music, and predicted values of generated stress, respectively.

In this figure, first, when the predicted startup completion time value 〒1 is calculated for the stress limit value σ1, it does not fall within the tolerance of the specified time TR, so the stress limit value is corrected and the startup completion time is predicted/measured again as σ2. . The predicted value for this σ2 is
is also not within the allowable error of TR, so the next stress limit value σ is determined. As shown in FIG. 4, the stress limit value correction method is based on the relationship between τ1 for σ1 and I2 for σ2, and proportional interpolation as shown in the following equation is used for TR. If the predicted start-up completion time value 1 for the corrected σ is within the allowable error with respect to the specified value Tml as shown in FIG. 3, then σ at this time becomes the final stress limit value,
Start-up control is actually started by the thermal stress prediction turbine control system.

The minimum life consumption schedule can be achieved by the procedure explained above, but next, the fifth
This will be explained using figures.

Among the processing functions of the startup completion time prediction 250 shown in FIG. 5, the processing functions excluding the ALR model 207, governor model 210, plant model 211, startup elapsed time storage 212, and target load attainment judgment 213 are shown in FIG. The thermal stress prediction turbine control system 50 shown in FIG. Identical functional blocks are indicated by the same numbers.

Therefore, the same functions will be explained in detail later, and other parts will be explained here. The optimum speed increase rate and load change based on the stress limit value σ are sequentially determined by the speed control system 160 and load control system 140, respectively, and the speed increase rate is determined by the governor model 210, and the load change rate is determined by AL.
Each of these values is set as a target value in the R model 207.

This target value is periodically optimally corrected until startup is complete. The output signals of the governor model 210 and the ALR model are input to the plant model 211 to simulate the dynamic characteristics of plant startup. The activation elapsed time storage 212 is for determining the time when the activation is completed, that is, the predicted activation completion time value 1. The function that determines the completion of this startup is the target load attainment determination 213. The principle of scheduling in the minimum life consumption scheduled start-up system 100 and the method for determining the stress limit value corresponding to the minimum life consumption amount have been described in detail above.

The above embodiment is characterized by minimizing the life consumption under scheduled start-up, but conversely, the life consumption remains the same, that is, the stress limit value is set to σ. If you leave it as is, the startup completion time can be brought forward, and you can also take advantage of this feature. That is, by notifying the central power dispatch center of the predicted start-up completion time, the central power dispatch center can modify the start-up and load commands of each plant for economical operation of the power system. In order to clarify the basis for the high feasibility of this startup system, the thermal stress predictive turbine control system will be explained in more detail.

FIG. 6 shows a thermal stress prediction turbine control system 50 of the present invention.
This shows the processing procedure in.

First, before turbine ventilation, the initial temperature distribution of the rotor is estimated in initial temperature distribution determination 101. Here, the temperature distribution is determined from the actual measured values of the temperatures of the inner and outer wall metals of the casing after the first stage in high-pressure turbines and the steam chamber in intermediate-pressure turbines, in areas where the rotor and metal wall thickness are about the same and the temperature distribution tends to be similar. presume. In the next stress limit value determination step 102, the stress limit value σ which has been finally corrected by the procedure described above is basically used to determine the value. To explain in more detail,
In order to compensate for an error in estimating the initial value of temperature distribution during rapid restart and when immediately bringing the computer online, a value lower than the given stress limit value σ5 is set at the beginning of startup. This specific method will be explained later with reference to FIG. In prediction time determination 103, it is determined how far in advance stress should be predicted and controlled.

This predicted time is determined to be an appropriate value depending on the steam conditions generated by the boiler and the turbine startup sequence. Steam condition change rate learning 104 is a function for grasping the state of the current boiler dynamic characteristics relative to the operating state of the turbine.

Specifically, it is to understand from actual measurements at what rate the turbine inlet steam conditions (main steam temperature, pressure, and reheat steam temperature) change with respect to changes in turbine speed or load. This learning result is used in steam condition prediction 106, which will be described later. In the driving mode determination 105,
It is determined whether the mode is the speed control mode or the load control mode based on the ON/OFF state of the circuit breaker 16, and if the former, the process flow is switched to the speed control system 160, and if the latter, the process flow is switched to the load control system 140.

In the speed control system 160, the current stress estimation 161 estimates the current value of the rotor stress.

This function has the following functions: first stage post steam condition calculation 107, rotor surface heat transfer coefficient calculation 108, rotor temperature distribution calculation 109, rotor thermal stress calculation 110, and rotor stress calculation 111 considering centrifugal stress. In the current stress level check 162, it is determined whether the estimated current stress is below a limit value.

At this time, if the stress exceeds the limit value at even one point, the speed is maintained as a general rule. In the next calculation mode judgment 163, it is judged whether the current calculation is the time to search for the maximum acceleration rate based on the predicted calculation, and if it is the time, the process is passed to the maximum acceleration rate search 170; This processing function is bypassed and the processing is passed to the next critical speed judgment 164.

In this case, the processing cycle of the current stress estimation 161 is τ1, the processing cycle of the maximum acceleration rate search 170 is τ2, and the processing cycle of the maximum acceleration rate search 170 is τ2.
= NTτ1 (nτ: integer)
It consists of each processing function.

Furthermore, the stress prediction 172 includes a steam condition prediction 106, a first stage post-steam condition calculation 107, a rotor surface heat transfer coefficient calculation 108,
Rotor temperature distribution calculation 109, rotor thermal stress calculation 110,
It is composed of each processing function of rotor stress calculation 111.
This maximum acceleration rate search 170 is performed using a previously prepared acceleration rate [good,
Role ・2, ... hindrance ... circle (Rpm/min)]
Among these, the acceleration rate assumption 171 is assumed in ascending order, and the future value of the stress that will occur in this case is predicted. First, the stress after τ1 is predicted, including the steam conditions after the first stage. If this predicted stress is less than the limit value, further prediction is made after τ1. By repeating this, if all the stresses do not exceed the limit values until the predicted time that has already been determined, the assumed speed increase rate is set as the maximum possible speed increase rate based on the stress. However, in the process of predicting stress in τ1 increments, if the predicted value exceeds the limit value before reaching the prediction time, a one rank lower acceleration rate is assumed, and the prediction calculation proceeds again in τ1 increments from the current time. As a result, if the predicted stress does not exceed the limit value, this acceleration rate is used. The dangerous speed judgment 164 is a function that judges whether the current speed is in the dangerous speed region.

The optimal speed increase rate determination 165 has a function of setting the maximum speed increase rate found by the maximum speed increase rate search 170 in the governor 10, but if the current speed of the turbine is in the critical speed region, the speed increase rate is not changed and the previous speed increase rate is set. It has a function to read the speed increase at the speed increase rate.

Furthermore, if the current stress estimate exceeds the limit value, the speed will be maintained regardless of the search result for the maximum acceleration rate, but if the current speed is in the critical speed range, the speed will be increased at the same acceleration rate as before. Continue speed. When the speed increase is completed, the circuit breaker 16 is closed, and the load is added, the operation mode changes to the speed control system 160.
from there to the load control system 140.

In the load control system 140, the current stress estimation 141 is a function of estimating the current value of rotor stress.

This function consists of the following functions: first stage post-steam condition calculation 107, rotor surface heat transfer coefficient calculation 108, rotor temperature distribution calculation 109, rotor thermal stress calculation 110, and rotor stress calculation 111, all of which are used to control speed. This is a function shared with the system 160. In the current stress level check 142, it is determined whether the estimated current stress is below a limit value.

At this time, if the stress exceeds the limit value at even one point, the load is maintained. In the calculation mode judgment 143, the current calculation is a predictor. Determine whether it is time to search for the maximum load change rate based on the calculation, and if it is the time, perform the maximum load change rate search 15.
If not, this processing function is bypassed and the process is passed to the next optimum load change rate determination 144.

In this case, the processing cycle of the current stress estimation 141 is τ1, the processing cycle of the maximum load change rate search 150 is τ2,
The relationship is τ2=NTτ1 (nτ: integer). Maximum load change rate search 150 uses load change rate assumption 151 and stress prediction 1
52, predicted stress level check 153, and predicted time arrival judgment 154.

Furthermore, the stress prediction 152 includes a steam condition prediction 106, a first stage post-steam condition calculation 107, a rotor surface heat transfer coefficient calculation 108,
Rotor temperature distribution calculation 109, rotor thermal stress calculation 110,
It consists of each processing function of rotor stress calculation 111,
All of these functions are shared with the speed control system. This maximum load change rate search 150 is performed using the load change rate prepared in advance [
Jin, F, 2, ... Jin... (.p (%/min)
], the load change rate assumption 151 is assumed in descending order,
Predict the future value of stress that will occur in this case. This maximum load change rate search procedure is similar to the maximum speed increase rate search procedure described above. The optimum load change rate determination 144 uses the maximum load change rate found by the maximum load change rate search 150 as AL. R7
However, if the main steam temperature or reheat steam temperature is below a specified value, a signal for load maintenance, that is, a load change rate of zero, is set in ALR7.

This function also has the function of holding the load regardless of the maximum load change rate search result if the current stress estimate exceeds the limit value. The search signal generation 145 is a function for quickly increasing the load by giving flexibility to the learning function in the steam condition change rate learning 104 in the load increase control at the time of turbine startup.

As outlined above, stress limit value determination 102, prediction time determination 103, speed control system 160 or load control system 140
If each function is operated at a period τ1, turbine startup control is executed.

This repetitive operation continues until the system stop judgment 112 requests a system stop. Next, details of the above functions will be explained step by step.

First, regarding the initial temperature distribution determination 101 of the rotor, FIG.
This will be explained with reference to FIG. FIG. 7 is a sectional view of the rotor 40 and casing 41 of the labyrinth packing section 1 viewed from the axial direction. However, regarding the intermediate pressure turbine, attention will be paid to the steam chamber wall instead of the casing 41, but since the concept is the same, only the high pressure turbine will be described here. THCO, THCl, T,,T, shown in Figure 7 now.
, Tj (j=1 to m) are the casing outer wall metal temperature, the casing inner wall metal temperature, the rotor surface temperature, the rotor bore temperature, and the temperature of each cylinder when the rotor is divided into m virtual coaxial cylinders. Although it is difficult to actually measure the temperature distribution of the rotor, it is particularly important to accurately determine its initial value for this system, which aims at rapid startup and rapid load changes. FIG. 8 shows the specific processing contents of this initial temperature distribution determination 101.

When this system starts operating, it estimates the radial temperature distribution inside the rotor from the actually measured casing inner and outer wall temperatures THC1 and THCO.

In this case, T,,Tb are considered as.

KT in the above equation (3) is a constant determined by the shape of the turbine. The internal temperature distribution is regarded as a value determined by linear interpolation of Ts and Tb, and is expressed by the following equation. Also, B in the same figure
is a variable indicating the magnitude of the temperature gradient in the radial direction inside the rotor.

When the gradient is large, the error in estimating the temperature distribution also becomes large. When the temperature difference between the inner and outer walls of the casing is larger than the specified value ΔT, B=1, and when it is smaller, B=0. Furthermore, if the current speed N1 is larger than the specified value N, the estimation error may become large even if the temperature difference is small, so B=1. This value of B is for reference in the next processing function, stress limit value determination 102. Stress limit value determination 102 is a function to determine limit values for rotor surface stress and bore stress.

This function will be explained using FIG. 9. Let us now assume that the turbine is started at Tokison.

If the slope of the initial rotor temperature distribution at t1 is small,
In other words, when B=0, the stress limit value is the value that was finally corrected in the stress estimation value correction step mentioned above (±σL for the rotor surface, ±σLB for the rotor bore). It is assumed to be constant. However, when the slope is large, that is, when B=1, the stress limit value is subtracted from the final corrected value by Δσ as shown in FIG. 9 to ensure safety, taking into account the estimation error of the initial temperature distribution. As this Δσ, a value necessary to compensate for the estimation error of the initial stress is selected. Since the temperature distribution estimation error becomes smaller as time passes after startup, Δσ is gradually decreased until Δσ=0 at the end of the time. Next, prediction time determination 103 will be explained.

What is important in determining the predicted time Tp is the behavior of the reheated steam temperature TRH immediately after the addition. When the boiler is joined, the amount of fuel in the boiler increases in a stepwise manner, so as shown in Figure 10, the reheat steam temperature in particular rises rapidly and tends to follow the main steam temperature TM with almost a first-order lag. . Therefore, even if the initial load is maintained after joining, the rotor stress of the intermediate pressure turbine may continue to increase for a while. Therefore, before annexation, this phenomenon should be quantitatively predicted, and after confirming that the resulting stress does not exceed the limit value, annexation permission order 15.
is given to the synchronous merging function 14. For this purpose, as shown in the same figure, the shortest prediction time required is time T2, at which the stress generated when the initial load is held after joining is at its peak, and predictions must be made at least up to this point. When this Tp is determined from the dynamic characteristics of the boiler and turbine, it can be expressed by the following equation. Here, A, b, c, d: Constants TMO: Current value of main steam temperature TRHA: Current value of reheat steam temperature If it is assumed, the relationship between T and ΔTMR will be as shown in FIG. 11.

The predicted time after merging will be explained using FIG. 12.

It is thought that the change in reheated steam temperature after incorporation shows the dynamic characteristics predicted before incorporation. Therefore, the predicted time Tp is also the predicted time TpO determined immediately before the annexation, as shown in FIG.
and decrease over time. TpO after annexation
At the point in time, the predicted time is shortened to TPL, which is the same as during normal load operation. Next, the steam condition change rate learning 104 will be explained. What is studied here is the amount of change in three state quantities: main steam temperature TM, main steam pressure PM, and reheat steam temperature TRH, with respect to speed or amount of change. Since both methods are learned in the same way, the case of main steam temperature will be explained as an example in FIG. 13. Further, this learning method is carried out in exactly the same manner when the speed is increased, but here, the case when the load is increased will be explained. Accurate stress prediction begins with highly accurate prediction of turbine inlet steam conditions. However, this steam condition is closely related to the operating state of the turbine, and it is not easy to uniquely express this correlation as a dynamic characteristic model. Therefore, as shown in FIG. 113, the ratio of the load σ and the main steam temperature ΔTMS that changed between the current time t and the past nτ1 (n: integer) is learned as a rate of change.

This makes it possible to predict the rate of change in main steam temperature for any rate of load change. Next, when the circuit breaker 16 is in the 0FF state, that is, each processing function of the speed control system 160 will be specifically explained.

First, the current stress estimation 161 will be explained.

As mentioned above, this processing function includes the first stage post-steam condition calculation 107, rotor surface heat transfer coefficient calculation 108, rotor temperature distribution calculation 109, rotor thermal stress calculation 110, and rotor stress calculation 111, which are also used in the load control system. It consists of processing functions. A step-by-step explanation will be given below. The first-stage post-steam condition calculation 107 is a function for calculating the first-stage post-steam temperature and pressure of the high-pressure and intermediate-pressure turbines from arbitrary main steam conditions, turbine speed, speed increase rate, load, and reheat-steam temperature.

It is difficult to measure the steam temperature and pressure after the first stage of high-pressure and intermediate-pressure turbines with high accuracy in operating conditions where the steam flow rate is small, such as during speed increase and low load. Also, if you rely on actual measurements, you cannot expect highly accurate predictions. FIG. 14 shows a calculation procedure for uniquely specifying the boiler generated steam conditions and the turbine operating conditions in order to solve this problem. This specification method can be used consistently from speed increase to load operation by using main steam temperature TMS1 pressure PMS1 reheat steam temperature TRH1 speed N1 within the increase rate and load L as input variables. However, for safety reasons, the steam temperature after the first stage of the intermediate pressure turbine is assumed to be the actual value of the reheated steam temperature assuming that there is no temperature drop due to the first stage. The meanings of the symbols used in FIG. 14 are as follows. N: Speed (Rpm) NO: Rated speed (Rpm) A: Speed increase rate (Rpm/min) L: Load (%) L″: Equivalent load under rated steam conditions (%) L1:
Boundary load between full-circle injection and mixed injection (%) L2: Boundary load between partial injection and mixed injection (%) TMS: Main steam temperature (
℃) TRH: Reheat steam temperature (℃) TMSR rated main steam temperature (%) T1″: Steam temperature after the first stage of high pressure turbine for L″
degree (°C) ΔTO: Temperature difference between the main steam temperature and the steam temperature in the high-pressure turbine bowl (°C) ΔTRO: ΔTO at rated steam conditions (°C) THL: Steam temperature after the first stage of the high-pressure turbine (°C) T, l: Steam temperature after the first stage of intermediate pressure turbine (°C) PM,: Main steam pressure (Ata) Pl.

: Steam pressure after the first stage of high pressure turbine equivalent to no-load operation (Ata) PHlR: First stage of high pressure turbine at rated load
Steam pressure after stage (Ata) P, lR: Steam pressure after first stage of intermediate pressure turbine at rated load (Ata) PHL:
Steam pressure after the first stage of the high pressure turbine (Ata) P, l: Steam pressure after the first stage of the intermediate pressure turbine (Ata) K
1: Adjustment valve throttling rate (K1=0 at all times during partial injection)
(Ata)K2: Temperature reduction coefficient due to the first stage of the high-pressure turbine KN: Steam pressure after the first stage of the high-pressure turbine equivalent to no-load loss at rated speed (Ata)KAC:
Steam pressure after the first stage of the high-pressure turbine equivalent to acceleration (A
ta/(Rpm)2/min)k: No-load loss index The processing content of the rotor surface heat transfer calculation 108 is as shown in FIG. Focus on.

However, although Fig. 15 shows a high-pressure turbine,
The heat transfer rate for the intermediate pressure turbine is determined using the same procedure. The meanings of the symbols used in this figure are as follows. N
:Speed (Rpm)THl
: Steam pressure after first stage of high pressure turbine (℃) PHL: Steam pressure after first stage of high pressure turbine (Ata) Xl,T: Steam conductivity after first stage of high pressure turbine Kcaf/
m●℃●Sec) ν1, τ: Steam kinematic viscosity coefficient after first stage of high-pressure turbine (d/Sec) γ1, T: Specific weight of steam after first stage of high-pressure turbine (Kg/771') F$L:
Labyrinth packing leakage flow rate (K9/Sec)F
SLV: Labyrinth packing volumetric leakage flow rate (I/
Sec) UAX: Labyrinth packing axial leakage flow velocity (m/Sec) URO: Labyrinth packing rotor surface speed (m/Sec) U: Labyrinth packing combined leakage flow velocity (m/Sec) Re Nozzle number Nu: Nutsselt number K : Rotor surface heat transfer coefficient (Kca'/d・℃・Se
c) KO: Constant determined by the shape of the turpin δ: Gap between labyrinth packing (m) d: Rotor surface diameter (m) Z: Number of fins of labyrinth packing A: Gap surface ratio of labyrinth packing (a) R,: Rotor surface Radius (m) PH2: Pressure after the second stage of the high pressure turbine (Ata) However, in Fig. 15, the pressure ratio after the first stage and after the second stage (P2/P1) is determined by the operating state of the turbine, that is, the speed, speed increase rate, Even if the load changes, it can be considered almost constant, so it is actually calculated as a constant.

Next, the rotor temperature distribution calculation 109 will be explained using FIG. 16. Since the heat movement inside the rotor can be regarded as a one-dimensional flow consisting only of the radial direction, the rotor is divided into m virtual cylinders as shown in Fig. 16, and the temperature distribution is calculated by focusing on the heat balance between each cylinder. demand. The time step size for heat balance calculation is τ
1, Ql,S is the amount of heat transferred from the steam to the rotor surface during τ1, Q,,l is the amount of heat transferred from the rotor surface to the center of the outermost cylinder, and Ql,,il is the amount of heat transferred from the j-th cylinder This is the amount of heat conducted to the J+1st cylinder. However, since the bore is in an insulated state, Q. n,. ．．
+1=0. Now, if the current time is t, then the time t-
The amount of heat transfer that occurs between each cylinder during τ1 from τ1 to t is expressed as follows. Here, λM is the thermal conductivity of the rotor material.

From the relationship Q,,s(t)=Qs,l(t), Tf3(t
``1-cho'' is expressed by the following formula.

Here, the amount of heat ΔQ,(t) accumulated in the j-th cylinder is expressed as 4r2+3ΔR, so the temperature T of the j-th cylinder is expressed by the following equation.

Here, ■j: Volume per unit length of the j-th cylinder
ρM: Density of the rotor CM: Specific heat of the rotor material Further, the rotor bore temperature Tb(t) is expressed by the following equation by approximating the temperature distribution with a quadratic equation.

FIG. 17 shows the detailed procedure of this processing function described above. Next, rotor thermal stress calculation 110 will be explained.

The rotor thermal stress, that is, the rotor surface thermal stress σST and the rotor bore thermal stress σ8, is expressed by the following equation based on the temperature distribution obtained by the rotor temperature distribution calculation 109 described above.

Here, E: Young's modulus of rotor material α: Linear expansion coefficient of rotor material ν: Poisson's ratio of rotor protection Ts: Rotor surface temperature Tb: rotor bore temperature TM: Rotor volume average temperature Note that the rotor volume average temperature TM is determined by the following equation.

L=. ΣT, (R, 2-R2, +1)/(Rs2-R,
2) J-1...(
19) Next, rotor stress calculation considering centrifugal stress 111
I will explain about it.

Since centrifugal stress is proportional to the square of the turbine speed N, the rated speed is N. If the bore centrifugal stress at the rated speed is σ8.8, the centrifugal stress σBO acting on the bore at the speed N is expressed by the following equation.

Therefore, the bore stress σ.

Hato becomes.

Note that stress concentration occurs on the rotor surface due to the surface shape, and the direction of action of thermal stress is the axial direction. Considering that the centrifugal stress is in the circumferential direction, both act in directions perpendicular to each other. Therefore, regarding the rotor surface stress, it is sufficient to consider only the thermal stress, which causes a problem of life consumption, and the rotor surface stress σ5 is as follows.

This completes the explanation regarding the current stress estimation 161.

The next current stress level check 162 is the above σ,,σ
8 is a function for determining whether or not the stress limit value set in the stress limit value determination 102 described above is exceeded.

The next calculation mode determination 163 is a function of determining whether or not the current calculation is the time to perform a search for the maximum acceleration rate based on the predicted calculation.

That is, when it is specified that the predictive calculation is to be performed once every n times, the present processing function 19 has the function of bypassing the maximum acceleration search 170 for n-1 times out of n diagrams. Next, the maximum acceleration rate search 170 will be explained.

This processing function uses the current time as a reference to determine the predicted time 1.
The stress generated on the rotor surface and rotor bore until after the prediction time Tp determined in step 03 is predicted with a time step width τ1, and each time, the stress generated is compared with the stress limit value, and the maximum stress during this period will not exceed the limit value. This is a function to search for the acceleration rate. The speed increase rate referred to here is selected from a plurality of speed increase rates prepared in advance based on the speed increase rate assumption 171. The plurality of acceleration rates are passed to the stress prediction 172 in order from the largest one based on the acceleration rate assumption 171. The processing function of the stress prediction 172 first predicts the stress on the rotor surface and bore after τ1 from the current time, and compares it with the stress limit value in the predicted stress level check 173. As a result of this comparison, if the stress of both is below the limit value, stress prediction 1
Return to step 72 and further predict the stress after step 71. In this way, what is the assumed value of the acceleration rate? T at intervals of T1 for x
The stress is predicted until after p and compared with the limit value, but if either the rotor surface stress or the bore stress exceeds the limit value, the process is returned to the acceleration rate assumption 171 and the acceleration rate assumption value is and similarly predict the stress. In this case, the assumption of the acceleration rate is made in descending order, and if the predicted stress value for the assumed acceleration rate value does not exceed the limit value over the entire prediction period TP, the prediction time arrival judgment 174 is made based on the assumption at this time. Determine the acceleration rate as the maximum acceleration rate and complete the search. If the stress exceeds the limit value for all the assumed values of the acceleration rate, the acceleration rate is set to zero as the maximum acceleration rate search result. In addition,
The processing content of the stress prediction 172 is the above-mentioned current stress estimation 161.
It is similar to that of . The difference is that the turbine inlet steam condition uses a predicted value instead of the current value, and the speed uses the predicted value corresponding to the assumed speed increase rate instead of the current value. In order to predict this turbine inlet steam condition, the result of learning the ratio of the amount of change in steam condition to the amount of load change is used, as explained in FIG. 13 and equation (7). That is, the time change rate of the main steam temperature with respect to the assumed value Hx of the speed increase rate is expressed by the following equation. The critical speed determination 164 is a function of determining whether the current turbine speed is in the critical speed region, and the result of this determination has an important meaning in the next optimal speed increase rate determination 165.

Note that this optimum acceleration rate determination 165 has already been described. As explained above, the optimum speed increase rate is set for the governor 10 every nτ1, but the current stress value is monitored at the cycle τ1, and if this exceeds the limit value, the speed is maintained. Turbine speed increase control can be performed safely even when the turbine inlet steam conditions change due to disturbances that were not taken into account at the time of prediction. Next, circuit breaker 1
6 is in the 0N state, that is, the load control system 140
Each processing function will be specifically explained.

In the load control system 140, the processing methods of current stress estimation 141, current stress level check 142, calculation mode judgment 143, and maximum load change rate search 150 are basically performed by the respective processing functions 161, 162, and 16 of the speed control system 160.
Same as 3,170. However, the only difference is that in the speed control system 140, the speed increase rate is the target of the maximum value search, whereas in the load control system 160, the load change rate is the target of the maximum value search. The load change rate assumption 151 in the maximum load change rate search 150 is assumed in order from the largest load change rate among a plurality of predetermined load change rates.

Next, the optimum load change rate determination 144 will be explained.

This processing function 144 has two functions. One is to set the load change rate found in the maximum load change rate search 150 at a cycle of γ1 to ALR7 and correct it. If the current stress exceeds the stress limit value during the period nτ1, Another function is to directly hold the load.The other is the load limiting function, which sets an upper limit of 4 loads depending on the main steam conditions. This function is to prevent erosion of the final stage blades of the turbine.

This load limiting method is as shown in Figures 18 and 19.
In this method, the lower limit values of the main steam temperature and the reheat steam temperature are determined, and if both of these limit values are not satisfied, the load is maintained.
That is, FIG. 18 shows the load limitation based on the main steam temperature, and the load is maintained unless the main steam temperature exceeds the lower limit value TMsL that meets the main steam pressure PMS. Further, FIG. 19 shows load limitation based on the reheat steam temperature, and if there is no reheat steam temperature equal to or higher than the lower limit value TRHL that meets the load L, the load is maintained. ] Next, search signal generation 145 will be explained.

As a method for predicting the rate of change in steam conditions, Figure 13 and (7)
) The rate of change in steam conditions is learned using the method shown in the formula below, and future values are predicted based on this. but,
If some kind of disturbance occurs in the boiler and the steam condition suddenly increases, as is clear from equation (7), the rate of change in the steam condition will be learned and stored to a greater extent than in normal times.
In such a case, the stress will be predicted to be higher than the actual one, and even though the actual stress is sufficiently small compared to the limit value, a long-term load retention phenomenon will occur, making it impossible to increase the load. There is a risk that it will happen. The search signal generation 145 is a function to prevent this phenomenon. This specific method is the 20th
By superimposing the search signal ΔLEX as shown in the figure on the load,
The change in steam conditions at that time is learned in the same way as equation (7).
In this case, the rate of change (ΔT
M, /ΔLEx), the already learned rate of change (Δ
TM3/ΔL). The correction method uses a weighting coefficient β as shown in the following equation. Next, this search signal Δ
LEX is generated at the same period as the maximum load change rate search period nτ1, but the change rate is determined as follows, and AL
Set to R7.

Now, among the values obtained by normalizing the current stress on the rotor surface and bore of the high/intermediate pressure turbine by the limit value, the one with the maximum absolute value is called σ. It is defined as That is, it is expressed by the following formula.
Here, σLS: Rotor surface stress limit value σLB: Rotor bore stress limit value σH,: High pressure turbine rotor surface stress σ1,: Medium pressure turbine rotor surface stress σHB: High pressure turbine rotor bore stress σ18: Medium pressure turbine rotor bore stress This σ1 Accordingly, the rate of change (.EXR) of the search signal as shown in FIG. 21 is determined.

The steam condition change rate learning 104 has a learning value correction function using a search signal as described above, but it also has a so-called forgetting characteristic in which the learning value is forgotten over time. .

That is, until new learning is performed, the learned value of the steam condition is forgotten according to the forgetting characteristic shown by the following equation. (2
7) and (28), if the learning result is corrected at a period τ1, a forgetting characteristic with a time constant γ1 is obtained.

The response of turbine inlet steam conditions to a load increase, especially the temperature increase characteristics of reheat steam temperature, changes significantly from the time of combustion at turbine startup until the low load range. Specifically, the time constant of temperature rise changes significantly. In order to effectively utilize the steam condition change rate learning function even in such a case, the operation cycle, that is, the AL. It is necessary to correct the setting cycle for R in response to changes in the time constant. In order to realize this, the calculation mode judgment 143 of the load control system 140 is
Provide the functions shown in Figure 2. That is, in the low load region, the maximum load change rate search period is made larger than nτ1 to ensure that the response of steam conditions with a large time constant is learned, and then the maximum load change rate search 150 is operated. . According to the present invention, (1) startup can be performed while faithfully observing the designated startup completion time, so it is possible to respond stably and promptly to the economic operation plan of the power system by the central power dispatch center.

(2) Since it is possible to start the rotor while minimizing the life consumption of the rotor by effectively utilizing the given start-up time, the rapid start-up performance is improved accordingly.

(3) Plant startup loss can be reduced by improving rapid startup performance.

(4) Stable control of rotor stress is possible, and rotor life can be managed with precision.

(5) The central power dispatch center can operate the power system more economically by correcting the start-up and load commands to each plant upon receiving notification of the predicted start-up completion time from each plant.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing the relationship between the minimum life consumption scheduled start-up system according to the embodiment of the present invention, the control system interlocked therewith, and the input/output signals with the plant, and Fig. 2 is the minimum life consumption scheduled start-up system according to the embodiment. A flowchart showing the basic functions of the system and its processing procedures. Figure 3 is a time chart illustrating how to predict the start-up completion time and minimize rotor life consumption. Figure 4 is a time chart illustrating how to predict the start-up completion time and minimize rotor life consumption. A characteristic diagram illustrating the method for determining stress limit values by convergence calculation, FIG. 5 is a flowchart showing the processing procedure of the start-up completion time prediction function in the minimum life consumption scheduled start-up system in the embodiment, and FIG. 6 is an embodiment of the present invention. Flowchart showing the processing procedure of the thermal stress prediction turbine control system in the minimum life consumption scheduled start-up system of A flowchart showing the method for determining the initial temperature distribution, Figure 9 is a characteristic diagram showing the stress limit values for the rotor surface and bore, and Figure 10 shows the relationship between the dynamic characteristics of the turbine inlet steam temperature and thermal stress immediately after load is added. Characteristic diagram, Figure 11 is a characteristic diagram showing the predicted time before merging, Figure 12 is a characteristic diagram showing the predicted time after merging, Figure 1
Figure 3 is a characteristic diagram showing the learning method for the steam condition change rate, No. 14.
The figure is a flowchart showing a method for estimating the steam conditions after the first stage.
Figure 15 is a flowchart showing the method for calculating the heat transfer coefficient of the labyrinth packing part, Figure 16 is a conceptual diagram showing the concept of heat balance between the virtual divided cylinders of the rotor, and Figure 17 is a concrete diagram of the rotor temperature distribution. A flowchart showing the calculation procedure, FIG. 18 is a flowchart showing the lower limit of main steam temperature for load limiting, FIG. 19 is a characteristic diagram showing the lower limit of reheat steam temperature for load limiting, and FIG. 20 is a flowchart showing the lower limit of main steam temperature for load limiting. FIG. 21 is a characteristic diagram showing a method of determining the rate of change of the search signal, and FIG. 22 is a characteristic diagram showing a method of determining the operation cycle. 100...Minimum life expectancy turbine starting system, 20
0...High pressure turbine, 300...Intermediate pressure turbine, 4
00...Low pressure turbine, 500...Generator, 1...
・High pressure first stage rear labyrinth packing part, 2... Medium pressure first stage rear labyrinth packing part, 3... Bore (center hole)
, 4... Optimum speed increase rate, 6... Optimum load change rate, 7.
...ALRl9...Momentary target load, 10...Governor, 11...Adjustment valve, 12...Actuator, 13
... Valve position command, 14... Synchronous addition function, 15...
- Combination permission command, 16... conversion or disconnection, 17... conversion or disconnection ON/OFF status signal, 18... load request signal, 16
...speed detection signal, 20...main steam, 21...reheat steam, 22...control valve position detection signal, 23...main steam pressure detection signal, 24...main steam temperature detection signal, 2
5... Reheat steam temperature detection signal, 26... Casing outer wall temperature detection signal after high pressure first stage, 27... Casing inner wall temperature detection signal after high pressure first stage, 28... After high pressure first stage Steam pressure detection signal, 29... Medium pressure steam room outer wall temperature detection signal, 30... Medium pressure steam indoor wall temperature detection signal, 31
...Start-up completion time instruction signal.

Claims (1)

[Claims] 1 Applicable to power generation equipment consisting of a working fluid generation source for driving a turbine, a valve that controls the flow rate of the working fluid generated from the source, a turbine operated by the fluid, and a generator mechanically connected to this. , comprising a control means for calculating the stress generated in the turbine due to a change in the state of the working fluid, and controlling the turbine while defining a rate of change in the state transition of the power generation equipment so that the calculated stress is equal to or less than a stress limit value. A method for starting power generation equipment includes a first step of predicting a start-up completion time corresponding to an assumed stress limit value, and a start-up predicted based on a preset expected start-up completion time and the first step. a second step of determining a stress limit value in relation to a completion time, the control means starting the turbine in accordance with the determined stress limit value; starting method.