JPH0434162B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an efficiency optimization control system for a power generation plant, and more particularly to an efficiency optimization control system for a power generation plant suitable for maintaining plant efficiency at the highest point.

Although the issue of highly efficient operational control of power plants has been a hot topic for some time, a control system that aims to improve the efficiency of a plant from a comprehensive perspective has not yet been put into practical use.

There are a few reports on optimizing boiler efficiency by manipulating the air-fuel ratio in thermal power plants (e.g., F. Moran et al; Development
and application of self−optimi Sing control
to coalfired steam-generating plant,
Proc IEE, VoL 115, No. 2 (1968-2)).

In these conventional control systems, as shown in FIG. 1, the optimal manipulated variable search means 3 determines the optimal manipulated variable 5 based on the status feedback signal 4 from the plant 100, and the power plant 100 is controlled based on this. It has the structural feature of controlling.

However, the conventional method has the following problems when applied to actual equipment.

() Since the optimal operating amount is searched based on the status feedback signal from the plant with a large thermal time constant, a large amount of time is required after finding the maximum efficiency point, that is, the optimal operating amount.

() Because it is difficult to actually measure efficiency with high precision, the efficiency index method, which uses behavior such as main steam pressure as a substitute for efficiency evaluation, is used, and its reliability is low.

() Because actual measured values are used, they are susceptible to noise and detection errors.

() Since the maximum efficiency point is searched based on the status feedback signal, the period during which optimization control is applied is limited to steady load operation, and control operation is forced to stop when the load fluctuates.

An object of the present invention is to eliminate the drawbacks of the conventional methods described above in a control system for a power generation plant, and also to improve the efficiency of the plant from an overall viewpoint, and in particular, to improve the efficiency of the power plant as the load demands change. We are developing a transient efficiency optimization control system for power plants that can efficiently optimize the manipulated variables of multiple operating parameters that affect efficiency, not only when the target settling load can be predicted by using the system, but even when the target settling load cannot be predicted. It is on offer.

In order to overcome the problems of the conventional system and achieve the object, the present invention uses a plant model 240 for calculating transient efficiency (hereinafter simply referred to as efficiency) built into the control system 200 shown in FIG. By using this system and applying model-based predictive control, it is possible to achieve fast and highly accurate efficiency optimization without being constrained by the response speed of the plant.

The basic method of the optimal manipulated variable search means 230 in the control system 200 shown in FIG. This is a method of searching for the optimal manipulated variable that will bring about the maximum efficiency point by repeating the procedure of outputting , and finding the corresponding plant efficiency 10.

In this case, the plant model 240 is capable of determining the transient efficiency of the plant by simulating plant dynamic characteristics corresponding to arbitrary load fluctuations and temporal changes in a plurality of operating parameters.

In the following, the present invention will be described in more detail by way of an example in which the present invention is applied to a thermal power plant.

FIG. 3 is a schematic block diagram showing the basic configuration of an efficiency optimization control system for a thermal power plant to which the present invention is applied. However, this figure only shows the functions specific to efficiency optimization, and the illustration of the minor control system and plant control system of each device to which the conventional method can be applied as is is omitted to avoid complication of the drawing.

There are many operating parameters that affect efficiency, but in this example, the O ₂ excess rate, which has a relatively large effect,
A control system that selects the parallel damper opening degree and the condenser vacuum degree and optimizes these manipulated variables will be described.

Before explaining the working principle of the entire control system,
An overview of the optimal operation amount search means 230 will be given below. The details will be explained later with reference to FIG. 5. Here, a case will be explained in which a well-known complex method, which is one of the extreme value search methods, is used as a basic algorithm for searching for the optimum manipulated variable.

In the illustration of the optimal operation amount search means 230 in FIG. 3, the operation parameters m ₁ ,
Only the relationship between m ₂ and efficiency η is shown.

First, the efficiency η ¹ when the trial operation amounts m ¹ ₁ and m ¹ ₂ corresponding to trial pattern 1 are given to the plant is determined.

In addition, in the above, the efficiency η ¹ when the trial operation amounts m ¹ ₁ and m ¹ ₂ are given to the plant is determined.

In the above, the trial manipulated variables ¹ ₁ , m ¹ ₂ are a set of manipulated variables m as a time function necessary to change the plant load from the current value to the target settled load at a planned rate of change. ₁ , _m2 .

As is clear, the combination of trial operation amounts m ₁ and m ₂ to change the plant load in the same way is
There may be many others. Let these combinations be [m ² ₁ , m ² ₂ ], [m ³ ₁ , m ³ ₂ ]... [m ^j ₁ , m ^j ₂ ]...,
These are referred to as trial patterns 2, 3...j...

Then, for trial patterns 2, 3...j..., the efficiencies η ² and η ³ are determined in the same way. Note that the (transient) efficiency η ^j of the plant in this case can be calculated, for example, using the following integral formula.

η ^j =∫ ^tb / _ta Q ₀ dt /∫ ^tb / _ta Q〓dt Here, Q〓 is the heat input to the plant (kcal/sec),
Q ₀ is the heat exchange power from the generator (kcal/sec),
[t _a /t _b ] is the integration interval of interest.

Among these trial patterns, the point corresponding to the pattern with the lowest efficiency (in this case, it is treated as trial pattern 1) and the center of gravity of the remaining trial patterns (in this case, trial patterns 2 and 3 are Select a new trial pattern 4 on the line that connects the connecting line segment 2-3 (on the connecting line ^segment _2-3 ) and extend it toward the center _of gravity ^. Find the corresponding efficiency η ⁴ .

The length of the extension line in this case can be determined experimentally or empirically depending on the target plant, but in general, the length of the extension line is determined at the point representing the trial pattern with the lowest efficiency (hereinafter referred to as It is said that the distance between the trial pattern (sometimes simply referred to as a trial pattern) and the center of gravity of the remaining trial patterns excluding the trial pattern should be approximately 1.3 times longer.

Next, a new trial pattern 5 is obtained in the same manner as described above from the newly created triangle 234, excluding the trial pattern 1 with the lowest efficiency. Yes,
If the new trial pattern 5 violates the constraint conditions of the plant state quantity, the process returns to the trial pattern 6 within the domain and determines the trial direction using a new triangle 346.

By repeating the trial method described above, the maximum efficiency pattern (in this case, trial pattern 7) is obtained.
can be reached. The operation amount corresponding to this last trial pattern is the optimal value m∧ _i (in this case m ⁷ ₁ ,
m ⁷ ₂ ), and in Fig. 2, the plant 100
This is the actual operation output 11 for .

By continuing this type of control operation, even if the maximum efficiency point moves due to changes in ambient conditions such as seawater temperature or atmospheric temperature, it is possible to follow this and determine the optimal operating amount. .

The above is an explanation of the principle of efficiency optimization control applying the known complex method.Next, the operating principle of the entire control system in an embodiment will be explained in detail. In addition, the manipulated variables in this case are m ₁ ,
It is assumed that there are three types: m ₂ and m ₃ .

From the load request value generation means 110 shown in FIG.
As shown in FIG. 4A, a load request value Lr(t) 13 is generated as a function of time and sent to the optimum manipulated variable search load point determining means 220.

The optimum operation amount search load point determining means 220 is configured to select one or more points within the load change region from the current load Lr (t ₀ ) immediately before the load request value changes to the target settling load LrS (=Lr (t ₃ )). Multiple optimal manipulated variable search load points L _S
(p) (where p=1 to l) is determined.

Here, L _S (p) is calculated by the optimum operation amount search means 230 using the plant model 240 for transient efficiency calculation.
This is the intermediate load point for searching for the maximum efficiency, that is, the optimum amount of operation. In the example shown in Fig. 4, the optimum operation amount search load points are shown as three points L _S (1), L _S (1), and L _S (3), but these numbers are determined by the load change width and the load point. It can be arbitrarily determined depending on the rate of change (dLr/dt).

The optimal operation amount search load point L _S (p) determined in this way is sent to the efficiency calculation load signal generation means 16, and the efficiency calculation load signal L _S (t ) to plant model 24
Sent to 0.

On the other hand, the optimal manipulated variable searching means 230 has its operation timing controlled by the operation control signal 15 to determine the optimal manipulated variable that maximizes the transient efficiency, and the optimal manipulated variable search means 230 has the operation timing controlled by the operation control signal 15 to determine the optimal manipulated variable that maximizes the transient efficiency. Trial operation amount by load point 19 - that is, m ^j ₁ (1), m ^j ₂ (1)...m ^j ₁ (2), m ^j ₂ (2),..., m ^j ₁ (p
),
m ^j ₂ (p) and m ^j ₃ (p), and receives from the plant model 240 the plant efficiency 10 - that is, η ^j for the trial pattern j.

In this way, the optimum operation amount searching means 230
are the optimal operating quantities 11 that maximize the transient efficiency while the load changes from the current value Lr (t ₀ ) to the target settling load Lrs, i.e., m∧ ₁ (p), m∧ ₂ (p), m∧ ₃
(p)
is determined in correspondence with the load point L _S (p).

In this case _{, in the optimum operation amount searching means 230, each operation amount m j is} determined in a predetermined relationship (for example _, Efficiency is calculated assuming that it changes (linearly).

FIG. 5 shows the processing procedure in the optimum operation amount search means 230. The control algorithm will be explained step by step below. However, the symbols used below are defined as follows.

m _j : Manipulated amount i=1 O ₂ excess rate (%) i=2 Parallel damper opening degree (%) i=3 Condenser vacuum degree (mmHg) Here, m ₂ directly indicates the parallel damper opening degree It is not defined by the following equation (1).

m ₂ = G _S / G _S + G _R × 100 (%) ...(1) In the above equation (1), G _S is the location of the primary superheater and economizer in the boiler of the thermal power plant. G _R represents the gas flow rate in the gas passage where the primary reheater and economizer in the boiler are located.

m _i.nax , m _i.nio : Upper and lower limits of operation amount G _{CL nax} , G _{CL nio} : Upper and lower limits of condenser circulating water flow rate (Kg/sec) D _Tnax : Condenser circulating water temperature rise range (℃) Y _nax : Upper limit of low-pressure turbine exhaust humidity (℃) G _{GR nax} , G _{GR nio} : Upper and lower limits of gas recirculation flow rate (Kg/sec) [Step 1] Initial trial of initial simplex formation Points m ^j ₁ (p) (i=1 to 3,
j=1~l) satisfies all of the above constraints, and in the (3×l)-dimensional space spanned by m _i (p),
A polygon (called a simplex) with k angles (in the case of Figure 6, k = 6 is used for convenience, but k is usually about 2 x 3 x l) as shown in Figure 6A. This is used as the initial simplex.

During the formation of this initial simplex, 1
The points are m ¹ _i (1) corresponding to the initial trial pattern, and the remaining (k-1) points are uniform random numbers r ^j (j
=2 to k), and is determined by the following equation (2).

m ^j _i (P) = m _i.nio + r ^j (m _i.nax −m
_i.nio )...(2) However, 0<r ^j 1 j is the trial pattern number p is the maximum efficiency (optimum operation amount) search intermediate load point number m ^j _i (p) determined in this way is the operation Although the quantity constraint (this is called an explicit constraint) is always satisfied, the process state constraint (this is called an implicit constraint) is not necessarily satisfied, so m ^j _i (p) (i= 1-3, p=1-l)
is sent to the trial operation signal conversion means 17, and the time function m ^j _i
(t) is given to the plant model 240 for confirmation.

If the shadow constraint condition is satisfied, the point corresponding to the trial pattern is moved in the direction of the center of gravity of the already determined points to the midpoint. In this way, points corresponding to all trial patterns are ultimately determined.

Through the above trials, each trial pattern j
The plant efficiency η ^j corresponding to the plant model 24
easily obtained from 0.

[Step 2] Calculating the center of gravity Here, among the trial points of the simplex,
Center of gravity m _Gj (p) of a simplex defined by (k-1) points excluding the lowest efficiency trial point
(i=1-3, p=1-l) - Figure 6B.
Now, assuming that the lowest efficiency point is the trial pattern with j=1, the center of gravity m _Gi (p) is expressed by the following equation (3).

Also, the lowest efficiency point (in this case, the point of j=1)
The distance Δm _Gj (p) from to the center of gravity is expressed by the following equation (4).

Δm _Gi (p)=m _Gi (p)−m ¹ _i (p) ...(4) [Step 3] Determine the point corresponding to the new trial pattern Set the direction of the new trial to the straight line L in Figure 6C. As shown, the point corresponding to the new trial pattern is a point taken in the direction of the center of gravity from the lowest efficiency point (point 1 in this example) and extended by α _i times the distance Δm _Gi (p) between the two points. In Figure 6C, point 7) is defined as m ^k+1 _i
(p), it is expressed as m ^k+1 _i (p)△=m _Gi (p)+α _i m _Gi (p)...(5).

In this case, as shown in FIG. 6C, if this point 7 violates the explicit constraint, a point corresponding to the new trial pattern is taken on the intersection of the constraint and the line L, and this is set as the regular trial pattern. I will set it to 7. In other words, when the upper limit m _i.nax is violated, m ^k+1 _i (p) = m _i.nax ...(6), and when the lower limit m _i.nio is violated, m ^k+1 _i ( p)=m _i.nio ……(7).

[Step 4] Efficiency calculation Plant model 240 for efficiency calculation (Figure 3)
The point m ^k+1 _i representing the new trial pattern is determined using
Find the efficiency η ^k+1 corresponding to (p). Here, the simulation range as a plant model covers all systems where energy balance is a problem among systems required for normal load operation.

That is, in the turbine system, not only the extraction system but also the seal steam is taken into consideration, and in the boiler system, not only heat loss due to exhaust gas but also heat radiation from the boiler wall is taken into consideration. Furthermore, actual measured values can be used for seawater temperature, atmospheric temperature, and wind speed as ambient conditions.

[Step 5] Constraint condition monitoring Plant model 240 for efficiency calculation (Figure 3)
Determine whether the process state calculated in violates the implicit constraint, and if so,
Point m ^k+1 _i (p) corresponding to trial pattern (k+1)
All information regarding [Point 7 in the above explanation] is invalid. Then, the process returns to the original step 3 and the point corresponding to the new trial pattern is determined.

In this case, in consideration of the causal relationship between the operation and the process state, α _i in the above equation (5) is corrected according to the following equation, and then the process returns to step 3.

α ₂ △ = α ₂ /2 ...(8) Condition (G _GR > G _{GR nax} ) ∪ (G _GR < G _{GR nio} ) α ₃ △ = α ₃ /2 ... (9) Condition (G _CL > G _{CL nax} ) ∪ (G _CL < G _{CL nio} ) ∪ (D _T > D _{T nax} ) ∪ (Y > Y _nax ) [Step 6] Convergence judgment In Step 5, the process state does not violate the implicit constraint condition. If it is determined that this is the case, proceed to step 6. Here, the maximum and minimum efficiencies among the efficiencies corresponding to the point corresponding to the new trial pattern and each point constituting the original simplex are defined as η _nax and η _nio , respectively, and the maximum efficiency point is converged (reached). Whether or not is determined is determined according to the following equation (10).

η _nax −η _nio /η _nax <ε (10) However, in equation (10), ε is the criterion for reaching the maximum efficiency point. If the above equation (10) is satisfied, it can be said that the practical maximum efficiency point has been reached.

When the maximum point is reached, in step 8, the manipulated variable corresponding to η _nax is changed to the optimal manipulated variable m∧ _i (p) [In Fig. 3, since j=7 is the optimal manipulated variable, m ⁷ _i ( p)]
and sends this to the operation command value determining means 250.

If the maximum point has not been reached, the process proceeds to step 7 to form a new simplex, and then returns to step 2 to calculate the center of gravity for the new simplex. Then, the same procedure as described above is repeated in the order of steps 3→4→5→6.

[Step 7] Formation of a new simplex Here, among the points composing the simplex, the points corresponding to the trial pattern showing the lowest efficiency are excluded, and the points corresponding to the new trial pattern are added. A new simplex is formed using the k points, and the process returns to step 2.

For each of the p intermediate load points for searching for the optimal manipulated variable, all the optimal manipulated variables 11, that is, m∧ _i (p), determined by the procedure shown in FIG. Determining means 250 (third
Figure).

In the operation command value determining means 250,
Perform the conversion process to a time function as shown in . That is, the optimal manipulated variable m∧ _i (p) is obtained as a discontinuous manipulated variable as shown by white circles in the figure. Therefore, linear interpolation is performed between each manipulated variable m∧ _i (p-1) and m∧ _i (p) and converted into a continuous manipulated variable as a time function, which is used as a manipulated command value m _i (t).

The operation command values 12 - that is, m ₁ (t), m ₂ (t), m ₃ (t) obtained as described above are supplied from the operation command value determining means 250 to the plant 100, and are supplied to the plant 100. Actual plant control is executed based on this.

Here, the load request value Lr(t) in this example is
The relationship between the load point LS(p) during the optimum operation amount search and the operation timing of the optimum operation amount search means 230 will be explained in more detail.

For the optimum operation amount search load point determining means 220, the load request value Lr(t) given from the load request value generating means 110 is initially given - that is, at the time _t0 shown in FIG. There are cases where the target settling load Lr _S is known and cases where it is unknown.

The known case is a case where the load change rate (dLr/dt) and the target settling load Lr _S are given to the optimum manipulated variable search load point determining means 220 at time t ₀ . On the other hand, the unknown case is a case where only the load change rate is known at time _t0 .

If it is known, as explained with reference to FIG _{. 4, multiple load points L S ( 1} ), L _S ( 2)…
In addition to the method of using .

One of them is a method in which only the target settling load Lr _S is used as the optimum operation amount search load point, and the other one is as follows.
One point on the way from Lr (t ₀ ) to Lr _S is determined as the optimal manipulated variable search load point, the optimal manipulated variable is calculated and output based on this, and then a new optimal manipulated variable is calculated as time passes. This method determines the search load point and executes similar calculations and outputs.

The former occurs when the load fluctuation width [Lr _S −Lr (t ₀ )] is relatively small or when the load change rate is relatively large - in normal control, when the load fluctuation width is small, the load change rate is large. - Valid when set. Moreover, the latter is effective when the load change rate is relatively small.

Furthermore, it is also possible to use the above three methods in combination and use them depending on the rate of change, fluctuation range, amount of operation, etc. of the load request value.

If unknown, assume that the load change rate obtained at time t ₀ continues for the subsequent scheduled time,
The same calculations and outputs as described above are performed using a load point after the scheduled time or a load point in the middle thereof as a load point in the middle of the optimal operation amount search.

In the embodiment described above, it is assumed that the optimum manipulated variable search load point determining means 220 operates after the load request value Lr(t) from the load request value generating means 110 has actually changed. did.

However, before the load request value Lr(t) changes, that is, while the plant is operating at a steady load, an optimal manipulated variable search load point is assumed in response to a predicted load change value prepared in advance, and these points are It is also possible to provide the optimum operation amount search means 230 to execute similar calculations.

That is, the optimal manipulated variable corresponding to the load point assumed as described above is determined by the optimal manipulated variable search means 230.
, obtained by the same method as above,
The actual load request value Lr(t) is determined by the operation command determining means 2.
50, this optimum operation amount can be corrected according to the difference between the predicted load value and the required load value Lr(t), and can be used as the actual operation command value.

This method will be explained in more detail below.

Now, as shown in FIG. 7, in the load point determination means 220 for searching for the optimum manipulated variable, it is assumed that there is a request for a load change from the steady load state Lr (t ₀ ) at time t ₀ , and this is determined by four load changes. Predicted value L _F1 (t),
Expressed as L _F2 (t), L _F3 (t), and L _F4 (t).

These predicted values are based on the load change range [La−Lr
(t ₀ )], [Lb−Lr(t ₀ )] and load change rates βa, βb.

The optimum operation amount search means 230 finds the optimum operation amount for each of the four predicted load change values using the method described above. The optimal manipulated variables obtained here are (a) For L _F1 (t), m∧ _i (1, 1), m∧ _i (1, 2), ... m∧ _i (1, p
),
…m∧ _i (1, l) (b) For L _F2 (t), m∧ _i (2, 1), m∧ _i (2, 2), …m∧ _i (2, p
),
…m∧ _i (2, l) (c) For L _F3 (t), m∧ _i (3, 1), m∧ _i (3, 2), … m∧ _i (3, p
),
...m∧ _i (3, l) (d) For L _F4 (t), m∧ _i (4, 1), m∧ _i (4, 2), ... m∧ _i (4, p
),
...m∧ _i (4, l).

In this way, if the prediction and calculation of the optimal manipulated variable for the predicted load change value is carried out during steady load operation of the plant, when the actual load request value Lr(t) occurs, the actual optimal manipulated variable m ∧ ^* _i (p) can be found quickly and easily.

FIG. 8 shows how to obtain m∧ ^* _i (p) by interpolation, as an example of a method for obtaining the actual optimum manipulated variable m∧ ^* _i (p).

Note that here, the actual load change is the load change width [Lr _S −
Lr (t ₀ )] is expressed by the load change rate β, and is assumed to be between the actual target settling load Lr _S , La and Lb, and the change rate β is between βa and βb. In other words, four load change predicted values having such a relationship are selected.

Under these conditions, the actual optimal manipulated variable m∧ ^* _i
(p) using the interpolation method, first, the optimal operating amounts m∧ _{i (1, p) and m∧ i ( 2} , p ), find the intermediate point m∧ _i (12, p) using equation (11) below.

Next, the optimal operation amounts m∧ _i ( ₃ , p) and m∧ _i ( ₄ ,
p), by the following equation (12), the intermediate point m∧ _i
Find (34, p).

m∧ _i (12, p) = m∧ _i (1, p) +
(∂m／∧ _i /∂β _La (β−βa)……(11) m∧ _i (34, p)=m∧ _i (3, p)+
(∂m/∧ _i /∂β _Lb (β−βa)……(12) Furthermore, the two intermediate points m∧ obtained by equations (11) and (12)
If _{a similar interpolation method is applied to i (} 12, p), m m∧ ^* _i (p) is found.

m∧ _i (p) = m∧ _i (12, p) + Lr _S −La/L
b−La[m∧ _i (34, p)−m∧ _i (12, p)]……(13
) Alternatively, the following method may be used.

(1) Optimal operation amount m∧ _i (1, p) and _{m∧ i ( 3} ,
p), the first intermediate point m∧ _i (13, p) is determined by the following equation (14).

m∧ _i (13, p) = m∧ _i (1, p) + ∂
_m /∧ _{i /} ∂Lr _S )〓 _a (Lr _S −La)……(14) (2) Optimal operation amount m∧ _i (2 , p), and m∧ _i (4,
p), the second intermediate point m∧ _i (24, p) is determined by the following equation (15).

m∧ _i (24, p) = m∧ _i (2, p) + ∂
m/∧ _i /∂Lr _S )〓 _b (Lr _S −La)……(15) (3) Two intermediate points m∧ _i found as above
Applying the interpolation method to (13, p) and m∧ _i (24, p), the actual load request value is obtained from the following equation (16).
Find the optimal manipulated variable m∧ _i (p) for Lr(t).

m∧ ^* _i (p) = m∧ _i (13, p) + β−βa/
βb−βa[m∧ _i (24, p)−m∧ _i (13, p)]……(
16) In the above, (∂m/∧ _i /∂β) _La = m/∧ _i (2, p) − m/∧ _i
(1, p) / βb − βa ... (17) (∂m / ∧ _i / ∂β) _Lb = m / ∧ _i (4, p) − m / ∧ _i
(3, p) / βb − βa ... (18) (∂m / ∧ _i / ∂Lr _S ) 〓 _a = m / ∧ _i (3, p) − m / ∧
_i (1, p)/Lb-La...(19) (∂m/∧ _i /∂Lr _S )〓 _b = m/∧ _i (4, p)-m/∧
_i (2, p)/Lb−La ……(20) Therefore, the right sides of equations (13) and (16) are equivalent, and in the end m∧ ^* _i (p) is expressed as the following equation (21). expressed.

m∧ ^* _i (p) = m∧ _i (1, p) + (β − βa) (Lr _S −L
a)/(βb−βa)(Lb−La)m∧ _i (1, p)+ m∧ _i (4, p)−m∧ _i (2, p)−m∧ _i (3, p
)...(21) The above method can be applied not only to the load increasing direction as shown in Fig. 7, but also to the load decreasing direction.

The optimal operation amount m∧ ^* _i (p) obtained using the above method is
The operation command value determining means 250 (Fig. 3) determines the time function.
The conversion to m _i (t) is as described above with respect to FIG.

When the target settling load is unknown as the load request value for the optimum operation amount search load point determining means 220, the optimum operation amount m∧ ^* _i (p) is determined as follows.

In other words, the predicted load change value is the predicted load change value L _Fa that changes at a rate of change γa up to a certain scheduled load.
(t) and the predicted load change value that changes with the change rate γb
Two L _Fb (t) are prepared in advance, and the optimal manipulated variable searching means 230 finds the optimal manipulated variable for each using the same method as described above. The optimal operation amount obtained here is (a) for L _Fa (t), m∧ _i (a, 1), m∧ _i (a, 2),...m∧ _i (a, p
),
…m∧ _i (a, l) (b) For L _Fb (t), m∧ _i (b, 1), m∧ _i (b, 2), … m∧ _i (b, p
),
...m∧ _i (b, l).

When the actual load request value Lr(t) occurs,
Assuming that the load changes by a predetermined width at the load change rate γ immediately after the load change starts, the following is calculated using the optimal manipulated variables m∧ _i (a, p) and m∧ _i (b, p) determined in advance. The optimum manipulated variable m∧ ^* _i (p) to be actually used can be determined by the interpolation method shown in the equation.

m∧ ^* _i (p) = m∧ _i (a, p) + γ−γa/
γb−γa(m∧ _i (b, p)−m∧ _i (a, p))……(
22) Although the present invention has been described above only when applied to a thermal power plant, the scope of application of the present invention is not necessarily limited to a thermal power plant. That is, it is also applicable to other power plants that are operated with load changes, such as nuclear power plants and hydroelectric power plants.

The present invention is also applicable to processes other than power plants, where efficiency is a criterion for evaluating the quality of operating conditions, and operating conditions such as production volume and number of devices in operation change from time to time.

Further, as explained with reference to FIG. 4, the operation command value determining means 250 in the embodiment of the present invention is configured to convert the optimal operation amount m∧ _i (p) given as a temporally discontinuous function to a continuous function. Operation command value m _i
(t), m∧ _i (p-1) and m∧ _i
(p), but the interpolation does not necessarily have to be linear interpolation and may be non-linear interpolation such as quadratic or cubic interpolation. Furthermore, the present invention can be implemented without any essential difference even by a method in which m∧ _i (p) is used as a representative manipulated variable between time t _p-1 − _{t p} without converting it into a continuous function. be.

Furthermore, in the embodiment of the present invention regarding the predicted load change value, a method of preparing two load change rates in the load increasing direction (or decreasing load direction) has been described.
The basic operating principle remains the same even when a method is used in which one load change rate is prepared in the downward direction.

Furthermore, in the present invention, it is not necessary to always consider both the upward and downward directions of the load to determine the optimal operation amount in advance. It is also possible to limit the predicted load change value to the direction of increase or decrease in load.

The first effect of the present invention is that the efficiency optimization function can be operated not only during steady load operation of the plant but also during load fluctuations by using the plant model built into the control system. Even in power plants where load changes occur, it is possible to significantly improve efficiency compared to conventional systems that only operate under steady load.

That is, there is a strong need for thermal power plants in recent years to be used for load adjustment in power systems, and it can be said that opportunities for steady load operation are rather small. Therefore, the efficiency can be significantly improved compared to the conventional method, which targets efficiency optimization only during steady load operation.
The effects of applying the present invention to thermal power plants are significant.

The second effect of the present invention is that when the target settling load of the load request value for the power generation plant is given to the control system, the current load immediately before the load change starts until the target settling load is reached. The point is that the optimum operation amount that can maximize the transient efficiency can be determined.

A second effect of the present invention is that the optimum operating amount that maximizes the transient efficiency can be determined in advance based on the predicted load change value before the load request value for the power generation plant is given to the control system.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the basic principle of a conventional efficiency optimization control system for a power generation plant, FIG. 2 is a block diagram showing the basic principle of a transient efficiency optimization control system for a power generation plant according to the present invention, and FIG. The figure is a block diagram of an embodiment of the transient efficiency optimization control system for a power plant according to the present invention, and FIG. 4 explains the processing method in the optimum operation amount search load point determining means and operation command value determining means in FIG. 3. Figure 5 shows the processing procedure for searching for the optimal operation amount in the present invention. Figure 6 shows an example of simplex changes in the processing procedure of Figure 5. Figure 7 shows load change prediction. FIG. 8 is a diagram for explaining one method of transient efficiency optimization control based on values, and FIG. 8 is a diagram for explaining how to obtain the optimal manipulated variable by interpolation. 9...Trial operation amount, 10...Plant transient efficiency, 11...Optimum operation amount, 12...Operation command value,
13...Load request value, 14...Optimum operation amount search load point signal, 15...Operation control signal, 100...Power plant, 200...Control system, 230...
...optimum operation amount search means, 240...plant model, 250...operation command value determination means.

Claims (1)

[Claims] 1. A system that includes a dynamic plant model that calculates the transient characteristics of a power generation plant, and predicts the dynamic behavior of the model by inputting a load request value and an operating amount into the plant model, and calculates the transient efficiency of the plant. a transient efficiency calculation means for calculating the transient efficiency; an optimum operation amount search means that repeats until an optimum operation amount is obtained for maximizing the transient efficiency corresponding to the required value; request value,
Optimal manipulated variable search load point determining means for determining at least one optimal manipulated variable search load point and supplying the signal thereof to the transient efficiency calculation means; 1. A transient efficiency optimization control system for a power plant, comprising operation command value determining means for determining an operation command value as a time function based on the determined optimal operation amount and the current value of the operation amount. 2. The change in the load request value is given as a target settling load after the scheduled time, and the optimum operation amount search load point determining means is configured to perform at least one set of load points including the target settling load.
The transient efficiency optimization control system for a power generation plant according to claim 1, characterized in that the control system determines a load point for searching for an optimum manipulated variable. 3. The change in the load request value is given as a load change rate, and the optimum manipulated variable search load point determining means then calculates at least one optimal manipulated variable assuming that the load request value changes at the same rate of change during the scheduled time. The transient efficiency optimization control system for a power generation plant according to claim 1, characterized in that a search load point is determined. 4 Transient efficiency calculation that includes a dynamic plant model that calculates the transient characteristics of the power plant, inputs the load request value and operation amount into the plant model, predicts the dynamic behavior of the model, and calculates the transient efficiency of the plant. means, and a step of outputting a trial manipulated variable to the transient efficiency calculating means to calculate the transient efficiency before outputting the manipulated variable to the plant, to calculate the transient efficiency corresponding to the load request value within the allowable manipulated variable range. means for searching for an optimal manipulated variable that is repeated until an optimal manipulated variable for maximizing efficiency is obtained; a means for predicting a load change during steady load operation of the power plant; and a schedule based on the predicted load change. Optimal manipulated variable search load point determining means for determining at least one load point for optimal manipulated variable search based on the predicted value of load change after a certain period of time, and supplying the signal to the transient efficiency calculation means; When the required load value changes, the change in the required load value and the predicted load change value are compared, and based on the difference between the two, the optimal operation amount determined in advance is corrected in accordance with the predicted load change value. , means for determining the optimal manipulated variable at a load point to search for an optimal manipulated variable corresponding to changes in the actual load request value, and an operation for determining an operational command value as a time function based on the optimal manipulated variable and the current location of the manipulated variable. 1. A transient efficiency optimization control system for a power generation plant, characterized by comprising command value determining means. 5. The predicted load change value is given as a target settling load after a scheduled time, and the optimum operation amount search load point determining means determines at least one optimum operation amount search load point including the target settling load. A transient efficiency optimization control system for a power plant according to claim 4, characterized in that: 6. The predicted load change value is given as a load change rate, and the optimum operation amount search load point determining means calculates at least one optimum operation amount, assuming that the load request value changes at the same rate of change during the scheduled time thereafter. 5. The transient efficiency optimization control system for a power generation plant according to claim 4, wherein a quantity search load point is determined.