JPH0895604A - Optimum operation system of power plant - Google Patents

Abstract

PURPOSE: To set a true optimum point by deciding the distribution of steam so that the largest total output of each turbine is secured for every quantity of supplied steam and also that the sum total of pressure-based supplied steam quantity is obtained at the minimum cost and then detecting a total energy cost minimum operation point. CONSTITUTION: A turbine output maximum point calculation part 1 decides the distribution of steam for every turbine based on the data on each supplied steam quantity that is given from a model of the turbine side so that the total output of each turbine is maximized. A minimization point calculation part 2 decides the steam distribution quantity for each boiler so that the sum total of pressure-based supplied steam quantity given from the boiler side is obtained at the minimum cost for every decided point. Furthermore, a total energy cost minimization point calculation part 3 searches a total every cost minimum operation point based on the steam distribution quantity decided for each boiler. Thus the cost is extremely reduced for a power station.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optimum operation system for a power plant equipped with a steam generating facility and a steam turbine (hereinafter referred to as "BTG plant"), particularly for optimally operating a boiler, a turbine and a generator. .

[0002]

2. Description of the Related Art What is the optimum operation system for a power plant?
In a factory having a regular steam turbine generator, the BTG plant is operated so as to supply the electric power / steam load required by the manufacturing process at a minimum cost.

FIG. 2 is a conceptual diagram of a BTG plant, and FIG. 3 is a diagram showing a system configuration example of the BTG plant. Figure 2
As shown in FIG. 4, steam is generated from Xn boilers B that use oil, gas, coal, recovered black liquor, etc. as fuel F to operate Xm turbines T, and each turbine T drives a generator G. To generate electric power, supply the electric power to the electric power load in the factory, and supply the steam extracted from each turbine T to the steam load in the factory. In this case, when the electric power generated by the generator G cannot cover the electric power load in the factory, the purchased electric power from the electric power company is supplied to the electric power load.

By the way, as a system configuration of such a BTG plant, as shown in FIG. 3, for example, steam generated from the boilers B1 and B2 using fuels having good controllability such as oil and gas is supplied to the low pressure side steam header SL. Supply to the coal,
Boilers B3, B that use fuel with high fuel efficiency such as black liquor
The steam generated from No. 4 is supplied to the high-pressure steam header SH,
The high pressure steam turbines T1 and T2 for driving the generator G are driven by the steam pressure flowing from the high pressure side steam header SH, and the low pressure steam turbine T3 for driving the generator G by the steam pressure flowing from the low pressure side steam header SL. I am trying to drive T4.

Further, the steam extracted from the high-pressure steam turbines T1 and T2 is supplied to the high-pressure load air supply header DH, and the steam extracted from the low-pressure steam turbines T3 and T4 is supplied to the low-pressure load air supply header DL. The high pressure side load air supply header DH and the low pressure side load air supply header DL supply high pressure steam and low pressure steam to the steam loads H and L of the factory.

The RPV 1 is a pressure reducing valve connected between the high pressure side steam header SH and the low pressure side steam header SL, and RP.
V2 is a pressure reducing valve connected between the high pressure side load air supply header DH and the low pressure side load air supply header DL.

In such a BTG plant, the fuel / steam generation amount characteristic of the boiler has a single gentle curve with almost flat peaks as shown in FIG. Therefore,
According to the boiler characteristics shown in FIG. 4, it can be seen that the evaporation amount from the boiler is almost proportional to the fuel amount in the lower limit range.

On the other hand, the steam amount / power generation output characteristic of the turbine is a curve having several depressions because it has non-convexity called a valve loop characteristic (this is called a valve point). In the characteristic diagram submitted by the equipment manufacturer as the characteristic curve of the turbine, the envelope is usually formed by connecting a plurality of peak points, but in reality it is different as shown in FIG. Therefore, according to the turbine characteristics shown in FIG. 5, it can be seen that the power generation output is different between the valve point portion of the steam amount and the recessed portion therebetween in the lower limit range.

By the way, in the optimized operation of the BTG plant, the steam distribution of the boiler and the turbine and the power distribution of the turbine and the power purchase are changed to operate at the point of minimizing the cost. 7 will be described.

Now, assume that two turbines having the same characteristics are operating as shown in FIG. 6, and the steam load at this time is 100 T / H. Here, if the load steam is evenly distributed to two units, it becomes 50 T / H each, but the amount of power generation at this time is 200 × 50 + 200 ×, as can be seen from FIG.
50 = 20,000 kW.

However, assuming that the distribution is 40 T / H on one side and 60 T / H on the other side, the power generation output is 210 ×.
40 + 218 × 60 = 21480 kW, which means that it is possible to generate as much as 7.4% by changing the steam distribution to the turbine.

Therefore, in the optimized operation system of the BTG plant, the BTG plant is mathematically modeled in order to find a cost-minimum one among the distribution combinations of the boilers and turbines capable of satisfying the electric power / steam load,
When the target function is the total energy cost,
By obtaining the distribution amount that minimizes (or maximizes) it as a mathematical solution using a high-speed computer, it is possible to operate a complicated BTG plant that is incomprehensible to humans at a minimum cost at all times.

FIG. 10 is a block diagram showing an example of this type of optimum operation system. In FIG. 10, the optimum operation system of the BTG plant is composed of a plant model PM and a total energy cost minimization point calculation unit 3.

In the plant model PM, the total energy cost minimization point calculation unit 3 is provided with the characteristics of the BTG plant, such as the characteristics of turbines, boilers, pressure reducing valves, operating restrictions, and the power balance and steam balance in each part of the system. .

Further, as the plant state of this plant model PM, the actual load amount (electric power / steam) is given to the total energy cost minimization point calculation unit 3 so that the linear programming method (LP) or the non-linear programming method (NL) is used.
P) is used to find the optimal allocation.

However, the linear programming method and the non-linear programming method which have been widely used in the past each have advantages and disadvantages. In other words, in the linear programming method, the target mathematical model can solve only linear (first-order) problems. Therefore, a plant model that is actually nonlinear must be linearly approximated and applied. Therefore, the accuracy of the solution is inevitably deteriorated due to the approximation, and the cost reduction effect is small.

On the other hand, the non-linear programming method can solve a multi-dimensional problem and therefore can handle a non-linear plant model, but has a drawback that only a local optimum solution can be obtained when the problem to be dealt with is non-convex.

Therefore, when the turbine valve has a non-convex characteristic such as a valve point characteristic, the objective function is multimodal, and thus the nonlinear programming method has a local optimum point and a global range as shown in a conceptual diagram in FIG. With respect to the statistical optimum point, only the optimum point near the initial value of the calculation can be obtained, and the true optimum point (global optimum solution) cannot be obtained.

Further, in order to obtain a solution that is as close to the true optimum point as possible using the nonlinear programming method, the initial value and the search method must be changed according to the problem, but the search logic becomes more complicated as the optimization is aimed, It takes time to tune the calculation parameters.

In the search for the optimum point using the non-linear programming (NPL), it is generally necessary to determine the initial value and the search range. The initial value decides where to start the calculation. The initial value is determined by using i) the current operating state, ii) linearly approximating each characteristic function, and using linear programming (LP). It is often used to use the solution obtained from the above.

Further, in the "Lagrange multiplier method + conjugate gradient method" which is often used as the NLP algorithm, it is necessary to tune the search range according to the problem so that there is one peak (valley) within the search range. is there.

FIG. 9 shows the relationship among the linear solution, the non-linear solution, and the true optimal solution. In FIG. 9, it is assumed that the optimum solution by LP is the “D” point. Further, in NLP, when the initial value is started from the “K” point, the lowest point “A” near the search is obtained as the optimum solution. This is a local optimum because it is larger than the true optimum "T".

"B" is a point which is moved from the local optimum point "A" by a distance L1 to find a better point. If the point "B" is lower than the point "A" as shown in this figure, the optimum point moves to the point "B". Then, similarly to the case of the “A” point, the locally optimum point “T” near the search is obtained as the optimum solution.

However, the point "C" separated by L2 is
Since it is located higher than point "A", there is no optimum point movement. In searching for the optimum point using this NPL, there is no way to correctly determine the search range, and there is no guarantee that the true optimum point "T" will be reached.

Since the nonlinear programming method is based on the premise that the problem to be handled is convex, the number of peaks is one (single peak).
It is not used unless the position of the peak is known. It can be seen that in the case of a problem including a plurality of non-convex characteristics such as the valve characteristic of the turbine, only the local optimum point can be searched.

Therefore, the conventional optimization system operates at the local optimum point, and it can be said that there is room for efficiency improvement until the true optimum point is obtained. As described above, in nonlinear programming, it is difficult to tune calculation parameters such as initial values and how to determine the search range, and users who do not have expertise in optimization calculation can change the system of the target energy plant or plant model corresponding to equipment changes. It was extremely difficult to perform system maintenance such as correction of.

[0027]

As described above, the conventional BT is used.
In the optimum operation system of the G plant, when the optimum distribution is calculated by the total energy cost minimization point calculation unit, only a linear solution with poor accuracy or a non-linear solution that can obtain only a local optimum solution is obtained, and in the case of the non-linear solution method, There is a drawback that it is difficult to tune calculation parameters such as initial values and search range.

In the present invention, the accuracy of the solution is higher than that of the conventional method,
A power plant that does not require tuning of calculation parameters, and that does not have expertise in optimization calculation and can perform system maintenance such as system modification of the target energy plant or modification of the plant model corresponding to equipment changes. The purpose is to provide an optimal operation system.

[0029]

In order to achieve the above object, the present invention constitutes an optimum operation system for a power plant by the following means. The invention corresponding to claim 1 operates a plurality of turbines by steam generated from a plurality of boilers, drives a generator by each turbine to generate electric power, and supplies this electric power to an electric power load, In a power plant that is configured to supply steam extracted from each turbine to a steam load, a plant model created separately for the turbine side and the boiler side,
Turbine output maximum point calculation means for determining the steam distribution of each turbine such that the total output of each turbine becomes maximum for each supply steam amount given from the turbine side model of this plant model, and this turbine output maximum point calculation Fuel cost minimization point calculation means that determines the steam distribution amount of each boiler so that the sum of the supply steam amount by pressure given from the boiler side model for each point obtained by the method occurs at the minimum cost, and this fuel A total energy cost minimization point calculation means for searching for a total energy cost minimum operation point for an electric load and a steam load based on the steam distribution amount of each boiler determined by the cost minimization point calculation means.

The invention corresponding to claim 2 is, in addition to the above configuration, a system graph for changing or modifying the plant model used in the calculation by the turbine output maximum point calculating means and the fuel cost minimizing point calculating means. And a plant model creating means for creating / editing a system model described by a mathematical expression, and generating data necessary for optimal calculation from the model created by this plant model creating means to a turbine side model and a boiler side model of the plant model. And a data generator to provide.

[0031]

In the optimum operation system of the power plant of the invention according to claim 1, the total output of each turbine is maximized for each supply steam amount given from the turbine side model of the plant model. Determine the steam distribution (points) for each turbine, and determine the steam distribution for each boiler so that the sum of the supply steam quantities by pressure given by the boiler side model at each point occurs at a minimum cost. By searching for the minimum operating point for the total energy cost including the power purchase cost for the above load, the true optimum point can be obtained, and a significant cost reduction can be achieved.

In addition, in the optimum operation system of the power plant of the invention according to claim 2, in addition to the above operation, the plant model used for the turbine output maximum point calculation and the fuel cost minimization point calculation is: By creating / editing a system model described by systematic graphics or mathematical expressions and generating data necessary for optimal calculation from this model, the model can be easily changed or modified.

[0033]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of an optimum operation system of a power generation plant (BTG plant) according to the present invention.

In FIG. 1, PM is a plant model created separately for the turbine side and the boiler side. This plant model PM is given turbine characteristics, turbine operation restrictions, steam balance, etc. to the turbine side model,
Boiler characteristics, boiler operation restrictions, steam balance, etc. are given to the model on the boiler side.

1 is a turbine output maximum point calculation unit for determining the steam distribution of each turbine so that the total turbine output is maximized based on the data for each supply steam amount given from the model on the turbine side. The minimum fuel cost that determines the steam distribution amount of each boiler so that the total sum of the supply steam amount by pressure given from the boiler side model for each point obtained by this turbine output maximum point calculation unit 1 is generated at the minimum cost It is a conversion point calculation unit.

Further, 3 is a plant state (electric load amount, steam load amount) including the unit price of power purchase as a plant state, and the steam distribution amount of each boiler determined by the fuel cost minimization point calculation unit 3 is shown. It is a total energy cost minimization point calculation unit that searches for a total energy cost minimum operation point.

On the other hand, 4 is an editor for changing the plant model PM, and this editor 4 changes or modifies the plant model used in the calculation of the turbine output maximum point calculation unit 1 and the fuel cost minimization point calculation unit 1. It is a plant model creating means for creating / editing a system model described by a systematic graph or mathematical expression.
Further, 5 is data for changing the data necessary for the turbine output maximum point calculation unit 1 and the fuel cost minimum operation point calculation unit 2 from the optimized model after being changed or modified by the editor 4 into a format suitable for the calculation method. It is a generator.

Next, the operation of the optimum operation system of the BTG plant configured as described above will be described. The turbine output maximum point calculation unit 1 determines the steam distribution of each turbine so as to maximize the turbine output for each total supply steam amount given from the model on the turbine side. In this case, in a plant in which a plurality of turbines having non-convex valve point characteristics are combined, the total supply steam amount is given, and the steam distribution combination to the turbine that maximizes the power generation amount at that time is obtained.

Here, in the case of obtaining the steam distribution combination to the turbine that maximizes the amount of power generation, the most primitive method for obtaining the global optimum point is to calculate the total amount of steam to be supplied by the step function and use the equation 1
00 steps (to prevent deterioration of calculation accuracy, the number of divisions of one turbine into a step function is sufficiently large compared to the number of valve points <about 3 to 5 for industrial steam turbines><e.g. 100 units>), and change the amount of steam distributed to each turbine for each step to obtain the total power generation output,
Among them, the distribution combination that maximizes the amount of power generation may be selected, but such an all-point check requires a huge number of calculations, which is actually difficult.

Therefore, in this embodiment, DP (Dynamic Programming: Sequential Optimal Method) or the like is used as a means for obtaining the same maximum points as all point trials with a small number of times. The calculation result by this DP is obtained, for example, for every several hundred to several thousand input steam amounts corresponding to the total of the step functions of all turbines between the minimum value and the maximum value that can be taken by the total of all turbines.

Further, in the fuel cost minimum operation point calculation unit 2, for each point of the total turbine input steam amount obtained by the turbine output maximum point calculation unit 1, the supply steam amount for each pressure given from the boiler side model is calculated. Determine the steam distribution for each boiler so that the total sum is generated at the minimum cost. In this case, the fuel unit price does not change in units of time unlike the electric power unit price, so it is given as a fixed value and can be manually changed when necessary.

As for the calculation of the fuel cost minimum operation point calculation unit 2, since the characteristic of the boiler is almost flat as described above, the linear approximation does not affect the accuracy of the solution, so that the calculation speed is high. Apply LP.

On the other hand, the total energy minimization point calculation unit 3 uses the distribution table of the fuel cost minimum / maximum power generation amount for each point of the total turbine input steam amount obtained by the fuel cost minimum operation point calculation unit 2, The total energy cost when the amount of load power is given is calculated by the following formula for each total steam amount.

C = Et (P1-Pg) + Cf C: Total energy cost (yen / H) Et: Purchased electricity unit price (yen / KWH) P1: Load power (KW) Pg: Generated power (KW) Cf: Fuel cost (Circle / H) Next, the total energy cost C is sorted in ascending order, and it is checked whether the total load supply steam amount at each pressure satisfies the total load steam amount. .

By the above calculation, the optimum power / steam distribution points for the power load and the steam load for each pressure can be obtained. Next, when the optimization model is modified to change the system of the target energy plant, add equipment, change equipment specifications, characteristics, etc., the user himself / herself creates a steam system diagram in the editor 4 as if it were an AD, and Input the rating, etc. from the steam system diagram and correct the optimization model.

In this case, the editor 4 can also edit driving constraint conditions, equality constraint conditions, and inequality constraint condition formulas at word processor intervals. Next, when the systematic diagram, device data, and constraint condition formulas created by the editor 4 are input to the data generator 5, the data generator 5 uses this information to generate data necessary for optimization calculation in a form suitable for the solution method. It is automatically generated and the plant model PM is modified.

In this case, the optimum calculation logic is the system configuration,
System maintenance is possible only by changing the model by separating from the data description part such as characteristics and constraints, and by using a method that does not require tuning parameters for the optimal solution.

As described above, in this embodiment, by separating the steam distribution calculation means for the turbine having the non-convex characteristic from the steam generation calculation means for the boiler having the convex characteristic and easily linearly approximated, the turbine Since it is possible to apply the sequential calculation method (DP) that obtains a global optimum solution, which is similar to the all-point search, which is much less than the number of calculations required for the all-point search in calculating the steam distribution of It is possible to improve the accuracy of the above, and a high-speed linear programming (LP) can be used for distribution of steam generated by the boiler.

Therefore, by adopting such a two-step process, a global optimum point including a non-convex turbine valve characteristic can be correctly captured, and a system that does not require a unique search logic or parameter adjustment can be obtained. Therefore, the optimum calculation using the model modified by the user becomes possible.

[0050]

As described above, according to the present invention, since the true optimum point can be obtained, the accuracy of the solution can be made higher than that of the conventional method, and the tuning of the calculation parameter is unnecessary. It is possible to provide an optimum operation system of a power plant in which a user who does not have specialized knowledge of optimization calculation can perform system maintenance such as system modification of a target energy plant or modification of a plant model corresponding to equipment change.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of an optimum operation system for a power plant according to the present invention.

FIG. 2 is a conceptual diagram showing a BTG plant.

FIG. 3 is a diagram showing a system configuration example of a BTG plant.

FIG. 4 is a characteristic curve diagram showing a relationship between fuel and steam generation amount of a boiler used in a BTG plant.

FIG. 5 is a characteristic curve diagram showing the relationship between the evaporation amount of a turbine used in a BTG plant and the power generation output.

FIG. 6 is a diagram showing an operation example of two turbines having the same characteristics.

FIG. 7 is a curve diagram showing a relationship between a main steam flow rate and a generated output by the operation of the turbine.

FIG. 8 is a diagram showing the concept of a local optimal solution and a global optimal solution for a non-convex valve point characteristic of a turbine.

FIG. 9 is a diagram showing the relationship among a linear solution, a non-linear solution, and a true optimal solution for non-convex valve-point characteristics of a turbine.

FIG. 10 is a block diagram showing a configuration example of a conventional optimum operation system of a power plant.

[Explanation of symbols]

MP ... Plant model, 1 ... Turbine output maximization point calculation unit, 2 ... Fuel cost minimization point calculation unit, 3 ... Total energy cost minimization point calculation unit, 4 ... Editor, 5 ... Data generator .

Claims (2)

[Claims] 1. A plurality of turbines are driven by steam generated from a plurality of boilers, each turbine drives a generator to generate electric power, and this electric power is supplied to an electric power load and extracted from each turbine. In a power plant that supplies the generated steam to the steam load, the plant model created separately for the turbine side and the boiler side, and each turbine for each supply steam amount given from the turbine side model of this plant model The maximum turbine output point calculation means that determines the steam distribution of each turbine so that the total output of the turbines is maximized, and for each point obtained by this turbine output maximum point calculation means, Fuel cost minimization that determines the steam distribution of each boiler so that the total sum of the supplied steam is generated at the minimum cost Int calculation means and total energy cost minimum point calculation for total energy cost minimum operation point for electric load and steam load based on steam distribution amount of each boiler determined by this fuel cost minimization point calculation means An optimal operation system for a power plant, comprising: 2. A system model created or edited by a system graph or a mathematical expression for changing or modifying a plant model used in the calculation of the turbine output maximum point calculating unit and the fuel cost minimizing point calculating unit. A plant model creating means and a data generator for creating data required for optimum calculation from the model created by the plant model creating means and providing the data to the turbine side model and the boiler side model of the plant model are provided. An optimal operation system for a power plant according to Item 1.