JP2005229657A - Method for determining consignment electric energy - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To determine a consignment electric energy pattern for maximizing the consignment electric energy under an optimal operating state of a plant satisfying a principle of same time same amount. <P>SOLUTION: The method for determining a consignment electric energy comprises an optimal operation means 20 for solving an optimization problem having the start/stop state and the fuel injection quantity of a constitutive apparatus as state variables for an objective function including minimization in the operation cost of schedule period and the amount of gas discharge; and a simulator 30 for calculating the amount of fuel used, the amount of gas discharge, and the like, on the basis of the start/stop state, or the like, of each apparatus and outputting the calculation results. The simulator 30 delivers a consignable electric energy to the optimal operation means 20 where one of a plurality of patterns given beforehand closest to the onsignable electric energy is determined as a scheduled value of consignment electric energy. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a consignment power amount determination method for determining an optimum consignment power amount pattern when surplus power is consigned in a cogeneration plant or the like.

  Conventionally, as a method of determining consignment power consumption, the plant load power fluctuation pattern is recorded in advance, and the load power of the current day is based on the previous day's load power fluctuation pattern or the past monthly load power fluctuation pattern. There is a method (first prior art) that assumes a quantity variation pattern and determines a consignment power amount pattern of the day based on this.

On the other hand, in the consignment of power, the total value of power demand of consumers and the total value of power generation are matched within a certain unit time (30 minutes, 1 hour, etc.) according to the so-called principle of the same amount. There is a need.
For this reason, when consigning power from a power station to a consumer, even when there is a large fluctuation in power demand, a simultaneous power control method for consigned power that adjusts the generator output command value so as to minimize the supply-demand balance error. (Second prior art) is described in Patent Document 1 below.
Similarly, for the purpose of observing the principle of the same amount at the same time, the supply and demand adjustment system of a specific scale electric power company receives the total load status of each consumer and the total power generation status of each generator via communication means. An electric power supply and demand method (third prior art) in which both the actual value and the planned value of these data are displayed on the screen to generate an accurate power generation amount control command is described in Patent Document 2 below. ing.

JP 2003-87971 A (Claim 1, [0012] to [0018], FIG. 2, FIG. 3, etc.) JP-A-2002-345154 (Claims 1 to 7, [0033] to [0044], FIGS. 10 to 12, etc.)

In the first prior art described above, when the past load power amount pattern used as the reference pattern is unique data based on various causes such as abnormal weather, the load power amount pattern of the day assumed based on this is also obtained. As a result, the reliability becomes low, and the actual load power amount pattern on the current day is greatly different. As a result, problems such as deviating from the simultaneous amount principle and the inability to effectively use the power generation capacity of the power generation facility arise.
In addition, when considering the overall efficiency of the plant, the optimum operating state of the plant is determined in consideration of not only the load power amount but also the heat load, etc., and the optimum entrusted power amount pattern is determined based on the operating state It is also desirable to do.

  Furthermore, in the second and third prior arts described above, it is possible to satisfy the demand of the same amount by adjusting the output of the generator and controlling the balance between supply and demand, but while maximizing the amount of power for consignment. It does not provide an optimal operation means for minimizing plant operation costs.

  Therefore, an object of the present invention is to determine the optimum operation state of the plant that minimizes the operation cost and gas emission amount during the operation plan period by the interaction between the optimum operation means and the steady plant simulator, and the generated power in the operation state. It is an object of the present invention to provide a consignment power amount determination method that enables selection of an optimum consignment power amount pattern that satisfies the simultaneous equal amount principle.

In order to solve the above problems, the invention described in claim 1 is a load predicting means for predicting various loads such as a power load, a thermal load, and an air load of a plant in an operation plan period;
Minimize plant operating costs and gas emissions during the operation planning period, satisfying the same supply and demand balance for various load forecast values and operating conditions of each plant component and maximizing consignment power consumption An optimal operation means for determining an optimal operation method of the plant by solving an optimization problem with the start / stop state of each plant component device and the fuel injection amount to each plant component device as a state variable for an objective function including optimization,
When the optimal operation means gives the start / stop state of each plant component device, the amount of fuel injected into each plant component device, the planned power transmission value, and various load prediction values, the received power amount and fuel consumption amount of the plant A steady plant simulator for sequentially calculating the gas discharge amount, the steady input / output state amount of each plant component device, and the like and outputting it to the optimum operation means,
The steady plant simulator calculates a consignable power amount from a difference between a power generation amount generated by a power generation facility and a load power amount predicted value based on the predicted load value, and sends the consignable power amount to the optimum operation means.
The optimum operation means determines a pattern closest to the consignable power amount among a plurality of consignment power amount patterns given in advance as a consignment power amount plan value.

  According to the present invention, the amount of consigned power is maximized while satisfying the supply and demand balance of the same amount within the power generation amount in the optimum operation state of the plant that minimizes the operation cost and gas emission amount in the planning period. And surplus power can be consigned to the customer based on the consigned power pattern.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing a schematic configuration of this embodiment. First, in this embodiment, in a cogeneration plant serving as a power consignment source, the load prediction unit 10 in FIG. Various loads such as a power load, a heat load, and an air load are predicted, and a load power amount obtained by combining them is predicted. For this prediction, for example, a plant load prediction method described in JP-A-2003-84805 can be used.

Here, FIG. 3 is a configuration diagram showing an example of a cogeneration plant. This plant includes a gas turbine 41, an exhaust gas boiler 42, a steam turbine 43, a compressor 44, a boiler 45, an absorption refrigerator 46, and a gas fired refrigerator 47. This is an energy plant that is composed of a heat storage layer 48 and obtains an air load, an electric power load, a steam load, and a heat (air conditioning) load. In this configuration, the load power amount of the entire plant can be predicted by predicting the various loads, and the difference between the generated power amount of the gas turbine 41 calculated by the steady plant simulator 30 described later and the load power amount is The amount of power that can be entrusted.
3 is merely an example, and the plant to which the present invention is applied is not limited to the example of FIG. 3 at all.

  FIG. 2 is a configuration diagram of a prediction model for predicting the thermal load of a plant, for example. This prediction model is a hierarchical neural network including a fully coupled portion 11 coupled to all input layer elements and a loosely coupled portion 12 coupled to some input layer elements.

Input factors that are input to the prediction model during learning include factors related to the output heat load, for example, weather conditions such as hourly temperature, maximum temperature, minimum temperature, minimum humidity, weather, and amount of solar radiation on the target date Factors, and whether the day of the week or weekday / holiday (Saturday is included in either weekday / holiday), the presence / absence and type of an event, the operating state of the plant, the operation pattern of the plant components, etc. Use all or some of the factors that determine it roughly. Actual values are used for these input data.
In addition, the actual value of the thermal load is used for the output, and the input factor and the output are given to the neural network for learning to construct a prediction model.

In the example of FIG. 2, the input factors given to the input layer are the maximum temperature, the minimum humidity, and the day of the week, and the heat load of the next day is predicted and output.
Further, the intermediate layer is coupled to only the input layer element to which the maximum temperature is input and outputs the temperature component, and the element is coupled to only the input layer element to which the minimum humidity is input and outputs the humidity component, It consists of an element that outputs only the day component by coupling only to the input layer element to which the day of the week is input, and an element that outputs the interaction component by coupling to all the input layer elements. Here, the output layer element is single.

When the above prediction model is used and, for example, the prediction calculation of the heat load for the next day is executed as the planning period, the forecast value for the next day is used as input data regarding weather conditions such as the maximum temperature and minimum humidity as input factors. Moreover, what is necessary is just to use those plan values, when inputting the operation state of a plant, the operation pattern of a plant component apparatus, etc.
It should be noted that power load, steam load, etc. during the planning period are also predicted by the load prediction means 10 using the same prediction model, and these load prediction values are input to the optimum operation means 20 in FIG.

  Next, the optimum operation means 20 and the steady plant simulator 30 are both realized by computer hardware and software. For example, the optimum operation method and the steady plant of the plant described in JP-A-2003-84805 are disclosed. It can be realized by a simulator.

First, when various load prediction values are obtained, the optimum operation means 20 satisfies the simultaneous supply and demand balance of the same amount with respect to the load prediction values for each load type, and the operation conditions of each device (such as the priority order of starting and stopping). As a constraint condition, for example, the operation cost of the next day plant (including fuel cost, power received (purchased) power charge, power consignment power charge) and gas emissions (eg CO ₂ emissions) The optimization problem with the start / stop state of each plant component and the fuel injection amount as the state variables is solved based on the objective function that minimizes the amount of deviation between supply and demand balance of various loads. Thereby, the optimal operation method of the plant which uses the start stop state of each plant component apparatus and the fuel injection amount as an operation plan value is determined.

The above optimization problem can be solved using optimization techniques such as Particle Swarm Optimization (PSO), genetic algorithm (GA), and tab search (TS).
Further, the consignment power charge that is a part of the operation cost is calculated based on a planned consignment power amount plan value to be described later.

The start / stop state and fuel injection amount of each device are sent to the steady plant simulator 30 as operation plan values, and the predicted load value by the load prediction means 10 is also sent to the steady plant simulator 30.
Further, the amount of power generated by the power generation facility calculated by the steady plant simulator 30 (the amount of power generated by the rating of the power generation facility, or the amount of power generated by the power generation facility in the optimum operation state of the plant) and the load prediction means 10 are used for prediction. The load electric energy prediction value based on various load prediction values is compared, and the difference between the two is sent to the optimum operation means 20 as the entrustable electric energy.

Further, the steady plant simulator 30 simulates a steady input / output state (plant dynamic characteristic) of each plant constituent device based on the operation plan value, and calculates the fuel usage amount and gas discharge calculated based on the state amount. Output quantity.
The steady plant simulator 30 also outputs data such as the amount of received (purchased) power in order to calculate the objective function (minimization of operation costs) in the optimum operation means 20.

  The optimum operation means 20 selects a pattern closest to the consignable power amount from among the consignment power amount patterns given in advance by a contract with the electric power company, and uses the pattern as a consignment power amount plan value as a steady plant simulator. Output to 30. This method of determining the planned power transmission amount is the gist of the present invention, and the determination method will be described later.

The method for determining the planned power transmission amount described above will be described below.
FIG. 4 is a diagram for explaining this determination method. In the figure, the part indicated by a solid line is a consignment power amount pattern given in advance by a contract with an electric power company, and the optimum operation means 20 is input from the steady plant simulator 30 from a plurality of consignment power amount patterns. The consignment power amount pattern closest to the consignable power amount (the maximum consignable power amount indicated by a broken line in the figure) is selected.
However, if the consignable power amount is smaller than the consigned power amount pattern based on the contract with the power company, the shortage is calculated, and the pattern with the smallest total value and the smallest overall deviation is selected. I decided to.

That is, if the shortage P _di is expressed by Equation 2, the shortage total value E ₂ expressed by Equation 3 is the smallest, and the total deviation E ₁ expressed by Equation ₁ is the smallest. Such a consignment power amount pattern is selected and determined as a consignment power amount plan value.

In the above equations,
i: Index indicating the control time in the consignment period (i = 1 to T, where T indicates the final control time)
P _f : Maximum consignable electric energy P _pn : Consigned electric energy pattern based on a contract with an electric power company.

  When the final planned power transmission amount is determined in this way, this planned value (that is, the power transfer amount pattern) is sent to the steady plant simulator 30 to calculate the plant dynamics, and again the power that can be transferred. And the amount of fuel used, the amount of gas discharged, etc. are calculated and input to the optimum operation means 20. The optimum operation means 20 performs the calculation of the start / stop state of the plant component equipment, which is the state variable, the calculation of the consignment power amount plan value, and the like by the same processing as described above, and thereafter repeats the same processing. Finally, the state variables and objective function values obtained by the optimum operation means 20 are output, and the optimum operation state of the plant is determined.

  As a modification of the present embodiment, the operation cost when power is consigned as described above is compared with the operation cost when power is not consigned, and determination is made as to whether or not consignment is performed according to the result. Means to do this may be added.

1 is a schematic configuration diagram of an embodiment of the present invention. It is a block diagram of the prediction model (hierarchical neural network) to which the embodiment of the present invention is applied. It is a block diagram which shows an example of an energy plant. It is a figure explaining the determination method of the final consignment electric energy pattern in embodiment of this invention.

Explanation of symbols

10: Load prediction means 11: Fully coupled part 12: Loosely coupled part 20: Optimal operation means 30: Steady plant simulator 41: Gas turbine 42: Exhaust gas boiler 43: Steam turbine 44: Compressor 45: Boiler 46: Absorption refrigeration machine 47 : Gas fired refrigerator 48: Thermal storage layer

Claims (1)

Load prediction means for predicting various loads such as power load, heat load, air load, etc. of the plant during the operation plan period;
Minimize plant operating costs and gas emissions during the operation planning period, satisfying the same supply and demand balance for various load forecast values and operating conditions of each plant component and maximizing consignment power consumption An optimal operation means for determining an optimal operation method of the plant by solving an optimization problem with the start / stop state of each plant component device and the fuel injection amount to each plant component device as a state variable for an objective function including optimization,
When the optimal operation means gives the start / stop state of each plant component device, the amount of fuel injected into each plant component device, the planned power transmission value, and various load prediction values, the received power amount and fuel consumption amount of the plant A steady plant simulator for sequentially calculating the gas discharge amount, the steady input / output state amount of each plant component device, and the like and outputting it to the optimum operation means,
The steady plant simulator calculates a consignable power amount from a difference between a power generation amount generated by a power generation facility and a load power amount predicted value based on the predicted load value, and sends the consignable power amount to the optimum operation means.
The optimum operation means determines a pattern closest to the consignable power amount among a plurality of consignment power amount patterns given in advance as a consignment power amount plan value.

Images