LOAD DISTRIBUTION AMONG THERMAL POWER PLANT UNITS USING OPTIMIZERS OF ANALOG COMPUTERS

A. A. TERESHCHENKO, A. I. SHUTKO

(Presented by the scientific-technical seminar of the Scientific Research Institute of Analog Computers)

Optimal distribution of thermal and electrical loads among different types of units of a thermal power plant (TPP) provides a certain saving of fuel.

Mathematically, the problem reduces to minimizing a function of a large number of independent variables, i.e., the total fuel consumption of the TPP,

\[ \Phi = B_1 + B_2 + \cdots + B_n = (Q_1, \ldots, Q_n; P_1, \ldots, P_i; D_{n1}, \ldots, D_{ni}; D_{\tau 1}, \ldots, D_{\tau i}) = \min, \]

where

\[ \begin{aligned} B &\text{— fuel burned in the boilers;}\\ Q &\text{— thermal load of the boilers;}\\ P &\text{— electrical load;}\\ D_n,\; D_\tau &\text{— thermal load of the turbines.} \end{aligned} \]

In implementing dependence (1), the main difficulty is the representation of the consumption characteristics of turbines with steam extractions. Thus, for example, in [1] a turbine with steam extractions is represented by an equivalent scheme in the form of combinations of back-pressure and condensing turbines—

Fig. 1

Linearization of turbine characteristics and failure to take into account actual operating conditions reduce the accuracy of this method.

In the literature [5] an attempt is made to avoid the shortcomings noted in [1]. Here the information necessary for optimizing the operating conditions is determined by a special computing device during the current operation of the thermal power plant [5].

In [3] the basic principles of optimal operating conditions of a thermal power plant are developed even when only approximate factory characteristics of the turbines are available.

In [2] a method is proposed that takes into account the technological features of a thermal power plant and operating conditions close to real ones. The method permits the use of both analog and digital computers.

Fig. 2

Works [6, 7], using the method of dynamic programming, are devoted to determining the global extremum of the objective function (1).

On the basis of an analysis of experience in the use of specialized computers and general-purpose digital computers, [4] makes an attempt to create a unified integrated system for regulating the power of a thermal power plant.

For a thermal power plant whose simplified thermal diagram is shown in Fig. 1, we used an analog computer (AVM) EMU-10, equipped with an automatic optimizer, to calculate optimal operating conditions. For this purpose, the thermal diagram (Fig. 1) is described by equations of the heat and electrical balances [1, 2], from which the electronic model shown in Fig. 2 was constructed. The consumption characteristics \(Q(P, Q_T, Q_n)\) of turbines with steam extraction are represented as the sum of the characteristics of parallel-connected high-, medium-, and low-pressure sections. These characteristics were obtained from operating-condition diagrams by the method of experimental design. In the electronic model the above characteristics are implemented on functional converters.

The essence of the method considered for calculating optimal operating conditions of a thermal power plant is as follows: from the components of the consumption characteristics expressed with respect to \(P\), \(Q_{\mathrm{T}}\), and \(Q_n\), values are obtained that are equivalent to the current total load on the thermal power plant \(\sum P\), \(\sum Q_{\mathrm{T}}\), \(\sum Q_n\). The quantity equivalent to the total load of the thermal power plant, expressed through the generalized parameter \(\sum P\), \(\sum Q_{\mathrm{T}}\), \(\sum Q_n\), is compared in absolute value, in a comparison element (CE), with the specified value. The signal from the comparison element is the control signal for the optimizer, which makes it possible to select the optimal operating conditions of the units. The operating condition will be the most advantageous if the condition

\[ \left| Q_{\mathrm{T}k}(\sum P, \sum Q_{\mathrm{T}}, \sum Q_n) \right| - \left| Q_{\mathrm{zad}}(\sum P, \sum Q_{\mathrm{T}}, \sum Q_n) \right| \to \min . \tag{2} \]

is satisfied.

An analysis of the literature on methods for distributing thermal power plant loads shows that the most promising is the method [2], which makes it possible to use both digital computers and analog computers.

In the present work, the method of direct optimization has been used, but analog computing equipment has been applied for its implementation. This made it possible, comparatively simply and with satisfactory accuracy, to provide recommendations for controlling the operating conditions of thermal power plants. In addition, the proposed optimization method can be used not only for issuing recommendations on operating-condition control, but also for controlling thermal power plant operating conditions in real time.

REFERENCES

1. V. M. Gornshtein. Method for calculating the most advantageous distribution of loads among units of a thermal power station. Electric Stations, 1962, No. 8.

2. K. Yu. Folkman. The most advantageous distribution of electrical and heat loads among turbogenerators as a basis for automatic optimization of the operating conditions of a thermal power station. Automation of Power Engineering, issue 1, 1964.

3. D. S. Robman. Method for determining the relative increments of heat consumption by turbine units of a thermal power plant. Problems of General Power Engineering and the Unified Power System, issue 1, 1965.

4. L. P. Fotin. On certain properties of the computing parts of complex systems for regulating the capacities of thermal power plants with a common steam header. Thermal Power Engineering, 1970, No. 8.

5. Complex Automation of Power Stations. Materials for the republican scientific and technical conference. Minsk, “Polymya,” 1966.

6. S. V. Shakhverdyan, D. M. Babayan. Application of three-dimensional dynamic programming to optimization of the operating conditions of a thermal power plant. Thermal Power Engineering, No. 2, 1969.

7. D. M. Babayan. Algorithm for the optimal distribution of loads among thermal power plant units. Thermal Power Engineering, 1970, No. 10.