JPH0777490B2 - Reactive power compensation method by optimal control of stochastic system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to a voltage fluctuation (flicker) generated from a fluctuating load, particularly in the vicinity of 10 Hz, by using at least a thyristor phase control reactor reactive power compensator (TCR). , Active filter (AF) consisting of a self-excited static power converter
The present invention relates to a compensation method for compensating with reactive power by applying the well-known optimal control theory of a stochastic system using and.

For the optimal control theory of this known stochastic system, refer to "Basic Manual for Automatic Control Handbook" (October 30, 1983).
P.553-557, "7.6 Optimal control of stochastic systems", published by Ohmsha, Japan, First Edition, Japan Society of Instrument and Control Engineers.

[Conventional technology]

Generally, a voltage fluctuation, especially a voltage fluctuation in the vicinity of 10 Hz (hereinafter, the voltage fluctuation is abbreviated as ΔV and a voltage fluctuation in the vicinity of 10 Hz is also abbreviated as ΔV ₁₀ ), is probabilistically generated by a load having a large fluctuation such as an arc furnace or a rolling mill. , This spreads to general consumers of the electric power system.

Therefore, conventionally, as shown in FIG. 3, for example, the output signal ΔQ _{TCR of the TCR 2} removes the disturbance signal from the variable load 1 (hereinafter, also simply referred to as ΔQ _F ) to suppress ΔV, particularly the ΔV ₁₀ signal. A so-called feedforward control method is adopted. In the figure, 3 is a system reactance, 4 is a parallel load, 5 is a visibility coefficient function, and 6 is an adder / subtractor.

[Problems to be solved by the invention]

However, in the method as described above, TCR is used as a reactive power compensator, and this has a delay time from the input to the output, so the disturbance signal ΔQ _F and the output signal ΔQ _{TCR of TCR are}
An error signal is likely to occur between and. In particular, the feed-forward control method has a problem that it has a control characteristic with high differential property and thus becomes sensitive to a noise signal, and an error signal is generated, so that the effect of suppressing ΔV, particularly ΔV ₁₀ , is low.

Therefore, it is an object of the present invention to provide a reactive power compensation method capable of significantly reducing ΔV, particularly ΔV ₁₀ component, as compared with the prior art.

[Means for solving problems]

At least thyristor phase control reactor reactive power compensator (TCR) and active filter (AF) against reactive power fluctuations generated in the grid due to large fluctuation load
In order to compensate reactive power by and, sudden stochastic reactive power fluctuation is used as disturbance signal (ΔQ _F ), output signal of _TCR (Q _TCR ), reactive power signal of parallel load other than variable load (ΔQ _F ).
Q _L ), the reactive power signal due to _AF (ΔQ _AF ), the characteristics of the visibility coefficient curve for evaluating the voltage fluctuation signal (ΔV ₁₀ ) near 10 Hz, and all signal systems are expressed by the state equation, and Δ
Suppression and reactive power operation command value to the TCR of the V ₁₀ (U _TCR) and
Focusing on the reduction of reactive power operation command value (U _AF ) to _AF 2
Using the next-form evaluation function, the optimal control calculation of the stochastic system is performed so that the TCR compensation amount and the AF compensation amount are properly shared, and U _TCR and U _AF are obtained. Make compensation.

[Action]

Based on the optimal control theory of the stochastic system, the reactive power operation command values U _TCR and U _AF for _TCR and _AF are _calculated and _calculated as TCR and
The ΔV, especially the ΔV ₁₀ signal is suppressed while adjusting the AF compensation amount.

Example of Invention

FIG. 1 is a system configuration diagram showing an embodiment of the present invention. This system consists of variable load 1, TCR 2, system reactance 3, parallel load 4, luminosity factor function 5, optimal gain calculator 7 of the stochastic system, active filter 8 and adder / subtractor 9, and the input and output of calculator 7 The relationship can be expressed by the stochastic differential equation shown in the following equation (1).

Next, from the standpoint of digital control, the equation (1) is converted from the continuous system to the discrete system to obtain the equation (2).

Here, x (k): discrete state variable matrix, which is indicated by a mark "*".

From the equation (2), the optimum gain for obtaining the TCR and the active filter optimal manipulated variable is obtained by minimizing the evaluation function J of the equation (3).

Here, E is an expected value Q; a weight matrix R; a weight matrix u (k); a discrete manipulated variable vector, and T is attached to indicate a transposed matrix.

By adjusting the load matrices Q and R in the equation (3), the TCR operation amount and the active filter operation amount that suppress the voltage fluctuation ΔV and ΔV ₁₀ signals are appropriately shared.
According to the optimal control theory of the stochastic system, the following equation (4), which is a Riccati equation, is calculated by the iterative method to obtain equation (5) indicating the optimal gain F.

The optimum operation signal of TCR and active filter using equation (5) is Becomes FIG. 2 shows a block diagram for obtaining the optimum gain F of the stochastic system based on the expressions (4) and (5) and performing the operation of the expression (6).

In FIG. 2, 71 is a feedback gain matrix, 72A-
72G, 73A to 73G are adder / subtractors, 74A, 74B are polarity reversals, 72 is an optimum gain F calculation unit according to equations (4) and (5), and 76 is a manipulated variable calculation unit for control. The feedback gain matrix of 71 is given by the optimum gain obtained by the calculation unit 75 by the equation (5).

The state variable vector x (k) in FIG. 2 is as follows, and is given as shown in FIG. Then, the reactive power operation command values U _TCR , U _AF calculated in FIG. 2 are TCR, AF
Is output to.

x ₁ (k); TCR internal variable x ₂ (k); TCR internal variable x ₃ (k); TCR output signal (ΔQ _TCR ) x ₄ (k); Voltage variable (ΔV) x ₅ (k); Visual sensitivity Internal variable of coefficient x ₆ (k); Luminous coefficient output signal (ΔV ₁₀ ) x ₇ (k); Disturbance signal (ΔQ _F ) x ₈ (k); Active filter output signal (ΔQ _AF ). Although TCR and TCR are used, TCR and fixed capacitor or TCR and capacitor switching reactive power compensator (TSC) can be used instead of TCR.
Further, although the case where the parallel load is provided in the system has been described here, when the parallel load is not provided, the reactive power signal ΔQ _L is ignored.

〔The invention's effect〕

According to the present invention, since the optimal control method of the stochastic system in which U _TCR , U _AF , and ΔV ₁₀ are used as the evaluation functions, the TCR and the active filter have an appropriate capacity sharing, and among the voltage fluctuations ΔV, especially ΔV ₁₀ is conventionally used. The effect that it can be suppressed to a greater extent is obtained.

[Brief description of drawings]

FIG. 1 is a system configuration diagram showing an embodiment of the present invention, and FIG.
FIG. 1 is a block diagram showing a calculation unit of the TCR manipulated variable and the active filter manipulated variable according to the optimal control system of the stochastic system of FIG. 1, and FIG. 3 is a system configuration diagram showing a conventional feedforward control system. Explanation of code 1 …… Variable load, 2 ... TCR, 3 ... System reactance, 4 ... Parallel load, 5 ... Visibility coefficient function, 6,9,72A
〜72G, 73A〜73G …… adder / subtractor, 7 …… optimal control arithmetic unit for stochastic system, 8 …… active filter (A
F), 71 ... Feedback gain matrix, 74A, 74B ... Polarity inverter, 75 ... Optimal gain calculation unit, 76 ... Manipulated amount calculation unit.

Claims (3)

[Claims] Claims: 1. A reactive power compensator (TCR) of a thyristor phase control reactor method and an active filter (AF) are used at least to compensate reactive power fluctuations generated in a grid due to a large fluctuation load, in order to compensate the reactive power. The random reactive power fluctuation is taken as the disturbance signal (ΔQ _F ), the output signal of the _TCR (ΔQ _TCR ), the reactive power signal of the parallel load other than the variable load (ΔQ _L ), the reactive power signal by the _AF (ΔQ _AF ) , And the characteristics of the visibility coefficient curve for evaluating the voltage fluctuation signal (ΔV ₁₀ ) near 10 Hz, all signal systems are expressed by a state equation, and ΔV ₁₀ is suppressed and reactive power operation command value (U _TCR ) and a quadratic form evaluation function focusing on reduction of reactive power operation command value (U _AF ) to _AF
The optimal control calculation of the stochastic system is performed so that the share with the compensation capacity and the AF compensation capacity is appropriate, and U _TCR and U _AF are obtained, and they are given to TCR and AF, respectively, to perform comprehensive reactive power compensation. Reactive Power Compensation Method by Optimal Control of Stochastic Systems. 2. The reactive power compensation method by optimal control of a stochastic system according to claim 1, wherein the TC
A reactive power compensation method by optimal control of a stochastic system characterized by using TCR and a fixed capacitor or TCR and a capacitor switching type reactive power compensator instead of R. 3. In the reactive power compensation method by the optimal control of the stochastic system according to claim 1 or 2, when the parallel load is not provided in the system, the reactive power signal (ΔQ _L ) is set. A reactive power compensation method by optimal control of a stochastic system characterized by ignoring.