RU2776423C1 - Method for controlling a filter compensating apparatus under non-steady non-linear loads and apparatus for implementation thereof - Google Patents

Abstract

FIELD: electrical engineering.

SUBSTANCE: area of application: in the field of electrical engineering and electric power engineering. Filter compensating apparatus (FCA) comprises a single-phase bridge inverter, a storage capacitor, an L-shaped filter including a throttle valve and a capacitor, a digital controller, a pulse-width modulator, analogue-to-digital converters, a network voltage sensor, a line current sensor, a load current sensor, and an FCA output current sensor. The control system of the apparatus includes an additional active power prediction unit which connects by the inputs to the output ports of the main digital master controller and by the outputs to the input ports of the main digital master controller.

EFFECT: increase in the efficiency of transmission and consumption and improvement of the quality of electrical power.

1 cl, 2 dwg

Description

The field of technology to which the invention belongs

The invention relates to electrical engineering and electric power industry, in particular to methods and devices for compensating inactive power components, and can be used in power supply systems for electrical and energy circuits with current and voltage distorting non-stationary non-linear loads in order to increase the efficiency of transmission, consumption and improve the quality of electricity.

State of the art

There is a known method for controlling the air conditioner of the mains (patent RU No. 2408122 C1), which consists in measuring the input voltage and current of a single-phase mains, the voltage signal is multiplied by the control signal and compared with the signal of the mains current sensor. The resulting difference signal is used to control the pulse-width modulator of the power channel of the air conditioner.

The disadvantages of this method is the inertia due to the operations of rectification and integration of the load current sensor signal, incorrect formation of the control signal for cases of non-stationary and rapidly changing loads.

The closest technical solution, chosen as a prototype of the method, is a method for generating compensation current into the supply network (patent RU No. 2183897), which consists in measuring the instantaneous values of the consumer current and the instantaneous values of the supply network voltage, determining their derivatives, the active power of the consumer and the effective value of the supply voltage. Then, the products of the reactor inductance and the derivative of the mains voltage, preliminarily multiplied by the quotient of dividing the active power of the consumer by the square of the effective voltage of the supply network, and by the derivative of the consumer current are calculated. The difference between the obtained values is found, and the instantaneous value of the mains voltage is added to the result. The voltage determined in this way is applied to the reactor connected to the supply network, thereby generating the required compensation current into the supply network.

The disadvantage of this method is the impossibility of compensating for inactive power components in non-stationary modes of operation of electrical loads due to the strict requirement that the active power of the load remain unchanged over the observation interval (mains voltage period).

A device is known using a single-phase active electric filter (patent RU No. 2458381 C2), containing a system for converting input information connected to the input of an error signal generation system synchronized with the network. The output of the grid-synchronized error signaling system is connected to the inverter. The output of the inverter is connected to the input of a pulse-width modulator (PWM). The input of the buffer stage is connected to the PWM output, the outputs of the buffer stage are connected to the inputs of the switching power amplifier, and the input circuit of the LC filter is connected to the output of the switching power amplifier. The outputs of the LC filter are connected to the mains phase.

The disadvantages of this single-phase active electric filter are: the need to change the operating mode of the filter when changing the spectral composition of voltages and currents in the supply network, the deterioration of the filtering quality under non-stationary loads.

The closest in technical essence to the proposed one is a single-phase active filter (patent RU No. 2670961), selected as a prototype device, containing an inverter, a storage capacitor, a ripple filter, a load current measuring transducer, a filter current measuring transducer, a filter voltage measuring transducer, a measuring transducer storage capacitor voltage, control system controller. The control system controller includes a storage capacitor voltage regulator, a reference current signal generator, a distortion current signal adder-computer, an adder, an error current signal calculator, a current regulator, and a pulse-width modulator.

The disadvantage of this device is limited functionality: the ability to work only with a dynamically changing non-linear load, but not for cases of non-stationary operating modes.

Disclosure of invention

The objective of the invention is to introduce into the method (algorithm) for controlling the device the operation of predicting the value of the active power of the load in the upcoming period of the supply voltage based on the results of monitoring the change in the active power of the load over previous periods and statistical processing of the data obtained.

The technical result of the claimed invention is to increase the efficiency and electromagnetic compatibility of the conversion of electrical energy when the shape of the current of the mains under non-stationary loads coincides with the shape of the mains voltage, provided that the voltage is sinusoidal.

The technical result is achieved by using a method for controlling a filter-compensating device under non-stationary non-linear loads, which consists in the fact that during a given number of periods of the supply voltage, the instantaneous values of the voltage and current of the load of a single-phase electrical network are measured, according to which, for each observation period, the active power is determined and stored. by the fact that, based on the obtained measurement results, the predicted value of active power is calculated in the subsequent period (control interval) of the supply voltage, while the predicted value of active power is used to determine the required instantaneous value of the current generated in the FKU network, connected in parallel to the load, as the difference between the instantaneous value of the load current and the product of the instantaneous voltage of the supply network, multiplied by the quotient of dividing the previously predicted active power of the consumer by the square of the effective value of the supply voltage networks.

The technical result is achieved using a device according to the method of controlling a filter compensating device for non-stationary non-linear loads, containing a single-phase bridge inverter, a storage capacitor, an L-shaped filter incorporating a choke and a capacitor, a digital controller, a pulse-width modulator, analog-to-digital converters, moreover, an additional active power prediction block is introduced into the device control system, which connects with its inputs to the output ports of the main digital control controller, and connects with its outputs to the input ports of the main digital control controller.

Brief description of the drawings

In FIG. 1 shows a functional block diagram of the FCU with the main elements for implementing the proposed control method.

In FIG. 2 shows the results of mathematical modeling of the predictive algorithm in the MathCAD environment.

Implementation of the invention

The essence of the proposed control method is that during a given number of supply voltage periods, without intervals between intervals, the instantaneous values of the voltage and current of a single-phase electrical network are measured, according to which, for each observation period (observation intervals), active power is determined and stored. Then, the predicted value of active power is calculated for the subsequent period (control interval) of the mains voltage. The predicted value of active power is used to determine the required instantaneous value of the current generated in the PKU network, connected in parallel to the load, as the difference between the instantaneous value of the load current and the product of the instantaneous voltage of the supply network, multiplied by the quotient of dividing the previously predicted active power of the consumer by the square of the effective value mains voltage.

The FKU control method is carried out as follows.

The main ratio of the proposed method for determining the output current of the PKU, provided that it is connected in parallel to the load, is the well-known expression for determining the inactive (passive) component of the consumer current:

where i _p (t) - instantaneous values of the inactive (passive) component of the current of the non-linear load;

i _n (t) - instantaneous values of the current of the non-linear load consumed from the network;

u _c (t) - instantaneous values of the mains voltage;

P is the active power of the load;

U is the operating voltage of the network.

The active power of the load is determined by the results of measurements of the instantaneous values of the mains voltage and load current:

The effective mains voltage is determined by the results of measurements of the instantaneous values of the mains voltage:

The operating voltage of the network will be considered unchanged.

Relation (3) has a significant limitation associated with the mandatory requirement that the active power remain unchanged over the averaging interval (mains voltage period). However, for non-stationary loads, this requirement is not feasible.

In order to eliminate the existing contradiction, the invention proposes to use the predicted value of the active power of the consumer in the coming period based on the results of monitoring the change in active power in previous periods (prehistory) and statistical processing of the initial data, namely:

The determination of the predicted value can be performed using well-known extrapolation methods: splines, wavelets, autocorrelation functions, and others.

For example, a relation is given for predicting the values of a time series based on a correlation analysis. In foreign literature, this relation is known as Burg's recursive algorithm:

where A _k is the original time series,

μ - calculated coefficient,

V _k - a copy of the original time series with the reverse order of the elements,

A _k+1 is the original time series, but with an additional element containing the value of the extrapolation coefficient for the k-th step of the recursion procedure.

The coefficient μ is calculated using autocorrelation functions under the condition of minimizing the sum of squares of the extrapolation error. The recursive formula looks like:

where N is the number of elements of the initial time series,

k - number of extrapolating elements,

ƒ _k - sequence of elements of the original series for the k-th extrapolating element,

b _k - a sequence of elements of the original series, but with the reverse order of the elements for the k-th extrapolating element.

The predicted value is determined by the expression:

P _forecast - predicted value;

m - the number of the last elements of the original time series P _N , according to which the forecast is built;

a _i - extrapolation coefficients;

P _N - initial time series containing N elements.

The results of mathematical modeling of the predictive algorithm in the MathCad environment using the built-in predict(y, m, n) function are shown in Fig. 2. Where y is a vector of real values taken at equal intervals of argument values; m - the number of consecutive values of the vector y, according to which the forecast is built; n is the number of elements of the forecast vector. The example uses the following predict function arguments: y=200, m=50, n=100.

For the example in FIG. Figure 2 shows 200 values of the original time series calculated from the function y(x)=exp(-x/100)*sin(x/10) (curve 1). The predicted values are calculated after 100 values of the original time series (curve 2). To compare the forecast results with the values of the original series, the latter is presented for the second hundred values in the form of curve 3.

The analysis of the results obtained indicates the effectiveness of the predictive algorithm with close proximity to the last values of the original time series. With regard to the proposed method of controlling the PKU, the predictive algorithm under consideration is applicable, since for the proposed control method it is sufficient to have no more than one or two predicted active power values for the upcoming control periods.

The device according to the method of controlling the filter compensating device for non-stationary non-linear loads (Fig. 1) consists of an electric power source 1, a load in the form of a non-linear resistance 2, FKU 3. The FKU receives signals from the mains voltage sensor 4, the line current sensor 5, the load current sensor 6, FKU output current sensor 7. The FKU includes a single-phase bridge inverter 8, a storage capacitor 9, an L-shaped filter that includes a choke 10 and a capacitor 11. The single-phase bridge inverter 8 is controlled using a digital controller 12 through a pulse-width modulator 13. The input signals of the controller are data from the predictor 14 and analog-to-digital converters 15, 16, 17, 18.

The device according to the control method FKU with non-stationary non-linear loads operates as follows. When switching the power switches of a single-phase bridge inverter 8 according to the signals of the control system, the storage capacitor 9 must be connected through the L-shaped filter 10, 11, in parallel to the load 2 so that the FKU current introduced into the network is equal to the inactive (passive) component of the current consumer, calculated by formula (4) with the opposite sign. It is under this condition that compensation will be provided for the inactive (passive) component of the consumer current as part of the network current.

The single-phase bridge inverter 8 is controlled by the signals of the pulse-width modulator 13, which receives control signals from the digital controller 12. The input signals of the digital controller 12 are the output signals from the analog-to-digital converters 15, 16, 17, 18 and the predictor 14. The inputs of the analog-to-digital converters connected to the outputs of the mains voltage sensor 4, the line current sensor 5, the load current sensor 6, the FKU output current sensor 7, respectively. The operation of the entire control system of the device is synchronized with the mains voltage signal.

The predictor 14 works as follows. During a given number of periods of the mains voltage, the predictor remembers the values of the active power of the non-stationary load, calculated by the digital controller 12 according to the formula (2) in a discrete form.

Then the predicted value of the active power of the non-stationary load is calculated according to the formula (5). The obtained value of the predicted value is returned to the digital control controller and is used in the formation of the control law for the output current of the PKU according to the formula (4). Further, the cycle of operation of the proposed device and its control system is repeated, but with an updated value of the active power of the non-stationary load in the last observation period.

The control system can be implemented on a general-purpose single-chip microcontroller (for example, ATMega168) or a specialized digital signal processor (for example, ADSP2104).

The use of the proposed method of active-type PKU allows to significantly improve the quality of electrical energy in electrical networks distributed over a large area with a non-linear non-stationary load by correcting the current shape and, as a result, the voltage of the electrical network.

The essence of the proposed device lies in the fact that the active power prediction block additionally introduced into the control system for the upcoming period of the supply voltage - the predictor - is connected with its inputs and outputs to the output and input ports of the main digital control controller.

From the above studies, we can conclude that the present invention has the following advantages in relation to known developments:

1. The presence in the method of controlling the device of the operation of predicting the value of the active power of the load in the upcoming period of the supply voltage;

2. The use of an additional block for predicting active power, connected to a digital control microcontroller, as part of the FCU.

3. The proposed PKU control method is universal, as it retains its performance not only for cases of non-stationary modes of non-linear loads, but also for cases of stationary modes of linear and non-linear loads.

Claims (1)

A filter-compensating device for non-stationary non-linear loads, containing a single-phase bridge inverter, which includes a storage capacitor, an L-shaped filter connected to the output of a single-phase bridge inverter, which includes a choke and a capacitor through which the device is connected to the mains in parallel with the load, digital controller, the inputs of which are connected to the outputs of analog-to-digital converters and to the input of a pulse-width modulator, the output of which is connected to the input of a single-phase bridge inverter, analog-to-digital converters, the inputs of which are connected to the outputs of the mains voltage sensor, line current sensor, load current sensor, FKU output current sensor, the input of the mains voltage sensor is connected to the line and neutral conductors of the network, the input of the line current sensor is connected to the break of the linear conductor of the network to the point of connection of the device, the input of the load current sensor is connected to the break of the linear conductor and the network after the point of connection of the device, the input of the output current sensor of the FKU is connected to the break of the output conductor of the FKU to the point of connection to the linear conductor, which differs in that an additional block for predicting active power has been introduced into the device control system, which, with its inputs, is connected to additional output ports of the main digital control controller, and connects with its outputs to additional input ports of the main digital control controller.

Images