RU2388037C1 - Method of widening range for stable operation of automatic control systems - Google Patents

Abstract

FIELD: physics; control. ^ SUBSTANCE: invention relates to automation and can be used in automatic control systems (ACS). The method of widening the range for stable operation of automatic control systems is based on correcting proportional and integral components of the control signal, where the corrective action is a continuous function of the current control error. For each control-action term ranging from the selected value of the current error to the zero value, the proportional component falls in accordance with the function, and the integral component rises as the absolute value of the current control error falls. ^ EFFECT: wider range for stable operation of automatic control systems when properties of the control object change and/or when considerable disturbances appear on the load or on the control channel. ^ 1 dwg

Description

The invention relates to the field of automatic control systems (ASR), and more particularly to regulators for stabilizing process parameters.

The vast majority of manufactured and operated regulators in the ASR implement the PID regulation law in their program. Regulators with the PID control law are essentially the only regulators used in practice in automatic process control systems. The widespread use of the PID law is due to the fact that the algorithm for their operation successfully simulates the work of an experienced human operator.

For microprocessor-based controllers, the PID control law is widely used (METAKON - 515 - manufacturer of NPF KontrAvt; UP350 - manufacturer of Etalon Pribor, Chelyabinsk). Such controllers provide such a service function as automatic parameter adjustment at the operator's command, followed by manual adjustment.

Optimization of regulator settings is carried out mainly by the three most common minimum search methods:

- standard deviation (RMS);

- Integrated Modular (IAE);

- time product integral of the module (ITAE).

Regulators configured using these methods have similar transient characteristics.

The regulators optimized by these tuning methods provide specified indicators of the quality of regulation within certain limits of the operating mode of the regulatory object and with disturbances limited in magnitude.

The quality indicators of the transition process deteriorate if the properties of the regulatory object and disturbing influences during operation go beyond these limits.

The disadvantage of these regulators is the immutability of the settings for both large and minor disturbances, which leads to a limited area of stable operation of ASR. The most significant deterioration in the quality of regulation is manifested when changing the parameters of the object of regulation.

When the properties of the control object change and when significant disturbances appear both in the load and in the control channel, the system tuning factors must change, however, the formation of the control action by the regulator remains unchanged in both steady-state and transient modes, which leads to an increase in oscillation, overshoot and regulation time.

In low-inertia objects (relative to the speed of movement of the regulatory body) with large disturbances close to stepwise, the controlled quantity changes quite quickly. A situation arises when the integral component of the control action accumulates faster than the regulatory body manages to work out. As a result, fluctuations or even loss of stability occur in the system.

The Proportional-Integral-Differential Regulator is known (Patent RU 2234116 C1), in which, to expand the stability region, a non-linear block with a dead zone is introduced, connected in parallel with the block forming a proportional part of the control action. Moreover, within the dead band of a nonlinear block, the proportional component of the control action remains small (weakened) (p.3, paragraph 35). This regulator does not eliminate the situation characteristic of low-inertia objects (relative to the speed of movement of the regulatory body), when the integral component of the control action accumulates faster than the regulatory body manages to work out.

A known method of correction of automatic control systems (Patent RU 2234116 C1), in which, when the current control error is reduced to a value equal to or less than the specified one, the signal in the feedback circuit changes, which is formed as the product of the rate of change of the output parameter of the object by a reduced fixed coefficient , which translates the transition process into a monotone. This technique increases the stability of the system within the established limits of the deviation of the stabilized parameter of the object from the set and accelerates the elimination of deviations in case it exceeds the set value, but weakens the proportional and differential components of the control signal, increasing the error of stabilization of the adjustable parameter in the steady state.

There is a method of expanding the boundaries of stable operation of the ACP, implemented when the controller is configured for the minimum regulation time (Industrial ACS and controllers. 2005, No. 5, p. 56).

Transient regulation processes of regulators tuned by the method of minimum regulation time become aperiodic. When disturbed by the "load" it is characterized by greater stability of the ASR. The duration of a single half-wave (when perturbed by the “load”) is significantly longer than the duration of the first half-wave with other tuning methods. Aperiodic transients are characterized by an increase in the error of stabilization of the controlled parameter in the steady state.

There is a method of expanding the boundaries of stable operation of the ASR by “weakening” the settings of the PI controller (Varlamov IG “In what cases is the“ weakening ”of the controller justified?” Industrial ACS and controllers. 2005, No. 9).

Reception of "weakening" of the controller settings allows you to expand the boundaries of the stable operation of the ACP, but at the same time the main indicators of the quality of regulation are deteriorating, the range of deviation of the adjustable parameter in the steady state is expanding.

The technical result to which the invention is directed is to expand the boundaries of the stable operation of the ASR and preserve acceptable indicators of the quality of regulation of the automatic system when changing the characteristics of the object, as well as when increased disturbing influences on the stabilization system of an adjustable value both during transient processes and in steady state mode.

To achieve a technical result, it is not a general “weakening” of the controller settings (reduction of the coefficients of the components of the PID control law), but correction of the proportional and integral components of the control signal as a function of the current control error in a selected range of values from a given value to zero.

A distinctive feature of the proposed method from the closest known above method is not a general weakening of the controller settings, but a correction of the proportional and integral components of the control signal depending on the current control error in the selected range of values of the current control error while keeping the controller coefficients unchanged.

To implement this method, two blocks of proportional and integral components correction are introduced into the structure of the controller between the comparison unit and each of the blocks forming the proportional and integral components of the control signal. In one, the operation of raising the value of the current control error to a power equal to or less than one is performed, and in the second to a power equal to or more than one.

The proposed method is characterized in that the control signal is generated by converting the signal of the current control error into a nonlinear function in front of each of the blocks forming the proportional and integral components of the controller control signal. Conversion of the current control error is performed in such a way that in the selected range of values from the set value to zero, it is weakened in front of the proportional unit and amplified in front of the integration unit. At the same time, the coefficient of adjustment of the controller during the operation of the controller after the settings are made remain unchanged.

The proposed method is illustrated in the block diagram of the ASR that implements the PI-regulation law with correction of the current regulation error by the proposed method. The system contains a comparison element 1, the inputs of which are connected to the stabilizer parameter setter 2 and to the output of the control object 3. The output of the comparison element 1 is connected to the input of the correction block 4 of the current channel control error, which forms a proportional component of the control signal, and connected to the input of the correction block 5 of the current control errors of the channel forming the integral component of the control signal.

The output from block 4 is connected to the input of the multiplication block 6, which forms the proportional component of the control signal, and the output from block 5 is connected to the input of the integrator 7, which forms the integral component of the control signal.

The outputs from blocks 6 and 7 are connected to the inputs of the adder 8. The output from the adder 8 is connected to the input of the control object 3.

The automatic system of the proposed method works as follows.

The input from the comparison element 1 receives a signal from the setter 2 of the stabilized parameter and the current value of the parameter from the output of the control object 3.

The signal of the current control error ε, taken from the comparison element 1, is sent to block 4 and block 5.

In block 4, the current control error ε is corrected (attenuated) within the selected range. The output signal ε _{n of} block 4 is a function of the variable ε in the selected range of ε values, starting from zero. Outside the selected range, the signal ε is transmitted to the output of block 4 without correction. In block 6, the signal ε _n is multiplied by the proportionality coefficient set when setting the ACP controller, while remaining optimal with increased values of the control error, which allows faster elimination of the deviation of the stabilized parameter from the reference when ε is greater than the upper limit of the attenuation range and reduces the transient oscillation with decreasing current control error within the attenuation range.

In block 5, the current control error ε is corrected (amplified) within the selected range for the integral component of the control signal. The output signal ε _{u of} block 5 is a function of the variable ε, in the selected range of ε, starting from zero.

Outside the selected range, the signal ε is transmitted to the output of block 5 without correction. In integrator 7, the signal ε _{u is} integrated.

The integral coefficient when setting the ACP controller is set to be slightly reduced (weakened), remaining unchanged with increased values of the regulation error.

Reducing the speed of exposure of the integral component with large deviations of the stabilized parameter from the given one avoids the situation when the integral component of the control signal accumulates faster than the regulator has time to respond.

At the same time, with a decrease in the regulation error after its predetermined value, the integral component is amplified, which provides greater accuracy in maintaining the preset value of the stabilized parameter in the steady state.

In more detail, the action of the correction blocks of the current regulation error is explained by the following examples.

1. For example, in block 4 of the correction of the proportional component of the control signal, the following correction function is implemented

Where

N is the upper limit of the correction range of the proportional component of the control signal, we take the attenuation range from zero to 1.5% of the measurement range of the stabilized parameter;

a is the attenuation coefficient, we take a = 1.2.

Then:

- at ε = 0.7% at the input of the block, at the output, the signal of the current control error will be ε _n = 0.6%,

- with ε = 0.2% at the input of the block, the output will be ε _n = 0.13%.

With a decrease in the current control error at the input of block 4, the signal attenuation is amplified. So, in the first case, the signal decreased 1.17 times, and in the second, 1.5 times.

If the input signal exceeds the set limit (in this case, 1.5%), the attenuation coefficient is set equal to unity, blocking the amplification of the output signal.

2. For example, in block 5 of the correction of the integral component of the control signal, the following correction function is implemented

Where

N is the upper limit of the correction range of the integral component of the control signal, we take the gain range from zero to 1.5% of the measurement range of the stabilized parameter;

in - gain, take in = 0.8.

Then:

- at ε = 0.7% at the input of the block, at the output, the signal of the current control error will be ε _u = 0.81%,

- with ε = 0.2% at the input of the block, the output will be ε _u = 0.3%.

With a decrease in the current control error at the input of block 4, the signal gain increases. So, in the first case, the signal increased 1.16 times, and in the second 1.5 times.

If the input signal exceeds the set limit (in this example, 1.5%), the gain is set equal to unity, blocking the gain of the output signal.

The initial adjustment of the controller is carried out using one of the generally accepted methods, while the power indices a and b must be equal to one.

With an increase in perturbing influences, in order to improve the transient performance, a power-law indicator, for example ( a ), is set to more than one in the correction block of the proportional law component, which leads to a “weakening” of the control action in the range from zero to the selected boundary and a decrease in the oscillation, and a power-law indicator , for example (c), in the correction block of the integral component of the law, less than unity is set while reducing the integrator coefficient, which leads to out "of the control action in the range from zero to a selected border and weakening control action outside the selected boundary. This method of correction of the integral component of the control law allows to reduce the integral accumulation rate outside the selected correction range, and at the same time, with a decrease in the control error within the correction range, the integral component is strengthened, which ensures greater accuracy in maintaining the set value of the stabilized parameter in the steady state.

Thus, the use of the proposed method of expanding the stability boundaries of the ASR reduces the likelihood of loss of stability while maintaining acceptable quality indicators of regulation of the automatic system in case of changes in the characteristics of the object, as well as when increased disturbing effects on the stabilization system of the adjustable parameter both in transient processes and in steady state .

Claims (1)

A method of expanding the range of stable operation of ASR with increased disturbing influences, including setting up a regulator that implements the proportional and integral components of the regulation law, in which the control signal depends only on the magnitude of the current regulation error and fixed values of the coefficients of the proportional and integral components of the regulation law established when adjusting the regulator , and the coefficients are reduced - "loosens the setting", characterized in that a control signal is generated by first converting the current control error before it is input to the inputs of the blocks of the proportional and integral components of the control law, and the conversion takes place in accordance with the values of the continuous function, the argument of which is the current control error, for each component of the control law ranging from the selected value of the current error to zero, the proportional component in accordance with the function of attenuating integral, increases as the absolute value of the current control error decreases.