阅读说明：本技术 用设定值加权来闭环控制控制器的方法 (Method for closed-loop control of a controller with setpoint weighting ) 是由 塞巴斯蒂安·吉恩·费尔南德·德诺芙 克里斯多夫·马克·亚历山大·勒布伦 于 2018-12-20 设计创作，主要内容包括：本发明涉及一种用于在闭环控制系统(3)的、特别是伺服阀-液压缸系统的仪器和控制装置(1)中闭环控制比例积分型控制器(2)的方法,所述控制器(2)包括设定值加权系数(β),所述闭环控制方法包括以下连续步骤：分配(11)单位值给设定值加权系数(β),优化(12)对控制器(2)的闭环控制以满足至少一个预定的性能标准,限定特征跟踪误差(ε<Sub>TC</Sub>)使得可以响应要被闭环控制的系统的性能限制,以及根据特征跟踪误差(ε<Sub>TC</Sub>)和控制器(2)的闭环控制来分配(132)设定值加权系数(β)值。(The invention relates to a method for closed-loop control of a proportional-integral controller (2) in an instrumentation and control device (1) of a closed-loop control system (3), in particular of a servo valve-hydraulic cylinder system, said controller (2) comprising setpoint weighting factors (β), said closed-loop control method comprising the successive steps of assigning (11) unit values to setpoint weighting factors (β), optimizing (12) the closed-loop control of the controller (2) to meet at least one predetermined performance criterion, defining a characteristic tracking error (c: (m:) TC ) So that it can respond to the performance limit of the system to be closed-loop controlled, and to track the error according to the characteristic(s) ((b)) TC ) And closed-loop control of the controller (2) to assign (132) values of the set-point weighting coefficients (β).)

1. Method for setting a corrector (2) of the proportional-integral type in a servo control system (3), in particular in a control command device (1) of a servo valve-hydraulic cylinder system, said corrector (2) comprising a set-point weighting factor (β), a proportional gain (K)_p) And integral gain (K)_i)，

The setting method comprises the following continuous steps:

-setting (11) the set value weighting factor (β) to a unity value;

setting (12) the proportional gain (K) of the corrector (2) that meets at least one predetermined performance criterion_p) And the integral gain (K)_i)；

Defining a characteristic tracking error (_TC) So that the performance limits of the servo control system (3) are met, and

tracking error according to the preset feature (a)_TC) And the proportional gain (K)_p) And the integral gain (K)_i) Setting (132) the set point weighting factor (β) to a set point weighting factor (β) value.

2. The setting method according to claim 1, wherein the proportional gain (K) of the corrector (2) is set_p) And the integral gain (K)_i) Further comprising the step of (12):

determining (121) and setting the proportional gain (K)_p) And the integral gain (K)_I) Is an initial value;

adjusting (122) the proportional gain (K) by iteration_p) And the integral gain (K)_I) To optimize at least one predetermined performance criterion.

3. The setup method according to claim 2, wherein the proportional gain (K) is determined (121)_p) And the integral gain (K)_I) The step of initial values of (a) is performed by a ziegler-nicols empirical method or a high bridge empirical method.

4. The setting method according to any of claims 1 to 3, further comprising the step of determining a safety margin (σ), and wherein the step of setting (132) the set point weighting factor (β) is based on a theoretical error of the control command device (1) (σ)_TH) And said safety margin (σ).

5. The setup method according to claim 4, wherein the safety margin (σ) is determined from a behavior difference between a real system and its linearized model.

6. The setup method according to any one of claims 1 to 5, wherein the corrector (2) does not comprise a derivative component.

7. The setting method according to any of claims 1 to 6, the method being automated by means of a setting module (14) comprising one or more storage units (15) in which set values are stored enabling an automatic setting method to be performed, the set values being performed by means of at least one processor (16).

8. A control command device (1) of a servo control system (3), in particular of a servo valve-hydraulic cylinder system, the control command device (1) comprising a set value (X) which is input into a corrector (2), the output signal of the corrector (2) being input into the servo control system (3), the servo control system (3) generating a response (Y) which is also input into the corrector (2), wherein the corrector (2) is a proportional integral type corrector comprising a set value weighter (7), the weighter (7) comprising a set value weighting coefficient (β), the set value weighting coefficient (β) of the corrector (2) being set by means of a setting method according to any one of claims 1 to 7.

9. The command device (1) according to claim 8, wherein the corrector (2) is configured to generate a command (U) corresponding to the sum of:

is gained by integration (K)_I) Integrated and modified error;

a difference between a setpoint (X') weighted by said setpoint weighting factor (β) and said response (Y) of said servo control system (3), said difference being determined by a proportional gain (K)_p) To be modified.

Wherein the integral gain (K)_I) The proportional gain (K)_p) And the set value weighting coefficient (β) is a settable parameter of the corrector (2).

Technical Field

The present invention relates to the general field of servo control systems.

In particular, the invention relates to the setting of a corrector using command weighting in a control command system.

Background

The invention is applicable to all types of servo control systems, in particular to control command systems of actuators of turbomachines, for example actuator servo control parameters such as the pitch angle of the blades, the fuel flow or the position of components with variable geometry.

These actuators typically comprise a servo valve-hydraulic cylinder assembly, the behavior of which is typically modeled in the field of servo control by a second order linear system with an integrator.

This type of operation can be converted by the following equation:

the behavior of the command loop is usually specified by a list of requirements of the command system, in particular criteria for response time, overshoot, stability or static and tracking errors.

It is also necessary to ensure command robustness to disturbances (e.g., measurement noise or resistance that occurs during monitoring of the set point) and modeling uncertainty.

In addition to performance objectives, the command methodology must be easily adjustable and of reasonable complexity. In fact, if the performance obtained on a real system does not meet the requirement list, the command method can be recalibrated on the test bench. These differences in behavior can be explained by poor system modeling Gsys. The settings can then be modified by adjusting Gsys to make it more representative of reality.

Finally, a gradient limiter is usually applied to the local loop set-point associated with the pitch adjustment to avoid too much dynamics (possibly generating over-torque on the propeller shaft). These limiters imply that the fastest set point sent to the local loop will have a ramp form with a known maximum gradient. Therefore, it is a crucial issue to set up the corrector when knowing the errors that will be obtained for these ramp settings.

Proportional-integral correctors are not always suitable in view of the strict requirements on these local loops. Therefore, more advanced corrector structures with more degrees of freedom are needed to meet the need.

However, conventional practice of using derivative components in the regulation may produce undesirable adverse effects, particularly in the event of acquisition noise or sudden changes in the set point.

One solution in the arrangement of the corrector to allow a third degree of freedom without increasing the derivative component is to use a PI-type corrector with set value weighting.

This type of corrector enables to modify the dynamics of the tracking PI setting while maintaining its stability and the characteristics of suppressing disturbances. In particular, this may enable the speed of the initial PI to be maintained while reducing overshoot.

However, unlike the PI corrector in which there are many setting methods, the setting of the PI corrector with the weighting of the set values is mainly to compare the proportional gain K_pIntegral gain K_iAnd set point weighting factor β, this heuristic may prove cumbersome.

Disclosure of Invention

A first object of the present invention is to overcome the drawbacks of the prior art by providing a method for automatically setting a corrector of the proportional-integral type with weighting of the set values.

Another object of the invention is to propose a simple setting method.

It is a further object of the invention to provide a simple setter structure.

It is another object of the invention to minimize overshoot without reducing the response speed.

It is a further object of the invention to improve the response time and overshoot without reducing stability and robustness, in particular immunity to interference.

It is another object of the invention to reduce the effect of noise on the measurement.

It is another object of the invention to optimize the overshoot/tracking error tradeoff.

When the set point is ramp-type, it is common to understand the difference between the set point of the controlled system and the response of the system by tracking error. The ramp type setting is typically a linear function.

It is a further object of the invention to estimate and define the expected tracking error based on the settings of the corrector.

For these purposes, the invention proposes a method for setting a corrector of the proportional-integral type in a control command device of a servo control system, in particular of a servo valve-hydraulic cylinder system, said corrector comprising a set value weighting coefficient, a proportional gain and an integral gain.

The setting method comprises the following successive steps:

setting the set point weighting factor to a unit value;

setting the proportional gain and the integral gain of the corrector that meet at least one predetermined performance criterion;

defining the characteristic tracking error such that the performance limit of the servo control system is met, an

The set value weighting coefficient is set as the set value weighting coefficient value according to the characteristic tracking error and the proportional gain and the integral gain set in advance.

The invention can be supplemented selectively but advantageously by the use of the following features, taken alone or in combination:

the step of setting the proportional gain and the integral gain of the corrector further comprises the steps of:

determining and setting the proportional gain and the integral gain as initial values;

adjusting the proportional gain and the integral gain by iteration to optimize at least one predetermined performance criterion.

-the steps of determining the proportional gain and the integral gain are carried out by a Ziegler-nicols (Ziegler Nichols) empirical method or a high bridge (Takahashi) empirical method;

-the method further comprises the step of determining a safety margin, and wherein the step of setting the set-point weighting factor is performed in dependence of a theoretical error of the control command means and the safety margin;

the safety margin is determined from the behavior deviation between the actual system and its linearized model;

-the corrector does not comprise a derivative component;

the method is automatically implemented by means of a setting module comprising one or more memory units in which set values are stored so that an automatic setting method can be performed, the set values being performed by means of at least one processor.

According to a second aspect, the invention also proposes a control command device of a servo control system, in particular of a servo valve-hydraulic cylinder system, the control command device comprising a set value which is input into a corrector, the output signal of which is input into the servo control system, the servo control system generating a response, the response also being input into the corrector, wherein the corrector is a proportional integral type corrector comprising a set value weighter, the weighter comprising set value weighting coefficients, the set value weighting coefficients of the corrector being defined by means of the setting method according to the invention.

Optionally but advantageously, in such a device, the corrector is configured to generate a command corresponding to the sum of:

the error integrated and corrected by the integral gain;

the difference between the setpoint weighted by the setpoint weighting factor and the response of the servo control system, which is modified by the proportional gain.

Where the integration gain, the proportional gain, and the set-point weighting factor are parameters of the corrector that can be set.

Drawings

Other characteristics and advantages of the invention will become more apparent from the following description, purely illustrative and non-limiting, which should be read with reference to the accompanying drawings. In the drawings:

FIG. 1 is a schematic diagram of a servo control chain according to the present invention;

FIG. 2 is a schematic diagram detailing a method for automatic setting of a corrector according to the present invention;

fig. 3 shows a setup module that can perform the method according to the invention.

Detailed Description

Summary of the invention

Referring to fig. 1, a control command chain 1 includes a corrector 2 and a servo control system 3.

In a preferred embodiment, the servo control system 3 comprises an integrator.

In the embodiment shown, the corrector 2 is of the proportional integral type with a set value weighting.

A set value X is input at the input of the corrector 2, the corrector 2 converting the set value X into a command U, which is input into the servo control system 3.

The servo control system 3 reacts according to the received command U, and the response Y of the servo control system is measured and returned to the corrector 2.

More specifically, corrector 2 performs in parallel a proportional action 4 and an integral action 5, which are input into adder 6.

Thus, the adder 6 generates a command U, which is input into the servo control system 3.

Both the set value X and the response Y are input into each of the proportional action 4 chain and the integral action 5 chain.

Proportional action 4 receives as input a set value X which is input to a weighter 7 to generate a weighted set value X'.

The weighter 7 applies a gain or set-point weighting factor β to the set-point X.

The weighted set value X ' and the response Y are input to a subtractor 9, which generates a weighted error e ', that is, a difference between the weighted set value X ' and the response Y.

Inputting the weighted error eTo proportional gain K_pThen, the signal is input to the adder 6.

The integrating action 5 receives as input the set value X, which is input to the subtractor 10 together with the response Y, and generates an error corresponding to the difference between the set value X and the response Y.

In this particular case, the error is a tracking error_TThe set value X is of a ramp type.

The error is then input to the integral gain K_IAnd then input into the integrator block 8. The output of the integrator 8 is input to the adder 6.

In other embodiments, the weighter 7 may be located on the integrating action 5, or upstream of the corrector, or each of the proportional action 4 and the integrating action 5 comprises a weighter 7, each weighter 7 having a weighting coefficient set point β, the coefficients being different from each other.

Corrector arrangement

As is well known, the proportional gain K_pAnd integral gain K_IHas an influence on the stability, response time and robustness of the control chain 1.

The setting of the degrees of freedom of the corrector 2 allows to optimize the criteria of stability, response time and robustness of the system, as well as to minimize overshoots and tracking errors.

The method for automatically setting these parameters includes a plurality of steps that are performed sequentially. This method is illustrated in fig. 2.

First, an assignment step 11 is performed during which the set value weighting coefficient β is fixed to a unity value. As such, the corrector has the behavior of a conventional proportional-integral corrector.

Then, an optimization step 12 is carried out, during which settings of the corrector 2 are made to optimize at least one performance criterion, which may be chosen from the following performance criteria: for example, robustness, response time, overshoot, or any other criterion or combination of criteria that makes it possible to quantify the performance and behavior of the servo control system.

In a determination step 121, an initial value is determined and assigned to the proportional gain K_pAnd integral gain K_I。

This is the first setting of the corrector 2 that can be achieved by conventional setting methods, such as empirical methods of Ziegler-nicols or high bridges (Takahashi), as described below.

In a determination step 121 according to the high-bridge method, the gain margin of the system to be adjusted is estimated by increasing the gain until a self-sustaining oscillating system is obtained.

The initial proportional and integral gain values are then defined according to the gain margin values given by the high-bridge method (there are available in the literature correspondence tables).

Any other conventional corrector setting method may be used to perform this step, the choice of another method causing the initial values of the proportional gain and the integral gain to be determined according to criteria other than gain margin, such as overshoot or response time, etc.

The proportional gain K is then compared in an adjustment step 122_pAnd integral gain K_IIs improved.

In an adjustment step 122, the proportional gain K is improved by iteration_pAnd integral gain K_ITo meet the stability, response time and robustness requirements specified by the requirement list of the control command chain 1. The value of the gain is increased or decreased until a setting is obtained which gives satisfactory results in the simulation.

Once the optimum proportional gain K is obtained_pAnd integral gain K_IThe value is fixed and then a weight setting step 13 is performed.

By applying the final value theorem to a system such as control command chain 1, the set value weighting coefficient β can be represented by the following relationship:

thus, in the depicted embodiment, the set point weighting factor β is a proportional gain K_pAnd integral gain K_IAnd a function of the systematic error.

Proportional gain K_pAnd integral gain K_IIs fixed, the value of the set point weighting factor β may be calibrated to achieve an error value corresponding to the criteria of the requirement list that define the performance to be achieved by the control command chain 1.

In order to obtain a behaviour that meets the criteria specified by the requirement list, it is necessary to determine the set size of the corrector for the most unfavorable operating situation.

In the case of a maximum gradient of the setpoint, the most unfavorable operating situation is encountered.

The set point gradient limit means that the most critical set point will have the form of a ramp with a gradient equal to that of the gradient limiter.

Thus, the type of error used to design the corrector size will be the tracking error, corresponding to the error caused by the most rigorous set point model (ramp).

Before the weighted distribution step 132, a modeling step 131 may be performed, during which the servo control system 3 is assimilated to a theoretical model 3' representative of its operation.

In selected embodiments, the servo control system model 3' is a perfect second order linear system associated with an integrator, which is influenced by a ramp-type setting of unity slope. For example, it may comprise an actuator of the servo valve-cylinder type.

Thus, the command chain 1 is modeled by a command chain model 1 ' comprising a corrector model 2 ' and a servo control system model 3 ' similar to the corrector 2.

During the modelling step 131, the corrector model 2' has the settings established during the assigning step 11 and the optimizing step 12.

The set value weight coefficient β is fixed to unity and the proportional gain K is_pAnd integral gain K_IIs fixed to the value obtained after the optimization step 12.

Theoretical error of the Command chain model 1_THA conventional derivation is performed which is then used to make the setting of setpoint coefficient β.

Theoretical error_THOr in a requirement listTechnical specifications and is extracted directly from the requirement list.

However, in other embodiments the set value may be applied on a ramp having a non-cell slope.

Thus, in the chosen ramp set point embodiment, the theoretical error of the model is the characteristic tracking error.

Thus, during the weight assignment step 132, the set value weight coefficient β may be defined for a value expressed according to the following equation:

the values thus represented will be assigned to the weighters 7 of the control chain 1.

Under the influence of the setting of the value of the set-point weighting factor β, command chain 1 will generate a tracking error_TThe tracking error_TWill tend to feature tracking errors_TCThe value of (c).

Alternatively, the theoretical error_THThe safety margin σ may be associated with a defined so as to account for non-linearities in the operation of the servo control system 3. it is necessary to account for imperfections within the overall aspects of the corrector 2. the set point weighting factor β is then defined according to the following formula:

in an embodiment in which the servo control system 3 is modeled as a perfect second order linear system with an integrator that is influenced by the ramp setpoint, the setpoint weighting factor β can then be defined by the following relation:

the structure of this embodiment of the corrector 2 makes it possible to modify the set value weighting factor beta without any impact on the performance of the corrector 2 in terms of stability and robustness.

Optimal setting of the set point weighting factor β allows optimization of response time, overshoot, and tracking error_TMore specifically, it makes it possible to target tracking errors_TLimiting the expected performance without reducing the pre-pass proportional gain K_pAnd integral gain K_iThe response time and overshoot performance obtained by the setup.

More specifically, the overshoot is highly contained while maintaining a response time similar to that of the corrector 2 without the weighter 7.

In this embodiment, the corrector 2 maintains its stability and robustness characteristics, whether or not it has a weighter 7.

By avoiding the addition of derivative components, the sensitivity of the system to measured noise is greatly limited.

In contrast to the specification of the requirement list, the theoretical error is taken into account in the setting of the corrector 2_THAn optimum overshoot/error trade-off can be obtained.

In an embodiment in which the servo control system 3 is modeled as a perfect second order linear system with an integrator that is affected by the ramp set point, the characteristic tracking error is taken into account in the setting of the corrector 2 compared to the specification of the requirement list_TCAn optimum overshoot/tracking error can be obtained_TIs determined.

The automatic setting method, besides simplifying the process, also greatly limits the duration of the operation.

The simple structure of the setter limits its development and maintenance costs.

The automatic setting method of the corrector 2 is implemented by means of a setting unit or module 14, which setting unit or module 14 comprises one or more storage units 15, wherein the set values are stored such that the automatic setting method can be performed.

The set-points are performed by means of at least one processor 16, said at least one processor 16 performing the automatic setting method of the corrector 2. The processor 16 and the memory 15 are typically part of the engine computer, but it is also possible to integrate them into one particular module, which is physically separated from the engine control unit.

Similarly, if the integrator system has a set-point weighted PI corrector setting, we can know the tracking error we will get for the ramp set-point due to the following relationship:

now, the servo control performance and hence the tracking error_TIs necessary in designing the rollback mode of the turbine.

When a fault is detected on the machine, rollback must be performed quickly to protect the machine by avoiding the machine remaining in a slowed down state for too long.

Conversely, too rapid a reduction in the propeller speed with respect to the power provided by the gas generator is also dangerous, since additional torque may be created on the propeller shaft.

Thus, since the torque that will occur in the propeller shaft during a fast transient can be known in advance, both from the driving (synchronization between the propeller and the gas generator) and from the mechanical design point of view, by tracking errors_TKnowledge of (a) knows the deceleration profile in advance so that the design of the propeller can be optimized.