CN102284505A - System for controlling thickness of fuzzy PI (Proportional Integral) based on ARM (Advanced RISC machines) - Google Patents

Abstract

The invention provides a fuzzy PI thickness control system based on ARM, comprising an entrance thickness gauge and an exit thickness gauge respectively located at the entrance and exit of a hydraulic rolling mill, and the entrance thickness gauge is connected to a compensator, which is characterized in that: the compensator and The exit thickness gauges are respectively connected to the fuzzy robust adjustment module, wherein the fuzzy robust adjustment module includes a fuzzy robust regulator realized by ARM, and the fuzzy robust regulator is respectively connected to the position signal acquisition module, pressure signal acquisition module, thickness signal acquisition module modules, execution modules, and signal conditioning modules. The invention proposes improvements aiming at a series of unstable problems occurring in traditional AGC control. In the actual cold rolling industry, due to the complexity of the rolling mill object, the traditional PID control scheme is difficult to achieve a good control effect. Due to the inherent nonlinearity of the fuzzy control strategy, the uncertainty of the object is not very prominent in the control process.

Description

A Fuzzy PI Thickness Control System Based on ARM

technical field

The invention relates to a thickness control system based on ARM and a fuzzy PI controller, which is an online parameter self-tuning thickness intelligent control system aimed at the longitudinal thickness difference of strips of hydraulic cold rolling mills in the cold rolling industry.

Background technique

At present, the general name of the strip longitudinal thickness difference control system in the hydraulic cold rolling industry is AGC (Automatic Gauge Control, that is, automatic thickness control). The control core of the hydraulic rolling mill AGC system lies in the servo control system of the oil cylinder position, referred to as APC (Automatic Position Control).

As shown in Figure 1, APC uses MTS or SSI series magnetostrictive displacement sensors to measure the position signal of the oil cylinder. The action of the oil cylinder is powered (energy) by the hydraulic station (high pressure) on the one hand, and its up and down movement is controlled by the electro-hydraulic servo valve on the other hand. The oil cylinder piston acts on the back-up roll, and the Newton force is transmitted to the work roll (actual rolling roll) through the back-up roll. In the constant roll gap rolling process, the computer sends a control command (given position signal), after a certain algorithm processing (such as PI control), and then the computer sends the output signal (execution command) to the electro-hydraulic servo valve. The purpose of thickness control is to make the actual rolling thickness of the strip equal to the set value. The strip AGC system must be compared with a target value preset by the computer and the measured actual thickness to obtain the deviation signal δh, or by changing the tension, The deviation signal δh is obtained by pressure, etc., and then the thickness automatic control device or computer function program is used to change the pressing position or strip tension or rolling speed to control the actual rolled thickness of the strip within the allowable range.

Objectively, it is required to take the actual thickness of the strip as the controlled quantity, rather than the actual position of the above-mentioned oil cylinder piston. In view of this consideration, on the basis of the position loop, a strip thickness monitoring link——AGC monitoring is introduced. Figure 2 is a control block diagram of AGC monitoring.

The above-mentioned AGC monitoring scheme, based on the original APC, is equipped with a thickness gauge at the outlet of the hydraulic rolling mill to measure the outlet thickness of the strip in real time. In this way, a true thickness feedback control system is formed. In fact, the oil cylinder of the rolling mill and its rolls, as mechanical components, have a large inertia, which leads to a certain lag in thickness control. Not only that, theoretically speaking, there is a certain distance between the thickness gauge and the work roll, and the part rolled by the strip is not the part where the thickness gauge collects the thickness signal, so there is a pure lag in practice, resulting in AGC monitoring does not achieve a very ideal control effect. Therefore, the improvement as shown in Figure 3 is proposed - the entrance and exit of the hydraulic rolling mill are equipped with thickness gauges at the same time.

If the average thickness of the strip entrance is represented by Gi, the average thickness of the exit is represented by Go, and △gi represents the floating value of the strip entrance thickness, and △go represents the floating value of the strip exit thickness, then:

Entrance thickness=Gi+△gi;

Outlet thickness=Go+△go;

Generally speaking, if △gi is positive, it indicates that the strip is too thick (compared with the average thickness), then the control algorithm Gc1 needs to be adjusted to ensure that the cylinder has a large enough reduction; otherwise, △gi is negative value, it indicates that the strip is too thin, and Gc1 also needs to be adjusted to ensure Gp (as shown in Figure 2, Gp represents the difference between the output of the controller Gc1 and the feedback value of the displacement sensor, and the actual input value of the controller PID). too small. Of course, △gi changes as a function of time and can be represented by △gi(t). Among them, △gi(t) changes with time. It can be seen that such compensation control is a kind of adaptive control in essence. Therefore, the research results of artificial intelligence can be used to design more intelligent control algorithms. All in all, AGC pre-control is a strategy of adding compensation control on the basis of AGC monitoring.

It is undeniable that conventional PI has played a historic role in traditional industrial control, especially thickness control in the cold rolling industry. However, with the continuous improvement of strip quality, especially the longitudinal thickness tolerance (2μ to 3μ variation), the traditional control method is greatly challenged. By analyzing the dynamic model of the hydraulic rolling mill, we will find that inherent nonlinear and slow time-varying factors exist in the position control link (for example, the servo valve has nonlinearity). Especially in the random interference that is ubiquitous in industrial sites (electromagnetic interference is an important type of interference), it is difficult for conventional PID control to exert a good control effect without the parameter self-tuning function. Through the analysis, it is found that it is difficult to obtain the precise mathematical model of the rolling mill system, so that the classical control theory based on the precise mathematical model of the controlled object loses its due effect.

Contents of the invention

The object of the present invention is to provide a control system with good robustness for the nonlinear, slow time-varying and uncertain disturbance characteristics of a hydraulic rolling mill.

In order to achieve the above object, the technical solution of the present invention is to provide a fuzzy PI thickness control system based on ARM, including the entrance thickness gauge and the exit thickness gauge respectively located at the entrance and exit of the hydraulic rolling mill, and the entrance thickness gauge is connected to compensate The device is characterized in that: the compensator and the outlet thickness gauge are respectively connected to a fuzzy robust adjustment module, wherein the fuzzy robust adjustment module includes a fuzzy robust regulator realized by ARM, and the fuzzy robust regulator is respectively connected to a position signal acquisition module , pressure signal acquisition module, thickness signal acquisition module, execution module and signal conditioning module;

The fuzzy robust regulator includes automatic compensation regulator, automatic thickness regulator and automatic position regulator, among which,

The automatic compensation regulator constitutes a feed-forward regulator, which adopts a conventional proportional regulator, and the automatic compensation regulator will appropriately amplify or reduce the deviation between the entrance thickness value obtained by the entrance thickness gauge and the thickness given value;

及该差值变化率的绝对值作为输入，并采用Fuzzy方法对常规PID调节器的比例常数

、积分常数

和微分常数

在线校正，通过常规PID调节器得到输出，

，其中，为由出口测厚仪得到的出口厚度值与厚度给定值的差值的积分； The automatic thickness regulator constitutes the outer ring regulator, which includes a Fuzzy module and a conventional PID regulator. The Fuzzy module converts the absolute value of the difference between the outlet thickness value obtained by the outlet thickness gauge and the thickness given value

and the absolute value of the rate of change of the difference As input, and using the fuzzy method to the proportional constant of the conventional PID regulator

, integral constant

and differential constant

On-line correction, the output is obtained through the conventional PID regulator ,

,in, is the integral of the difference between the outlet thickness value obtained by the outlet thickness gauge and the given thickness value;

之和与通过液压轧机用以测量油缸的位置信号的位置传感器得到的位置信息的差值，通过自动位置调节器的输出来控制伺服阀进而达到控制液压轧机油缸的目的。 The automatic position regulator constitutes the inner loop regulator, which adopts a conventional PID regulator. The input of the automatic position regulator is the output of the automatic compensation regulator and the output obtained through the automatic thickness regulator.

The difference between the sum and the position information obtained by the position sensor used to measure the position signal of the oil cylinder in the hydraulic rolling mill is used to control the servo valve through the output of the automatic position regulator to achieve the purpose of controlling the oil cylinder of the hydraulic rolling mill.

The invention proposes improvements aiming at a series of unstable problems occurring in traditional AGC control. In the actual cold rolling industry, due to the complexity of the rolling mill object, the traditional PID control scheme is difficult to achieve a good control effect. Due to the inherent nonlinearity of the fuzzy control strategy, the uncertainty of the object is not very prominent in the control process.

Description of drawings

Fig. 1 is APC block diagram;

Fig. 2 is a control block diagram with AGC monitoring;

Fig. 3 is a schematic diagram of rolling mill control;

Fig. 4 is a block diagram of the improved AGC control system;

Fig. 5 is a block diagram of fuzzy PI parameter adjustment;

Figure 6 shows the components of the ARM-based intelligent AGC thickness control system;

Fig. 7 is a block diagram of fuzzy PI thickness control of hydraulic rolling mill.

Detailed ways

In order to make the present invention more comprehensible, a preferred embodiment is described in detail below with accompanying drawings.

The anti-interference ability of the traditional AGC system cannot meet the needs of modern metallurgical automation. Therefore, how to improve the stability and anti-interference ability of the system is the core of the design of the robust controller. The root of the problem is that conventional PID does not have the function of parameter self-tuning, and cannot make adaptive parameter adjustments for complex environments. The scheme proposed by the present invention adopts the concept of modular design, and establishes an embedding system surrounded by "satellite" with the ARM chip as the core unit of the underlying hardware, the fuzzy PI algorithm as the core unit of the upper control software, and peripheral modules (collecting data and its output conditioning). The solution is to separate the functions of each part to the greatest extent, so as to achieve the purpose of minimizing the coupling degree of each module.

The hardware structure of the present invention is shown in Fig. 4 and Fig. 6, comprises the inlet thickness gauge and the outlet thickness gauge that are respectively positioned at the inlet and outlet of the hydraulic rolling mill, the inlet thickness gauge is connected to the compensator, and the compensator and the outlet thickness gauge are respectively Connect the Fuzzy Robust Conditioning Module.

As shown in Figure 6, the fuzzy robust adjustment module includes a fuzzy robust regulator implemented by ARM, and the fuzzy robust regulator is respectively connected to the position signal acquisition module, pressure signal acquisition module, thickness signal acquisition module, execution module and signal conditioning module .

Combined with Figure 7, the core unit of the upper control software includes automatic compensation adjuster (ACA in English, full name Automatic Compensation Adjuster), automatic thickness adjuster (ATA in English, full name Automatic Thickness Adjuster) and automatic position adjuster (English abbreviation It is APA, the full name is Automatic Position Adjuster).

It can be seen from Figure 7 that the core unit of the upper layer control software is a composite control system composed of ACA, ATA, and APA. Among them, APA constitutes the inner loop regulator; ATA constitutes the outer loop regulator; ACA constitutes the feedforward regulator. The inner and outer loops constitute two-stage cascade control, together with feed-forward control, together constitute the above-mentioned composite system. Each regulator is explained separately below.

The ACA uses the deviation of the inlet thickness gauge from a given thickness as its input signal. Based on the following idea: No matter the inlet thickness gauge or the outlet thickness gauge, there is a certain spatial distance between its installation position and the position of the roll system. In order to meet the real-time control, a compensation strategy must be adopted, and ACA is designed based on this. ACA uses a proportional regulator to properly amplify and reduce the deviation between the given thickness value and the inlet thickness value (this value is measured by the inlet thickness gauge). The feed-forward regulator ACA is designed to improve the control precision of the system. As shown in Figure 3, there is still a considerable distance between the installation position of the thickness gauge and the actual rolling position of the strip. The value measured by the instrument is lower than the reference value. As a result, the inaccuracy of the measured value will reduce the control accuracy of the entire hydraulic rolling mill thickness control system. Take the average of the inlet thickness and outlet thickness and compare with the thickness given value. If the former is greater than the latter, the output of the controller should be increased; otherwise, the output of the controller should be reduced. This is the design guideline of ACA.

ATA is the central regulator of the system. In view of the fact that the parameters of the outer ring regulator depend on the specific process, it is considered to design a fuzzy PID parameter self-tuning thickness regulator. The key to this design is the online self-adjustment of PID parameters, and to ensure that the system has industrial-grade stability. In terms of improving the stability of the system, it is considered to design a filter unit in the forward channel to filter out the noise in the collected signal. The filter unit here also needs to be able to flexibly adjust its cut-off frequency to adapt to different complex scenes. In fact, the problem of filtering is the problem of restoring the information flow disturbed by noise as much as possible, that is, the problem of predicting a time series confused by noise, so the problem of filtering is a prediction problem. The so-called prediction, mathematically speaking, is to estimate the statistical parameters of the entire ensemble from the past data of a time series. Considering the above problems, the system is called an intelligent robust servo regulation system.

，其中，

代表PID的输出，

、及

分别代表由出口测厚仪得到的出口厚度值与厚度给定值的差值、该差值的变化率及该差值的积分，

、

和

分别代表PID调节器的比例常数、积分常数和微分常数。 As shown in Figure 5, ATA is an automatic thickness regulator that uses the Fuzzy method for online correction of PID parameters. The conventional PID control formula is

,in,

represents the output of the PID,

, and

respectively represent the difference between the outlet thickness value obtained by the outlet thickness gauge and the given thickness value, the rate of change of the difference and the integral of the difference,

,

and

Represent the proportional constant, integral constant and differential constant of the PID regulator respectively.

、

和

分别为

和

的二元函数，亦即：

，

，

。 From the essence of Fuzzy control, the three undetermined parameters of the APA regulator

,

and

respectively

and

A binary function of , that is:

,

,

.

较大时，为使系统具有较好的快速跟踪性能，应取较大的

和较小的

，同时为避免系统的响应出现较大的超调，应对积分作用加以限制，通常取

=0。当

处于中等大小时，为使系统具有较小的超调，

应取得小一些；在这种情况下，

的取值对系统的影响较大，

的取值要适当。当

较小时，为使系统具有较好的稳态性能，

、

均应取得大些，同时为避免系统在设定值附近出现震荡，

的选择很重要。 A general analysis of it is as follows:

When the value is larger, in order to make the system have better fast tracking performance, a larger value should be taken

and smaller

, at the same time, in order to avoid a large overshoot in the response of the system, the integral action should be limited, usually take

=0. when

At a moderate size, in order to make the system have a small overshoot,

should be made smaller; in this case,

The value of has a great influence on the system,

The value should be appropriate. when

When is small, in order to make the system have better steady-state performance,

,

should be larger, and at the same time, in order to prevent the system from oscillating near the set value,

The choice is very important.

、

，先讨论

。 As mentioned earlier, there are two language variables for the Fuzzy controller, namely

,

, discuss first

.

的基本论域，亦即厚度误差的绝对值，在实际的冷轧领域，总希望带材的厚差控制的愈小越好，但是误差是客观存在的，达到误差为零亦是不可能的。实际中要求成品厚差控制在|10|μm，或者更好的情形|5|μm。但是，一般来讲来料公差比较大，这里以|60|μm进行分析。那么|e|的基本论域为[0,60]( μm)。下面与|e|相对应的

的论域为X={0,1,2,3,4,5,6,7,8,9,10,11,12}，从而得出量化因子

。 (1)

The basic field of discussion, that is, the absolute value of the thickness error, in the actual field of cold rolling, it is always hoped that the thickness difference of the strip can be controlled as small as possible, but the error exists objectively, and it is impossible to achieve zero error . In practice, the thickness difference of the finished product is required to be controlled at |10|μm, or better |5|μm. However, generally speaking, the tolerance of incoming materials is relatively large, and the analysis is performed here with |60|μm. Then the basic domain of discourse of |e| is [0,60] (μm). The following corresponds to |e|

The domain of discourse is X={0,1,2,3,4,5,6,7,8,9,10,11,12}, thus obtaining the quantization factor

.

这一语言变量选择六个语言值：极大（Extremely）、非常大（Very）、相当大（Largely）、比较大（Some）、略大（A little）、稍微大（Slightly）。分别定义为：EB、VB、LB、SB、AB、SLB。 (2) is the absolute value of the error

This linguistic variable selects six linguistic values: Extremely, Very, Largely, Some, A little, and Slightly. They are respectively defined as: EB, VB, LB, SB, AB, SLB.

，不妨令

； (3) According to practical experience, it is more reasonable to determine that the membership functions used to describe the fuzzy subsets EB, VB, LB, SB, AB, and SLB on domain X are represented by normal distribution. Membership function expression:

, may wish to order

;

的赋值表便可以表示成如下表形式： (4) Linguistic variables

The assignment table of can be expressed in the form of the following table:

the 0 1 2 3 4 5 6 7 8 9 10 11 12 SLB 0.9394 1.0000 0.9394 0.7788 0.5698 0.3679 0.2096 0.1054 0.0468 0.0183 0.0063 0 0.0005 AB 0.7788 0.9394 1.0000 0.9394 0.7788 0.5698 0.3679 0.2096 0.1054 0.0468 0.0183 0.0000 0.0019 SB 0.3679 0.5698 0.7788 0.9394 1.0000 0.9394 0.7788 0.5698 0.3679 0.2096 0.1054 0.0000 0.0183 LB 0.1054 0.2096 0.3679 0.5698 0.7788 0.9394 1.0000 0.9394 0.7788 0.5698 0.3679 0.0000 0.1054 VB 0.0183 0.0468 0.1054 0.2096 0.3679 0.5698 0.7788 0.9394 1.0000 0.9394 0.7788 0.0000 0.3679 EB 0.0019 0.0063 0.0183 0.0468 0.1054 0.2096 0.3679 0.5698 0.7788 0.9394 1.0000 0.0000 0.7788

In the above table it is assumed that when the domain of the fuzzy set of the absolute value of the error is {SLB, AB, SB, LB, VB, EB}={1, 2, 4, 6, 8, 10}.

的赋值表。至此，Fuzzy模块的两个输入语言变量的赋值表已经建立。采用同样的原理，也可以得到Fuzzy模块的三个输出语言变量

、

和

的赋值表。至此，亦即建立了Fuzzy模块知识库中的数据库，它包含输入、输出语言变量的赋值表。 In the same way, it can be concluded that the linguistic variable

assignment table. So far, the assignment tables of the two input language variables of the Fuzzy module have been established. Using the same principle, the three output language variables of the Fuzzy module can also be obtained

,

and

assignment table. So far, the database in the knowledge base of the Fuzzy module has been established, which contains the assignment table of input and output language variables.

Inference rules of the Fuzzym module: the rule statement adopts a relatively simple IF——THEN statement, namely:

;

；

;

；

;

The corresponding inference results are:

，

、

和

分别是集合

、

施加相应的模糊规则后得到的关于比例、积分

和微分

的模糊关系，它们本质上是模糊集合的一种。其中

为误差绝对值的模糊集，为误差变化率绝对值的模糊集，

代表模糊集合的合成运算，它与普通矩阵的乘法运算过程相同，只是运算符号不同。模糊控制规则具有相同的结构，但是针对的是不同的对象

、

和

，因此所得结果也就不同。

,

,

and

are collections

,

After applying the corresponding fuzzy rules, the ratio ,integral

and differential

They are essentially a kind of fuzzy sets. in

is the fuzzy set of the absolute value of the error, is the fuzzy set of the absolute value of the error rate of change,

Represents the synthesis operation of fuzzy sets, which is the same as the multiplication operation process of ordinary matrices, but the operation symbols are different. Fuzzy control rules have the same structure, but for different objects

,

and

, so the results are different.

），便可以对参数进行在线调整。 Fuzzy judgment: Finally, the weighted average method is used to dequantize the obtained fuzzy data (that is, divide by the quantization factor

), the parameters can be adjusted online.

之和与通过液压轧机用以测量油缸的位置信号的位置传感器得到的位置信息的差值，通过自动位置调节器的输出来控制伺服阀进而达到控制液压轧机油缸的目的。 APA adopts the traditional PID adjustment technology, and its input is the output of the automatic compensation regulator and the output obtained through the automatic thickness regulator.

The difference between the sum and the position information obtained by the position sensor used to measure the position signal of the oil cylinder in the hydraulic rolling mill is used to control the servo valve through the output of the automatic position regulator to achieve the purpose of controlling the oil cylinder of the hydraulic rolling mill.

Anti-electromagnetic interference analysis of ARM system:

The modern hydraulic rolling mill is an automation system integrating mechanical equipment, electric transmission, AGC thickness control, and hydraulic transmission. Power lines, control lines, and signal lines are in the large environment of the factory, so it is particularly important to improve the anti-interference ability of control equipment, especially embedded controllers. Anti-interference measures for ARM chips:

The first step is to add storage capacitors and decoupling capacitors to the power supply part of the chip, which can improve the electromagnetic compatibility of the circuit. The storage capacitor serves to stabilize the power supply voltage of the chip, while the decoupling capacitor provides an AC evacuation channel for system noise (interference).

In the second step, the following methods can be used to reduce or weaken the earth loop interference:

--- Shorten the ground wire or increase the cross-sectional area of the ground wire;

---Optical isolation;

---Add a filter network to the signal input circuit.

In the third step, the common impedance interference can be reduced or weakened by the following methods:

---Reduce the impedance value of the common ground wire of the circuit;

---Adopt appropriate circuit connection method (best match) in sensitive circuit parts, such as using different ground wires for high power lines and low signal lines; separate analog ground from digital ground, etc.

Claims (2)

1. PI type Fuzzy thickness control system based on ARM, comprise the inlet calibrator and the outlet calibrator that lay respectively at hydraulic mill inlet and exit, the inlet calibrator connects compensator, it is characterized in that: compensator and outlet calibrator connect fuzzy robust adjustment module respectively, wherein, fuzzy robust adjustment module comprises the fuzzy robust regulator of realizing by ARM, and fuzzy robust regulator is link position signal acquisition module, pressure signal acquisition module, thickness signal acquisition module, Executive Module and signal condition module respectively;

Fuzzy robust regulator comprises automatic compensated regulator, automatic thickness adjuster and automated location adjuster, wherein,

Automatically compensated regulator constitutes feedforward regulator, and it adopts conventional proportional controller, and the deviation of compensated regulator inlet thickness value that will obtain by the inlet calibrator and thickness set-point is carried out suitable amplification or dwindled automatically;

The automatic thickness adjuster constitutes the outer shroud adjuster, and it comprises Fuzzy module and conventional PID adjuster, the absolute value of the exit thickness value that will be obtained by the outlet calibrator by the Fuzzy module and the difference of thickness set-point

And the absolute value of this difference rate of change

As input, and adopt the proportionality constant of Fuzzy method to conventional PID adjuster , integral constant

And derivative constant

On-line correction obtains output by conventional PID adjuster ,

, wherein,

The integration of the exit thickness value that the outlet calibrator of serving as reasons obtains and the difference of thickness set-point;

Ring adjuster in the automated location adjuster constitutes, it adopts conventional PID adjuster, the output that is input as automatic compensated regulator of automated location adjuster and the output that obtains by the automatic thickness adjuster The difference of sum and the positional information that obtains in order to the position sensor of the position signalling of measuring oil cylinder by hydraulic mill is controlled servo valve and then is reached the purpose of control hydraulic mill oil cylinder by the output of automated location adjuster.

2. a kind of PI type Fuzzy thickness control system based on ARM as claimed in claim 1 is characterized in that: the method that improves the antijamming capability of the described fuzzy robust regulator of realizing by ARM is:

The first step, partly add savings electric capacity and decoupling capacitance at the power supply of ARM chip, improve the Electric Circuit Electromagnetic Compatibility ability, wherein save electric capacity and play the effect of stablize chip power supply voltage, decoupling capacitance then is to provide the passage that exchanges evacuation for system noise;

The cross-sectional area of second step, shortening ground wire or increase ground wire, output adopts photoelectricity to isolate, and reduces or weaken the loop interference of beating at signal input circuit interpolation filter network;

The 3rd goes on foot, adopts and reduce or weaken the common impedance interference.

Images