阅读说明：本技术 冷轧带材板形仪检测信号挠曲附加分量的消除方法 (Method for eliminating deflection additional component of detection signal of cold-rolled strip shape meter ) 是由 于华鑫 张桐源 高心成 赵耕 廖霜 杨昇 刘宏民 于 2020-11-30 设计创作，主要内容包括：本发明提供一种冷轧带材板形仪检测信号挠曲附加分量识别和消除方法,其包括以下由计算机执行的步骤：1、获取计算需要的板形仪的数字量参数；2、获取计算需要的特定采样点序号及其范围参数；3、获取当前检测单元当前周期的采样点对应AD值,并根据误差最小原理,使用采样数据与附加波形曲线的误差计算公式倒推出各检测单元的挠曲附加信号在当前周期内的波形函数具体形式,并得到当前周期各单元消除影响后的波形。本发明提供了一种方便快速的整辊板形检测辊在线消除挠曲附加信号的方法,提高了获取真实板形信号过程的准确性、快速性。(The invention provides a method for identifying and eliminating deflection additional components of a detection signal of a cold-rolled strip shape meter, which comprises the following steps executed by a computer: 1. acquiring digital quantity parameters of the shape meter required by calculation; 2. acquiring the serial number of a specific sampling point required by calculation and a range parameter of the serial number; 3. and acquiring the AD value corresponding to the sampling point of the current period of the current detection unit, and according to the principle of minimum error, using an error calculation formula of the sampling data and the additional waveform curve to back-out the specific form of the waveform function of the bending additional signal of each detection unit in the current period, and acquiring the waveform of each unit in the current period after influence elimination. The invention provides a method for conveniently and quickly eliminating the deflection additional signal of the whole-roller plate shape detection roller on line, and the accuracy and the rapidity of the process of acquiring the real plate shape signal are improved.)

1. A method for eliminating deflection additional components of a detection signal of a cold-rolled strip shape meter is characterized by comprising the following steps: which comprises the following steps:

s1: acquiring digital quantity parameters of a required plate shape instrument, and comprising the following substeps:

s11: acquiring the number N of detection units;

s12: acquiring the number m of sensors circumferentially arranged on a single detection unit of the plate-shaped roller;

s13: calculating the number n of sampling points according to the sampling frequency f of the plate-shaped signal processor, the diameter D of the plate-shaped roller and the rotating linear velocity v of the plate-shaped roller, wherein,the number n of sampling points of each period of each detection unit is the same;

s2: acquiring the sequence number and the range of the required boundary sampling point, and comprising the following substeps:

s21: acquiring the number r of sampling points which do not participate in calculation in one rotation of the plate-shaped roller, namely a signal period according to the angle theta of the strip-coated plate-shaped roller_θWherein the number of sampling points corresponding to the wrap angle theta is as follows:

s22: obtaining the number m.n of sampling points which do not participate in calculation in one signal period and are formed by one rotation of the plate-shaped roller_θThe number of the boundary sampling point and its range, n₁～n₂The number of the sampling points at the boundary of the first peak of the original signal and its range, n₃～n₄The sequence number and the range of the boundary sampling point of the second peak of the original signal are adjusted according to the number of the peaks in a period, namely the number of the sensors arranged in one detection unit;

s3: the method comprises the following steps of obtaining an AD signal value corresponding to a sampling point of a current detection unit in a current period, and according to the principle of minimum error, using an error calculation formula of sampling data and an additional waveform curve to back-out a specific form of a waveform function of a bending additional signal of each detection unit in the current period, and obtaining a waveform of each unit in the current period after influence elimination, wherein the method comprises the following substeps:

s31: let i equal to 1 and j equal to 1 to n, and obtain the AD value y corresponding to each sampling point in each detection unit signal through the electrical signal transmitted from the plate-shaped roller_ij；

S32: using sampled data and additional wavesCalculating the error equation Q of the current unit in the current period by a shape curve_i：

Where y is the true signal function, a_iFor the i-th detection unit deflection waveform amplitude, b_iZero-shift of deflection waveform for the i-th detection cell, y_ijThe value of the sampling point AD of the ith detection unit is the value of the sampling point AD of the ith detection unit;

s33: according to the principle of error minimization, Q_iAre respectively paired with a_iAnd b_iAnd solving a partial derivative to obtain a control equation of the current unit in the current period:

namely:

matrix inversion operation is carried out on the above formula to obtain a of each unit in the current period_i、b_iThe value of (a) is:

s34: obtaining the specific form of the additional waveform curve of the current unit j in the current periodThen, the signal is removed from the original signal to obtain the current periodShape detection signal waveform y after unit elimination_ei：

If i is not greater than N, i is made equal to i +1, and step S3 is repeated N times with the initial value of i being 1 until i > N, at which time i is made equal to N +1, and the above-described steps are ended.

2. The method of eliminating the additional component of the deflection of the detection signal of the cold rolled strip shape gauge of claim 1, wherein: the AD value of the ith detection unit in one period is y_i1,y_i2,…,y_ij,…,y_in。

3. The method of eliminating the additional component of the deflection of the detection signal of the cold rolled strip shape gauge of claim 1, wherein: the plate-shaped detection roller is provided with two rows of through holes on the circular side surface along the axial direction of the roller body, a plurality of sensors are arranged in the through holes at intervals, and a detection unit is formed by the plurality of sensors on each circumference.

4. The method of eliminating the additional component of the deflection of the detection signal of the cold rolled strip shape gauge of claim 3, wherein: the sensors are piezoelectric sensors, and the output end of each sensor is connected with the input end of a plate-shaped signal processing computer.

5. The method of eliminating the additional component of the deflection of the detection signal of the cold rolled strip shape gauge of claim 4, wherein: and finally, displaying the waveform of the obtained plate shape detection signal in the plate shape monitoring panel in the form of a relative length difference histogram and a plate shape target curve.

6. The method of eliminating the additional component of the deflection of the detection signal of the cold rolled strip shape gauge of claim 4, wherein: when the strip is tensioned to cover the surface of the plate-shaped detection roller, the residual stress distribution of the cold-rolled strip caused by flatness error is quantitatively output in a relative length difference mode according to the magnitude of the radial force which is subjected to the strip along the axial direction and is not uniformly distributed, and the obtained distribution is the strip plate-shaped distribution detected by each detection unit of the plate-shaped roller.

Technical Field

The invention relates to the technical field of metallurgical rolling, in particular to a method for quickly identifying and eliminating influence components generated by bending deformation of a plate-shaped roller in a detection signal of a cold-rolled strip shape meter on line.

Background

Due to the wide application of the cold-rolled strip as a raw material, strong support is continuously injected for national economic construction and the continuous development of various industries. With the advancement of the innovation of the traditional industry in China, under the environment of no quality demand, high-end products such as thin strips and foils with large width-thickness ratio are increasingly pursued in the production of cold-rolled strips. On the other hand, the popularity of intellectualization and de-manual work makes more and more enterprises realize the advantages of cost reduction in automated production, yield improvement and reduction of manual errors. In the field of metallurgical rolling, a plate shape instrument is one of auxiliary production tools for high-end plate and strip materials, and is very suitable for the new environment.

It has been found in field commissioning and applications that different plate-like roll configurations have different effects on the sensors installed therein. Under the action of gravity, the plate-shaped roller naturally generates slight flexural deformation when rotating. This deflection adds an additional sinusoidal-like waveform influencing signal to the force signal waveform received by the roll shape gauge, which is a type of axially through-hole sensor-mounted, during its operation (tension is transmitted from the strip to the roll surface). And because the effect of tension, the deflection influence is enlarged by several times compared with the condition that only the self-weight is received when idling, and the detection unit with weak signal strength in the roller can generate large interference to cause detection errors, and the detection errors need to be eliminated by seeking means.

In general, methods including fourier transform and digital filtering are often used to remove such additional signal effects, but similar methods each have short panels: the Fourier transform method has low speed of eliminating influence signals and more required sampling points, can be used as an off-line analysis method and is not suitable for on-line application of plate shape detection; the digital filtering method can be applied to the condition that the speed is not changed, namely the frequency of the influence signal and the original signal is fixed, but the speed is continuously changed when the plate-shaped roller works actually, the parameter of the filter is changed at any time, the original signal is distorted, and the digital filtering method is not suitable for being applied under the actual working condition. The rest filtering methods are different, and various inapplicable problems exist. At present, no relevant research exists in the metallurgical rolling field at home and abroad.

Disclosure of Invention

The invention aims to provide an online eliminating method for a bending additional component of a detection signal of a cold-rolled strip whole-roller type shape meter. The invention fully researches two signals collected by the plate-shaped roller: the correlation and difference between the original detected signal and the additional signal. When such an additional signal is found to be present, it is characterized with respect to its characteristics in that the roll body is deflected by the influence of its own weight and the tension of the strip, causing the sensor mounting hole to be deformed and thus pressing or releasing the sensor, so that the signal period is related to the roll speed. In practice, in the actual detection process, the period and the phase are the same as those of the detection signal, the amplitude is deviated, and a certain zero drift exists, so that the detection result is inaccurate. The patent provides a method for eliminating deflection additional components of a detection signal of a cold-rolled strip shape meter, which can eliminate deviation and ensure the accuracy of the detection signal of the shape meter.

Specifically, the invention provides a method for eliminating the deflection additional component of a detection signal of a cold-rolled strip shape meter, which comprises the following steps:

s1: acquiring digital quantity parameters of a required plate shape instrument, and comprising the following substeps:

s11: acquiring the number N of detection units;

s12: acquiring the number m of sensors circumferentially arranged on a single detection unit of the plate-shaped roller;

s13: calculating the number n of sampling points according to the sampling frequency f of the plate-shaped signal processor, the diameter D of the plate-shaped roller and the rotating linear velocity v of the plate-shaped roller, wherein,the number n of sampling points of each period of each detection unit is the same;

s2: acquiring the sequence number and the range of the required boundary sampling point, and comprising the following substeps:

s21: acquiring the number r of sampling points which do not participate in calculation in one rotation of the plate-shaped roller, namely a signal period according to the angle theta of the strip-coated plate-shaped roller_θWherein the number of sampling points corresponding to the wrap angle theta is as follows:

s22: obtaining the number m.n of sampling points which do not participate in calculation in one signal period and are formed by one rotation of the plate-shaped roller_θThe number of the boundary sampling point and its range, n₁～n₂The number of the sampling points at the boundary of the first peak of the original signal and its range, n₃～n₄The sequence number and the range of the boundary sampling point of the second peak of the original signal are adjusted according to the number of the peaks in a period, namely the number of the sensors arranged in one detection unit;

s3: the method comprises the following steps of obtaining an AD signal value corresponding to a sampling point of a current detection unit in a current period, and according to the principle of minimum error, using an error calculation formula of sampling data and an additional waveform curve to back-out a specific form of a waveform function of a bending additional signal of each detection unit in the current period, and obtaining a waveform of each unit in the current period after influence elimination, wherein the method comprises the following substeps:

s31: let i equal to 1 and j equal to 1 to n, and obtain the AD value y corresponding to each sampling point in each detection unit signal through the electrical signal transmitted from the plate-shaped roller_ij；

S32: calculating the error equation Q of the current unit in the current period by using the sampling data and the additional wave curve_i：

Where y is the true signal function, a_iFor the i-th detection unit deflection waveform amplitude, b_iZero-shift of deflection waveform for the i-th detection cell, y_ijThe value of the sampling point AD of the ith detection unit is the value of the sampling point AD of the ith detection unit;

s33: according to the principle of error minimization, Q_iAre respectively paired with a_iAnd b_iAnd solving a partial derivative to obtain a control equation of the current unit in the current period:

namely:

matrix inversion operation is carried out on the above formula to obtain a of each unit in the current period_i、b_iThe value of (a) is:

s34: obtaining the specific form of the additional waveform curve of the current unit j in the current periodThen, the waveform is removed from the original signal to obtain the waveform y of the plate-shaped detection signal after the influence of the current unit in the current period is eliminated_ei：

If i is not greater than N, i is made equal to i +1, and step S3 is repeated N times with the initial value of i being 1 until i > N, at which time i is made equal to N +1, and the above-described steps are ended.

Preferably, the size of the AD value in one period of the ith detection unit is y_i1,y_i2,…,y_ij,…,y_in。

Preferably, the plate-shaped detection roller is provided with two rows of through holes along the axial direction of the roller body on the circular side surface, a plurality of sensors are arranged in the through holes at intervals, and a detection unit is formed by the plurality of sensors on each circumference.

Preferably, the sensors are piezoelectric sensors, and the output end of each sensor is connected with the input end of a plate-shaped signal processing computer.

Preferably, the finally obtained shape detection signal waveform is displayed in the shape monitoring panel in the form of a relative length difference histogram and a shape target curve.

Preferably, when the strip is stretched to cover the surface of the shape detection roller, the residual stress distribution of the cold-rolled strip caused by flatness error is quantitatively output in the form of relative length difference through the magnitude of the unevenly distributed radial force applied along the axial direction of the strip, and the obtained result is the strip shape distribution measured by each detection unit of the shape detection roller.

Compared with the prior art, the invention has the beneficial effects that:

the invention only relates to a small derivation calculation amount, and provides a general scheme for conveniently and quickly identifying the shape of the whole-roller plate detection roller on line and eliminating a deflection additional signal. Compared with the conventional filtering and other modes, the method has the advantages that the process of acquiring the real plate shape signal is quicker, so that the real-time requirement of the plate shape detection closed loop under the actual working condition is met; the identified deflection waveform almost completely coincides with the phase range outside the peak section of the original signal, thereby ensuring that the result is more accurate.

Drawings

FIG. 1 is a flow chart of the calculation of the present invention;

FIG. 2 is a schematic view of the force deflection and influence on the sensor of the whole roll type plate-shaped roll according to the present invention;

FIG. 3 is a schematic diagram of the principle and details of a minimum error method;

FIG. 4 is a graph comparing a set of raw signals, deflection signals, and de-deflected signals for a real signal in an embodiment of the present invention;

FIG. 5 is a graph comparing the distribution of the sheet shape before and after the removal of the deflection signal.

Detailed Description

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

The method for eliminating the deflection component of the detection signal of the cold-rolled strip shape meter mainly comprises the following steps executed by a computer, and the flow of the method is shown in the attached figure 1 in the specification.

Specifically, the invention provides a method for eliminating the deflection additional component of a detection signal of a cold-rolled strip shape meter, which comprises the following steps:

s1: acquiring digital quantity parameters of a required plate shape instrument, and comprising the following substeps:

s11: acquiring the number N of detection units;

s12: acquiring the number m of sensors circumferentially arranged on a single detection unit of the plate-shaped roller;

s13: calculating the number n of sampling points according to the sampling frequency f of the plate-shaped signal processor, the diameter D of the plate-shaped roller and the rotating linear velocity v of the plate-shaped roller, wherein,the number n of sampling points of each period of each detection unit is the same;

s2: acquiring the serial number and the range of a specific sampling point required by calculation;

s21: acquiring the number r of sampling points which do not participate in calculation in one rotation of the plate-shaped roller, namely a signal period according to the angle theta of the strip-coated plate-shaped roller_θWherein the number of sampling points corresponding to the wrap angle theta is as follows:

s22: obtaining the number m.n of sampling points which do not participate in calculation in one signal period and are formed by one rotation of the plate-shaped roller_θThe number of the boundary sampling point and its range, n₁～n₂The number of the sampling points at the boundary of the first peak of the original signal and its range, n₃～n₄The number and range of the sampling point at the boundary of the second peak of the original signal are determined according to the number of the peaks in a periodThe number and the range of the boundary sampling points are increased by installing sensors in one detection unit;

s3: the method comprises the following steps of obtaining an AD signal value corresponding to a sampling point of a current detection unit in a current period, and according to the principle of minimum error, using an error calculation formula of sampling data and an additional waveform curve to back-out a specific form of a waveform function of a bending additional signal of each detection unit in the current period, and obtaining a waveform of each unit in the current period after influence elimination, wherein the method comprises the following substeps:

s31: let i equal to 1 and j equal to 1 to n, and obtain the AD value y corresponding to each sampling point in each detection unit signal through the electrical signal transmitted from the plate-shaped roller_ij；

S32: calculating the current unit error equation of the current period by using the sampling data and the additional waveform curve:

wherein, a_iFor the i-th detection unit deflection waveform amplitude, b_iZero-shift of deflection waveform for the i-th detection cell, y_ijThe value of the sampling point AD of the ith detection unit is shown.

S33: according to the principle of minimum error, obtaining a current unit control equation of the current period:

namely:

matrix inversion operation is carried out on the above formula to obtain a of each unit in the current period_i、b_iThe value of (a) is:

s34: obtaining the specific form of the additional waveform curve of the current unit j in the current periodThen, the waveform of the plate shape detection signal after the influence of the current unit in the current period is eliminated can be obtained by removing the waveform from the original signal:

if i is not greater than N, i is made equal to i +1, and step S3 is repeated N times with the initial value of i being 1 until i > N, where i is N +1, and the above-described steps are ended.

The working principle of the present invention is further explained below with reference to the following examples:

s1: acquiring digital quantity parameters of the plate shape instrument required by calculation, specifically comprising the following steps:

s11: and acquiring the number N of the detection units.

S12: and acquiring the number m of sensors circumferentially arranged on a single detection unit of the plate-shaped roller. Here, m is 2, that is, one detection unit and two sensor blocks. As shown in fig. 2 in the specification, the sensor a and the sensor B together constitute a detection unit.

S13: the sampling frequency f of the plate-shaped signal processor is 5000HZ, the diameter D of the plate-shaped roller is 340mm, the rotating linear speed v of the plate-shaped roller is 11.89m/s, and the number of sampling points is calculatedEach unit has the same period n.

S2: acquiring the serial number of a specific sampling point and the range parameter thereof required by calculation, wherein the method comprises the following steps:

s21: the angle theta of the wave signal generated by the strip affecting the plate-shaped roller is taken to be 28.67 DEG, and the angle theta is calculatedThe number of sampling points corresponding to the wrap angle θ:obtaining the number r of sampling points which do not participate in calculation in one signal period, namely one rotation of the plate-shaped roller_θ＝72。

S22: obtaining the number m.n of sampling points which do not participate in calculation in one signal period and are formed by one rotation of the plate-shaped roller_θThe number of the boundary sample points and their ranges. Since m is 2, two peaks exist in one period of the original waveform. Here, first peak sample signal point number: i.e. i_ASecond peak sample number 118: i.e. i_B345. The boundary sampling numbers of the original signal (two peak values) obtained by the method are respectively: the details are shown in figure 3 of the specification.

S3: let i equal 1 and j equal 1 to n. Obtaining the corresponding AD value of the sampling point of the current period of the current detection unit, and according to the principle of minimum error, using the error calculation formula of the sampling data and the additional waveform curve to back-out the specific form of the waveform function of the bending additional signal of each detection unit in the current periodAnd obtaining the waveform of each unit in the current period after the influence is eliminated. i ≦ N, i ═ i +1, and the process returns to S3 again. The method comprises the following steps:

s31: obtaining AD value y corresponding to each sampling point of each detection unit through electric signals transmitted from the plate-shaped roller_ij(the size of the AD value in one cycle of the i-th detection unit is y_i1,y_i2,…,y_ij,…,y_in). In this example, the AD value of one detection unit for one sampling period of a set of 449 sampling points at the current rotation speed and sampling frequency is obtained, as shown in table 1 below.

TABLE 1 original signal sampling point number and corresponding AD value

S32: calculating the current unit error equation of the current period by using the sampling data and the additional waveform curve:

wherein, a_iFor the i-th detection unit deflection waveform amplitude, b_iZero-shift of deflection waveform for the i-th detection cell, y_ijThe value of the sampling point AD of the ith detection unit is shown.

S33: according to the principle of minimum error, obtaining a current unit control equation of the current period:

namely:

solving a of each unit in the current period by matrix inverse operation_i、b_iThe value of (c):

s34: obtaining the specific form of the additional waveform curve of the current unit j in the current periodThen, the waveform of the plate shape detection signal after the influence of the current unit in the current period is eliminated can be obtained by removing the waveform from the original signal:based on the minimum error method, the following data are calculated by using the sampling point data in the table 1: 189.0348, b-321.591. That is, the waveform of the deflection-added signal in this example is embodied in the form:

and the original signal waveforms before and after the deflection signal is removed by the current period of the plate-shaped roller current detection unit can be obtained by removing the original signal waveforms (subtracting point values). The comparison of the waveforms of the original signal, the real signal and the deflection signal is shown in figure 4 of the specification. The actual plate shape distribution without deflection influence can be obtained by continuously executing the process, and the description in the attached figure 5 is a plate shape comparison diagram before and after the deflection signal is eliminated according to the process.

The bending signal obtained by fitting almost completely coincides with the waveform in the phase range outside the wave crest section of the original signal (the two waveform curves are obviously separated only at the phase position near the wave crest section), which shows that the waveform of the additional signal obtained by calculation by the minimum error method is very accurate. Compared with a series of methods for removing signal interference, the minimum error method overcomes the short plate of the method, the obtained real signal is more straight in the phase range except the peak signal, and the accuracy of the real signal is greatly improved while the calculation speed meets the application requirement. The sawtooth characteristics are eliminated by the plate shape distribution calculated by the signals after deflection removal, and the method is ideal in effect of improving the detection precision of the whole-roller plate shape meter.

Preferably, as shown in fig. 2, two rows of through holes 1 are formed in the circular side surface of the plate-shaped detection roller along the axial direction of the roller body, a plurality of sensors 2 are arranged in the through holes 1 at intervals, and the plurality of sensors 2 on each circumference form a detection unit 3.

Preferably, the sensors 2 are piezoelectric sensors, and the output of each sensor 2 is connected to the input of a plate-shaped signal processing computer.

The shape meter is commonly used in the online detection process of the shape (flatness) error of the cold-rolled strip, and the detection equipment is a shape detection roller. When the strip is tensioned to wrap the surface of the strip, the residual stress distribution of the cold-rolled strip caused by flatness error is quantitatively output in a relative length difference mode through the magnitude of the radial force which is subjected to uneven distribution along the axial direction of the strip, and the obtained distribution is the strip shape (flatness) distribution of the strip measured by each detection unit of the plate-shaped roller.

The calculated AD value data is displayed in the form of a relative length difference histogram and a plate shape target curve in the plate shape instrument matching software of the plate shape monitoring panel, and an operator or a closed loop system can sequentially adjust the rolling mill control means to optimize the product plate shape index.

The above-mentioned embodiments are merely illustrative of the preferred embodiments of the present invention, and do not limit the scope of the present invention, and various modifications and improvements made to the technical solution of the present invention by those skilled in the art without departing from the spirit of the present invention shall fall within the protection scope defined by the claims of the present invention.