JPH03258410A - Abnormality diagnostic method for plate thickness on tandem rolling machine - Google Patents

Abstract

PURPOSE:To allow to control an abnormality of plate thickness to the minimum by comparing the statistical moment of the exit side of each rolling stand of the normal time with the abnormal time generated of the plate thickness, and extracting the rolling stand expressing that the plate thickness variance of the exit side of the above among each rolling stand is a same state with the plate variance of the exit side of the final rolling stand. CONSTITUTION:At the generating time of the abnormality of the plate thickness, the statistical moment of the variance range of the plate thickness with the plate thickness measuring device 3 of the exit side of each rolling stand 1 is calculated with the arithmetic processing and deciding device 5. These are compared with the statistical moment calculated at the normal time of the plate thickness by every rolling stand 1. The rolling stand 1 to generate the abnormality of the plate thickness can be identified by extracting the rolling stand expressing the same state with the plate thickness variance of the exit side of each rolling stand 1. Therefore, the rolling stand to generate the abnormality of the plate thickness can be accurately identified and diagnosed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to diagnosis for identifying an abnormal rolling stand when an abnormality occurs in the operation of a tandem rolling mill composed of a plurality of rolling stands.

[Conventional technology]

In recent years, requirements for sheet thickness accuracy in hot and cold open rolling have become increasingly strict, and in the event of an abnormality in operation, it is necessary to immediately identify the rolling stand that caused the abnormality and take measures.

As a method for determining abnormalities in chemical plants, the causal relationships between process variables are represented in a signed directed graph, and logical processing is performed based on information on whether each process variable exceeds an upper or lower limit. A method for determining the origin of an abnormality is disclosed in Japanese Patent Publication No. Sho 62-53760.

However, in tandem rolling mills, the causal relationship between process variables is not clear. Further, since there is a close relationship between the variation pattern of the process variable and the cause of the abnormality, it is impossible to accurately determine the origin of the abnormality based on whether the upper limit value or the lower limit value is exceeded.

Therefore, in the end, there is no established method for identifying the source of an abnormality, and the operator still relies on his or her five senses.
Currently, this is done by visually following the thickness chart.

[Problem to be solved by the invention]

The problem to be solved in the present invention is to establish a method that can accurately grasp the singularity of the plate thickness abnormality that occurs in the operation of a tandem rolling mill and accurately identify the stand where the abnormality occurs. The objective is to find a method that can accurately identify stands where abnormalities occur online.

[Means to solve the problem]

The present invention calculates the first to fourth order statistical moments of the plate thickness variation width on the outlet side of each rolling stand when a plate thickness abnormality occurs,
These and the statistical moment calculated when the plate thickness is normal are compared for each rolling stand, and the rolling stands in which the plate thickness variation at the outlet side of each rolling stand is the same as the plate thickness variation at the outlet side of the final rolling stand are extracted. As a result, the above objectives were achieved.

[Effect]

The present invention makes it possible to extract the characteristics of plate thickness variation by calculating the statistical moment of plate thickness variation, thereby making it possible to effectively identify abnormal rolling stands.

〔Example〕

When diagnosing plate thickness abnormalities, the result of determining the origin of the abnormality greatly influences subsequent detailed diagnosis. In this example, a statistical moment was applied as the logic for determining the starting point.

With reference to FIG. 1 showing a specific method of the present invention, (a)
indicates a sequentially extracted time series waveform. This time series waveform can be expanded into the following first- to fourth-order statistical moments.

First moment rTz= f x p(x)dxSecond moment J= f (x-m+)”p(x)dxThird moment m3= f (x-m,)'p(x)d
x fourth moment m4= f (x-m,)'p(x
)dx where X is the amplitude of the time series waveform, p (x) is
shows the expected value of

The first moment is the average value, and the square root of the second moment is the standard deviation (m 2°゛5). The third and fourth moments are normalized to skewness (m, 7m2'・5) and koutosis (rn, /m
22). These are effective in determining the occurrence of plate thickness abnormalities. For example, the average value (b-1) is effective for determining fluctuations with an offset, the standard deviation (b-2) is effective for determining wide fluctuations, and the koutosis (b-3) is effective for determining shocking fluctuations.

Furthermore, the present invention applies the above-mentioned method of moments to the abnormality origin determination logic. In other words, it is based on the logic that the mode of fluctuation in the final gauge continues from the abnormal stand.

Therefore, the following logic can be formed based on this assumption.

■ Calculate the average value, standard deviation, and cuttosis for the plate thickness signal at the exit side of the final stand, compare these with normal values, and select parameters that show changes.

■ Calculate the parameters selected in (■) for the exit side plate thickness signal of each stand except the final stand, and compare it with the normal value at each stand to check whether there is any change.

■ The part where the stand abnormality occurs has the largest change in ■.

FIG. 2 shows a time-series direct reading fluctuation waveform of the plate thickness at the exit side of each rolling stand of stand #1, stand #2, and stand #6 of a specific tandem rolling mill.

Expanding this to the above-mentioned -th to fourth-order statistical moments,
Calculate the mean value, standard deviation, and koutosis for each,
FIG. 3 is a diagram showing the change.

The horizontal axis of the figure shows data numbers, 1 to 4 indicate normal times, and 5 indicates abnormal times. First, in stand #6, which is the final stand, among the average value, standard deviation, and kutosis, only the kutosis significantly increased compared to the normal time. So, when we look at the changes in #1 stand and #2 stand for Kutsis, we see a large increase in both stands. Therefore, it was found that the origin of the abnormality was the #1 stand, which was the most upstream side.

As a result, we were able to draw the conclusion that it is possible to determine the origin of an abnormality simply by applying the statistical moment.

FIG. 4 is a block diagram showing the configuration of a diagnostic device for carrying out the diagnostic method of the present invention. In the figure, 1 is a rolling stand, 2 is a plate (material to be rolled), 3 is a plate thickness gauge, 4 is a signal acquisition device, 5 is a calculation/judgment device, and 6 is a display device.

The steps of the diagnostic process in this diagnostic device will be explained according to the flowchart shown in FIG.

First, the plate thickness signal of each rolling stand 1 is read by the plate thickness gauge 3 and collected by the signal collecting device 4. Next, the calculation/judgment device 5 calculates each statistical moment of the final rolling stand plate thickness signal. Each calculated statistical moment is compared with the normal state, and the statistical moment with the largest change is extracted. The extracted statistical moments are calculated for the thickness signals of rolling stands other than the final rolling stand. Next, each rolling stand is compared with the statistical moment under normal conditions, and the rolling stand with the largest change is extracted.

The above extraction results are displayed on the display device 6.

〔Effect of the invention〕

According to the present invention, the following effects can be achieved.

(1) Even during operation, it is possible to accurately identify and diagnose rolling stands that cause plate thickness abnormalities.

(2) Since it uses statistical moments,
No particularly large-scale diagnostic equipment is required.

(3) Therefore, plate thickness abnormalities can be minimized.

[Brief explanation of drawings]

The accompanying drawings illustrate embodiments of the invention. 1 to 3 are diagrams showing an embodiment when statistical moments are applied, FIG. 4 is a block diagram showing the configuration of a diagnostic device for carrying out the present invention, and FIG. 5 is a flow chart. 1: Rolling stand 2: Plate (material to be rolled) 3: Plate thickness gauge
4 Double sign collection device 5; Calculation/judgment device 6: Display device Fig. 4 Fig. 5

Claims (1)

[Claims] 1. When a plate thickness abnormality occurs, calculate the statistical moment of the width of plate thickness variation on the outlet side of each rolling stand, and combine these and the statistical moment calculated when the plate thickness is normal for each rolling stand. By comparing and extracting the rolling stands where the plate thickness variation on the outlet side of each rolling stand shows the same pattern as the plate thickness variation on the final rolling stand outlet side, the rolling stand that causes the plate thickness abnormality is identified. How to diagnose thickness abnormalities. 2. In the description of claim 1, the statistical moment of the width of plate thickness variation at the exit side of each rolling stand is developed into first to fourth order statistical moments, and each statistical moment is calculated based on each factor of plate thickness variation. Diagnosis method for plate thickness abnormalities in tandem rolling mills as basis for abnormality diagnosis. 3. In the description of claim 2, the first-order to fourth-order statistical moments are expanded, and the first-order statistical moment is used as a basis for determining offset fluctuation, and the second-order statistical moment is used as a basis for determining fluctuation width. A method for diagnosing plate thickness abnormalities in a tandem rolling mill that uses cutosis as the basis for determining shock fluctuations. 4. A method for diagnosing a plate thickness abnormality in a tandem rolling mill according to claim 1, wherein a statistical moment is applied to the abnormality origin determination logic.