JPH03160314A - Thickness measuring method using radiation and apparatus therefor - Google Patents

Abstract

PURPOSE:To measure the thickness with high accuracy by detecting the radioactive rays generated when the radioactive rays are irradiated to an object to be measured, converting the radioactive rays to the thickness, and then presuming a true value by a Kalman filter. CONSTITUTION:Radioactive rays generated from an object S to be measured after being projected from a source 1 to the object S are detected by a radiation detector 2 consisting of a scintillator 21 and a photomultiplier tube 22, thereby to obtain a required electric signal. The electric signal is amplified to a predetermined level by an amplifier 3, and only a detecting signal is extracted by a pulse height selector 4 and counted by a counter 5. A microcomputer 6 reads an output of the counter 5 every cycle of a predetermined time and converts the same to the thickness. The latest (n) points are linearly fed back by the method of least squares, and the true thickness is presumed by a Kalman filter.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a radiation thickness measurement method for measuring the thickness of steel materials, etc. by transmission or reflection of radiation. ``Prior art and problems to be solved by the invention'' In steel plate rolling lines and steel pipe OrA building lines, radiation is projected onto the object to be measured as a cleaning means, and the transmitted or reflected radiation is collected. What is counted and converted into thickness, i.e. radiation 1 {
A method of measurement is used that utilizes the decay of a chemical substance, but in this case, noise is included, which is thought to be mainly due to statistical fluctuations in the dose. One example is a noise removal means using an analog method, which is disclosed in Japanese Patent Application Laid-Open No. 62-238404. This analog noise removal means passes the radiation detection signal from the detector through a 4iF band filter for noise removal, subtracts the filtered signal from the unfiltered raw signal to obtain a difference signal, and The difference signal is compared with a reference signal, and if it is above the reference level, the raw signal is output as is as a thickness signal, and if it is below the reference level, the filtered signal is output as a thickness measurement signal. In other words, if the raw signal is uniformly passed through an 11-band filter, the thickness signal component existing in the same frequency range as the noise will be removed along with the noise, so the thickness signal component in the noise frequency range will be at the reference level. When the difference is larger than the standard level, the raw signal is output as it is, since the S/N does not decrease much even if the noise is not removed, and when it is smaller than the reference level, the thickness signal in the noise frequency range is output. Since the component is not substantially included in the raw signal, the filtered tS signal with noise removed is output. However, this analog noise removal means has the following problems: i) Radiation fluctuations are statistical, and when they are sampled in a time series, they are normally distributed regardless of frequency, but for example, when DC current MW3 is If detected with an analog type detector such as a scintillation counter, the characteristics of a specific frequency range will be added due to the time constant of the circuit, and if detected with a digital type detector such as a scintillation counter, the characteristics of a specific frequency range will be added due to the measurement time (interval, period). Therefore, among the statistical fluctuations of radiation, noise can be reduced (Ii) only in a certain frequency band, but it cannot be reduced in that peak frequency range. The phase will shift and the waveform of the original signal will change.ii)9
If a thickness signal component exists in the gi sound frequency range, it will still contain noise. iii) When there is a steep wear change at the tip, t&end, such as a steel pipe rolled with a small-diameter seamless stretch reducer, if this is detected at high speed, it is difficult to distinguish whether it is a signal component or noise. It is impossible to process with fixed standard values. Due to these problems, highly accurate thickness measurements are still not possible. Therefore, the present invention aims to solve the above problems and enable highly accurate measurement by processing statistical fluctuations in radiation using a Kalman filter using its statistical characteristics. "A Thousand Steps to Solve the Problem" In order to achieve the above object, the invention of claim 1 projects radiation onto the object to be measured whose thickness is to be measured, and emits radiation from the object through transmission and reflection. A method for measuring radiation abrasion, characterized in that radiation is detected by a radiation detector, the detected signal is counted and converted into thickness, and then a true value is estimated by a Kalman filter. The invention consists of a radiation source that projects radiation onto an object to be measured whose thickness is to be measured, a radiation XIA detector that converts radiation emitted from the object through transmission and reflection into a predetermined electrical signal, and a radiation a wave height selector that extracts only the required detection signal from the amplitude signal, a counter that counts the detection signal, and a Kalman filter that reads and processes the output of the counter at a fixed time period. ``Function'' of a radiation thickness measuring device characterized by comprising a microcomputer with a JII'Ii program. Processed and from the Kalman filter it depends on the thickness ^ ^
^ is output. In Heko's equation, Xk-+ is the previous estimated value, and Xi is ill
The edge value, K is the gain, and C1 is the coefficient. ``Example'' FIG. 1 shows the circuit configuration of a device according to the present invention. In this device, radiation is projected from a radiation source l toward an object to be measured S, and the radiation transmitted from the object S is detected by a radiation detector 2 consisting of a scintillator 21 and a photomultiplier tube 22. Detect and obtain the required electrical signal, increase the width of the signal to the required level with the amplifier 3, and pulse height selector 4.
Only the detection signal is extracted, the detection signal is counted by a counter 5, and the result is subjected to arithmetic processing by a microcomputer 6 equipped with a Kalman filter. As shown in Figure 2, the processing performed by the micro combinator 6 is to read the output of the counter 5 at a fixed time period, convert it into thickness observation data, and then calculate the number of counts. fi and calculate the coefficient Co. required for the Kalman filter. Calculate C+ and r&V, and from these and the above observed data, use a Kalman filter to calculate the true numerical value I.
I! and {IL1 ask. In the above least squares method, the III constant value x-co +CI ・t is obtained at several points before the i-final n point of the observation data, but here, in order to use it for the Kalman filter,
Calculate the coefficients Co, C+ related to the constant value X and the component 11IiV found from these. In addition, the formula X-CO→-
C+-L is a function of time t when {11 constant value X is locally regarded as a straight line, Go represents the initial condition, and C1 represents the slope of the straight line. ■ is unbiased loss, meaning statistical noise, errors due to Iril regression, etc. For example, if the latest n point is n-5 and the time pitch Δt=1, (n-1)/2-(5-1
) / 2 -211! The previous quasi-definite value
, 2 111i1 observation points i- before and after this center
2, each observation value at i-1, i+l, +4-2 is Xi
-z, Xi-wealth, Xi++ X1+2, then the coefficients CD, CI, and division ■ regarding the estimated value X can be obtained by the following formula. 4+1+1+4 Ca×CI−L=5 Next, in the Kalman filter, the gains K,
The 11 constant value X1 of the true value related to rg-sensitivity and the error portion P (same as next) are calculated, and the estimated value Xi is output. Note that the subscript k-+ indicates the previous value. 5 Ptomi V ``Effects of the Invention'' According to the present invention, statistical fluctuations in radiation can be accurately eliminated using its statistical characteristics, without any unreasonableness unlike in the case of conventional analog noise removal means. It is possible to ensure the specified measurement accuracy. In addition, even in cases where the wall thickness changes rapidly at high speeds, such as steel pipes rolled with a small-diameter seamless stretch reducer, the thickness can be measured accurately and with high precision. Furthermore, the device can be simply constructed using a microcomputer.

[Brief explanation of the drawing]

The drawings show embodiments of the present invention; FIG. 1 is a Pronk diagram showing the circuit configuration of a device according to the present invention, and FIG. 2 is a flowchart of the microcomputer in the block diagram. l...Radiation source 2...Radiation detector 2l...
・Scintillator 22... Photomultiplier tube 3... Amplifier 4... Wave height selector 5... Counter
6...Microcomputer S...Object to be measured

Claims (2)

[Claims] (1) Radiation is projected onto the object to be measured whose thickness is to be measured, the radiation emitted from the object through transmission and reflection is detected by a radiation detector, the detected signals are counted and converted into thickness, and then the Kalman filter is applied. A radiation thickness measurement method characterized by estimating a true value. (2) A radiation source that projects radiation onto the object to be measured whose thickness is to be measured, a radiation detector that converts the radiation emitted from the object through transmission and reflection into a predetermined electrical signal, and a radiation detector that converts the electrical signal to a predetermined level. A microcomputer consisting of an amplifier that amplifies up to A radiation thickness measuring device comprising: a computer;