Processing unit

Abstract

To appropriately discriminate an abnormal value from among measurement values of the thickness of an object by a pulse eddy current flaw detection.SOLUTION: A processor 5 includes: a collector 54 for continuously collecting measurement values of the thickness of an object 9 measured by pulse eddy current flaw detection; and a discriminator 55 for discriminating an abnormal value from among measurement value groups as sets of measurement values collected by the collector 54. The discriminator 55 executes first discrimination for discriminating a first abnormal value from among measurement value groups on the basis of measurement values during a first predetermined period from among measurement value groups, and second discrimination for removing the first abnormal value from measurement values during a second period longer than the first period from among measurement value groups and discriminating a second abnormal value from among measurement value groups on the basis of a change with time of measurement values during the second period when the first abnormal value is removed.SELECTED DRAWING: Figure 5

Description

Technical Field

[0001] The technology disclosed herein relates to a processing device.

Background Art

[0002] Conventionally, a technique for discriminating an outlier included in a measurement value has been known. For example, the device disclosed in Patent Document 1 removes specific measurement data from the population of measurement data so that the variation of the measurement data in the measurement data group is smaller than the variation of the measurement data in the population based on the difference between the measurement data in the population of measurement data.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in the measurement of the thickness of an object by pulsed eddy current testing, the measurement is carried out over a long period. It is difficult to appropriately discriminate outliers for such measurement values collected over a long period.

[0005] The technology disclosed herein has been made in view of such a point, and its object is to appropriately discriminate outliers from the measurement values of the thickness of an object by pulsed eddy current testing.

Means for Solving the Problems

[0006] The apparatus disclosed herein comprises a collector that continuously collects measurement values ​​of the thickness of an object measured by pulsed eddy current testing, and a discriminator that identifies abnormal values ​​from a group of measurement values ​​which is a collection of the measurement values ​​collected by the collector. The discriminator performs a first discrimination, which identifies a first abnormal value from the group of measurement values ​​based on the measurement values ​​during a predetermined first period, and a second discrimination, which removes the first abnormal value from the measurement values ​​during a second period which is longer than the first period, and identifies a second abnormal value from the group of measurement values ​​based on the change in the measurement values ​​over time during the second period from which the first abnormal value has been removed. [Effects of the Invention]

[0007] According to the processing apparatus of this disclosure, abnormal values ​​can be appropriately identified from the thickness measurements of an object obtained by pulsed eddy current testing. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is an explanatory diagram showing the configuration of the measurement system. [Figure 2] Figure 2 is a flowchart of the eddy current measurement process using the control device. [Figure 3] Figure 3 is a block diagram showing the configuration of the control system of the controller of the processing unit. [Figure 4] Figure 4 is a graph showing the time evolution of the voltage signal V(t) detected by the first detector, i.e., the time evolution of the eddy current. [Figure 5] Figure 5 is a graph showing an example of a group of measured values ​​before the first discrimination is performed. [Figure 6] Figure 6 is a graph showing an example of a group of measured values ​​before the second discrimination is performed. [Figure 7] Figure 7 is a graph showing an example of a set of measurement values ​​used to determine the rate of change of the measured values. [Modes for carrying out the invention]

[0009] The following describes exemplary embodiments in detail with reference to the drawings. Figure 1 is an explanatory diagram showing the configuration of the measurement system 100.

[0010] -Outline of the measurement system- The measurement system 100 comprises a probe 1, a control device 3 that controls the probe 1, and a processing device 5 that obtains a measurement value from the detection result of the probe 1 and processes the measurement value. The measurement system 100 measures the thickness of an object by pulsed eddy current testing (PEC). The probe 1 generates eddy currents in the object 9 and detects the generated eddy currents. In this example, the object 9 is a pipe. The pipe is a metal pipe through which fluid flows.

[0011] -probe- Probe 1 is a non-contact probe and is positioned in close proximity to the object 9. Note that "non-contact" means that it can be used without contact, but does not exclude its use in a contact state. Probe 1 generates eddy currents in the object by forming a fluctuating magnetic field. Probe 1 also detects the change in the eddy currents generated in the object as an induced voltage.

[0012] Probe 1 comprises an excitation coil 11 that generates eddy currents in an object 9 using magnetic flux from an excitation current, and a detection coil 12 that detects the eddy currents in the object 9. Probe 1 may further comprise a casing 13 that houses the excitation coil 11 and the detection coil 12. Probe 1 generates eddy currents in the object 9 using the excitation coil 11 and detects the generated eddy currents with the detection coil 12. In this example, probe 1 has two sets of excitation coils 11 and detection coils 12, with each set of excitation coils 11 and detection coils 12 forming one set. The excitation coils 11 and detection coils 12 are arranged such that the axis of each set of excitation coils 11 and the axis of each set of detection coils 12 are in a straight line. In this case, the detection coil 12 is positioned closer to the object 9.

[0013] When a current is applied to the excitation coil 11, a magnetic field is formed in the direction of its axis. Currents are applied to one excitation coil 11 and the other excitation coil 11 so as to form magnetic fields opposite to each other in the direction of the axis. For example, magnetic flux is generated from one excitation coil 11 toward the object 9, and magnetic flux is generated from the object 9 toward the other excitation coil 11. As a result, most of the magnetic flux emitted from one excitation coil 11 exits in the direction of the axis of one excitation coil 11 and enters the object 9, passes through the object 9 in a substantially arc shape, and heads toward the other excitation coil 11 in the direction of the axis of the other excitation coil 11 and enters the other excitation coil 11. By varying the current applied to the excitation coil 11, the magnetic field generated in the object 9 varies, and eddy currents are generated in the object 9.

[0014] On the other hand, eddy currents generated in the portion of the object 9 near the detection coil 12 form magnetic flux that penetrates the detection coil 12. When the magnetic flux penetrating the detection coil 12 changes, an induced electromotive force is generated in the detection coil 12. The detection coil 12 detects the eddy currents in the object 9 by detecting this induced electromotive force.

[0015] The probe 1 may further include a temperature sensor 21 that detects the temperature of the object 9. The temperature sensor 21 may be disposed within the casing 13. The temperature sensor 21 is, for example, a thermocouple.

[0016] -Control device- The control device 3 includes an exciter 31 that applies an excitation current to the excitation coil 11, a first detector 32 that detects transient changes in the eddy currents of the object 9, a second detector 33 to which the output of the temperature sensor 21 is input, a communicator 34 that communicates with external devices, a memory 35 that stores various information, and a controller 36 that controls at least the exciter 31, the first detector 32, the second detector 33, the communicator 34, and the memory 35.

[0017] The exciter 31 supplies a pulsed excitation current to the excitation coil 11. The exciter 31 includes a pulse generator 31a that generates a pulse signal, and a transmission amplifier 31b that amplifies the pulse signal from the pulse generator 31a and outputs it as an excitation current.

[0018] The first detector 32 detects the induced electromotive force generated in the detection coil 12 in response to the eddy current of the object 9. The transient change in the induced electromotive force generated in the detection coil 12 is related to the transient change in the eddy current generated in the object 9. The first detector 32 has at least a receiving amplifier 32a that amplifies the voltage generated in the detection coil 12. The first detector 32 may further have a filter that performs a filtering process on the voltage signal.

[0019] The second detector 33 receives the output from the temperature sensor 21. The second detector 33 may have an amplifier that amplifies the output from the temperature sensor 21.

[0020] The communicator 34 performs wireless communication with an external device. For example, the communicator 34 transmits the voltage signal (i.e., the detection signal) detected by the first detector 32 and the detection signal detected by the second detector 33 to the processing device 5.

[0021] The storage 35 stores programs and various data executed by the controller 36. For example, the storage 35 stores a control program. The storage 35 stores the detection signal of the first detector 32 and the detection signal of the second detector 33. The storage 35 is formed of a non-volatile memory, a HDD (Hard Disc Drive), an SSD (Solid State Drive), or the like.

[0022] The controller 36 controls the entire control device 3. The controller 36 performs various calculation processes. For example, the controller 36 is formed by a processor such as a CPU (Central Processing Unit). The controller 36 may also be formed by an MCU (Micro Controller Unit), MPU (Micro Processor Unit), FPGA (Field Programmable Gate Array), PLC (Programmable Logic Controller), system LSI, etc.

[0023] For example, the controller 36 causes the exciter 31 to output an excitation current for a predetermined period, and then acquires a detection signal from the first detector 32 after the output of the excitation current stops. The controller 36 also acquires a detection signal from the second detector 33. The controller 36 stores the detection signals from the first detector 32 and the second detector 33 in the memory 35, and transmits the detection signals stored in the memory 35 to the processing unit 5 via the communicator 34 at a predetermined timing.

[0024] The measurement of eddy currents by the control device 3 will be explained in detail using Figure 2. Figure 2 is a flowchart of the eddy current measurement by the control device 3.

[0025] In step S101, the controller 36 measures the temperature of the object 9. That is, the controller 36 causes the second detection unit 33 to acquire the output of the temperature sensor 21, and stores the acquired detection signal in the memory unit 35. As will be described later, the temperature measurement of the object 9 may be repeated, and the memory unit 35 accumulates the measured temperatures.

[0026] Next, in step S102, the controller 36 measures the eddy currents of the object 9 (more specifically, measures the transient changes of the eddy currents). The controller 36 performs one set of eddy current measurements for a predetermined number of repetitions n. Specifically, the controller 36 causes the exciter 31 to output an excitation current to the excitation coil 11. The excitation coil 11 forms a magnetic field in the direction of its axis when the excitation current is applied. One excitation coil 11 and the other excitation coil 11 form magnetic fields that are opposite to each other in the direction of their axes. For example, a magnetic flux is generated from one excitation coil 11 toward the object 9, and a magnetic flux is generated from the object 9 toward the other excitation coil 11. The controller 36 stops the output of the excitation current and causes the first detector 32 to detect the eddy currents generated in the object 9. The controller 36 continues the detection of voltage signals by the first detector 32 for a predetermined period and stores the detected electrical signals in the memory 35. As a result, the controller 36 measures the transient change (change over time) of the induced electromotive force of the detection coil 12, that is, the transient change of the eddy current generated in the object 9. The controller 36 repeats this eddy current measurement a number of times n.

[0027] Subsequently, in step S103, the controller 36 measures the temperature of the object 9. The process in step S103 is the same as the process in step S101.

[0028] In step S104, the controller 36 determines whether the temperature change of the object 9 is greater than a predetermined standard. For example, the controller 36 determines whether the temperature difference between the temperature measured in step S103 and the previously measured temperature is greater than a predetermined value α. The controller 36 makes this determination by calculating the temperature difference from the two most recent measured temperatures stored in the memory 35.

[0029] If the temperature change of the object 9 is not greater than a predetermined standard, that is, if the temperature difference is less than or equal to a predetermined value α, the controller 36 determines in step S105 whether the number of sets of eddy current measurements for the number of repetitions n has reached a predetermined number of sets m. If the number of sets of measurements has not reached the number of sets m, the controller 36 returns to step S102 and repeats the process from step S102 onward.

[0030] Steps S102 to S105 are repeated until the number of measurement sets reaches the set count m. In step S106, the controller 36 transmits the eddy currents (detection signals) and temperatures (i.e., detection signals) stored in the memory 35 to the processing unit 5. In other words, the controller 36 transmits the eddy currents for the number of repetitions n × the number of sets m, and the temperatures for the number of sets m, to the processing unit 5.

[0031] Once the controller 36 has finished transmitting the eddy current and temperature data, it terminates the eddy current measurement.

[0032] On the other hand, in step S104, if the temperature change of the object 9 is greater than a predetermined standard, that is, if the temperature difference is greater than a predetermined value α, the controller 36 stops measuring the eddy currents and proceeds to step S106. In step S106, the controller 36 transmits the eddy currents and temperature stored in the memory 35 up to the time the measurement was stopped to the processing unit 5. At this time, the controller 36 also notifies the processing unit 5 that the measurement of the eddy currents was stopped because the temperature change of the object 9 was large.

[0033] The controller 36 repeats these eddy current measurements at a predetermined measurement cycle. This allows for the periodic acquisition of measurement results for multiple eddy currents. The temperature of the object 9 is monitored during the measurement of multiple eddy currents. Specifically, one temperature measurement is performed for every n eddy current measurements. The transient changes in eddy currents depend on the temperature of the object 9. One purpose of measuring the temperature of the object 9 is to correct the thickness of the object 9, which is determined from the transient changes in eddy currents, by the temperature of the object 9. Another purpose of measuring the temperature of the object 9 is to remove eddy currents from the thickness measurement of the object 9 when the temperature change of the object 9 is too large. If the temperature change of the object 9 is too large, there is a risk that the thickness of the object 9 cannot be accurately corrected by the measured temperature due to delays in detection by the temperature sensor 21, etc. Therefore, if the temperature change of the object 9 is too large, the controller 36 stops the eddy current measurement in order to remove the eddy currents measured at that time from the thickness measurement of the object 9. Furthermore, the controller 36 formally transmits the eddy currents and temperatures measured up to that point to the processing unit 5. The processing unit 5 does not use the received eddy currents and temperatures to measure the thickness of the object 9.

[0034] Furthermore, if the measurement system 100 is equipped with multiple probes 1, the controller 36 measures the eddy currents for each probe 1 sequentially or in parallel. If the measurement system 100 is equipped with a mover for moving the probe 1 on the surface of the object 9, the controller 36 moves the probe 1 to another measurement position and performs eddy current measurement once it has finished measuring the eddy currents at one measurement position. The controller 36 continues measuring the eddy currents by having the probe 1 scan the surface of the object 9.

[0035] - Processing Unit - The processing unit 5 is formed by a computer or a computer network (a so-called cloud). As shown in Figure 1, the processing unit 5 includes a communicator 51 that communicates with external devices, a memory 52 that stores various information, and a controller 53 that determines the thickness of the object 9 based on the transient changes of the detected eddy currents.

[0036] The communication device 51 communicates wirelessly with external devices. For example, the communication device 51 receives detection signals from the control device 3 (i.e., detection signals from the probe 1 and the temperature sensor 21).

[0037] The controller 53 controls the entire processing unit 5. The controller 53 performs various calculations. For example, the controller 53 is formed by a processor such as a CPU (Central Processing Unit). The controller 53 may also be formed by an MCU (Micro Controller Unit), MPU (Micro Processor Unit), FPGA (Field Programmable Gate Array), PLC (Programmable Logic Controller), system LSI, etc.

[0038] The memory unit 52 stores the program executed by the controller 53 and various data. For example, the memory unit 52 stores the control program. The memory unit 52 is made of non-volatile memory, an HDD (Hard Disk Drive), or an SSD (Solid State Drive), etc. The memory unit 52 stores eddy currents measured by the probe 1, temperature measured by the temperature sensor 21, and information necessary to calculate the thickness of the object 9.

[0039] Figure 3 is a block diagram showing the configuration of the control system of the controller 53 of the processing unit 5. The controller 53 implements various functions by reading a control program from the memory 52 into memory and loading it. Specifically, the controller 53 functions as a collector 54 that measures the thickness of the object 9 from eddy currents and collects the measured values, a discriminator 55 that identifies abnormal values ​​from the group of measured values, and a calculator 56 that determines the rate of change of the thickness of the object 9 based on the group of measured values.

[0040] =Collector= The collector 54 determines the thickness of the object 9 as a measured value based on the eddy currents measured by the probe 1. However, if the collector 54 stops measuring the eddy currents due to a large temperature change in the object 9, it will not calculate the thickness of the object 9. Known techniques are used to measure the thickness based on eddy currents. For example, eddy currents attenuate as they penetrate the object 9. Eddy currents gradually attenuate from the surface of the object 9 (the surface facing the probe 1) until they reach the back surface, and then attenuate rapidly once they reach the back surface. The voltage detected by the detection coil 12 also changes in a similar manner to the eddy currents. There is a correlation between the thickness of the object 9 and the time it takes for the detected voltage to attenuate rapidly. The controller 53 calculates the thickness of the object 9 based on the time it takes for the detected voltage to attenuate rapidly.

[0041] Let's explain the thickness measurement in more detail. Figure 4 is a graph showing the time evolution of the voltage signal V(t) detected by the first detector 32, i.e., the time evolution of the eddy current. The graph in Figure 4 is a log-log graph. In Figure 4, the voltage signal V0(t) is the voltage signal of the object 9 having a thickness d0, and the voltage signal V1(t) is the voltage signal of the object 9 having a thickness d1 that is thinner than the thickness d0. The eddy current measured by the probe 1, which is transmitted from the control device 3 to the processing device 5, is the voltage signal V(t) shown in Figure 4.

[0042] Eddy currents attenuate as they penetrate the object 9. The eddy currents gradually attenuate from the surface of object 9 (the surface facing probe 1) until they reach the back surface, where they attenuate rapidly. The voltage signal V(t) also shows a similar change to the eddy currents. In other words, the transient change in the voltage signal V(t) corresponds to the transient change in the eddy currents. The change in the voltage signal V(t) until the eddy currents reach the back surface of object 9 is represented linearly on a log-log graph. After that, the voltage signal V(t) attenuates rapidly. The voltage signal V(t) that changes in this way can be expressed by the following equation (1).

[0043]

number

[0044] As can be seen from equation (1), the mode of change of the voltage signal V(t) switches over time τ. For the sake of explanation, τ will be referred to as the "decay time" below. The decay time τ is expressed by the following equation (2).

[0045] τ = σμd 2 ...(2) Here, σ is the conductivity of object 9, μ is the magnetic permeability of object 9, and d is the thickness of object 9.

[0046] In other words, the decay time τ changes depending on the thickness d of the object 9. Assuming that the conductivity σ and permeability μ of the object 9 are constant, the decay time τ changes depending on the thickness d of the object 9. Also, even if the decay time τ and thickness d change, τ / d 2 This is constant. Therefore, if we know the decay time τ0 for a known thickness d0 and the decay time τx for an unknown thickness dx, we can determine the unknown thickness dx based on the following equation (3).

[0047]

number

[0048] The processing unit 5 uses the known thickness of the object 9 as the reference thickness d0, and has previously acquired the transient change of the voltage signal V0(t) at the reference thickness d0 and the temperature T0 of the object 9 at that time, and stores them in the memory unit 52. Hereinafter, the voltage signal V0(t) will be referred to as the reference voltage signal V0(t). The memory unit 52 stores multiple voltage signals Vx(t) transmitted from the control unit 3 along with the temperature Tx. The collector 54 determines the thickness dx of the object 9 by comparing the transient change of each voltage signal Vx(t) with the transient change of the reference voltage signal V0(t). Specifically, the collector 54 uses equation (3) to determine the thickness dx after the change from the reference thickness d0, the reference decay time τ0, and the decay time τx.

[0049] Furthermore, the collector 54 corrects the calculated thickness dx based on the temperature Tx of the object 9. Specifically, in the calculation described above, it is assumed that the conductivity σ and permeability μ of the object 9 are constant. However, the conductivity σ and permeability μ of the object 9 are temperature-dependent. Therefore, the collector 54 uses the temperature T0 of the object 9 when the voltage signal V0(t) at the reference thickness d0 is obtained and the temperature Tx of the object 9 when the voltage signal Vx(t) at thickness dx is obtained to determine the corrected thickness dx'. Specifically, the collector 54 determines the corrected thickness dx' based on the following equation (4).

[0050] dx' = dx - α × ΔT × dx ... (4) Here, α is the temperature correction coefficient, and ΔT = T0 - Tx. The temperature correction coefficient α is predetermined based on the change in the voltage signal V(t) when the temperature of the object 9 is changed, or the change in thickness d obtained from equation (3) when the temperature of the object 9 is changed, and is stored in the memory 52.

[0051] As mentioned above, the temperature of object 9 is measured once for each set of eddy current measurements, that is, once for every n measurements of eddy currents. In this example, the temperature Tx included in one set is used as the temperature Tx for all voltage signals Vx(t) in the same set.

[0052] In this way, the collector 54 determines the thickness dx' of the object 9, taking into account its temperature. The collector 54 determines the thickness dx' of the object 9 for multiple voltage signals Vx stored in the memory 52. ​​However, the collector 54 does not determine the thickness dx' of the object 9 for voltage signals Vx for which it has been notified that the eddy current measurement has been stopped. The collector 54 stores the determined thickness dx' of the object 9 in the memory 52. ​​The memory 52 accumulates the measured values. The determined thickness dx' of the object 9 is the measured value of the thickness of the object 9. In this way, the collector 54 collects the measured values ​​of the thickness of the object 9. Since the eddy currents are measured periodically by the probe 1, the collector 54 collects the measured values ​​of the thickness of the object 9 periodically, i.e., continuously.

[0053] =Discriminator= Next, we will explain the discriminator 55 in detail.

[0054] The discriminator 55 identifies abnormal values, i.e., abnormal values, from a group of measurement values, which is a collection of multiple measurement values, compared to other measurement values. The discriminator 55 performs a first discrimination to identify a first abnormal value from the group of measurement values, and a second discrimination to identify a second abnormal value from the group of measurement values. The first and second discriminations differ in which period of measurement values ​​are referenced, the discrimination method, and how abnormal values ​​are handled.

[0055] First, regarding the first discrimination, the discriminator 55 identifies a first outlier from the group of measured values ​​based on the measured values ​​taken during a predetermined first period. The discriminator 55 determines a representative value of the measured values ​​taken during the first period and identifies any measured value that falls outside the first range determined based on the representative value as a first outlier. The first range is defined by an upper limit greater than the representative value and a lower limit less than the representative value. In other words, the first range includes the representative value and extends above and below the representative value.

[0056] The first outlier is permanently removed from the set of measured values. In other words, if the first classification is performed repeatedly, the measured value identified as the first outlier is not used in another first classification. For example, if the first outlier is included in the first period of another first classification, the first outlier is removed from the derivation of the representative value. Furthermore, the set of measured values ​​from which the first outlier has been removed is used in the calculation of the second classification and the rate of change of the measured values.

[0057] For example, the classifier 55 performs a first discrimination for each target measurement value in the group of measurement values. The target measurement value for the first discrimination is one or multiple measurement values ​​that are consecutive in time series. The first period is a predetermined period in the past with respect to the target measurement value, more specifically, the predetermined period immediately preceding it. In other words, the classifier 55 discriminates a first outlier from the target measurement value based on the measurement values ​​included in the immediately preceding first period with respect to the target measurement value. If the target measurement value changes, the first period also changes. The representative value is a value that represents the measurement values ​​included in the first period. The representative value is, for example, the mean or the median. The first range is, for example, a range defined based on the representative value. The first range is the range that includes the representative value. The upper limit of the first range is the representative value plus a predetermined upper limit, and the lower limit of the first range is the representative value minus a predetermined lower limit. The lower limit (the width between the representative value and the lower limit) may be greater than the upper limit (i.e., the width between the upper limit and the representative value).

[0058] Figure 5 is a graph showing an example of a group of measurements before the first classification is performed. In Figure 5, the horizontal axis represents time, and the vertical axis represents the thickness of object 9. The vertical lines (dashed lines) in the graph represent the division into daily intervals. In Figure 5, for the sake of explanation, each measurement is shown at equal intervals in the time direction. This is also the case in Figures 6 and 7, which will be described later.

[0059] In the example in Figure 5, the target measurement is the measurement included in the target day. The first period is the day immediately preceding the target measurement, i.e., the previous day. The representative value is the mean. The upper and lower limits of the first range are both greater than the variance relative to the mean.

[0060] The discriminator 55 separates the measurement values ​​included in the measurement value group by day, and designates the measurement values ​​included in each day as the target measurement values ​​for the first discrimination. The discriminator 55 calculates the average value of the measurement values ​​for the first period of the target measurement values, i.e., the measurement value of the previous day, and sets the first range of the target measurement values ​​based on the average value. In Figure 5, the dashed line represents the average value of the measurement values ​​of the target measurement values ​​for the previous day, and the dashed line represents the upper and lower limits of the first range of the target measurement values. The discriminator 55 discriminates measurement values ​​that fall outside the first range as first anomalies. In Figure 5, normal measurement values ​​are represented by circles, and first anomalies are represented by triangles. The discriminator 55 stores the first anomalies in the memory 52.

[0061] The classifier 55 repeats the first classification by changing the target measurement value. At this time, if the first period includes a measurement value that has already been classified as the first outlier, the classifier 55 removes the first outlier and calculates the average value. In other words, the first outlier is not used in the repeated first classification.

[0062] Thus, in the first discrimination, the discriminator 55 identifies a measurement value that deviates significantly from a relatively recent past measurement value as a first outlier. As can be seen from Figure 5, in the first range, the spread from the mean to the lower limit is greater than the spread from the mean to the upper limit. The measurement value is the thickness of the object 9, and the thickness of the object 9 can decrease over time. In other words, a measurement value below the past average value may be affected by the decrease in thickness. By making the spread from the mean to the lower limit greater than the spread from the mean to the upper limit, it is possible to prevent misidentification of a measurement value that has become smaller due to a decrease in thickness as a first outlier.

[0063] Next, regarding the second discrimination, the discriminator 55 identifies the second outlier from the group of measured values ​​based on the measured values ​​during the second period. The measured values ​​referenced for identifying the second outlier are those obtained by removing the first outlier from the measured values ​​during the second period, which is longer than the first period. The second period overlaps with or is the same as the period targeted for the second discrimination. The discriminator 55 finds an approximate straight line of the temporal change of the measured values ​​(excluding the first outlier) during the second period, and identifies any measured values ​​that fall outside the second range based on the approximate straight line as the second outlier. The second range is a range that extends above and below the approximate straight line.

[0064] The second outlier is removed from the set of measured values ​​when the rate of change is calculated by the calculator 56, which will be described later. However, when the discriminator 55 repeatedly performs the second discrimination, measured values ​​that were previously identified as second outliers are returned to the set of measured values, and the second discrimination is performed on the set of measured values ​​that includes the second outlier.

[0065] For example, the discriminator 55 performs a second discrimination for each target measurement value in the set of measurement values. The target measurement values ​​for the second discrimination are multiple measurement values ​​that are consecutive in time series. The second period is the period that includes the target measurement values ​​for the second discrimination. In other words, the measurement values ​​included in the second period are the target measurement values ​​for the second discrimination. The approximation line is a straight line that represents the relationship between the measurement values ​​and time. For example, the approximation line can be obtained by the least squares method. The second range is a range that extends above and below the approximation line by a predetermined threshold.

[0066] Figure 6 is a graph showing an example of a group of measured values ​​before the second discrimination test. The first outlier has been removed from the measured value group in Figure 6. In Figure 6, the thick solid line is the approximate straight line, and the dashed lines represent the upper and lower limits of the second range. In the example in Figure 6, the second period is ten days.

[0067] The discriminator 55 sets the measurement values ​​included in the group of measurement values ​​over a continuous period of ten days as the target measurement values ​​for the second discrimination. In other words, the second period is also ten days. The discriminator 55 finds an approximate straight line of the change in the target measurement values ​​over time and sets the second range based on the approximate straight line. The discriminator 55 sets the upper and lower limits of the second range by shifting the approximate straight line up or down by a predetermined threshold. For example, the threshold is set based on the variance with respect to the approximate straight line. For example, the threshold is the value obtained by multiplying the variance by a predetermined multiplier. The discriminator 55 discriminates the target measurement values ​​that fall outside the second range as second outliers. Specifically, the discriminator 55 discriminates the target measurement values ​​whose difference (absolute value) from the approximate straight line is greater than the threshold as second outliers. In Figure 6, normal measurement values ​​are represented by circles, and second outliers are represented by rectangles. The discriminator 55 stores the second outliers in the memory 52.

[0068] The discriminator 55 repeats the second discrimination on the target measurement values ​​from the group of measurement values, while changing the target measurement values, i.e., the target period. For example, the discriminator 55 repeats the second discrimination by shifting the target period by one day at a time. The interval for shifting the target period is not limited to one day, but may be two days at a time, or ten days at a time (i.e., the same interval as the target period). The target measurement values ​​for the second discrimination do not include measurement values ​​that have already been identified as first outliers, but include measurement values ​​that have already been identified as second outliers. In other words, if a measurement value that has already been identified as a first outlier is included in the second period, the discriminator 55 removes the first outlier and finds an approximate straight line. On the other hand, if a measurement value that has already been identified as a second outlier is included in the second period, the discriminator 55 finds an approximate straight line including the second outlier. In other words, in the repeated second discrimination, the first outlier is not used, but the second outlier is used.

[0069] In this way, in the second discrimination, the discriminator 55 identifies measurement values ​​that deviate significantly from the changes in measurement values ​​over a relatively long period as second outliers. In the second discrimination, measurement values ​​that change differently from the patterns of change in measurement values ​​are identified as second outliers.

[0070] =Calculator= Next, the calculator 56 will be explained in detail. The calculator 56 calculates the rate of change of the measured values ​​in the group of measured values ​​from which the first and second outliers have been removed. The calculator 56 calculates the rate of change of the measured values ​​during the calculation period. The calculator 56 finds an approximate straight line of the change in the measured values ​​over time during the calculation period and calculates the slope of the approximate straight line as the rate of change of the measured values. The approximate straight line is a straight line that represents the relationship between the measured values ​​and time. For example, the approximate straight line can be obtained by the least squares method.

[0071] Figure 7 is a graph showing an example of a set of measured values ​​used to determine the rate of change of the measured values. The first and second outliers have been removed from the set of values ​​in Figure 7. In Figure 7, the thick solid line is the approximate straight line of the set of measured values ​​from which the first and second outliers have been removed, the dashed line is the approximate straight line of the set of measured values ​​including the first and second outliers, and the double dashed line is the approximate straight line of the set of measured values ​​from which only the first outlier has been removed. In the example in Figure 7, the calculation period is ten days.

[0072] The calculator 56 sets the measurement values ​​included in the calculation period from the set of measurement values ​​as the target measurement values ​​for calculating the rate of change. In this example, the calculation period is the same as the second period of the second discrimination. The calculator 56 finds an approximate straight line of the change over time of the target measurement value and finds the slope of the approximate straight line. The slope of the approximate straight line is the rate of change of the measurement value, that is, the rate of change of the thickness of the object 9. The calculator 56 stores the slope of the approximate straight line as the rate of change of the measurement value in the memory 52.

[0073] The calculator 56 repeatedly calculates the rate of change of a target measurement while changing the target measurement from the group of measurement values, i.e., the calculation period. For example, the calculator 56 repeatedly calculates the rate of change by shifting the calculation period by one day at a time. The interval for shifting the calculation period is not limited to one day, but may be two days at a time, or ten days at a time (i.e., the same interval as the calculation period).

[0074] In this way, the calculator 56 calculates the rate of change of the measured values ​​in the group of measured values ​​from which the first and second outliers have been removed, thereby excluding measured values ​​that deviate significantly from the other measured values. As a result, the calculator 56 can accurately determine the rate of change of the measured values. Furthermore, as can be seen from Figure 7, even by removing only the first outlier from the group of measured values ​​(see dashed line), the accuracy of calculating the rate of change of the measured values ​​is improved compared to not removing both the first and second outliers (see dashed line).

[0075] As described above, the processing unit 5 performs a first discrimination and a second discrimination to identify abnormal values ​​from the measurement value group in two stages. Specifically, the discriminator 55 performs a first discrimination, which identifies a first abnormal value from the measurement value group based on the measurement value during the first period, and a second discrimination, which removes the first abnormal value from the measurement value group during the second period, which is longer than the first period, and identifies a second abnormal value from the measurement value group based on the change over time of the measurement value during the second period from which the first abnormal value was removed. In the first discrimination, a measurement value that shows a large change in value relative to a reference measurement value is identified as a first abnormal value. In the second discrimination, a measurement value that changes differently from the change pattern of the reference measurement value is identified as a second abnormal value. Since the first discrimination and the second discrimination have different perspectives for identifying abnormalities, the measurement values ​​used as the basis for discrimination are different from each other. In the first discrimination, the measurement value during the relatively short first period is used as the basis. In the second discrimination, since abnormalities are identified based on the change pattern of the measurement value, the measurement value during the second period, which is longer than the first period, is used as the basis. In this way, outliers are identified from two different perspectives, improving the accuracy of identifying outliers within the measurement data set. As a result, the accuracy of the measurement data improves.

[0076] In more detail, in the first determination, the first outlier is identified based on the representative value of the measurements during the first period. In other words, if multiple measurements are included in the first period, the first outlier is identified based on a value that represents the multiple measurements, such as the mean or median. This makes it possible to easily determine whether or not the change in the target measurement is large for multiple measurements included in the first period.

[0077] Specifically, in the first discrimination, the first range is determined based on the representative value of the measured values ​​during the first period, and measured values ​​that fall outside the first range are identified as first outliers. The gap between the lower limit of the first range and the representative value is larger than the gap between the upper limit of the first range and the representative value. In other words, in identifying first outliers, the discrimination criterion for values ​​below the representative value is more relaxed than the discrimination criterion for values ​​above the representative value. As a result, when the measured values ​​decrease over time, outliers can be appropriately identified by relaxing the discrimination criterion for values ​​below the representative value.

[0078] Furthermore, in the second discrimination step, an approximate straight line of the temporal change in measured values ​​during the second period is determined, and the second outlier is identified based on this approximate straight line. Specifically, a second range is set based on the approximate straight line, and measured values ​​that fall outside this range are identified as the second outlier. The approximate straight line represents the temporal change pattern of measured values ​​during the second period. Therefore, by identifying measured values ​​that fall outside this range as the second outlier, it is possible to easily identify measured values ​​whose changes differ significantly from the temporal change pattern of measured values ​​during the second period.

[0079] Furthermore, the second discrimination is repeatedly performed by changing the target measurement value within the set of measurement values. Here, the target measurement value for the second discrimination does not include measurement values ​​already identified as the first outlier, but includes measurement values ​​already identified as the second outlier. In other words, measurement values ​​identified as the first outlier are also removed in the subsequent second discrimination. On the other hand, measurement values ​​identified as the second outlier are used as measurement values ​​in the subsequent second discrimination. For example, measurement values ​​identified as the second outlier are used in the calculation of the approximation line in the subsequent second discrimination. This improves the accuracy of the approximation line of the time-dependent change in measurement values ​​during the second period. In other words, if the number of measurement values ​​used in the calculation of the approximation line is small, the accuracy of the approximation line decreases. The degree of anomaly in the second outlier is smaller than that of the first outlier. Therefore, when calculating the approximation line, the accuracy of the approximation line can be improved by increasing the number of measurement values ​​used in the calculation by also using measurement values ​​already identified as the second outlier.

[0080] Then, the calculator 56 of the processing unit 5 calculates the rate of change of the measured values ​​in the group of measured values ​​from which the first and second outliers have been removed. In short, the processing unit 5 can improve the accuracy of calculating the rate of change of the measured values ​​by identifying the first and second outliers in the group of measured values.

[0081] Furthermore, the control measure 3 of the measurement system 100 measures the temperature of the object 9 when measuring eddy currents, and if the temperature change is large, it removes the measured eddy currents from the calculation of the thickness of the object 9. In other words, the measurement system 100 identifies abnormal measurement values ​​based on the temperature change of the object 9 before the first and second discrimination. As a result, the measurement system 100 discriminates the measurement values ​​in three stages.

[0082] Other embodiments As described above, the embodiments described herein have been presented as examples of the technology disclosed in this application. However, the technology in this disclosure is not limited thereto and can be applied to embodiments that have been modified, replaced, added, or omitted as appropriate. Furthermore, it is possible to combine the components described in the embodiments above to create new embodiments. In addition, the components described in the attached drawings and detailed description may include not only components essential for solving the problem, but also components that are not essential for solving the problem, in order to illustrate the technology. Therefore, the mere presence of such non-essential components in the attached drawings and detailed description should not be immediately assumed to mean that those non-essential components are essential.

[0083] The above embodiment may also have the following configuration.

[0084] For example, the object 9 used for thickness measurement by pulsed eddy current testing is not limited to pipes. Any object can be used as the object 9, as long as eddy currents are generated.

[0085] Furthermore, the probe 1 is not limited to the configuration described above. For example, the probe 1 includes two sets of excitation coils 11 and detection coils 12, but there may be one set of excitation coils 11 and detection coils 12, or three or more sets. The excitation coils 11 and detection coils 12 do not have to be arranged so that their axes are in a straight line. A core may be provided inside the excitation coil 11 or the detection coil 12. The excitation coil 11 may be positioned closer to the object 9 than the detection coil 12. Furthermore, the detection unit of the probe 1 is not limited to the detection coil 12. The detection unit may be any device that can directly or indirectly detect the eddy currents of the object 9, for example, a Hall element.

[0086] The temperature sensor 21 is not limited to a thermocouple. The temperature sensor 21 only needs to be able to detect the temperature of the object; for example, it may be a thermistor. The number of temperature sensors 21 is not limited to one. There may be two or more temperature sensors 21.

[0087] Furthermore, the configuration of the measurement system 100 is merely an example. For instance, the processing unit 5 may incorporate some or all of the functions of the control unit 3. Also, if the measurement system 100 is equipped with multiple probes 1 and multiple control units 3 corresponding to the probes 1, the processing unit 5 may process detection signals from the multiple control units 3.

[0088] The measurement of eddy currents by the control device 3 is not limited to the method described above. For example, repeated measurements of eddy currents do not need to be divided into multiple sets, and furthermore, the temperature of the object 9 does not need to be measured for each set. In repeated measurements of eddy currents, the temperature of the object 9 may be measured only once. Also, the frequency of temperature measurement may be the same as the frequency of eddy current measurement, or it may be more frequent than the frequency of eddy current measurement.

[0089] In step S104, the determination of whether the temperature change of object 9 is greater than a predetermined standard is not limited to whether the temperature difference between the current measurement temperature and the previous measurement temperature is greater than a predetermined value. For example, the standard may be whether the temperature difference between the moving average of the measurement temperatures and the current measurement temperature is greater than a predetermined value.

[0090] Furthermore, measuring the temperature of object 9 is not essential when measuring eddy currents. Even if the temperature of object 9 is measured, it is not necessary to use the temperature of object 9 as a criterion for deciding whether or not to stop measuring eddy currents. In other words, the measurement of eddy currents may continue regardless of changes in the temperature of object 9. The temperature of object 9 may only be used to correct the calculated thickness.

[0091] Furthermore, the measurement of eddy currents by probe 1 only needs to be performed intermittently and does not need to be performed periodically. In other words, the measurement of the thickness of object 9 also only needs to be collected intermittently and does not need to be collected periodically.

[0092] Furthermore, the transmission of data from the control device to the processing device 5 is not limited to the examples described above. For example, the control device 3 may calculate the average of multiple eddy currents (e.g., eddy currents measured repeatedly n times) and transmit the average of the eddy currents for m sets. Also, if the control device 3 stops the measurement due to a large temperature change in the object 9, it does not need to transmit the eddy currents up to the point of termination to the processing device 5.

[0093] Furthermore, any method can be used to derive the thickness in the processing apparatus 5. Also, correction of the thickness due to the temperature of the object 9 is not mandatory.

[0094] The first period in the first discrimination does not have to be the day before the period targeted for the first discrimination (i.e., the period in which the target measurement is included). The first period may be two days before the target period. The first period does not have to be a period earlier than the period targeted for the first discrimination. Since the first discrimination determines whether or not something is abnormal based on the magnitude of the change in the measurement, it is sufficient to determine whether or not the target measurement is abnormal based on measurement values ​​from a period that is temporally close to the target measurement. Therefore, the first period may be a period later than the period targeted for the first discrimination. Also, the length of the first period does not have to be the same as the period targeted for the first discrimination. In other words, in the example above, the period targeted for the first discrimination is one day, and the first period is the day before the target period. For example, the period targeted for the first discrimination may be one hour, and the first period may be the ten hours before (for example, immediately before) the target period. Thus, the length of the period targeted for the first discrimination and the length of the first period do not have to be the same.

[0095] The representative value in the first discrimination decision does not have to be the mean or median. The representative value may be, for example, the minimum value of the measurements taken during the first period.

[0096] Regarding the first range in the first discrimination, the width between the lower limit of the first range and the representative value may be the same as the width between the upper limit of the first range and the representative value. Alternatively, the width between the lower limit of the first range and the representative value may be smaller than the width between the upper limit of the first range and the representative value. For example, this is effective when the thickness of object 9 tends to increase. The width between the lower limit of the first range and the representative value, or the width between the upper limit of the first range and the representative value, may be less than or equal to the variance with respect to the mean.

[0097] The second period in the second discrimination is not limited to ten days. The second period can be set arbitrarily as long as it is longer than the first period. In the second discrimination, the change in the measured values ​​over time during the second period does not have to be represented by an approximate straight line. For example, the change over time may be represented by a regression line. Also, with respect to the second range, the upward and downward spread from the approximate straight line can be set arbitrarily.

[0098] The flowchart is merely an example. The steps in the flowchart may be changed, replaced, added, or omitted as appropriate. The order of the steps in the flowchart may also be changed, or serial processes may be processed in parallel. For example, step S104 may be omitted, and the measurement of eddy currents for set number m may always be performed regardless of the temperature change of the object 9.

[0099] The functions realized by the components described herein may be implemented in circuits or processing circuits, including general-purpose processors, application-specific processors, integrated circuits, ASICs (Application Specific Integrated Circuits), CPUs (a Central Processing Unit), conventional circuits, and / or combinations thereof, programmed to realize the functions described herein. A processor includes transistors and other circuits and is considered a circuit or processing circuit. A processor may be a programmable processor that executes a program stored in memory.

[0100] In this specification, circuits, units, and means are hardware programmed to perform or execute the functions described herein. Such hardware may be any hardware disclosed herein, or any hardware known to be programmed to perform or execute the functions described herein.

[0101] If the hardware is a processor that is considered to be a type of circuit, then the circuit, means, or unit is a combination of hardware and software used to constitute the hardware and / or processor.

[0102] The technology disclosed in this disclosure can be summarized as follows:

[0103] [1] The processing device 5 includes a collector 54 that continuously collects measurement values ​​of the thickness of an object 9 measured by pulsed eddy current testing, and a discriminator 55 that identifies abnormal values ​​from a group of measurement values ​​which is a collection of the measurement values ​​collected by the collector 54. The discriminator 55 performs a first discrimination, which identifies a first abnormal value from the group of measurement values ​​based on the measurement values ​​during a predetermined first period, and a second discrimination, which removes the first abnormal value from the measurement values ​​during a second period which is longer than the first period, and identifies a second abnormal value from the group of measurement values ​​based on the change in the measurement values ​​over time during the second period from which the first abnormal value has been removed.

[0104] In this configuration, first, the first anomaly is identified from the group of measurement values ​​based on the measurement values ​​during the first period. Next, the second anomaly is identified based on the change over time of the measurement values ​​from which the first anomaly has been removed during the second period, which is longer than the first period. Since the anomaly is identified in two stages—the first identification, which identifies the first anomaly based on the measurement values ​​of the relatively short first period, and the second identification, which identifies the second anomaly based on the measurement values ​​of the longer second period—the accuracy of anomaly identification is improved. In the second identification, by identifying the second anomaly based on the change over time of the measurement values ​​during the second period, anomaly measurement values ​​can be appropriately identified for measurement values ​​over a relatively long period. Furthermore, since the first anomaly has been removed from the measurement values ​​during the second period in the second identification, the accuracy of the second identification is improved. As a result, anomalies can be accurately identified from the measurement values ​​of the thickness of the object 9.

[0105] [2] In the processing device 5 described in [1], the discriminator 55 determines a representative value of the measured values ​​during the first period in the first discriminator, and determines that the measured values ​​that fall outside the first range determined based on the representative value are the first abnormal values.

[0106] With this configuration, when multiple measurement values ​​are included in the first period, a representative value can be used to identify the first outlier, making it easy to determine whether the measurement value in question is abnormal or not among the multiple measurement values ​​in the first period.

[0107] [3] In the processing apparatus 5 described in [1] or [2], the predetermined first range is defined by an upper limit greater than the representative value and a lower limit less than the representative value, and the range between the representative value and the lower limit is greater than the range between the upper limit and the representative value.

[0108] This configuration allows for the proper identification of the first outlier when the thickness of object 9 decreases over time. In other words, when the thickness of object 9 decreases over time, the measured value tends to decrease over time. A measured value that changes below the representative value may include the effect of the decrease in the thickness of object 9. By making the range between the lower limit of the first range and the representative value larger than the range between the upper limit of the first range and the representative value, the discrimination criterion for values ​​below the representative value becomes more lenient than the discrimination criterion for values ​​above the representative value. This prevents misidentification of measured values ​​that have decreased due to changes in the thickness of object 9 as the first outlier.

[0109] In the processing device 5 described in any one of [1] to [3], the discriminator 55, in the second discriminator, determines an approximate straight line of the change over time of the measured value during the second period, and determines that the measured value that falls outside the second range based on the approximate straight line is the second abnormal value.

[0110] In this configuration, the approximation line represents the change in measured values ​​over time during the second period. By identifying the second outlier based on the second range using the approximation line as a reference, it is possible to easily identify the second outlier based on the change in measured values ​​over time during the second period.

[0111] In the processing device 5 described in any one of [1] to [4], the discriminator 55 repeatedly performs the second discrimination on the target measurement while changing the target measurement from the group of measurement values, and the target measurement does not include the measurement values ​​that have already been determined to be the first abnormal value, but includes the measurement values ​​that have already been determined to be the second abnormal value.

[0112] This configuration allows for the appropriate determination of the time-series changes in measured values, which serve as the criteria for the second discrimination, in the repeated second discrimination process. In other words, the first outlier is determined based on measured values ​​during a relatively short first period. In contrast, the second outlier is determined based on measured values ​​during a relatively long second period, excluding the first outlier. Therefore, the degree of abnormality in the second outlier is smaller than that of the first outlier. In the second discrimination process, by using measured values ​​that have already been identified as second outliers in past second discrimination processes to derive the time-series changes, the number of measured values ​​used to derive the time-series changes can be increased, thereby improving the accuracy of the time-series changes.

[0113] [6] The processing device 5 described in any one of [1] to [5] further comprises a calculator 56 for calculating the rate of change of the measured values ​​in the group of measured values ​​from which the first abnormal value and the second abnormal value have been removed.

[0114] With this configuration, the rate of change of the measured values ​​can be accurately determined by removing the first and second outliers identified as described above from the group of measured values. [Explanation of symbols]

[0115] 5 Processing Unit 54 Collector 55 Discriminator 56 Calculator

Claims

1. A collector that continuously collects measurement values ​​of the thickness of an object measured by pulsed eddy current testing, The collection device includes a discriminator that identifies abnormal values ​​from a group of measurement values, which is a collection of measurement values ​​collected by the aforementioned collector. The aforementioned discriminator is A first determination is made to identify a first abnormal value from the group of measurement values ​​based on the measurement values ​​taken during a predetermined first period from the group of measurement values, A processing device that performs a second determination, which involves removing the first abnormal value from the measured values ​​during a second period that is longer than the first period among the group of measured values, and determining a second abnormal value from the group of measured values ​​based on the change in the measured values ​​over time during the second period from which the first abnormal value has been removed.

2. In the apparatus according to claim 1, The discriminator is a processing device that, in the first discriminant, determines a representative value of the measured values ​​during the first period, and determines that the measured values ​​that fall outside a first range determined based on the representative value are first abnormal values.

3. In the apparatus according to claim 2, The predetermined first range is defined by an upper limit greater than the representative value and a lower limit less than the representative value. A processing device in which the range between the representative value and the lower limit is greater than the range between the upper limit and the representative value.

4. In the apparatus according to claim 1, The discriminator is a processing device that, in the second discriminant, finds an approximate straight line of the change over time of the measured value during the second period, and determines that the measured value that falls outside the second range based on the approximate straight line is the second abnormal value.

5. In the apparatus according to claim 1, The discriminator repeatedly performs the second discrimination on the target measurement while changing the measurement value of the target from the group of measurement values. The aforementioned processing apparatus includes a measurement value which does not include a measurement value which has already been identified as the first abnormal value, but includes a measurement value which has already been identified as the second abnormal value.

6. In the processing apparatus according to any one of claims 1 to 5, A processing apparatus further comprising a calculator for calculating the rate of change of the measured values ​​in the group of measured values ​​from which the first and second abnormal values ​​have been removed.

Images