Defect detection device and defect detection method

Abstract

To improve inspection accuracy in defect detection based on a temperature distribution.SOLUTION: A defect detection device comprises: a heat source 2 that applies thermal action to a location at a predetermined distance apart from a reference point set on a surface of an inspected object; a temperature distribution measuring instrument 3 that measures a temperature distribution in a surface region that has received the thermal action including the reference point; a temperature gradient calculation part 52 that calculates a temperature gradient in the temperature distribution; and a defect information generation part 53 that generates defect information from the behavior of the temperature gradient. The temperature gradient calculation part 52 calculates the temperature gradient by performing smoothed differential processing on measured values at a plurality of temperature measurement points in a predetermined range of the temperature distribution obtained by applying the thermal action.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a defect detection device and a defect detection method for detecting defects present in an object to be inspected based on a temperature distribution on the surface of the object to be inspected, which is generated by applying a thermal effect to the surface of the object to be inspected using a hot or cold source. [Background technology]

[0002] Patent Document 1 discloses a surface defect detection device that applies a heating or cooling source to the surface of a test piece, measures the temperature distribution that occurs on the surface of the test piece, analyzes the temporal and spatial fluctuations of the temperature distribution, and calculates the position and depth of the defect from the non-uniformity in the defect area.

[0003] Patent Document 2 discloses an automatic detection device for internal defects of an object, which is characterized by processing thermal images of the object measured as the object surface temperature rises or falls, determining the inflection points of each of the obtained pixel lines, and detecting the area surrounded by these inflection points. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 4-331360 [Patent Document 2] Japanese Patent Application Laid-Open No. 2001-50921 Summary of the Invention [Problem to be solved by the invention]

[0005] The technologies of Patent Documents 1 and 2 attempt to detect defects based on the temperature distribution that appears on the surface of the object being inspected. However, because the change in the surface temperature distribution caused by a minute defect is small, there is a problem in that the inspection accuracy deteriorates due to noise that occurs in the measurement values ​​depending on the surface condition of the object being inspected and the environmental condition of the object being inspected. For this reason, there is a demand for improved inspection accuracy when detecting defects based on temperature distribution. [Means for solving the problem]

[0006] The defect detection device according to the present invention comprises a heat source that applies thermal action at a position a predetermined distance away from a reference point set on the surface of an object to be inspected, a temperature distribution measuring instrument that measures the temperature distribution of the surface area that has been subjected to the thermal action and that includes the reference point, a temperature gradient calculation unit that calculates the temperature gradient in the temperature distribution, and a defect information generation unit that generates defect information from the behavior of the temperature gradient, and the temperature gradient calculation unit calculates the temperature gradient by performing smoothing differential processing on a plurality of temperature measurement points within a predetermined range in the temperature distribution obtained by applying the thermal action.

[0007] According to this configuration, a heat source applies heat to a position a predetermined distance away from a reference point (e.g., a location where a defect is suspected) set on the surface of the object to be inspected, and the temperature distribution on the surface of the object to be inspected resulting from this heat application is measured. Measurement values ​​from multiple temperature measurement points located within a predetermined range of the temperature distribution are subjected to smoothing and differential processing to calculate a temperature gradient, and defect information is generated from the behavior of the calculated temperature gradient. Because the temperature gradient is calculated by smoothing and differential processing of the measurements from multiple temperature measurement points in this manner, a temperature gradient in which noise that may be mixed in during defect detection is smoothed is obtained, improving inspection accuracy. The heat source used in this inspection is a hot or cold source. It is preferable to use a lower-temperature cold source when the object to be inspected is at a high temperature, and a higher-temperature hot source when the object to be inspected is at a low temperature.

[0008] When the point of thermal action is changed, the line of thermal action of the thermal action to the defect will be different, and as a result, the structure of the inspection object through which heat conduction passes will also be different, and defect detection will be performed from multiple angles. Therefore, in the present invention, it is proposed that the defect information be generated from the behavior of the temperature gradient of the multiple temperature distributions obtained by applying the thermal action multiple times at different thermal action positions.

[0009] Setting a predetermined range in the temperature distribution that defines the multiple temperature measurement points used to calculate the temperature gradient is important for defect detection. Because the behavior of heat transfer from a heat source varies depending on the surface condition and internal structure of the object being inspected, it is preferable to use temperature measurement points that exist within a certain range in size for highly accurate defect detection. Therefore, in the present invention, it is proposed that the temperature gradient calculation unit sets the predetermined range to an area between the reference point and the point of application of the thermal effect along a predetermined line passing through the reference point.

[0010] If the predetermined range is too large, the number of temperature measurement points to be processed will be large, which will increase the burden on the temperature gradient calculation process. To avoid this problem, it is proposed that the temperature gradient calculation unit calculates the temperature gradient by performing smoothing difference processing on the temperature distribution at multiple temperature measurement points in a direction along a predetermined straight line passing through the reference point.

[0011] Since the temperature distribution obtained using a temperature distribution measuring device such as an infrared thermograph is a temperature distribution on a two-dimensional surface, a temperature gradient curve is calculated along countless directional lines. To reduce the inspection burden, it is preferable to limit the directional lines used to calculate the temperature gradient curve to predetermined effective straight lines, for example, at least a straight line passing through a reference point. Therefore, the present invention proposes that the temperature gradient calculation unit calculates a temperature gradient curve for the temperature distribution along the direction of the predetermined straight line passing through the reference point, and the defect information generation unit generates the defect information from the behavior of the temperature gradient curve. In this case, it is particularly preferable to adopt a temperature gradient curve along the direction of a straight line connecting the reference point and the point of application of the thermal effect.

[0012] Based on the experimental and empirical knowledge of the inventors of the present invention, as an embodiment for obtaining a particularly suitable temperature gradient curve, the present invention proposes that measurement values ​​of 5 to 15 temperature measurement points in a direction along the direction of the straight line connecting the reference point and the point of application of the thermal effect be used to calculate the temperature gradient curve.

[0013] In the technology of the present invention for generating defect information from temperature gradient behavior, it is important to determine what temperature gradient behavior is associated with the presence of a defect, and this is preferably determined experimentally and empirically. Based on the inventor's experimental and empirical knowledge for improving inspection accuracy, the present invention proposes that the temperature gradient behavior is the position of an inflection point of the temperature gradient curve and the shape around the inflection point, and that the defect information generation unit estimates defect information, such as defect position and defect size, from the position of the inflection point and the shape around the inflection point. The defect size is a value indicating the spatial extent of the defect, including the defect depth. In particular, since the shape around the inflection point is considered to be affected by the defect structure, including the defect size, it is possible to construct an estimation model for estimating the defect structure from the shape around the inflection point (inflection point shape).

[0014] The smoothing difference process is a process of smoothing a plurality of temperature distributions and calculating a temperature gradient from the smoothed temperature distributions, or a process of smoothing a plurality of temperature gradients calculated from a plurality of temperature distributions, but it is also possible to simultaneously calculate the smoothing and the difference (differential), which is efficient. For this reason, in a preferred embodiment of the present invention, the smoothing difference process uses the following equation, which simultaneously smooths the temperature distribution and calculates the temperature gradient:

number

[0015] If the object to be inspected is large, the inspection range will be wide, so in order to perform an efficient defect inspection, it is preferable to limit the inspection to areas where there is a high possibility of defects. Areas where there is a high possibility of defects (estimated defect areas) can be estimated from the surface condition and structure of the object to be inspected. Furthermore, in order to improve inspection efficiency, it is also important to set a small area including the estimated defect area as the inspection target area. For this reason, the present invention proposes that the reference point is the estimated defect area, and that the predetermined distance is several mm.

[0016] This application covers not only the above-described defect detection device, but also a defect detection method based on the operating principle of the defect detection device. Such a defect detection method for detecting defects present in an object to be inspected includes the steps of: installing a heat source at a surface position a predetermined distance from a reference point set on the surface of the object to be inspected; applying thermal action from the heat source for a predetermined time (preferably 3 to 10 minutes); measuring a temperature distribution in the surface region subjected to the thermal action, including the reference point, after the predetermined time has elapsed since the start of the thermal action from the heat source; calculating a temperature gradient by performing smoothing differential processing on a plurality of temperature measurement points within a predetermined range in the temperature distribution obtained by applying the thermal action; and generating defect information from the behavior of the temperature gradient. The functions and effects described in the above-described description of the defect detection device can also be applied to this defect detection method.

[0017] Other features, operations, and advantages of the present invention will become apparent from the following description of the invention using the accompanying drawings. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is an external view showing a specific example of use of the defect detection device. [Figure 2] FIG. 1 is a schematic diagram illustrating a basic configuration of a defect detection device. [Figure 3] FIG. 2 is a functional block diagram showing each functional unit of the defect detection device. [Figure 4] FIG. 4 is a schematic diagram showing an example of a temperature distribution. [Figure 5] FIG. 2 is a schematic diagram showing an example of a temperature gradient curve. [Figure 6] FIG. 10 is a flowchart showing the flow of a defect detection process using the defect detection device. DETAILED DESCRIPTION OF THE INVENTION

[0019] FIG. 1 shows a specific example of the use of the defect detection system according to the present invention. This example shows an inspector inspecting defects on-site in wind turbine blades for wind power generation facilities. The procedure generally involves applying a heat source (hot or cold) 2 to the surface of the turbine blade (object of inspection 1), and capturing an image of the temperature distribution on the surface of the turbine blade due to the thermal effect of the heat source 2 using an infrared camera unit 3, which serves as a temperature distribution measuring device. The captured infrared image is converted into temperature distribution image data showing the temperature distribution using an infrared thermography measurement algorithm, and is sent to a controller 5 (not shown in FIG. 1) (see FIG. 2) to determine the presence or absence of defects. This inspection system does not detect relatively large defects, such as broken or debonded blades, which are the target of detection by typical defect detection systems, but rather detects defects that contribute little to the temperature change of the entire blade, such as minute defects occurring on the surface of the blade substrate (fine defects near the surface).

[0020] 2 and 3 show one embodiment of a defect detection device according to the present invention. This defect detection device comprises a heat source 2, an infrared camera unit 3, a holder 4, and a controller 5. The heat source 2 is a device that can be used as a cold heat source (electric cooler or refrigerant) or a hot heat source (electric heater or heat medium), and has a heat release surface that can be brought into close contact with the surface of the object to be inspected 1. The holder 4 comprises a first mount 4a for mounting the infrared camera unit 3, a second mount 4b for supporting the heat source 2, and a base 4c. The first mount 4a and the second mount 4b are connected to the base 4c. In other words, the heat source 2, the infrared camera unit 3, and the holder 4, which function as a defect detection operating tool, are configured as a single unit that can be operated with one hand.

[0021] The holder 4 has a first mount 4a and a second mount 4b designed so that, with the heat emitting surface of the heat source 2 in close contact with the surface of the inspection object 1, the imaging center of the infrared camera unit 3 is located a few millimeters away from the heat emitting surface. The holder 4 may be configured so that the distance between the heat emitting surface and the imaging center of the infrared camera unit 3 is variable.

[0022] 3, the controller 5 includes an input / output data processing unit 50, a control unit 51, a temperature gradient calculation unit 52, and a defect information generation unit 53. The controller 5 is essentially a computer system, and can be configured as a dedicated machine, a laptop computer, a tablet computer, or the like.

[0023] The input / output data processing unit 50 performs preprocessing and data conversion of input data input to the controller 5, and preprocessing and data conversion of output data output from the controller 5. Examples of input data include temperature distribution image data, which is data relating to temperature distribution from the infrared camera unit 3. Examples of output data include control signals for controlling the operation of the heat source 2, and notification data given to a notification device 9 such as a monitor or printer.

[0024] The control unit 51 generates control signals for internal functional units and external devices based on stored programs and operation signals from the operation device 8.

[0025] The temperature gradient calculation unit 52 receives temperature distribution image data (temperature distribution) from the infrared camera unit 3 and calculates the temperature gradient from the temperature distribution shown in the received temperature distribution image data. The temperature distribution image data (hereinafter simply referred to as temperature distribution) shows the temperature distribution in the field of view of the infrared camera unit 3 using color information on a pixel-by-pixel basis, an example of which is shown in FIG. 4. In the temperature distribution of FIG. 4, the direction extending from the left end (origin) to the right end of the horizontal axis (x-axis) coincides with the direction extending from the point of application of the heat source 2 to the inspection object 1 through the center of the infrared camera unit 3, and the vertical axis (y-axis) is perpendicular to the horizontal axis. In other words, the temperature distribution of FIG. 4 shows the temperature distribution in the field of view of the inspection object surface measured by the infrared camera unit 3.

[0026] When calculating a temperature gradient from a temperature distribution that spreads two-dimensionally (on a plane), it is preferable to calculate the temperature gradient along a straight line, so a temperature gradient calculation line is set. Here, the temperature gradient calculation line is a line that connects the point of application of the heat source 2 to the inspection object 1 and the estimated defect location where the defect is estimated to exist. If the estimated defect location has a wide range, heat sources are installed in multiple locations, the temperature distribution at each location is measured, and the temperature gradient (temperature gradient curve) is calculated from each temperature distribution.

[0027] In this embodiment, the temperature gradient along the set temperature gradient calculation line is calculated by smoothing difference processing. For this purpose, the temperature gradient calculation unit 52 includes a smoothing difference processor 52a. To calculate a smoothing difference value at one coordinate position (a pixel position or a representative pixel position of multiple pixels) in the temperature distribution, the smoothing difference processor 52a uses color information of each pixel group (or representative color information of multiple adjacent pixels) of the temperature distribution image data showing the temperature distribution and the following equation:

number

[0028] The temperature gradient calculation unit 52 calculates a temperature gradient curve, which represents the behavior of the temperature gradient, using the smoothed difference values ​​calculated sequentially at each coordinate position (temperature measurement point) by the smoothed difference processor 52a as the temperature gradient value. The intervals between each coordinate position are set according to the resolution and shooting distance of the temperature distribution image data. An example of this temperature gradient curve is shown in FIG. 5.

[0029] Here, the temperature gradient can be calculated in various ways, which are listed below. (1) When multiple different temperature distributions (multiple temperature distributions obtained with different thermal action positions, different imaging center positions, or both) are generated, smoothing difference processing is performed on these multiple temperature distributions. (2) According to the inventor's knowledge, the number of temperature measurement points (measured values) suitable for obtaining the temperature distribution is 5 to 15 points, and preferably around 9 points. (3) The temperature gradient calculation unit 52 calculates a temperature gradient curve for the temperature distribution in a direction along a predetermined straight line passing through a reference point as an estimated defect location. (4) When the estimated defect location is used as a reference point, the suitable distance from the reference point to the point of thermal action of the heat source 2 is several mm, preferably around 5 mm.

[0030] The defect information generation unit 53 generates defect information from a temperature gradient curve, which represents the behavior of the temperature gradient. Specifically, the characteristic behavior of the temperature gradient is considered to be the shape of the temperature gradient curve, and the defect information generation unit 53 includes a shape evaluation unit 53a that evaluates this shape. The shape evaluation unit 53a considers the characteristic behavior of the temperature gradient to be the positions of inflection points on the temperature gradient curve and the shape around the inflection points (inflection point shape). The shape evaluation unit 53a estimates the defect structure, including the defect position and defect size, which are defect information, from the positions and shapes of the inflection points. The evaluation algorithm of the shape evaluation unit 53a may be a known evaluation method such as a curvature analysis method or a matching method, or an evaluation model using machine learning. In the temperature gradient curve of FIG. 5, which shows the temperature gradient when y = 20 mm in FIG. 4, it is estimated that a defect exists 5 mm above the point of application of the heat source 2, based on the position and size of the inflection point, which appears 5 mm away from the origin.

[0031] Next, an example of a defect inspection procedure using the defect inspection device configured as described above will be described with reference to the flowchart in Fig. 6. In this defect inspection procedure, the surface defect state is estimated by the inspector's vision or touch, and the estimated defect is classified into a case where it extends in only one direction and a case where it extends in two directions.

[0032] First, the inspector observes the surface condition of the inspection object 1 and finds a location where the presence of a defect (surface defect) is suspected (#01). Furthermore, based on the appearance of the suspected defect location, it is checked whether the surface defect has occurred in only one direction or in multiple directions (here, two directions) (#02). If it has occurred in only one direction (#02 Yes branch), simple defect detection processing is performed in steps #31 to #38, and if it has occurred in multiple directions (#02 No branch), complex defect detection processing is performed in steps #21 to #26 (applying heat multiple times at different heat application positions).

[0033] In the complex defect detection process, a heat source (cooling in this case) 2 is placed sequentially at two estimated defect locations (the distance between the estimated defect locations and the heat source 2 is set to several mm), and the temperature distribution is acquired at each location (#21). At this time, the thermal action time of the heat source 2 is set to 3 to 10 minutes depending on the characteristics of the inspection object 1 and the inspection environment.

[0034] A temperature gradient curve is calculated from the multiple temperature distributions obtained by applying heat multiple times (#22).The peak shape (shape of the inflection point) at the characteristic peak value (prominent inflection point) is confirmed as the characteristic behavior of the calculated temperature gradient curve (#23).

[0035] A check is made to see if there is a broad (gradually inflected) peak shape (#24). If there is no broad peak shape (#24 No branch), the defect is assumed to be a surface defect, and the defect depth is estimated from the temperature gradient curve and output (#25). If there is a broad peak shape (#24 Yes branch), the defect is assumed to be a surface defect and an internal defect, and the depth of the surface defect and the position and depth of the internal defect are estimated from the temperature gradient curve and output (#26).

[0036] In the simple defect detection process, a heat source (cooling in this case) 2 is placed at one estimated defect location. Here too, the distance between the estimated defect location and the heat source 2 is a few mm. The thermal exposure time is set to 3 to 10 minutes, and a smoothed temperature distribution is obtained from the obtained temperature distribution (#31).

[0037] A temperature gradient curve is calculated from the acquired temperature distribution (#32). The peak shape (shape of the inflection point) at the characteristic peak value (prominent inflection point) is confirmed as the characteristic behavior of the calculated temperature gradient curve (#33).

[0038] Here again, a check is made to see if there is a broad (gently inflected) peak shape (#34). If there is no broad peak shape (#34 No branch), the defect is assumed to be a surface defect, and the defect depth is estimated and output (#35). If there is a broad peak shape (#34 Yes branch), a check is made to see if the direction of the internal defect in the broad peak shape differs from that of the surface defect (#36). If the direction of the internal defect in the broad peak shape differs from that of the surface defect (#36 Yes branch), heat source 2 is installed in a different direction, heat is applied from the different direction, the temperature distribution is obtained, and the temperature gradient curve is recalculated (#37). The defect is assumed to be a surface defect and an internal defect, and the depth of the surface defect and the position and depth of the internal defect are estimated from the recalculated temperature gradient curve and output (#38). On the other hand, if the direction of the internal defect with a broad peak shape and the surface defect are the same (#36 No branch), the depth of the surface defect and the position and depth of the internal defect are estimated from the currently obtained temperature gradient curve and output (#38).

[0039] [Another embodiment] (1) In the above-described embodiment, the inspector holds the holder 4 and checks the surface of the inspection object 1 while applying heat to the suspected defect location to perform defect inspection. However, the holder 4 may be mounted on an autonomous vehicle to automatically perform defect inspection, or at least temperature distribution measurement. (2) In the block diagram of Fig. 3, the defect information generating unit 53 is shown as generating defect information from the behavior of a given temperature gradient by computer processing, but the defect information generating unit 53 may function in cooperation with a manual operation command. For example, the defect information generating unit 53 may display data representing the behavior of a temperature gradient curve on a monitor or the like as support information, and an inspector may determine the defect information. (3) The functional units shown in the block diagram of FIG. 3 are for illustrative purposes only, and the functional units may be further divided or integrated. (4) In the above-described embodiment, an infrared camera type temperature distribution measuring device is used. However, instead of this, an optical fiber infrared type or a radiation infrared detection type may be used. (5) In the above-described embodiment, the measurement values ​​of the multiple temperature measurement points used to calculate the temperature gradient were selected along a predetermined straight line passing through the reference point, but they may also be selected from a quadratic surface area such as a triangle or trapezoid between the reference point and the heat source 2.

[0040] The configurations disclosed in the above embodiments (including other embodiments, the same applies below) can be applied in combination with configurations disclosed in other embodiments, as long as no contradiction arises. Furthermore, the embodiments disclosed in this specification are examples, and the embodiments of the present invention are not limited to these, and can be modified as appropriate within the scope that does not deviate from the purpose of the present invention. [Industrial Applicability]

[0041] The present invention is applicable to a defect detection technique that detects defects present in an object to be inspected based on the temperature distribution on the surface of the object to be inspected that is generated by applying heat to the surface of the object to be inspected. [Explanation of symbols]

[0042] 1: Inspection item 2: Heat source 3: Infrared camera unit (temperature distribution measuring device) 4: Holder 5: Controller 50: Input / output data processing section 51: Control unit 52: Temperature gradient calculation section 52a: Smoothing differential processor 53: Defect information generation unit 53a: Shape evaluation section

Claims

1. a heat source that applies heat to a position a predetermined distance away from a reference point set on the surface of the object to be inspected; a temperature distribution measuring device for measuring the temperature distribution of the surface area subjected to the thermal effect, including the reference point; a temperature gradient calculation unit that calculates a temperature gradient in the temperature distribution; a defect information generating unit that generates defect information from the behavior of the temperature gradient; Equipped with The temperature gradient calculation unit is a defect detection device that calculates the temperature gradient by performing smoothing and differential processing on measured values ​​of multiple temperature measurement points within a predetermined range in the temperature distribution obtained by applying the thermal action.

2. 2. The defect detection device according to claim 1, wherein the defect information is generated from behavior of the temperature gradient of a plurality of the temperature distributions obtained by applying the heat treatment a plurality of times at different heat treatment positions.

3. 2. The defect detection device according to claim 1, wherein the temperature gradient calculation unit defines the predetermined range as a region between the reference point and the point of application of the thermal effect along a predetermined straight line passing through the reference point.

4. The defect detection device according to claim 1, wherein the temperature gradient calculation unit calculates the temperature gradient by performing smoothing differential processing on a plurality of temperature measurement points in a direction along a predetermined straight line passing through the reference point for the temperature distribution.

5. 2. The defect detection device according to claim 1, wherein the temperature gradient calculation unit calculates a temperature gradient curve for the temperature distribution in a direction along a predetermined straight line passing through the reference point, and the defect information generation unit generates the defect information from the behavior of the temperature gradient curve.

6. 6. The defect detection device according to claim 5, wherein the temperature gradient calculation unit calculates the temperature gradient curve in a direction along a straight line connecting the reference point and the point of application of the thermal effect.

7. 7. The defect detection device according to claim 6, wherein measurement values ​​at 5 to 15 temperature measurement points in a direction along the direction of the straight line connecting the reference point and the point of application of the thermal effect are used to calculate the temperature gradient curve.

8. The defect detection device of claim 5, wherein the behavior of the temperature gradient is the position of an inflection point of the temperature gradient curve and the shape around the inflection point, and the defect information generation unit estimates the defect position and defect size, which are the defect information, from the position and size of the inflection point.

9. The smoothed difference process is performed using the following equation: [Equation 1] The defect detection device according to claim 1, wherein the defect detection is performed using a

10. The defect detection device according to claim 1 , wherein the reference point is an estimated defect location, and the predetermined distance is several millimeters.

11. A defect detection method for detecting defects present in an inspection object, comprising: A step of placing a heat source at a surface position a predetermined distance away from a reference point set on the surface of the test object; a thermal action step of applying thermal action by the heat source for a predetermined time; a temperature distribution measuring step of measuring a temperature distribution in a surface area subjected to the thermal action, including the reference point, after the predetermined time from the start of the thermal action by the heat source; a temperature gradient calculation step of calculating a temperature gradient by performing smoothing difference processing on a plurality of temperature measurement points within a predetermined range in the temperature distribution obtained by applying the thermal action; a defect information generating step of generating defect information from the behavior of the temperature gradient; A defect detection method comprising:

Images