JP4595077B2 - Prediction method for strength deterioration of corroded structures - Google Patents

Description

  The present invention relates to a method for predicting the strength deterioration of a corroded structure for predicting future strength changes in corroded structures such as corroded steel structures such as bridges where the surface shape changes due to corrosion and the strength decreases. It is.

  In recent years, in Japan, many existing steel structures such as steel bridges have a service period of 40 to 60 years or more, resulting in a problem of strength reduction due to corrosion of the structure. Therefore, appropriate maintenance management such as repair of corrosive structures has become necessary.

  The maintenance management includes renewing the corroded structure when the risk of collapse of the corroded structure is increased. Therefore, not only the current strength measurement of the corroded structure but also the previous repair. In order to plan the renewal in advance and minimize the life cycle cost, a method for predicting the future strength deterioration of the corroded structure is required.

  As such a future strength deterioration prediction method, a virtual function is applied to the current strength (for example, the measured numerical value of the strength is divided by the route T (T is time)), and the future strength is virtually determined. And a method for predicting the future intensity by applying a temporal change in intensity to a higher order function (3rd order or 4th order) is known.

  However, since the above conventional method predicts not the future corrosion state of the corroded structure but its future strength change (deterioration), it is between the strength due to the actual actual corrosion state and the predicted strength. There is a problem that deviation is likely to occur and strength deterioration cannot be predicted accurately.

  In addition, in the past, future corrosion surface shapes are now being evaluated in order to evaluate the life cycle cost of the target corroding structure and to predict strength deterioration suitable for maintenance and management of structures with economic rationality. A surface shape creation model that can be predicted more accurately from the measurement data of the corroded surface shape up to now has not been known.

  The present invention has been made in view of the above-mentioned problems, and its purpose is to measure the current corroded surface shape with a relatively simple configuration, and from the current corroded surface shape to the future corroded surface. An object of the present invention is to realize a strength deterioration prediction method for a corroding structure including a surface shape creation model that can predict a shape more accurately and evaluate a future strength from the predicted corroded surface shape.

  In order to solve the above problems, the method for predicting the strength deterioration of a corroded structure according to the present invention includes (1) non-contact and three-dimensional measurement of the corroded surface shape of the corroded structure, and (2) changing the parameters. By using the surface shape creation model for reproducing various uneven surface shapes and calculating the parameters related to the surface shape creation model so as to match the unevenness of the corrosion surface shape measured in three dimensions, 3) Calculate the future corroded surface shape based on the surface shape creation model based on the calculated parameters, and (4) predict the future strength deterioration in the corroded structure by the future corroded surface shape. Yes.

  According to the above method, by calculating the parameters related to the surface shape creation model so as to match the unevenness of the corroded surface shape actually measured three-dimensionally, the environment where the corroded surface shape is placed A surface shape creation model capable of showing the progress of corrosion in the state including the material of the corrosive structure and the presence or absence of the coating film can be created from the calculated parameters.

  Thus, in the above method, by using the surface shape creation model that can indicate the progress of corrosion according to the actual situation, the future corrosion surface shape is calculated based on the surface shape creation model based on the calculated parameters. Thus, it is possible to more accurately predict future strength deterioration in the corroded structure based on the future corroded surface shape.

  As a result, the method can measure the strength in a non-contact manner from a position away from the corroded surface, so that the measurement can be performed non-destructively and includes the actual amount of unevenness up to now. The strength of future corroded structures can be calculated more accurately based on the corroded surface shape, so the strength deterioration prediction suitable for the maintenance and management of structures with economic rationality can be calculated more accurately from the measured data of the corroded surface shape so far. Therefore, the life cycle cost of the target corroding structure can be more accurately evaluated.

  In the strength degradation prediction method, the surface shape creation model assumes a plurality of lattice points at a predetermined interval, converts each of the unevenness amounts at each lattice point, and measures each unevenness amount as a corrosion shape. There may be. In the strength deterioration prediction method, the predetermined interval may be set to 5 mm or less. The predetermined interval is preferably 2 mm or less, and if the predetermined interval is less than 0.1 mm, the amount of calculation increases, but no improvement in accuracy can be recognized, so 0.1 mm or more, more preferably 0.5 mm. The above is mentioned.

  In the surface shape creation model, virtual corrosion holes having a depth F as a parameter may be randomly formed on a target corrosion surface, and a predetermined number per unit time as a parameter may be virtually formed. In the strength deterioration prediction method, an influence radius of the radius r as a parameter may be set in the virtual corrosion hole.

  In the strength deterioration prediction method, in the surface shape creation model, when measuring the unevenness amount at each lattice point considering the influence of each virtual corrosion hole on the lattice point by a spatial autocorrelation model as a corrosion shape, You may use the volume of the virtual corrosion hole which affects the said.

  In the strength deterioration prediction method, the corrosion surface shape is measured in a non-contact and three-dimensional manner by projecting a light pattern on the surface of the corrosion structure and imaging the light pattern on the corrosion surface. After obtaining the imaging pattern, the surface shape of the corroding structure may be measured in three dimensions using the light pattern and the imaging pattern.

  In the strength deterioration prediction method, the corrosion surface shape of the corrosion structure may be measured a plurality of times at different times. In the strength deterioration prediction method, the corroded structure may be a corroded steel structure.

  In the strength deterioration prediction method, the corroded surface shape is a state of a paint film, and the deterioration of the paint film is quantified at a lattice point on the paint film surface to obtain an anticorrosion ability at the lattice point. Calculate and calculate the coating deterioration over time as the anti-corrosion ability that changes with the passage of time at the lattice point, and even when the anti-corrosion ability is below a predetermined value, it is recognized that corrosion progress has occurred. Good.

  As described above, the method for predicting the strength deterioration of a corroded structure according to the present invention measures the corroded surface shape of the corroded structure in a non-contact and three-dimensional manner, and reproduces various uneven surface shapes by changing parameters. The surface shape creation model is used to calculate the parameters related to the surface shape creation model so as to match the unevenness of the corrosion surface shape measured three-dimensionally, and the surface shape creation model based on the calculated parameters Is a method for calculating a future corroded surface shape based on the above and predicting future strength deterioration in the corroded structure based on the future corroded surface shape.

  Therefore, the above method is a surface shape creation model that can show the progress of corrosion in the environment where the corroded surface shape is placed, and in the state including the presence or absence of the material and coating film of the corroded structure. It can be created from the calculated parameters.

  Thus, in the above method, by using the surface shape creation model that can indicate the progress of corrosion according to the actual situation, the future corrosion surface shape is calculated based on the surface shape creation model based on the calculated parameters. In addition, it is possible to more accurately predict future strength deterioration in the corroded structure based on the future corroded surface shape.

  As a result, the method can measure the strength in a non-contact manner from a position away from the corroded surface, so that the measurement can be performed non-destructively and includes the actual amount of unevenness up to now. Since the corrosion surface shape can calculate the strength deterioration of future corrosive structures more accurately, the strength deterioration prediction suitable for the maintenance of structures with economic rationality can be calculated from the measurement data of the corrosive surface shape so far. It is possible to predict accurately, and there is an effect that the life cycle cost of the target corroding structure can be more accurately evaluated.

  An embodiment of a method for predicting strength deterioration of a corroded structure according to the present invention will be described with reference to FIGS. The above strength degradation prediction method can improve the accuracy of damage assessment, residual strength assessment, life cycle cost assessment, etc. in line with corrosion structures such as steel structures such as bridges where corrosion has progressed over time. It is suitable for systematization of the maintenance, management, and renewal techniques for the corrosive structures. Damage evaluation (soundness evaluation) is to evaluate to what extent the found corrosion damage reduces the performance required for the target structure (for example, the strength of the structure).

  As shown in FIG. 1, the strength deterioration prediction method includes an inspection process (step 1, hereinafter, abbreviated as S, S2), an evaluation process (S3, S4), and a maintenance planning process (S11-13). is doing. In addition, as a method for evaluating the strength of the corroded structure at the present time, it is only necessary to include the inspection step, the evaluation step, and the coping step (S5, S7, S8, S9, and S10).

  In the inspection process, first, the coating film and rust on the corroded surface of the corroded structure to be subjected to strength deterioration prediction are removed (step 1, hereinafter, step is abbreviated as S). The uneven shape on the corroded surface is measured non-contactly and three-dimensionally on the corroded surface (S2). For the measurement of the uneven shape on the corroded surface, a surface shape measuring method described later can be used, but other measuring methods, for example, an articulated three-dimensional measuring device using a laser line probe can also be used. In the above measurement, a non-contact type is preferable, but a contact type can also be used.

  In addition, said S1 can be omitted if there is no trouble in measurement, and when measuring the deterioration of the protective function of the coating film, the surface shape of the coating film (deterioration of the protective function of the coating film) is predicted. You may measure as an object as it is. In addition, as for the corroded surface to be subject to the above, for example, it is already known that the corrosion of the steel structure is concentrated in the place where the water performance is poor. Measurements may be made only at points necessary for analysis.

  Thereafter, the surface shape is evaluated by statistical evaluation (S3) of the measurement result. In the statistical evaluation, the measured corrosion surface shape is analyzed as an irregular waveform. In other words, 10 or more, usually 1000 or less, measurement lines parallel to each other are virtually set on the measured virtual corrosion surface. Thereafter, a power spectrum and a cross spectrum are obtained from a Fourier analysis for an irregular waveform based on each corrosion amount at 256 points or more, usually 25600 points or less on each side line.

In the power spectrum analysis, S _Z : power spectral density (mm ³ ), Z _avg : average corrosion amount (mm), σ _Z : standard deviation (mm), λ: wave number (1 / mm),
{[ΛS _Z (λ)] / σ _Z ² } = f (λ · Z _avg )
It is represented by In the above power spectrum analysis, it is known that the proportional relationship between the average corrosion amount and the dominant wavelength, that is, the wavelength is short when the average corrosion amount is small, whereas the wavelength is long when the average corrosion amount is large.

  In the evaluation of the surface shape, a surface shape creation model for reproducing various uneven surface shapes by changing each variable parameter is used. The surface shape creation model can reproduce or restore the shape of the target corroded surface by changing each variable parameter. Using the surface shape creation model, parameters relating to the surface shape creation model are calculated so as to match the unevenness of the corroded surface shape actually measured three-dimensionally. This calculation method will be described later. The above parameters include the depth (strength) F, a predetermined number of units per unit time (usually one year), and the surroundings (surface direction of the corroded surface) of each virtual corrosion hole set randomly at a position. The influence radius r indicating the degree of the spatial influence on each is given.

  In addition, the corroded surface having the same statistics as the corroded surface measured by the Spatial Auto-Correlation method on the basis of each spectrum obtained is an appropriate mesh (grating) interval (usually 5 mm or less, more preferably 2 mm or less). The above-mentioned spatial autocorrelation method is an irregular waveform of a corroded surface shape, regarded as a steady normal stochastic process, and is not determined by the statistical values (average value, standard deviation, autocorrelation function, crosscorrelation function) of the corrosion amount. This is a method for generating a regular wavefront. In the statistical evaluation, a histogram indicating the appearance frequency corresponding to each corrosion depth is also used. Based on the statistical evaluation, the evaluation of the surface shape is analyzed (S4). At this time, the wavelength of the power spectrum corresponds to the radius of influence r of the virtual corrosion hole, and the radius r can be calculated from the wavelength.

  Subsequently, from the analysis result described in S4 above, the target corrosive structure is in a healthy range, is in the middle of an observation required, that is, a healthy range and a dangerous range, and a dangerous range. It is discriminated whether it is within (S5). As a result of the determination, if the target corroding structure is within a healthy range, the use is continued (S6). At this time, the strength deterioration due to the measurement of the surface shape of the corrosion object is sequentially measured over time at intervals during normal inspection.

  On the other hand, if the result of the determination is within the range of observation required, that is, between a healthy range and a dangerous range, the target corrosive structure is monitored, that is, the surface measurement interval is set shorter than the interval during normal inspection. Are sequentially measured (S7).

  Further, when the determination result is within the dangerous range, the degree of the risk is determined (S8). In S8, it is determined for the time being that the repair / reinforcement can be outside the dangerous range. When this occurs, the corroded structure is repaired and reinforced (S9) so that it is within a healthy range or within the range of observation. When it is determined in S8 that it cannot be out of the dangerous range due to repair / reinforcement, the corroded structure is discarded or replaced or renewed (S10). S5, S7, S8, S9 and S10 are referred to as coping steps.

  Subsequently, each parameter of the surface shape formation model is adjusted from each surface shape, which is a plurality of results data obtained over time by the analysis of the surface shape described in S3, so as to match the unevenness amount of each surface shape. The future surface shape indicating the degree of progress of the corrosion deterioration of the corroded structure is predicted by extrapolating the future surface shape based on the surface shape formation model based on the respective parameters and calculating the future time (S11).

  Thereafter, based on the future surface shape, future strength deterioration of the corroded structure is predicted over time (S12). Based on the strength deterioration prediction, an inspection plan and a repair plan are prepared for the corroded structure (S13). Thereafter, the process returns to the above-described S1 and repeats S1 to S13 every predetermined period. Said S11, S12, and S13 are called a maintenance plan process.

  As described above, in the method for predicting the strength deterioration of the corroded structure of the present invention, the change in the degree of the corrosion is measured over time in detail as much as possible with respect to the initial corrosion found in the target structure, The result data is statistically processed, and the measurement results are accumulated for each inspection process.

  From the obtained measurement results, the required performance analysis of strength is performed by FEM (finite element) analysis, etc., and it is determined whether the current possessed performance satisfies the required performance, and appropriate measures are taken. Using the surface shape creation model for reproducing various uneven surface shapes by changing parameters, and further matching the unevenness of the corroded surface shape measured in three dimensions, which is the accumulated result data, Parameters relating to the surface shape creation model are calculated.

  Subsequently, a future corrosion surface shape based on the surface shape creation model based on the parameters is calculated, and a future corrosion progress change (strength deterioration) in the corrosion structure is predicted from the future corrosion surface shape. Based on the prediction of deterioration of possession performance (strength).

  Examples of the method for predicting the deterioration of the retained performance (strength) from the predicted future corrosion surface shape include a finite element method analysis or a simple strength estimation method for the predicted future corrosion surface shape. The finite element method analysis uses a plate element or a solid element, and when performing the finite element method analysis on a plate element, the size of the element is preferably 20 mm or less per side, and the plate thickness is the predicted surface shape in the element For the average plate thickness of the plate thickness at the lattice point and the position of the center plane, the average value of the eccentricity at the lattice point of the predicted surface shape may be used. The simple strength estimation method uses a statistical index such as an average thickness value or standard deviation obtained from the predicted future corrosion surface shape.

  After that, based on the above-mentioned prediction of the deterioration in possession performance, the maintenance management plan is reviewed and revised in consideration of social conditions (such as the economy) and the latest technology (development of new and inexpensive construction methods), and the life of the target structure. The cycle cost can be set as small as possible, that is, the life cycle cost can be optimized (minimized).

  Next, an example of the surface shape measuring method used for the strength deterioration predicting method of the corroded structure of the present invention will be described. First, the structure of the surface shape measuring apparatus used for the surface shape measuring method will be described. In the surface shape measuring apparatus, as shown in FIG. 2, a liquid crystal projector (projection unit) for projecting a projection pattern (light pattern) 3a based on preset projection pattern information onto a corroded surface 1 to be measured. 3 is provided with its optical base point (the position of the light source) arranged on the virtual reference line 2.

  Although not shown, the liquid crystal projector 3 includes a light source and a liquid crystal panel for forming a projection pattern 3a based on the projection pattern information with respect to light from the light source. The projection pattern information may be analog information or digital information, but digital information including color information together with position information related to the projection pattern is preferable for convenience of screen processing described later.

  The projection pattern 3a is different from the color of the surface of the corroded surface 1 based on the projection pattern information, and is a light spot of a measurement color (preferably a contrast color (complementary color)) that can be distinguished from the color of the surface of the corroded surface 1. For example, a mesh-like (square grid, grid-like) pattern having a repetitive structure is set. Therefore, the color of the projection pattern 3a is preferably white when the corroded surface 1 is black or dark gray like the corroded surface. The pattern may be a triangle or pentagon other than a square lattice, or may be a collection of points with predetermined intervals. When the projection pattern 3a has grid-like marks, the distance between the marks adjacent to each other varies depending on the state of the corroded surface, but is 10 mm or less, more preferably 5 mm or less, and 2 mm to further improve reproducibility. The following is desirable.

  Further, the liquid crystal projector 3 uses a memory (storage unit) for storing detailed position information such as the shape and interval of the projection pattern as digital information, and the projection pattern 3 a is liquid crystal based on the digital information of the projection pattern. Although not shown, a driver for forming each pixel of the panel is provided.

  Further, the projection pattern information includes, on the corroded surface 1, optically distinguishable from a pattern image, for example, a singular mark that is a dot-like light spot that differs in at least one of luminance, chromaticity, and saturation. Further, it may be included at a preset position in the projection pattern 3a. Further, it is desirable that the singular mark has a different chromaticity among luminance, chromaticity, and saturation (when the projection pattern 3a is blue, for example, red or yellow), and is also identified from the surface of the corroded surface 1. It is preferable that the chromaticity is set to be possible (for example, red or yellow).

  In addition, the singular mark may be plural (A to C in FIG. 2) in a preset arrangement (for example, a triple star arrangement at equal intervals on a straight line). It is desirable that the plurality of specific reference points A to C have a repeating structure. By providing such a peculiar mark, the correlation and correspondence between the projection pattern 3a based on the projection pattern information and the later-described shooting pattern (imaging pattern) 6a can be determined more reliably and accurately. The reproducibility of the measured surface shape, which is a measurement result close to the actual surface shape of the corroded surface 1, can be improved.

  In the surface shape measuring apparatus, the projection for measuring or setting the projection angle α, which is the acute angle of the optical axis 4a, which is the central axis of the projection light 4 from the liquid crystal projector 3 with respect to the corroded surface 1, with respect to the virtual reference line 2 described above. An angle setting section (not shown) is preferably provided.

  Further, in the surface shape measuring apparatus, the image light 5 from the surface of the corroded surface 1 on which the projection pattern 3a is projected is received and photographed, and the obtained photographing pattern 6a is output as a digital image signal. A camera (camera unit) 6 is provided with its optical base point (lens position) arranged on the virtual reference line 2. An example of the digital camera 6 is a color CCD camera having 1 million to 10 million pixels, for example, 3 million, and may further include an optical zoom of about 2 to 10 times.

  The digital image signal includes digital information of the photographing pattern 6 a in which the projection pattern 3 a is deformed according to the surface shape of the corroded surface 1. Further, the digital image signal includes digital information of each of the marks A ′ to C ′ in which the positional relationship between the marks A to C is deformed according to the surface shape of the corroded surface 1.

  A light reception angle setting unit (not shown) for measuring or setting a light reception angle β that is an acute angle of the optical axis 5a of the image light 5 to the digital camera 6 with respect to the virtual reference line 2 may be provided. Furthermore, an interval setting unit (not shown) is provided for measuring or setting an interval l between the base point on the virtual reference line 2 of the liquid crystal projector 3 and the base point on the virtual reference line 2 of the digital camera 6. Also good.

  Such a virtual reference line 2 is for determining the positional relationship between the liquid crystal projector 3 and the digital camera 6, and therefore, the virtual reference line with respect to the averaged virtual plane of the corroded surface 1. 2 is not preferable because it is difficult to measure when set to be vertical, and is preferably parallel, but measurement is possible even if it is not particularly parallel. In other words, even when it is not parallel, it can be made parallel by correction from the data of the photographing pattern 6a included in the obtained digital image signal.

  Further, the arrangement of the liquid crystal projector 3 and the digital camera 6 with respect to the corroded surface 1 is projected from the liquid crystal projector 3, and a projected pattern 3a is formed on the corroded surface 1, and the projected pattern 3a thus formed is digitally displayed. Although it is only necessary to be able to shoot with the camera 6, it is desirable that the optical axis 4a of the light 4 from the liquid crystal projector 3 and the optical axis 5a of the image light 5 to the digital camera 6 are arranged so as to intersect each other. Is more distant than the surface of the corroded surface 1.

  In addition, it is desirable that the liquid crystal projector 3 and the digital camera 6 are arranged so that the projection angle α and the light receiving angle β are substantially equal to each other. The fact that the projection angle α and the light receiving angle β coincide with the normal direction of the averaged virtual measurement plane of the corroded surface 1 indicates that the projected pattern 3 a imaged on the corroded surface 1 is This is not preferable because it hardly deforms depending on the shape (unevenness). Therefore, it is desirable that the angles of the projection angle α and the light receiving angle β are set in a state where the optical axes 4a and 5a are not in agreement with the direction of the normal line and are inclined with respect to the normal line. The angle range of the projection angle α is preferably about 45 degrees in terms of accuracy, but may be less than 90 degrees. The angle range of the light receiving angle β is preferably about 45 degrees in terms of accuracy, but is less than 90 degrees. Good.

  Further, it is desirable that each elevation angle of the liquid crystal projector 3 and the digital camera 6 with respect to the corroded surface 1 is zero. Each elevation angle being zero means that the virtual optical axis plane including the optical axes 4a and 5a is parallel to the normal direction of the virtual measurement plane.

  Each digital information of the projection pattern 3a and the photographing pattern 6a is input, and a personal computer (measurement unit) 7 for measuring the surface shape of the corroded surface 1 based on the digital information is used to display the display unit 7a such as a liquid crystal panel. It is provided. In the present embodiment, in the measurement of the surface shape by the personal computer 7, the projection angle α, the light receiving angle β, and the interval l are added as digital data, if necessary, by user input or input by the above-described measurement. Also good. In addition, the personal computer 7 stores in advance functions as a control unit for controlling the projection angle setting unit, the light receiving angle setting unit, and the interval setting unit in order to set the projection angle α, the light receiving angle β, and the interval l. You may control according to a program or a user's input.

  In the above embodiment, the liquid crystal projector 3 is used as an example. However, any projection pattern 3a may be used as long as the projection pattern 3a can be formed on the corroded surface 1, and the projection pattern 3a is reflected by a polygon mirror that reflects laser light in a desired direction. May be a laser projector that can be formed on the corroded surface 1. The polygon mirror has a polygonal column shape with a mirror-finished peripheral surface, and rotates with the central axis as a rotation axis and displaces the rotation axis to apply a laser to the rotating peripheral surface. Light can be irradiated at a predetermined timing, reflected in a desired direction, and irradiated to a desired position to form a predetermined pattern.

  As an embodiment of the surface shape measuring apparatus according to the present invention, for example, as shown in FIG. 3, the distance l between the liquid crystal projector 3 and the digital camera 6 is 50 cm, the focal length of the liquid crystal projector 3 is 2.62 cm, 6 has a focal length of 11.7 cm, a projection angle α of the liquid crystal projector 3 with respect to the corroded surface 1 is set to 75 degrees, and a light receiving angle β of the digital camera 6 from the corroded surface 1 is set to 75 degrees. The number of pixels (pixels) of the digital camera 6 is 3 million.

  Next, a surface shape measurement method by image measurement using the surface shape measurement apparatus described in the present embodiment will be described based on the flowchart shown in FIG. As shown in FIG. 3, first, the liquid crystal projector 3 and the digital camera 6 are fixed by setting the interval l, the projection angle α and the light receiving angle β between the liquid crystal projector 3 and the digital camera 6 as in the above-described embodiment. (S31).

  Subsequently, with the liquid crystal projector 3 with respect to the corroded surface 1 which is a corroded steel surface, the blue projected pattern 3a including the red singular target point is set to a distance of 2 mm between the grid-like target points adjacent to each other. To project (S32). Thereafter, the projected projection pattern 3a is photographed by the digital camera 6 (S33). At this time, in the digital camera 6, the projection pattern 3a is deformed according to the surface shape of the corroded surface 1 as shown in FIG. The red singular mark is a red dot 6b continuous on a straight line.

  Next, a digital image signal of an image including the photographing pattern 6a photographed by the digital camera 6 is converted by the personal computer 7 to a target point (for example, a lattice point of the lattice) or a specific target point pixel on the photographing pattern 6a. The position of (pixel) is determined using, for example, preset intervals and chromaticity at each target point such as a lattice point of the projection pattern 3a and each specific target point (S34).

  Subsequently, the personal computer 7 performs the following (1) projection angle α, light receiving angle β, (2) focal lengths fL, fR, (3) interval l with respect to the surface coordinate analysis program stored in the personal computer 7 in advance. (4) The coordinate value of the target point on the projection pattern 3a and (5) the data of the pixel position of the target point on the shooting pattern 6a are input (S35).

  Thereafter, the personal computer 7 calculates the three-dimensional coordinates of each target point based on the above data, the coordinates of the projection pattern 3a of the liquid crystal projector 3, and the coordinates of the CCD photographing pattern 6a of the digital camera 6. The unevenness on the corroded surface 1 is measured. As the generated corroded surface, as shown in FIG. 6, a reproduced surface 8 that reproduces the surface shape of the corroded surface 1 is obtained (S36). On the reproduction surface 8, a reproduction photographing pattern 8 a corresponding to the photographing pattern 6 a is also shown for comparison.

  Below, the surface shape reproduced by the image measurement with respect to the corroded surface 1 of the corroded surface was compared with the surface shape by high-precision but expensive laser measurement with respect to the same corroded surface 1. As shown in FIG. 7, a high correlation was observed between the above results. That is, it was found that the surface shape measuring device and the surface shape measuring method according to the present invention are non-contact measurement by an optical method and the reproducibility of the corroded surface 1 is excellent.

  Therefore, the surface shape measuring apparatus and the surface shape measuring method of the present invention can be made inexpensive and easy to carry because the liquid crystal projector 3 and the digital camera 6 can be reduced in cost and convenience of measurement. It can be seen that the improvement is possible and the excellent reproducibility of measurement can be secured.

  By reproducing the surface shape as described above with respect to the corroded surface 1, the degree of corrosion on the corroded surface can be measured, and the remaining load capacity of the actual structure having the corroded surface 1 is evaluated according to the degree of corrosion. Is possible.

  In the above description, the corroded surface 1 is the surface of a structure mainly composed of steel. However, the present invention is not limited to the measurement of the surface shape of the structure. For example, precision parts such as gears, surface measurement of a vehicle, and surface measurement of a face or a fingerprint can be used.

  Furthermore, in the present invention, since the position information of the measurement surface can be measured digitally, it is possible to perform fixed point observation over time for buildings (structures) such as bridges, buildings, and dams to which loads are constantly applied. Thus, the degree of deformation of the building can also be measured. Therefore, the present invention can be suitably used to detect the deformation of the building over time.

  Next, an example of a method for determining the same point corresponding to each other on the projection pattern 3a and the photographing pattern 6a will be described. For the determination of the same point, RGB values (text-formed) included in each pixel (pixel) are used. The RGB value of the pixel of the projection pattern 3a is LR (K, L), and the RGB value of the pixel of the shooting pattern 6a is RR (I, J). When the point in the positional relationship shown in FIG. 8 is a corresponding point, the point (I, J) is obtained when the sum of squares of the difference between the RGB values of the corresponding points is minimized as in the following equation (1). And the point (K, L) are the same point. This operation is called matching. In the above matching, how many pixels around the sum of squares can be set can be freely set in the program.

  Next, an example of a method for calculating the unevenness on the corroded surface 1 will be described with reference to FIGS. 3 and 9 to 13. In FIG. 9, point A is the light source position of the liquid crystal projector 3, point B is the lens position of the digital camera 6, point C is the intersection of the optical axis 4a of the liquid crystal projector 3 and the optical axis 5a of the digital camera 6, and point D is The point on the corroded surface 1, the point E is the origin of the projection pattern 3a, the point F is the origin of the imaging pattern 6a, the point G is a position corresponding to the point D on the projection pattern 3a, and the point H is on the imaging pattern 6a. The position corresponding to the point D is represented.

  For such imaging conditions, the conditions shown in FIG. 3 are used, and the intervals l, α, β, fL, and fR are known. The elevation angles of the liquid crystal projector 3 and the digital camera 6 were set to zero.

  In FIG. 10 showing the projection pattern 3a, if the X and Y coordinates of the point G are XL and YL, γ and δ are expressed by the following equations.

γ = tan ⁻¹ (XL / fL) (2)
δ = tan ⁻¹ (YL / (fL ² + XL ² ) ^1/2 ) (3)
In FIG. 11 showing the photographing pattern 6a, if the X and Y coordinates of the point H are XR and YR, ε and φ are expressed by the following equations.

ε = tan ⁻¹ (XR / fR) (4)
φ = tan ⁻¹ (YR / (fR ² + XR ² ) ^1/2 ) (5)
In FIG. 12, θ1 and θ2 are α, β, γ, and ε, respectively. Θ1 = α−γ (6)
θ2 = β + ε (7)
It can be expressed as. Therefore, θ3 is
θ3 = π− (θ1 + θ2) (8)
It is represented by

  From the sine theorem,

Therefore, b and c are

It is represented by

Therefore, x and z use θ1 and b,
x = b cos θ1 (12)
z = bsin θ1 (13)
It is represented by

In FIG. 13, y uses b and δ,
y = btan δ (14)
It can be expressed as.

  Using equations (2) to (14) to rewrite x, y, z,

It is expressed. Among the symbols in the equations (15) to (17), l, α, β, fL, and fR are known, so that the coordinates (XL, YL) and (XR, YR) of the points G and H are If known, the coordinates of the point D when the point A is the origin, that is, the coordinates at an arbitrary point on the corroded surface 1 are obtained. Therefore, if the same point on the projection pattern 3a and the photographing pattern 6a can be determined, the unevenness of the photographed corroded surface 1 can be calculated.

  In the method for predicting the strength deterioration of a corroded structure according to the present invention, following the generation of the reproduced surface 8, the residual load capacity analysis model of the corroded structure is created based on the reproduced surface 8, and the remaining load capacity is evaluated. . Evaluation of the load bearing capacity of a corroded structure using the residual load bearing capacity analysis model is performed in either of the following a) and b).

  a) If the part that dominates the load bearing force and the fracture mode can be estimated, the load bearing force can be simply measured by limiting to that part. That is, if the shape of the specific part is a rod shape, a rectangular plate shape, or a cylindrical shell shape, the load bearing force can be obtained from the past analysis result or the like only with the shape size or the shape ratio.

For example, in the case of a steel plate garter (used for bridge girder etc.)
a-1) If the part is a flange subjected to compression, buckling may be evaluated by determining the plate thickness by (average plate thickness + 2 × standard deviation).

  a-2) If the part is a flange that is subject to tension, the average plate thickness may be examined in several cross sections orthogonal to the load acting direction, and the strength evaluation may be performed using those minimum values.

  a-3) If the part is a web subjected to shearing, the average plate thickness may be used for analysis.

  b) When there are multiple sites and failure modes that control the load bearing capacity, and which one controls the failure of the entire structure, a corroded surface is generated in the entire structure, and the generated corroded surface is used to make a finite What is necessary is just to calculate a load bearing capacity by an element model (FEM) analysis and to obtain a load bearing capacity.

  Next, the relationship between the corrugated shape of the corroded surface of the corroding structure and its statistical characteristics will be described with reference to FIGS. Conventionally, the retained strength of corroded steel materials is often estimated using the average plate thickness obtained from the plate thickness measurement results at several corroded locations. It has been used as a standard.

  In this way, it is thought that it is simple and effective in practice to estimate the holding strength using statistics such as average plate thickness, but in order to evaluate the holding strength of corrosive steel with higher accuracy. Therefore, it is necessary to measure the unevenness of the corroded surface in more detail and reflect the statistical characteristics in the strength evaluation.

However, there have been few examples of the detailed measurement of the corrugated surface of the corroded surface and the statistical characteristics of the corroded surface. Therefore, the present inventors obtain a plate garter that is a main girder of a corroded bridge that has been shared and disposed of for about 100 years, and measure the unevenness of the corroded surface of each part of the plate garter. The relationship between the corrugated shape of the corroded surface and its statistical characteristics was investigated. The following abbreviations described in FIGS. 14 to 23 are as follows. t _avg is the average thickness, στ is the standard deviation of the thickness, t ₀ is the maximum thickness (thickness before corrosion), and t _min is the minimum thickness.

  The web, edge stiffening plate, upper flange and lower flange in the plate garter were used as shear test specimens, and the measurement results of the plate thickness histograms are shown in FIGS. 14 to 17, respectively.

  The distribution form of the web thickness histogram shown in FIG. 14 is almost symmetrical, and the maximum frequency generation thickness and the average thickness are almost the same. Further, the average corrosion depth (= 1 mm) and the standard deviation (= 0.58 mm) were the smallest among the four specimens shown in FIGS.

  In the plate thickness histogram of the end stiffening plate shown in FIG. 15, the maximum frequency generation plate thickness is generated as the average plate thickness, but the standard deviation (= 0.83 mm) is larger than that of the web. This is presumably because large pitting corrosion and layered rust were generated visually on the end stiffener plate, and larger irregularities were formed.

  In the plate thickness histogram of the upper flange shown in FIG. 16, the maximum frequency occurrence plate thickness does not match the average plate thickness, and the standard deviation (= 2.49 mm) is 4 shown in FIGS. It was the largest of the two specimens. This was thought to be due to the fact that local corrosion by sleepers is dominant in the upper flange.

  In the plate thickness histogram of the lower flange shown in FIG. 17, it can be seen that the frequency near the average plate thickness is extremely high. In this lower flange, it is considered that the entire upper surface of the lower flange on the girder north side is corroded with small irregularities, and the thickness of this portion is dominant. In the lower flange, since the thickness is locally reduced in the vicinity of the vertical stiffener and the end stiffener plate, the standard deviation is relatively large.

  Next, the web, flange A1 and flange A2 in the plate garter were each used as a bending test specimen A, and the measurement results of their plate thickness histograms are shown in FIGS. 18 to 20, respectively. Each of these specimens A had a rivet, but the measurement result of the thickness of the rivet part was not taken into account in calculating the statistics.

  18 to 20, the same statistical characteristics as the web of the shear test specimen are observed in the web in which slight corrosion is distributed over the entire surface, the maximum frequency appears in the vicinity of the average plate thickness, and the histogram has a small standard deviation. It became the distribution shape. Further, in the flange, both A1 and A2 are remarkably corroded, and in particular, since there are many large pitting corrosions in A1, the distribution has a large standard deviation, and the maximum frequency does not necessarily occur near the average plate thickness. That is, the maximum frequency occurs at a position different from the vicinity of the average plate thickness.

  Finally, the web, flange B1 and flange B2 in the plate garter were each used as a bending test specimen B, and the measurement results of their plate thickness histograms are shown in FIGS. 21 to 23, respectively. In the specimen B, corrosion other than the flange B2 is relatively slight as shown in FIGS. 21 and 22, and from FIG. 21 and FIG. 22, a distribution form having the maximum frequency plate thickness near the average plate thickness is confirmed. It was. On the other hand, the flange B2 is corroded to the same average plate thickness as the flange of the specimen A, but since there is almost no local corrosion, the standard deviation value of the flange B2 is the standard of each flange A1, A2. It turns out that it is smaller than each value of deviation.

  From the results shown in FIGS. 17 to 23, the standardized average value obtained by dividing the difference between the average plate thickness and the maximum frequency plate thickness by the average plate thickness or the maximum frequency plate thickness, and the standard deviation is the maximum plate. It can be seen that the strength of the corroded structure can be measured by the standardized deviation value standardized by dividing by the thickness or the average plate thickness. That is, it can be seen that the degree of corrosion and the degree of local corrosion can be grasped based on the standardized average value and standardized deviation value based on the above results.

  Next, the above-described surface shape creation model will be described.

  In this model, the actual structure corrodes under the influence of local environmental factors such as structural features and installation locations, and the environmental factors include the effects of climate such as temperature and humidity. Since the influence of flying salinity is large, the external factor that causes corrosion such as flying salty is called the attack factor, and the corrosion progress of the corroded surface in steel is modeled by the attack factor. That is, it is assumed that the attack factor is randomly dropped to each two-dimensional position on the corroded surface, and one attack factor gives the same amount of corrosion to the corroded surface. Corrosion on the corroded surface should be handled continuously, but it is preferable to consider it discretely in order to express the corrugated shape of the corroded surface and perform statistical processing thereof because the amount of calculation can be reduced.

  Therefore, in this model, the continuous corroded surface is divided into lattice points as shown in FIG. 24, and the attack factor is generated on the lattice points having a square mesh, while the corrosion occurs only in the depth direction of the corroded surface. However, if the spatial distribution in the surface direction of the corroded surface is taken into consideration, the accuracy is considered to be higher, and correction is made as described later in consideration of the spread caused by the attack factor.

Assume that n attacks randomly fall on lattice points on a corroded surface such as steel during a period ΔT. It is assumed that the amount of corrosion at a lattice point due to one attack factor is due to the strength F of the attack factor. The strength F of the attack factor is
F (mm / piece) = q (mm / (piece · mm ² )) · A (mm ² ) (Equation 18)
It is represented by

  Here, q is the corrosion strength per unit area, and A is one area of the lattice points. As shown in FIG. 24, the grid point area A represents the area of the gray portion and is determined by the grid point width.

In addition, the total attack amount in a certain period ΔT is
Q (mm) = F (mm / piece) · n (piece) (Formula 19)
It becomes.

  It is assumed that the corrosion of the corroded structure in this model proceeds with this total attack amount given every ΔT. Further, it is considered that it is sufficient for the corrosion progress to represent a change in units of years, and ΔT = 1 year.

  The concept of the time degradation model is shown in FIG. Here, the elapsed time T = 10 (year), the number of attack factors n = 3 (pieces / year), and the unit time ΔT = 1 (year) are given. The circles in FIG. 25 represent attack factors. Corrosion depths (virtual corrosion holes) corresponding to the strength F (mm) of the attack factors are formed at the lattice points hit by the attack factors. Also, the change in the color of the attack factor in FIG. 25 indicates that the attack factor hits the lattice point a plurality of times. Thus, the independent corrosion depth of each lattice point at the elapsed time T is obtained. In order to give a spatial correlation to the independent corrosion depth of this adjacent point, a spatial autocorrelation model (Equation 20) is applied to generate an integrated corrosion depth.

  Next, a spatial autocorrelation model according to the strength deterioration prediction method of the present invention will be described. First, by observing the surface shape of a steel plate that has undergone considerable corrosion, it can be seen that the areas where the corrosion is severely spread. Apply Okumura, Fujii et al. (Makoto Okumura, Ken Fujii, Makoto Tsukai: Modeling steel plate corrosion shape considering spatial autocorrelation, JSCE Proceedings pp109-116, 2001.) Therefore, it represents the corrosion phenomenon that occurs continuously in space and assumed the surface shape of the corroded structure.

  It seems that it is better to assume the corrosion depth at each point in consideration of not only that point but also the influence from adjacent points. In the spatial autocorrelation model, by considering the independent corrosion (independent corrosion depth) at that point, it is assumed that they act spatially between the points of interest and form a corrosion depth. By applying spatial autocorrelation, it is possible to determine the corrosion depth with spatial relevance. The spatial autocorrelation model is expressed by equation (20). When this is represented in a matrix, it is as follows.

  FIG. 26 shows the concept of the spatial autocorrelation model on the corroded surface. Divide the corroded surface into grids and discretize them. In the spatial autocorrelation model, the corrosion depth Zi at the i point is the independent corrosion depth vi of the i point itself and the independent corrosion depth vj of the grid point j that is a certain distance (dij distance between the i and j points) from that point. Assuming that ρ is a parameter representing the strength of distance attenuation, and indicates the degree of influence from adjacent points. According to previous studies (the above-mentioned JSCE paper pp109-116, 2001), the attenuation parameter ρ is preferably 0.3 to 0.5, for example, 0.4.

  Further, the influence from the point j varies depending on the distance between the two points. If the distance between the point j and the point of interest i is close, the influence is greatly affected, and if the distance is increased, the mutual influence is reduced. In FIG. 27, the difference in the influence of distance attenuation is shown by using three points color-coded for points having different distances from the point i in FIG. In the above equation (21), Ω is a region having spatial influence. Previous studies have assumed a 50 mm radius region for each point.

  The spatial autocorrelation model described in the above study creates a corroded surface by replacing the corroded surface with a discrete grid and applying spatial autocorrelation to the independent corrosion depth generated at each grid point. .

  However, in the conventional spatial autocorrelation models, when the mesh width is changed, there is a problem that even if the same independent corrosion depth is given to the same point, the generated corrosion depth is different. As an example, a case where a steel plate of 100 mm × 100 mm is divided into meshes as corrosion structures and a corrosion depth is generated using a spatial autocorrelation model will be considered. The mesh width is changed to 10 mm, 5 mm, 4 mm, 2 mm, and 1 mm, but the independent corrosion depth on the same lattice point is made equal. FIG. 28 shows the change of the cross section having the corrosion depth generated by changing the mesh width.

  As can be seen from FIG. 28, the finer the mesh width, the greater the corrosion depth. This is because, as shown in FIG. 29, the 1 mm mesh model in FIG. 29B is affected by more points at the i point than the 5 mm mesh model in FIG. 29A. is there. In other words, in the conventional spatial autocorrelation model, a discrete corrosion depth is given to the lattice points, and it cannot be applied to a continuous corrosion surface. That is, when the conventional spatial autocorrelation is applied to a continuous corrosion surface, the corrosion depth at point i is affected by the independent corrosion depth at point j having an area dA as shown in FIG. .

  Therefore, in order to consider a continuous corroded surface, the independent corrosion depth needs to be continuous. Therefore, the corrosion depth Zi at point i is expressed by the following equation (22).

It is expressed. That is, in the spatial autocorrelation model of the present invention, in order to calculate the volume of the corrosion depth Zi as a continuous one from each of the discrete grid-like corrosion depths, It is obtained by approximation using volume. The volume is approximately obtained by multiplying the depth (strength) of the independent corrosion depth by the area of the independent corrosion depth. For example, as shown in FIG. 31 (a), the influence of each independent corrosion depth in the vicinity is taken into account by the volume obtained by multiplying the independent corrosion depth by the area of the lattice point and the distance from the i point. The volume of the continuous independent corrosion depth at the point i can be approximated. Further, as shown in FIG. 31 (b), if the grid point width is made fine, the corrosion volume by each independent corrosion depth approaches the actual corrosion volume more accurately than the conventional one. That is, the i-point corrosion depth Zi is affected by the j-point corrosion volume (the product of the independent corrosion depth v and the lattice point area dA). In the spatial autocorrelation model of the present invention, By calculating using the volume, the influence on the i point in a certain area can eliminate the dependency on the mesh width.

  By modifying the spatial autocorrelation model of the present invention as described above, it can be understood that even if it is divided into arbitrary meshes and discretized, it can be treated as a continuous corroded surface. The depth of corrosion according to this modified model converges on a continuous corrosion surface as the mesh width is reduced. As an example, a 100 mm × 100 mm steel plate is divided into a mesh as a corroded structure in the same manner as described above, a corrosion depth is generated using a spatial autocorrelation model, and a change in the corrosion depth in a certain cross section is shown in FIG. As can be seen from FIG. 32, as the lattice point width is made finer, it is understood that the value converges to a certain value. When the mesh width is set to 10 mm, the difference in the generated corrosion depth is large and the accuracy is low, but if the mesh width is 5 mm, the difference in the corrosion depth from the mesh width of 1 mm is about 0.3 mm. It can be judged that a mesh width of 5 mm or less used in the spatial autocorrelation model of the present invention is almost appropriate.

  As described above, the independent corrosion depth v can be calculated from the equation (23), which is an inverse matrix of the spatial autocorrelation model, using the data of the calculated actual corrosion depth z.

Next, the surface shape creation model of the present invention will be described. First, it was assumed as a surface shape creation model that the independent corrosion depth at each lattice point acts only in the depth direction at that lattice point. At this time, as shown in FIGS. 33A to 33D, when the lattice width of the lattice 11 assumed on the corrosion surface 1 is reduced, the area 11a of each independent corrosion depth at the lattice point of the lattice 11 is also reduced. Become. 33A and 33C show the grating 11 having a width of 5 mm, and FIGS. 33B and 33D show the grating 11 having a width of 1 mm. Therefore, the area 11a at the lattice point is 25 mm ² in FIG. 33A and 1 mm ^{2 in} FIG. If the width of the lattice points is reduced, the number of lattice points needs to be increased if the total corrosion amount is constant.

Therefore, in the surface shape creation model, the number of attack factors is changed in accordance with the change in the number of lattice points. That is, assuming that the ratio n / N of the number n of attack factors and the total number N of lattice points is the influence area ratio of the attack factor on the surface, the number of attack factors is changed so that this is constant. In this change in the number of attack factors, the volume of the independent corrosion depth in one grid point area of 25 mm ² of 5 mm mesh (FIG. 33A) is 1 mm ² of grid point area of 1 mm mesh in the same region (FIG. 33 ( b)) Must be the same as the volume of 25 independent corrosion depths. Accordingly, FIG. 33 (c), as shown in (d), 1 single attack factor in 5mm width becomes 25 5 ² times in 1mm wide, this is the independent corrosion depth obtained if touch points each The volume will be the same.

  For example, in the case of a steel plate of 100 mm × 100 mm, if the lattice point width is 5 mm, the total number N of lattice points is 441 (21 × 21). If the grid point width is 1 mm, the total grid point number N is 10201 (101 × 101). Assuming that n is 50 when the grid point width is 5 mm, when the grid point width is 1 mm, (1 mm N = 10201) × (5 mm n = 50) / (5 mm N = 441). N in the width is 1156. FIG. 34 shows a plot in which 1156 hit points at 1 mm and 441 hit points at 5 mm are randomly given to lattice points. Although the total volume of the independent corrosion depth is the same at 5 mm and 1 mm, as can be seen from FIG. 34, when the attack factor is randomly decreased, the surface shape of the independent corrosion depth is completely different from FIG. 34. I understand.

In order to solve this problem, in the surface shape creation model of the present invention, a model that does not depend on the width of the lattice 11 is constructed by giving an influence radius r as one parameter to the attack factor. Here, the radius of influence of one attack factor is r (mm). The strength F of the attack factor and the number n of the attack factors do not depend on the mesh width. The total attack volume of the entire corroded structure per unit time is
Q = F (mm / piece) × πr ² (mm ² ) × n (piece) (Formula 24)
It becomes.

  In the first assumed model described above, it is assumed that the attack factor falls on each lattice point of the lattice 11. However, in the surface shape creation model of the present invention, not each lattice point as shown in FIG. In this case, the x and y coordinates generated by random numbers are randomly dropped. As a result, the number and coordinates of attack factors can be determined without depending on the lattice width of corrosion.

  It is assumed that one attack factor causes corrosion of F (mm) within the influence radius r (mm) of the attack factor. Therefore, the independent corrosion depth of F (mm) is given to the lattice points existing within the influence radius r. As shown in FIG. 35 (d), the independent corrosion depth □ is generated at the lattice point within the influence radius r by the attack factor of the first year.

  Moreover, the independent corrosion depth (circle) is similarly produced | generated by the attack factor of the 2nd year. Grid points where □ and ○ overlap indicate that they have been attacked for two consecutive years. When this influence radius r (mm) is increased, corrosion occurs at many lattice points due to one attack factor. When this independent corrosion depth is integrated over a certain period as in the previous model, the amount of attack is accumulated at each lattice point, and the independent corrosion depth at the elapsed time T is generated.

  Thus, in order to consider the spatial spread in the obtained independent corrosion depth, in the surface shape creation model of the present invention, the depth of corrosion is obtained by incorporating the idea of the corrected spatial autocorrelation (Equation 25). ing.

Next, a method for setting the attack point of the attack factor will be described. As a method for generating uniform integer random numbers, in the present invention, a uniform real random number sequence between (0, 1) is created by a multiplication congruence method. Multiplication congruential method is
Recurrence formula: Nj + 1 = K * Nj mod M (Formula 26)
This is a method of creating a number sequence. This gives a sequence of numbers, where K and M are integer constants, and the remainder obtained by dividing the number obtained by multiplying the integer Nj by K by M is the next Nj + 1. However, only a completely identical number sequence is given from the same initial value, and if it is repeated M times, the same value as the initial value appears and becomes a periodic number sequence. So, to make this cycle as long as possible,
K = 48828125, M = 2 ³¹ (Formula 27)
Pseudo-uniform random numbers were used. In order to make this a uniform integer random number of 1, 2, 3,... N, the integer part was extracted and used by multiplying by N.

  Next, the generation of hit points will be described. First, in the surface shape creation model of the present invention, the coordinate axis of the hit point is considered at a pitch of 0.1 mm. For example, when the x coordinate is from 0 to 100, this is set to 1000 values of 0.1 mm pitch from 0.0 to 100.0. This is multiplied by 10 and considered as 0 to 1000. Within this range, after generating a uniform integer random number, the numerical value obtained by dividing by 10 again was taken as the x coordinate value. In this way, a random number sequence with a pitch of 0.1 mm is obtained. Further, if the same method is performed by changing the initial value, the y-coordinate can be generated as a random number sequence. The point that can be represented by this coordinate value was the coordinate value of the hit point falling on the plane of the corroded surface. FIG. 36 shows the frequency distribution when the hit points are generated with uniform random numbers in the x coordinate (0 to 100) and the y coordinate (0 to −100). Originally, uniform random numbers must have the same appearance frequency within a range. As shown in FIG. 36, although the frequency of the created uniform random numbers has some variation, it seems that they are generated evenly on average.

Then, the examination result about the parameter in a surface shape creation model is shown below.
(1) Influence radius r of attack factor and mesh width of lattice 11 As shown in FIG. 37 (a), in the method of generating the independent corrosion depth described in the previous section, within one radius r (mm) per attack factor. F (mm) deep corrosion is generated. As shown in FIG. 37 (b), the cumulative total attack volume after the elapsed time T is expressed by the following equation (28).

It can be expressed. dA is a lattice point area. Independent corrosion depth is continuous, but this model treats it as a collection of discrete points. As shown in FIG. 37 (c), if the mesh width is dense, the cumulative total attack volume is closer to the value obtained by performing the quadrature calculation of the actual corrosion depth than when the mesh width is roughly taken as shown in (c). Therefore, in order to investigate the relationship between the mesh width and the influence radius r, an independent corrosion depth was given to the 10 cm square corrosion surface while changing the mesh width. The mesh width is set to 1, 3, 5, 10 mm in addition to the standard 0.5 mm. The parameters are n = 100 (pieces), F = 0.01 (mm / pieces), T = 20 (years), and the influence radius r is changed to 1.0 mm, 3.0 mm, 5.0 mm, 10.0 mm. It was. FIG. 38 is a diagram in which the independent corrosion depth is plotted at each lattice point on the x coordinate axis. As shown in FIG. 38, when r = 1.0 mm or r = 3.0 mm, when the mesh width is 5 mm or 10 mm, the shape of this independent corrosion depth is not sufficiently captured, and a large difference appears. I understand that.

  On the other hand, as shown in FIG. 39, if the influence radius is r = 5.0 mm or r = 10.0 mm, it is possible to roughly grasp the change in the independent corrosion depth even if the mesh width is large. is there. When the mesh width is 2 mm, the shape of the independent corrosion depth is captured when r = 3.0 mm or more and the mesh width is 0.5 mm, and when r = 5.0 mm or more when the mesh width is 5 mm, the shape It can be seen from FIG.

  That is, if (influence radius r / mesh width)> 1.0, a value close to the total attack volume can be extracted. However, it can be seen that when the mesh width is roughly 10 mm, the difference from the total attack volume increases.

  Next, the relationship between the attack amount F, the attack number n, and the influence radius r will be described. First, the corrosion surface shape of the model of the present invention is determined by the strength F of the attack factor, its radius r, and the number n of the attack factors. When the total attack volume per unit time (formula 24 described above) is constant, the average corrosion amount is almost the same if the elapsed time is the same. As an example, the size of the corroded structure (steel plate) was 100 mm × 100 mm, the mesh width was 2 mm, the total attack amount was constant, and n, F, and r were changed as shown in Table 1. Based on n = 100 (pieces), F = 0.001 (mm / year), and r = 5 (mm), this is defined as (a) in Table 1. Tables (b) and (c) in which n and F are changed while r is constant are shown. Tables (d) and (e) in which r and F are changed with n being constant are shown. Furthermore, what changed n and r by making F constant is shown to (f) and (g) of Table 1.

  If the number n of attack factors is increased or the influence radius r of the attack factors is reduced, the standard deviation is reduced. On the other hand, it can be seen from Table 1 that the standard deviation increases when the number n of attack factors is reduced or the influence radius r is increased.

  Next, a comparison result between the predicted corrosion depth and the actually measured corrosion depth according to the surface shape creation model of the present invention will be described. First of all, in the case of modeling without a coating film, it should be compared with corrosion data such as weather resistant steel materials without a coating film, but data on the shape of such steel material could not be obtained. The above-mentioned corroded surface of the plate girder bridge after 100 years is compared. A part (100 mm × 100 mm) was cut out from the actual corrosion data of the members of the plate girder bridge and compared. The method of determining parameters is determined by the following method.

  As described in Equation (23) above, the independent corrosion depth at each point can be obtained by multiplying the measured corrosion depth by the inverse matrix of the spatial autocorrelation model. The total integrated value (sum) of the independent corrosion depth is the attack volume Q in the equation (28), and Q is determined by F, n, and r.

  So far, it is known from the spectrum that the influence radius r can be 3 mm to 8 mm, so set r = 3 mm, 5 mm, and 8 mm, respectively, and change F and n according to the measured value Q. Then, when the average corrosion depth and the standard deviation (index value of the unevenness amount) are obtained, they are as shown in FIGS. Since the standard deviation based on the actually measured value varies depending on n, n corresponding to the intersection of the standard deviation of the actually measured value and the curve in the figure is obtained, and the strength F of the attack factor can be obtained from the equation (24).

  42 to 45 show a comparison between the measured thickness histogram and contour lines and the thickness histogram and contour lines obtained by the above method. As can be seen from FIG. 42 to FIG. 45, it can be seen that the plate thickness created by the model expresses the measured value well.

  That is, as described in the above equation (23), the independent corrosion depth at each point is obtained by multiplying the actually measured corrosion depth by the inverse matrix of the spatial autocorrelation model. The sum of the independent corrosion depths is the total attack volume (FIG. 37 (b), Equation 28). From this total attack volume Q value, n and F are changed, and n and F are determined so as to be close to the standard deviation of the actual corroded surface (index value of the unevenness).

  46, 47, and 48 show contour maps of the flange, web, and end plate, and a histogram of corrosion depth. (A) of FIGS. 46 to 48 are measured corrosion depths, respectively, and (c) and (e) of FIGS. 46 to 48 represent predicted corrosion depths (corrosion surface models), respectively.

  As can be seen from the respective histograms (b), (d), and (f) of FIGS. 46 to 48 showing the measured corrosion depth and the predicted corrosion depth for each depth (mm), the average value / standard If the deviations are made equal, the histograms of the actually measured corrosion depth and the predicted corrosion depth based on the model are generated approximately the same.

  However, the maximum frequency differs greatly on the web. However, as long as the contour map is seen, it can be said that a similar shape can be expressed. In addition, when (c) and (d) in FIGS. 46 to 48 are compared, it can be said that the generated corroded surface differs depending on the influence radius r. It is considered that a shape close to the waveform of the actual corroded surface can be expressed by changing the influence radius r. The value of the influence radius r can be determined in the future using the above-mentioned power spectrum showing the waveform of the actual corrosion surface and the corrosion surface by the model.

  Next, the applicability of the surface shape creation model of the strength deterioration prediction method of the present invention in consideration of coating film deterioration will be described. First, while the coating film is in a state where it has a sufficient anticorrosion effect, the progress of corrosion becomes slow. In order to model the progress of corrosion, it is necessary to consider deterioration of the anticorrosion performance of the coating film (that is, deterioration due to occurrence of pinholes, cracks, and peeling, which is one state of the coating film). Therefore, a model capable of predicting the progress of corrosion is created by considering the influence of the coating film on the surface shape creation model described above. Furthermore, the applicability of this model was examined by comparing with the measured corrosion depth of corroded bridge members.

Modeling of corroded surface considering paint film In actual bridges, the surface is treated with paint film, and the paint plays a role in preventing corrosion. Therefore, it is necessary to consider the anticorrosive effect of the coating film when considering the actual progress of corrosion. If there is a coating film, the speed of corrosion will become slow, and corrosion will progress easily as a coating film deteriorates. In other words, in order to make a comparison with the corroded surface of an actual plate girder bridge, it is necessary to consider the difference in the corrosion speed due to such a coating film. Moreover, the coating film is periodically repainted, and the progress of corrosion changes each time. By considering the influence of the coating film, it is possible to represent the progress of corrosion in a form close to the actual phenomenon.

  The durability of the coating film varies depending on the type of coating, but is generally as shown in Table 2 below.

  Therefore, the assumption method of coating film deterioration of this model is shown below. First, the coating film applied to the corroding structure is assumed to gradually peel off over time, and the deterioration of the coating film is considered to be caused by fine pits (corrosion holes) generated in the coating film. Corrosion proceeds under the influence of corrosion factors in the pits of the coating film. If the coating film deteriorates with the passage of time, the number of pits of this coating film will increase and it will be strongly influenced by corrosion. Moreover, it is thought that the surrounding coating film also peels off gradually in the place where the coating film peeled off.

  In this surface shape creation model, the coating film capability TM is given to all lattice points to express the deterioration of the coating film. As shown in FIG. 49, the coating film ability (anticorrosion ability) is decreased by one for each certain period (in FIG. 49, the time between the circles adjacent to each other) within the service life period of the coating film. In other words, when the TM value becomes 0 (predetermined value) or less, it indicates that the coating film disappears after the first circle mark time if TM = 1, that is, it is recognized that there is no anti-corrosion ability. Yes.

  Further, it is assumed that the coating film hardly deteriorates in the initial state, but rapidly deteriorates with time. Accordingly, when the coating film ability is first given to the lattice points, the ratio is changed so that there are few lattice points having a coating ability of TM = 0 and many lattice points having a coating ability of TM = 5. If the paint film is repainted, the paint film ability is given to all grid points again and repeated. In addition, when a coating film at a certain point is peeled off, the surrounding coating film is gradually peeled off.

  In the surface shape creation model, the spread of the coating film is considered as shown in FIG. Assume that the attack factor falls into the pit and that the pit gradually spreads every year. In this pit spread, when the adjacent lattice points enter the spread, the coating ability becomes zero. However, there is no prior art for examining how the pits spread, and the present invention assumes that the pits spread by 1 mm per year after the attack factor drops.

  Then, the comparison result of the prediction corrosion depth by the surface shape creation model in consideration of the coating film and the actually measured corrosion depth is shown below. Based on the above, the surface shape creation model is created so that the average value and the standard deviation of the measured corrosion depth after 100 years are the same. This time, the specimen used for the measurement of rotted surface seems to have been repainted several times, but the actual film history is not fully understood. Therefore, in the present invention, the repainting of the coating film is assumed to be 30 years. Although the service life of a coating film changes with kinds of coating film, as shown in Table 2, it is about 20 years in the case of the corrosive structure arrange | positioned at a coast part. Therefore, the service life of the coating film was 20 years.

  Moreover, the ratio of the coating film ability to be given at the beginning and at the time of repainting is as follows: TM = 1 is 2% of the whole, TM = 2 is 5%, TM = 3 is 8%, TM = 4 is 10%, TM = 5 was assumed to be 75%. The result of modeling in this way is shown. FIG. 51 compares the actual corrosion depth of the flange, the contour map of the predicted corrosion depth obtained by the surface shape creation model of the present invention, and the histogram. Compared to the histograms of measured corrosion depth and model predicted corrosion depth, it can be said that they are very similar.

  FIG. 52 shows the deterioration of the coating film by this model as a coating film deterioration rate curve. For the first 10 years, the number of pits is small and the coating film deterioration rate is small. However, when the useful life is close to 20 years, the coating film spreads and the coating film capacity decreases, and the coating film deterioration is rapidly progressing. This seems to be a good representation of the coating deterioration curve as shown in FIG.

  Further, by changing the elapsed years while keeping the parameters constant, it is possible to create predicted corrosion surfaces at each elapsed time (after 20 years and after 60 years), respectively, as shown in FIG.

  FIG. 54 shows the result of reproducing the corroded surface by the above method for the end plate. FIG. 55 shows models of predicted corrosion surfaces predicted for 20 years and 60 years from the 100 year model. In the end plate, it is judged that a shape similar to the histogram of the actual corroded surface can be captured visually in the comparison with the contour map and also in the histogram.

  As described above, according to the method for predicting the strength deterioration of a corroded structure according to the present invention, a corrugated surface that changes spatially and temporally using the concept of an attack factor is used to create a surface shape creation model in consideration of the anticorrosive effect of coating. Turned into. The results of examining the applicability by comparing the predicted corrosion depth by this surface shape creation model with the measured corrosion depth of the corroded structure (steel material) are shown below.

  (1) The surface shape creation model can be made more general by correcting the spatial autocorrelation model so that the corrosion depth is affected by the surrounding corrosion volume, and by giving the attack factor an influence radius. It was. As a result, it was shown that this surface shape creation model can reduce the dependence on the mesh interval and improve reproducibility even if the corroded surface is divided into meshes and discretized.

  (2) By using the strength F, the number n, and the influence radius r of the attack factor as parameters, a surface shape creation model that can take into account deterioration with time was developed. In this surface shape creation model, various corrosion surface shapes can be created by changing parameters. As a result of comparing the obtained predicted corrosion depth and the surface shape of the measured corrosion depth, it was found that the predicted corrosion depth generated by the surface shape creation model can create a corrosion state close to the actual corroded surface. This modeling of corrosion progress is effective in estimating a decrease in yield strength of a corroded structure by strength analysis.

  (3) The decrease in the anticorrosion performance due to the coating film was expressed by using the coating film ability TM. It has been found that the deterioration of the coating film due to the coating film ability can sufficiently express the actual coating film deterioration.

  Since the strength deterioration measuring method in the corroding structure of the present invention can predict the future surface shape of the corroding surface related to the target corroding structure with good reproducibility, the future strength deterioration in the corroding structure will be detected. Predict more accurately from shape.

  As a result, the strength deterioration measuring method of the present invention can measure the surface shape of a target structure such as a steel bridge, various iron pipes, a steel sluice gate, a steel tower, a storage tank, a building, etc. with good reproducibility and non-contact. Therefore, it is possible to provide future data on the surface shape that is useful for the diagnosis of aging and corrosion degree of the above target structures, and for the judgment of repair / reinforcement measures, renewal, etc. based on the diagnosis results. It can utilize suitably for the use of.

It is a flowchart which shows the strength deterioration prediction measuring method of the corrosion structure which concerns on this invention. It is a schematic block diagram which shows the surface shape measuring apparatus used for the said intensity degradation prediction method. It is a schematic block diagram which shows one Example of the said surface shape measuring apparatus. It is a flowchart which shows one Embodiment of the strength degradation prediction method which concerns on this invention. It is a top view of the corroded surface where the projection pattern was projected which shows one step in the said intensity degradation prediction method. It is the perspective view which showed the reproduction | regeneration surface which reproduced the surface shape of the corrosion surface 1, and the reproduction | regeneration photography pattern corresponding to the photography pattern for a comparison on a reproduction | regeneration surface which show the other step in the said strength degradation prediction method. . It is the graph which compared the surface shape reproduced by the said image measurement with the surface shape by the laser measurement with respect to the same corrosion surface. It is a front view of a projection pattern and an imaging pattern showing an example of a method for determining the same point corresponding to each other on the projection pattern and the imaging pattern in the intensity deterioration prediction method. It is a schematic block diagram which shows the method of measuring the unevenness | corrugation of the corrosion surface in the said strength degradation prediction method. It is a principal part schematic block diagram of the said FIG. It is another principal part schematic block diagram of the said FIG. It is a schematic block diagram for the calculation formula in the said strength degradation prediction method. It is a schematic block diagram for another calculation formula in the said intensity degradation prediction method. It is a plate | board thickness histogram of the object site | part (web) as a shearing test specimen of the plate garter as a corrosion structure in the said strength degradation prediction method. It is a plate | board thickness histogram of the other object part (edge stiffening board) as a shearing test specimen in the said plate garter. It is a plate | board thickness histogram of the further another object site | part (upper flange) as a shearing test specimen in the said plate garter. It is a plate | board thickness histogram of the further another object site | part (lower flange) as a shear experiment specimen in the said plate garter. It is a plate | board thickness histogram of the object site | part (web) as the bending test specimen A in the said plate garter. It is a plate | board thickness histogram of the other object part (flange A1) as the bending test specimen A in the said plate garter. It is a plate | board thickness histogram of the further another object site | part (flange A2) as the bending test specimen A in the said plate garter. It is a plate | board thickness histogram of the object site | part (web) as the bending test specimen B in the said plate garter. It is a plate | board thickness histogram of the other object part (flange B1) as the bending test specimen B in the said plate garter. It is a plate | board thickness histogram of the further another object site | part (flange B2) as the bending test specimen B in the said plate garter. It is a top view which shows the area of a lattice point in the said intensity degradation prediction method. It is a top view which shows the time-dependent concept of the surface shape formation model in the said strength degradation prediction method, (a) shows the 1st year, (b) shows the 2nd year, (c) shows the 10th year. It is a top view which shows the concept of the spatial autocorrelation model in the said intensity degradation prediction method. It is a graph which shows the influence by the distance of one corrosion depth and the other corrosion depth of the circumference | surroundings in the said strength deterioration prediction method. It is a graph which shows the depth of the prediction corrosion depth by the difference in mesh width in the said strength degradation prediction method, respectively. It is a graph which shows the influence from the surrounding other corrosion depth by the difference in the said mesh width, respectively, (a) shows when the mesh width is 5 mm, (b) shows the time when the mesh width is 1 mm. It is a top view which shows the concept of the said continuous spatial autocorrelation model. The schematic cross-sectional view which shows the approximation of the volume of the independent corrosion depth by the difference in mesh width in the said spatial autocorrelation model is shown, (a) When the mesh width is 5 mm, (b) is when the mesh width is 1 mm. Show. It is a graph which shows the depth of the prediction corrosion depth by the difference in mesh width in the said spatial autocorrelation model, respectively. In the strength deterioration prediction method, the influence of the attack factor due to the difference in mesh (lattice point) width is shown, (a) is a plan view when the mesh width is 5 mm, (b) is a plan view when the mesh width is 1 mm. (C) is sectional drawing when a mesh width is 5 mm, (d) shows sectional drawing when a mesh width is 1 mm. It is a top view which shows each plot of the hit point by each attack factor with respect to the corrosion plane in the said strength degradation prediction method. It is a top view which shows the time-dependent concept of the surface shape formation model in the said intensity degradation prediction method. FIG. 4 is a histogram of a uniform random number sequence of coordinate values for random placement of attack factors in the intensity deterioration prediction method, where (a) shows an x coordinate and (b) shows a y coordinate. The formation of the predicted corrosion depth by the attack factor in the strength deterioration prediction method is shown, (a) is a conceptual cross-sectional view of the attack factor, (b) is a cross-sectional view of the predicted corrosion depth by the attack factor, (c) is the mesh width (D) shows a cross-sectional view of the predicted corrosion depth when the mesh width is large. It is a graph which shows the change of the independent prediction corrosion depth by the change of the parameter influence radius r in the said strength degradation prediction method, (a) shows the time of r = 1.0 mm, (b) shows r = 3. When 0.0 mm is shown. It is a graph which shows the change of the independent prediction corrosion depth by the change of the parameter influence radius r in the said strength degradation prediction method, (a) shows the time of r = 5.0mm, (b) shows r = 10mm. Indicates the time. It is a graph which shows the change of the average corrosion amount when changing n at each influence radius (r = 3mm, 5mm, 8mm) in the said strength degradation prediction method. It is a graph which shows the change of a standard deviation when changing n at each influence radius (r = 3mm, 5mm, 8mm) in the said strength degradation prediction method. The measured value of the corroded surface is shown, (a) is a corrosion depth histogram, (b) is a contour map of the corrosion depth. The predicted corrosion amount when the radius of influence (r = 3 mm) in the strength deterioration prediction method is shown, (a) shows a histogram of corrosion depth, and (b) shows a contour map of corrosion depth. The predicted corrosion amount when the radius of influence (r = 5 mm) in the strength deterioration prediction method is shown, (a) shows a histogram of corrosion depth, and (b) shows a contour map of corrosion depth. The predicted corrosion amount when the radius of influence (r = 8 mm) in the strength deterioration prediction method is shown, (a) shows a histogram of corrosion depth, and (b) shows a contour map of corrosion depth. The degree of corrosion on the corroded surface of the flange used in the strength deterioration prediction method is shown, (a) is a contour map of the actual corrosion depth, (b) is a histogram of the actual corrosion depth, and (c). Is a contour map of the predicted corrosion depth by modeling, (d) is a histogram of the predicted corrosion depth by the above modeling, (e) is a contour map of the predicted corrosion depth by other modeling, and (f) is the above other model. (G) is a bar diagram showing the relationship between the brightness and height of the contour map. The degree of corrosion on the corroded surface of the web used in the strength deterioration prediction method is shown, (a) is a contour map of the actual corrosion depth, (b) is a histogram of the actual corrosion depth, and (c). Is a contour map of the predicted corrosion depth by modeling, (d) is a histogram of the predicted corrosion depth by the above modeling, (e) is a contour map of the predicted corrosion depth by other modeling, and (f) is the above other model. (G) is a bar diagram showing the relationship between the brightness and height of the contour map. The degree of corrosion on the corroded surface of the used end plate in the strength deterioration prediction method is shown, (a) is a contour map of the actual corrosion depth, (b) is a histogram of the actual corrosion depth, (c ) Is a contour map of the predicted corrosion depth by modeling, (d) is a histogram of the predicted corrosion depth by the above modeling, (e) is a contour map of the predicted corrosion depth by other modeling, (f) is the above other Histogram of predicted corrosion depth in modeling, (g) is a bar diagram showing the relationship between brightness and height of the contour map. It is a graph which shows the concept of coating-film degradation in the said strength degradation prediction method. It is a top view which shows the spread of the coating-film degradation by an attack factor in the said strength degradation prediction method. The degree of corrosion on the corroded surface of the flange having a coating film in the strength deterioration prediction method is shown, (a) is a contour map of the actual corrosion depth, (b) is a histogram of the actual corrosion depth, (c) is a contour map of the predicted corrosion depth by modeling, (d) is a histogram of the predicted corrosion depth by modeling, and (e) is a bar diagram showing the relationship between the brightness and height of the contour map. It is a graph which shows prediction with time of the coating-film degradation repainted every 30 years in the said strength degradation prediction method. The predicted corroded surface based on the surface shape creation model by each parameter from the corroded surface after 100 years in the flange is shown, (a) is a contour map of the predicted corroded surface after 20 years, and (b) is after 60 years. (C) is a bar diagram showing the relationship between the lightness and height of the contour map. In the strength deterioration prediction method, shows the degree of corrosion on the corroded surface of the end plate having a coating film, (a) is a contour map of the corrosion depth measured, (b) is a histogram of the corrosion depth measured above, (C) is a contour map of the predicted corrosion depth by modeling, (d) is a histogram of the predicted corrosion depth by the above modeling, and (e) is a bar diagram showing the relationship between the brightness and height of the contour map. The predicted corrosion surface based on the surface shape creation model by each parameter from the corrosion surface after 100 years in the end plate is shown, (a) is a contour map of the predicted corrosion surface after 20 years, and (b) is 60 years. It is the contour map of the said prediction corrosion surface after.

Explanation of symbols

S2 Corrosion surface measurement S3 Statistical evaluation of measurement results S4 Strength evaluation analysis S11 Corrosion degradation progress prediction S12 Strength degradation prediction

Claims (8)

Non-contact and three-dimensional measurement of the corroded surface shape of corroded structures,
Using the surface shape creation model to reproduce various uneven surface shapes by changing the parameters, calculate the parameters related to the surface shape creation model to match the corrugations of the corroded surface shape measured three-dimensionally And
Calculate the future corrosion surface shape based on the surface shape creation model by the calculated parameters,
In the method for predicting the strength deterioration of a corroding structure in which future strength deterioration in the corroding structure is predicted by the shape of the future corroding surface ,
Assuming a plurality of lattice points at predetermined intervals where virtual corrosion holes of depth F as parameters are virtually formed on the corrosion surface of the corrosion structure,
An attack factor as an external factor that causes corrosion is set on the corroded surface at a predetermined number n per unit time as a parameter at random.
A method for predicting a deterioration in strength of a corroding structure, characterized in that an influence radius of a radius r as a parameter is set in the attack factor so that the virtual corrosion hole is generated at a lattice point within the influence radius . The grid points have a square mesh,
The method according to claim 1 , wherein the influence radius is larger than a mesh width of the square mesh .   The method for predicting deterioration in strength of a corroded structure according to claim 1, wherein the predetermined interval is set to 5 mm or less. In the surface shape creation model, when the unevenness amount at each lattice point taking into account the influence of each virtual corrosion hole on the lattice point by a spatial autocorrelation model is measured as a corrosion shape, the affected virtual corrosion hole The method for predicting strength deterioration of a corroded structure according to any one of claims 1 to 3 , wherein the volume of the corroded structure is used. Measuring the corroded surface shape in a non-contact and three-dimensional manner is to project a light pattern on the corroded surface of the corroded structure, obtain an image pattern by imaging the light pattern on the corroded surface, The method for predicting the strength deterioration of a corroded structure according to any one of claims 1 to 4 , wherein the corroded surface shape of the corroded structure is measured in three dimensions by the light pattern and the imaging pattern. The method for predicting the strength deterioration of a corroded structure according to any one of claims 1 to 5 , wherein the measurement of the corroded surface shape of the corroded structure is performed a plurality of times at different times. The strength deterioration prediction method for a corroded structure according to any one of claims 1 to 6 , wherein the corroded structure is a corroded steel structure. The corroded surface shape is the state of the paint film,
Deterioration of the coating film is quantified at the grid points on the coating film surface to calculate the anticorrosion ability at the grid points,
Calculate the coating deterioration over time as the anti-corrosion ability that changes over time at the lattice points,
When the anticorrosion ability is equal to or less than a predetermined value, is identified as corrosion progress has occurred, the strength deterioration prediction methods corrosion structure according to any one of claims 1 to 7, characterized in that.

Images