Method for estimating shape of structure

Abstract

Provided is a method for estimating the shape of a cylindrical structure. This method comprises: a step (a) for obtaining first point cloud data along the shape of a wall surface over the entire area of a region of interest of a structure; a step (b) for obtaining second point cloud data by fitting a cylinder to the first point cloud data; a step (c) for setting a reference coordinate system including a reference surface formed of a portion of the second point cloud data, and a reference axis orthogonal to the reference surface and substantially parallel to the longitudinal direction of the structure; a step (d) for converting the second point cloud data or data based on the second point cloud data into a plurality of pieces of divided region data for respective unit regions each defined by a reference length in a direction along the axis of the fitted cylinder and a reference angle in the circumferential direction centered on the axis of the cylinder; and a step (e) for obtaining, on the basis of the divided region data, an approximate ellipse for each axis coordinate position, which is a coordinate position in a direction along the reference axis, by performing elliptical approximation of a cross section orthogonal to the reference axis for each axis coordinate position.

Description

Method for estimating the shape of a structure

[0001] The present invention relates to a method for estimating the shape of a cylindrical structure.

[0002] When producing clinker, which is the raw material for cement, a rotary kiln capable of high-temperature firing is used. A rotary kiln has a cylindrical kiln shell. Because the temperature inside the kiln shell can reach a maximum of 1300 to 1400°C, the inner wall of the kiln shell is generally lined with firebricks.

[0003] The firebricks (hereinafter simply referred to as "bricks") have a slightly trapezoidal cross section and are laid along the inner wall of the kiln shell. The arrangement of the firebricks is basically maintained by uniform compressive force between adjacent bricks. However, as the rotary kiln operates, localized stresses are applied to the kiln shell wall, which can cause deformation of the kiln shell. Furthermore, the central axis of the kiln shell can also deform due to its own weight.

[0004] If the kiln shell is deformed or its center axis is misaligned, the bricks may be worn, peeled off, or their alignment may be disrupted (brick failure). Therefore, it is important to grasp the shape of the kiln shell using a simple method and to periodically manage it.

[0005] On the other hand, it is difficult to grasp the overall shape of a large structure such as a kiln shell. In particular, even if the local deformation or misalignment of the central axis described above occurs, the magnitude of the displacement is small relative to the size of the kiln shell, making it relatively difficult for workers to grasp the deformation during periodic inspections.

[0006] Conventionally, a method disclosed in Patent Document 1 below is known as a method for roughly grasping the shape of a kiln shell.

[0007] JP 2011-21805 A

[0008] The kiln shell of a rotary kiln is supported by rollers at multiple locations spaced apart along its length. Tires (kiln tires) are attached to the outer periphery of the kiln shell while in contact with the rollers, and rotate together with the kiln shell by receiving the rotational force from the rollers while evenly distributing the weight of the kiln shell.

[0009] The method described in Patent Document 1 merely calculates the circumference and diameter of the tire. Information such as the circumference and diameter of the tire is only useful for estimating the local shape (size) of the rotary kiln, and it is difficult to grasp the overall shape of the rotary kiln based on this information.

[0010] In view of the above problems, an object of the present invention is to provide a method for estimating the shape of any location of a cylindrical structure such as a rotary kiln, and even the shape of the entire structure.

[0011] The present invention is a method for estimating the shape of a cylindrical structure, comprising: (a) a step of obtaining first point cloud data that follows the shape of a wall surface of a target area of ​​the structure over the entire area; (b) a step of fitting a cylinder to the first point cloud data to obtain second point cloud data; (c) a step of setting a reference coordinate system including a reference plane formed by a portion of the second point cloud data and a reference axis that is perpendicular to the reference plane and substantially parallel to the longitudinal direction of the structure; (d) a step of converting the second point cloud data or data based on the second point cloud data into a plurality of divided area data, which are data for each unit area defined by a reference length in a direction along a cylindrical axis that is the central axis of the fitted cylinder and a reference angle in a circumferential direction about the cylindrical axis; and (e) a step of ellipsically approximating a cross section perpendicular to the reference axis for each axial coordinate position that is a coordinate position related to a direction along the reference axis, based on the divided area data, to obtain an approximate ellipse for each axial coordinate position.

[0012] By the step (a), a massive amount of point cloud data (first point cloud data) is obtained that represents the actual state of the site, reflecting the shape of the wall surfaces throughout the entire target area of ​​the structure. This step (a) can utilize a method of scanning the interior walls of the structure using, for example, a 3D laser scanner.

[0013] In step (b), a cylinder corresponding to the shape of the wall surface of the structure is fitted to the first point cloud data to obtain point cloud data (second point cloud data) based on a new coordinate system. Since this second point cloud data essentially represents a cylindrical shape, in step (c), a reference coordinate system can be set based on this point cloud data.

[0014] In step (d), the second point cloud data, which is a plurality of data groups, or data based on the second point cloud data is converted into data for each of a plurality of divided regions.

[0015] The second point cloud data is data obtained by fitting a cylinder to the first point cloud data, and the positions of each point included in the second point cloud data depend on the positions of each point included in the first point cloud data. While any method for acquiring the first point cloud data is possible, it is typically obtained by scanning or capturing an image of the wall surface of a structure, and therefore there is usually variation in the distance between each point included in the obtained first point cloud data. Therefore, it is considered that the distance between each point in the second point cloud data also varies, just like in the first point cloud data.

[0016] Step (d) is performed for the purpose of converting the second point cloud data into data for each unit area, thereby replacing it with a data group corresponding to positions that are substantially equally spaced. Note that "data based on the second point cloud data" refers to data converted and generated from the second point cloud data, and corresponds to, for example, third point cloud data described below.

[0017] As an example, a large amount of second point cloud data is converted into data (divided region data) for each unit region, where the reference length along the cylinder axis is 5 cm and the reference angle in the circumferential direction around the cylinder axis is 1°. Then, based on the second point cloud data, the separation distance R from the cylinder axis at the position of each point included in the point cloud data is calculated. Here, when multiple second point cloud data are arranged in the same unit region, the average value of the separation distances R at the positions of the second point cloud data can be used as the divided region data corresponding to the unit region. Furthermore, when a single second point cloud data (point data belonging to the second point cloud data) is arranged in the same unit region, the value of the separation distance R in the second point cloud data (point data belonging to the second point cloud data) can be used as the divided region data corresponding to the unit region.

[0018] This step (d) obtains a data set (divided area data) for each equally spaced area (divided area) in both the axial direction (direction along the cylindrical axis) and the circumferential direction (direction of rotation around the cylindrical axis), which reflects the shape of the wall surface of the structure.

[0019] Therefore, by analyzing the divided region data arranged in the circumferential direction around the reference axis at the same coordinate position (axial coordinate position) along the reference axis, the shape of the cross section perpendicular to the reference axis at that coordinate position can be grasped. In step (e), the cross section perpendicular to the reference axis is approximated to an ellipse based on the multiple divided region data arranged in the circumferential direction around the reference axis for each identical axial coordinate position.

[0020] By performing step (e) for each axis coordinate position, it is possible to grasp continuous changes in the axial direction of the approximate ellipse that simulates the shape of the wall surface of the structure when viewed in the longitudinal direction (axial direction) of the structure. For example, by comparing the shapes of the approximate ellipses at different axis coordinate positions, it is possible to use the results to estimate the shape of the structure.

[0021] The above-described steps (b) to (e) may be performed by a processor executing a predetermined software program using the first point cloud data obtained by a predetermined method. In particular, step (e) may be performed using, for example, the least squares method. From the viewpoint of improving accuracy, a robust estimation process may also be used.

[0022] The reference axis set in step (c) is aligned with the longitudinal direction of the elongated structure, but does not need to be completely parallel to it. In other words, the reference direction being "substantially parallel to the longitudinal direction" means that the reference direction is close enough to the longitudinal direction of the structure to be recognized as being aligned with the longitudinal direction. Typically, the absolute value of the angle between the reference axis and the longitudinal direction of the structure is 10° or less, preferably 5° or less, and more preferably 2° or less.

[0023] In a typical example, the structure is a rotary kiln, and in this case, the target area may be an area where the inner wall of the rotary kiln has been exposed by removing refractory bricks.

[0024] The above-described method for estimating the shape of a structure may further include a step (f) of comparing, for each axis coordinate position, the outer edge shape of the approximate ellipse with the outer edge shape indicated by the divided area data.

[0025] By comparing the shape of the approximate ellipse with the outer edge shape indicated by the actual divided area data for each axial coordinate position, it is possible to grasp in more detail the manner in which the cross-sectional shape of the structure changes in the axial direction for each axial position.

[0026] Specifically, the step (f) may include a step of comparing, for each axis coordinate position, the diameter of the approximate ellipse with the distance between the divided area data and the cylinder axis.

[0027] The above-described method for estimating the shape of a structure may further include a step (g) of calculating the flattening or overriding ratio of the approximate ellipse for each of the axis coordinate positions.

[0028] The above-mentioned method for estimating the shape of a structure may include a step (h) of measuring the progress of displacement of the center position of the approximate ellipse for each axis coordinate position from a virtual reference line that is substantially parallel to the longitudinal direction of the structure.

[0029] Here, the virtual reference line may be any virtual line that is substantially parallel to the longitudinal direction of the structure. As a detailed example, the virtual reference line may be a line that virtually connects the centers of the approximation ellipses at two different locations, or the reference axis included in the reference coordinate system set in step (c).

[0030] According to the above, the direction and amount of deviation of the center position of each approximate ellipse from the virtual reference line can be detected, and therefore, based on the detection results, the state of deflection or bending of the structure can be recognized.

[0031] As an example, a coordinate system (first evaluation coordinate system) is prepared, with the virtual reference line as the origin and the plane perpendicular to the virtual reference line as the plane, and the center positions of the approximate ellipses are plotted on this evaluation coordinate system. By identifying the quadrant with the highest plotting frequency among the first to fourth quadrants on the evaluation coordinate system, the direction of deflection or bending can be detected.

[0032] Furthermore, a coordinate system (second evaluation coordinate system) is prepared in which the virtual reference line is set as one of the vertical and horizontal axes and a first direction constituting a plane perpendicular to the virtual reference line is set as the other axis, and the center positions of each approximate ellipse are plotted on this evaluation coordinate system. For example, the second evaluation coordinate system is set as a coordinate system in which the virtual reference line is set as the horizontal axis and the vertical direction is set as the vertical axis. This allows intuitive recognition of how the center position of the approximate ellipse is displaced in the vertical direction depending on the longitudinal position of the cylindrical structure.

[0033] Similarly, a coordinate system (third evaluation coordinate system) is prepared in which the virtual reference line is set as one of the vertical and horizontal axes, a plane perpendicular to the virtual reference line is formed, and a second direction different from the first direction is set as the other axis, and the center positions of each of the approximate ellipses are plotted on this evaluation coordinate system. For example, the third evaluation coordinate system is set as a coordinate system in which the virtual reference line is set as the horizontal axis and the left-right direction when the virtual reference line is viewed in the depth direction is set as the vertical axis. This allows intuitive recognition of how the center position of the approximate ellipse is displaced horizontally depending on the longitudinal position of the cylindrical structure.

[0034] The structure may be configured to be rotatable around a central axis along the longitudinal direction of the structure, and step (a) may include a step of obtaining the first point cloud data for the structure in a first stopped state, and then rotating the structure to move the position of the top of the structure to a position different from the top, and then obtaining the first point cloud data for the structure in a second stopped state, steps (b) to (d) may be performed based on the first point cloud data obtained in each of the first stopped state and the second stopped state, and step (e) may include a step of obtaining the approximate ellipse for each axial coordinate position in the first stopped state and the approximate ellipse for each axial coordinate position in the second stopped state.

[0035] If there is a discrepancy between the outer edge shape of the approximate ellipse and the outer edge shape indicated by the divided area data at a specific axis coordinate position, it is possible that deformation has occurred on the wall surface of the structure at that location. Here, particularly in large structures, temporary deformation of the wall surface may occur due to some factor. In this case, even at the same axis coordinate position, the shape of the approximate ellipse may change depending on the rotation angle of the structure. In contrast, if permanent deformation has occurred, the shape of the approximate ellipse is considered to be approximately the same at the same axis coordinate position, regardless of the rotation angle of the structure.

[0036] According to the above method, an approximate ellipse is obtained for each axis coordinate position under a plurality of stationary states with different rotation angles of the structure. Therefore, by comparing the approximate ellipses under each stationary state, it is possible to estimate the shape of the structure in more detail.

[0037] The above-mentioned method for estimating the shape of a structure may include a step (f) of comparing, for each of the axial coordinate positions, the outer edge shape of the approximate ellipse with the outer edge shape indicated by the divided area data in each of the first and second stopping states; and a step (h) of estimating that there is a local deformation at the specific axial coordinate position of the structure if a discrepancy is confirmed between the outer edge shape of the approximate ellipse and the outer edge shape indicated by the divided area data at the specific axial coordinate position in both the first and second stopping states in the step (f).

[0038] The above-mentioned method for estimating the shape of a structure may include a step (g) of calculating the flattening or overriding ratio of the approximate ellipse for each of the axis coordinate positions for each of the first stop state and the second stop state.

[0039] The step (d) may include a step (d1) of dividing the second point cloud data into a plurality of sections along the reference axis and fitting a cylinder to each of the second point cloud data for each section to obtain third point cloud data, and a step (d2) of converting the third point cloud data into a plurality of divided area data defined by the reference length and the reference angle based on the cylinder axis, which is the central axis of the cylinder fitted when obtaining the third point cloud data.

[0040] In long structures, as time passes after construction, parts of the structure may sag due to its own weight. Also, depending on the site, the possibility of ground subsidence or fault displacement cannot be denied. In other words, the central axis of the cylinder fitted to the second point cloud data may differ from the central axis of the actual structure.

[0041] According to the above method, the second point cloud data is divided into a plurality of sections along a direction (reference axis) substantially parallel to the longitudinal direction of the structure, and a cylinder is newly fitted to each of the divided point cloud data (third point cloud data). Each cylinder fitted to each of the third point cloud data obtained by this fitting process corresponds to a small section obtained by virtually dividing the structure in the longitudinal direction.

[0042] In other words, each third point cloud data is data simulating a small section formed by virtually dividing a long structure in the longitudinal direction. When a structure is long, the central axis may be shifted depending on the position in the longitudinal direction (i.e., the axial coordinate position). However, the third point cloud data obtained by the above method is fitted to each small section obtained by essentially dividing the structure in the longitudinal direction. Therefore, even if the central axis of a long structure is shifted, there is almost no shift in the central axis on a small section basis. In other words, it is possible to set the central axis of one small section to a different position from the central axis of another small section.

[0043] According to the above method, in step (d2), the third point cloud data is converted into data for each unit area (plurality of divided area data) defined by the reference length and the reference angle based on the cylindrical axis, which is the central axis of the cylinder fitted when obtaining the third point cloud data. Then, in step (e), an approximate ellipse for each axis coordinate position is obtained based on the divided area data.

[0044] In other words, with this method, even if the structure is long and the central axis of the structure is slightly shifted depending on the axial coordinate position, divided region data is obtained based on the third point cloud data obtained by fitting a cylinder having an appropriate central axis depending on the axial coordinate position. In other words, divided region data is obtained that reflects the distance from the central axis that is appropriately set depending on the axial coordinate position. Therefore, by elliptical approximation of a cross section perpendicular to the reference axis for each axial coordinate position based on this divided region data, it is possible to achieve elliptical approximation that more accurately reflects the shape of the wall surface of the structure for each axial coordinate position.

[0045] According to the present invention, the overall shape of a cylindrical structure can be estimated.

[0046] 1 is a flowchart showing a procedure of a first embodiment of a method for estimating the shape of a structure according to the present invention;

[0023] FIG. 1 is a diagram showing a schematic diagram of a scanning operation of the inner wall of a kiln shell;

[0024] FIG. 2 is a photograph showing an example of a scanning operation, viewed in the longitudinal direction of the kiln shell;

[0025] FIG. 3 is a block diagram showing a schematic diagram of an example of the configuration of a processing device;

[0026] FIG. 4 is an example of point cloud data (second point cloud data) in which the coordinate space has been transformed;

[0027] FIG. 5 is a diagram showing a reference coordinate system superimposed on the second point cloud data shown in FIG. 5;

[0028] FIG. 6 is a schematic diagram for explaining divided area data;

[0029] FIG. 7 is an example of a distribution map mapped based on divided area data acquired for a kiln shell;

[0030] FIG. 8 is an example of a diagram showing the shape of a cross section at a specific axial coordinate position based on divided area data acquired for a kiln shell;

[0031] FIG. 9 is a diagram showing a 3D scatter plot based on divided area data and measurement results using a straight line for a region of a kiln shell at the same axial coordinate positions as shown in FIG. 10;

[0032] FIG. 11 is a diagram showing the shape of a measured outer edge derived based on divided area data and the shape of an approximate ellipse superimposed at a certain coordinate position in the Z-axis direction. 13A is a diagram showing a 3D contour diagram based on divided area data acquired for a kiln shell, color-mapped in accordance with the amount and direction of deviation from the shape of an approximate ellipse for each coordinate position in the Z-axis direction. FIG. 13B is a diagram showing results obtained under three different stop angles, color-mapped on the 3D contour diagram in the same illustrative manner as FIG. 12 . FIG. 13C is a diagram showing results obtained under three different stop angles using a kiln shell different from that of FIG. 13A , color-mapped on the 3D contour diagram in the same illustrative manner as FIG. 13A . FIG. 13D is a graph showing the change in the flattening ratio of an approximate ellipse based on divided area data acquired for a kiln shell in the Z-axis direction, showing the change in flattening ratio under three different stop angles. FIG. 13C is a flowchart showing a schematic representation of another procedure of a first embodiment of a method for estimating the shape of a structure according to the present invention. FIG. 13D is a flowchart showing a schematic representation of a procedure of a second embodiment of a method for estimating the shape of a structure according to the present invention. This is a graph plotting the center positions of each approximate ellipse on a coordinate system with the virtual reference line BL as the origin and a plane perpendicular to the virtual reference line BL as the plane, based on the data obtained by executing step #12.1 is a graph plotting the center positions of each approximate ellipse on a coordinate system with the virtual reference line BL as the horizontal axis and the vertical direction as the vertical axis, based on data obtained by executing step #12. FIG. 1 is a graph plotting the center positions of each approximate ellipse on a coordinate system with the virtual reference line BL as the horizontal axis and the horizontal direction when the kiln shell is viewed in the reference measurement direction, based on data obtained by executing step #12. FIG. 1 is a flowchart schematically showing another procedure of a second embodiment of a method for estimating the shape of a structure according to the present invention. FIG. 1 is a flowchart schematically showing a procedure of a third embodiment of a method for estimating the shape of a structure according to the present invention. FIG. 1 is a block diagram schematically showing an example of the configuration of a processing device. FIG. 1 is a schematic diagram illustrating a situation in which the central axis of a fitted cylinder is misaligned with the central axis of the actual kiln shell. FIG. 1 is a schematic diagram illustrating fitting point cloud data to a cylinder for each subsection. FIG. 1 is a schematic diagram illustrating fitting point cloud data to a cylinder for each subsection. FIG. 1 is a flowchart schematically showing another procedure of a third embodiment of a method for estimating the shape of a structure according to the present invention.

[0047] Hereinafter, embodiments of a method for estimating the shape of a structure according to the present invention will be described with reference to the accompanying drawings. However, the drawings are merely schematic illustrations, and the dimensional ratios in the drawings do not necessarily match the actual dimensional ratios. Furthermore, the dimensional ratios between the drawings do not necessarily match.

[0048] This method can be used to estimate the shape of cylindrical structures, and is more typically suited to the evaluation of large structures, such as rotary kiln shells. Other examples of structures include tunnels, sewer pipes, and industrial pipes.

[0049] This method makes it possible to estimate the shape of the cylindrical structure exemplified above. The estimation results can be used, for example, to evaluate the surface condition of the wall surface of the structure. The surface condition here refers to whether or not there are irregularities on the surface, and what kind of irregularities exist. More specifically, the existence of irregularities refers to the location where the irregularities occur, the size of the irregularities (area, depth), the distribution of the irregularities, etc. For example, if part of the inner wall of a cylinder has peeled off, the peeled off area is recessed compared to the surrounding area, and therefore irregularities exist in that area.

[0050] In the following embodiment, an example will be described in which the structure to be estimated in shape is a kiln shell of a cement kiln.

[0051] [First Embodiment] Fig. 1 is a flowchart schematically showing the procedure of a first embodiment of a method for evaluating the shape of a structure according to the present invention. The evaluation method of this embodiment will be described below with reference to the flowchart shown in Fig. 1.

[0052] (Step #1) First, point cloud data is acquired that conforms to the shape of the wall surface of the target cylindrical structure. This point cloud data corresponds to "first point cloud data."

[0053] As an example of a method for acquiring the first point cloud data, scanning using a 3D laser scanner can be used, as shown in Figure 2. Figure 2 is a schematic diagram of a scanning operation. Figure 3 is a photograph of an example of a state during scanning, viewed in the longitudinal direction of the kiln shell 1.

[0054] According to the method shown in Figure 2, the inner wall of a kiln shell 1 of a cement kiln is scanned by a 3D laser scanner. The kiln shell 1 has a long cylindrical shape and is usually lined with refractory bricks. However, in this embodiment, the area to be scanned (target area) may be an area where the refractory bricks have been removed and the inner wall of the kiln shell 1 is exposed.

[0055] In this embodiment, the 3D laser scanner 11 irradiates a measurement object (here, the inner wall of the kiln shell 1) with laser light and receives the light reflected from the measurement object. The 3D laser scanner 11 calculates the distance between the light emission window and the measurement object by detecting the phase difference between the emitted light and the reflected light. This type of method is generally called a "phase difference method."

[0056] Typically, the 3D laser scanner 11 has a main body that can rotate 180° along a horizontal plane, and a measurement unit including a laser light emission window and a light receiving unit that can rotate 360° vertically along a plane formed by one axis constituting the horizontal plane and a vertical axis. In this way, since the entire main body can be rotated horizontally 180° while the measurement unit is rotated 360°, it is possible to obtain highly accurate data of the entire circumference in a short period of time. As a specific example, the 3D laser scanner 11 can measure the coordinates of 40 million points per 3 minutes.

[0057] When the longitudinal length of an object is long, such as the kiln shell 1, installing the 3D laser scanner 11 in one location and measuring the coordinates of all point clouds will result in low accuracy of the measurement results for distant locations. In such cases, the installation location of the 3D laser scanner 11 may be moved and measurements may be taken multiple times.

[0058] 2, scanning is performed using the 3D laser scanner 11 at the installation location a1, and then the 3D laser scanner 11 is moved to the installation location a2 and scanning is performed in the same manner. The movement of the installation location of the 3D laser scanner 11 and scanning are repeatedly performed according to the longitudinal length of the kiln shell 1.

[0059] When adopting this method, it is preferable to install a reference ball 13 as a marker at an arbitrary position from the viewpoint of combining point cloud data obtained by scanning at each installation location. By including the reference ball 13 in the scanning target, point cloud data obtained at adjacent installation locations can be combined with high accuracy. From this viewpoint, the reference ball 13 is installed at a position that is included in both the area that can be scanned by the 3D laser scanner 11 from an adjacent installation location on the rear side and the area that can be scanned by the 3D laser scanner 11 from an adjacent installation location on the front side.

[0060] As an example, the installation locations (a1, a2, a3, ...) of the 3D laser scanners 11 can be spaced at intervals of 10 m to 20 m. Note that, although scanning is performed while the installation location of the 3D laser scanner 11 is moved in the above description, multiple 3D laser scanners 11 may be prepared and each 3D laser scanner 11 may be installed at a different location (a1, a2, a3, ...). In this case, the point cloud data obtained by each 3D laser scanner 11 may be output to a separate arithmetic processing device, and synthesis processing may be performed on the arithmetic processing device side.

[0061] The interval (measurement pitch) of the point cloud data may be set appropriately depending on the dimensions of the kiln shell 1. As an example, the measurement pitch of the point cloud data when scanning over a longitudinal distance of 10 m to 20 m is preferably 5 mm to 15 mm, and more preferably about 12 mm.

[0062] Step #1 obtains a large number of point cloud data (first point cloud data) that simulates the condition of the wall surface of the kiln shell 1. This step #1 corresponds to process (a).

[0063] (Steps #2 and #3) Next, a process of fitting a cylinder to the first point cloud data dc1 obtained in step #1 is performed. Specifically, the following steps are executed: converting the coordinate space of the first point cloud data dc1 into a coordinate space that facilitates comparison with the cylindrical shape; and fitting a cylinder whose central axis coincides with the Z axis and whose radius is d in the converted coordinate space. The function for performing the fitting process may be included in the 3D laser scanner 11 itself, or may be included in a processing device different from the 3D laser scanner 11. Here, the first point cloud data whose coordinates have been converted is referred to as "second point cloud data dc2." Although the coordinates of the second point cloud data dc2 (see FIG. 5) change due to the conversion of the coordinate space, their relative positions remain unchanged from those of the first point cloud data dc1.

[0064] FIG. 4 is a block diagram that schematically shows the configuration of a processing unit when the processing unit performs the fitting process.

[0065] 4 includes a fitting processing unit 21, a divided area data conversion unit 22, an approximate ellipse generation unit 23, a comparison processing unit 24, a storage unit 27, and an output unit 29. The fitting processing unit 21, the divided area data conversion unit 22, the approximate ellipse generation unit 23, and the comparison processing unit 24 are each configured by software or dedicated hardware capable of executing predetermined calculation processes. Details of the divided area data conversion unit 22, the approximate ellipse generation unit 23, and the comparison processing unit 24 will be described later from step #4 onwards.

[0066] The storage unit 27 is a storage area for temporarily recording the calculation results. The output unit 29 is functional means for outputting the calculation results. The output unit 29 is composed of a monitor for outputting information to the arithmetic processing device 20 itself, an interface for outputting information to other devices via wireless or wired connections, and the like. The arithmetic processing device 20 may be composed of a server.

[0067] The fitting processing unit 21 performs a process of fitting a cylinder to the first point cloud data dc1 obtained in step #1 using the 3D laser scanner 11. However, in the present invention, the fitting method is not limited.

[0068] As described above, when performing the fitting process to a cylinder, first, a coordinate space transformation process is performed to facilitate comparison with the cylindrical shape. That is, the fitting processing unit 21 performs coordinate transformation by performing arithmetic processing on the first point cloud data dc1 obtained by the 3D laser scanner 11, and calculates the second point cloud data dc2.

[0069] The Z axis is defined in a direction parallel to the direction of the central axis of the cylinder to be fitted to the first point cloud data dc1, and the X and Y axes are defined so as to be perpendicular to the Z axis.

[0070] First, to provisionally determine the direction of the Z axis, point A(x A , y A , z A ), point B(x B , y B , z B ) are assumed. The distance between any measurement point and a line segment AB connecting points A and B is calculated geometrically. Hereinafter, the distance between the i-th measurement point and the line segment AB is referred to as Di.

[0071] Next, the radius d of the cylinder fitted to the first point cloud data dc1 is determined so that the least square sum ε of the error defined by the following formula (1) is minimized. As an example of a specific calculation method, a method can be adopted in which the value of ε is calculated while gradually changing the value of d in the following formula (1), and d at which this value of ε is minimized is searched for. The minimum value of ε derived through such calculation will be referred to as ε below. AB It is written as follows.

[0072]

[0073] Next, the above-described calculation process is repeated while changing the coordinates of points A and B. By this calculation, the least squares sum ε AB is obtained.

[0074] The respective least square sums ε obtained in this way AB Comparing them, ε AB The combination of point A and point B where the value of d is the smallest is identified, and the value of d at that time is then identified. The direction of the line segment AB connecting point A and point B thus identified is determined as the direction of the Z axis. The value of d at this time becomes the radius of the cylinder fitted to the first point cloud data dc1 (see FIG. 5).

[0075] As shown in Figure 6, for example, a cylindrical coordinate system (θ, r, z) can be established, with the origin O being the intersection of the end face 18 (corresponding to the reference plane) of the kiln shell 1 on the kiln burner side (inlet side) and the Z axis, and the central axis of the fitted cylinder being the Z axis. Figure 6 is a diagram in which the θ-r-z coordinate system is added to Figure 5. r is the radial distance from the Z axis to the position of the second point cloud data dc2. θ is the clockwise rotation angle around the origin O, with the top of the kiln shell 1 set as 0°. z is the coordinate position in the Z axis direction. In other words, in this cylindrical coordinate system, the Z axis corresponds to the reference axis, and the θ-r plane corresponds to the reference plane.

[0076] As a result, the fitting processing unit 21 performs a calculation process to fit a cylinder to the first point cloud data dc1, thereby obtaining second point cloud data dc2 (corresponding to step (b)). Furthermore, a reference coordinate system (θ-r-z coordinate system) is identified (corresponding to step (c)). Information about the central axis of the fitted cylinder (here, the Z axis), i.e., information about the reference direction, may be recorded in the storage unit 27.

[0077] (Step #4) Next, the second point cloud data dc2 is divided into unit regions in the axial direction (Z direction) and the circumferential direction (the direction of rotation based on the Z direction), thereby obtaining a plurality of divided region data df. This step #4 is executed by the divided region data conversion unit 22.

[0078] 7 is a schematic diagram for explaining the divided region data. In this step #4, the second point cloud data dc2 is converted into data for each unit region 41 defined by a reference length za in the direction (Z direction) along the central axis (cylinder axis 31) of the fitted cylinder and a reference angle θa in the circumferential direction (rotation direction) around the cylinder axis 31. This data for each unit region is referred to as "divided region data."

[0079] The reference length za is, for example, 1 cm to 10 cm, and more specifically, 5 cm. The reference angle θa is appropriately set according to the inner diameter of the kiln shell 1 so that the length of the arc portion of the outer edge of the unit area 41 corresponds to the reference length za. For example, the reference angle θa is 0.5° to 3°.

[0080] Several pieces of second point cloud data dc2 exist within each unit area 41, but the number varies depending on the location. When multiple pieces of second point cloud data dc2 exist within a certain unit area 41, the average value of the r values ​​described in the coordinate information in the θ-r-z coordinate system of each piece of second point cloud data dc2, i.e., the average value of the distance from the central axis of the fitted cylinder (cylinder axis 31), is calculated and used as the data corresponding to that unit area 41. When a single piece of second point cloud data dc2 exists within a certain unit area 41, the value of r described in the coordinate information in the θ-r-z coordinate system of that second point cloud data dc2 is used as the data corresponding to that unit area 41.

[0081] The data generated for each unit area 41 corresponds to "divided area data." In other words, step #4 is performed to convert the second point cloud data dc2, in which the spacing between adjacent points varies, into data (divided area data) with substantially equal spacing. For example, if the reference length za is 5 cm, the second point cloud data dc2 acquired in large quantities corresponding to the wall surface of the kiln shell 1 is converted into divided area data df, which are arranged for each unit area 41 within a 5 cm x 5 cm range on the wall surface of the kiln shell 1. If the area to be estimated for the shape of the kiln shell 1 is within the kiln length range of 0 m to 40 m and the reference angle θa is 1°, the multiple divided area data df will be grid-like data (matrix data) with 800 points in the Z-axis direction and 360 points in the circumferential direction.

[0082] 8 shows the distribution of the r value (distance from the cylinder axis 31) for each unit area 41 based on the divided area data df for each unit area 41 obtained by performing steps 1 to 4 on an actual kiln shell 1. As illustrated in FIG. 8, by color-coding (or shading) the r value for each unit area 41, it is possible to roughly recognize the cross-sectional shape of each position on the kiln shell 1.

[0083] 8, it was confirmed that the horizontal dimension is larger than the vertical dimension. In other words, the cross-sectional shape of the kiln shell 1 perpendicular to the Z axis appears to be a perfect circle, but in reality it is deformed into an ellipse. This is thought to be because, in the case of a large structure such as the kiln shell 1, its vertical length becomes shorter than its horizontal length due to its own weight.

[0084] Figure 9 shows the cross-sectional shape of the same kiln shell 1 as in Figure 8 in the region between 13.5 m and 15 m on the Z-axis. Note that Figure 9 is drawn so that the distance r = 2.70 m is positioned at the center, making the deformation more noticeable. Figure 9 shows that, when viewed in the +Z direction, the kiln shell 1 is locally bulging inward at the lower right and upper left.

[0085] Conventionally, when such deformation occurs, the extent of deformation is determined by setting reference points at two healthy locations on either side of the deformed area, stretching a vertical line at the same height from these reference points, and measuring the height from the vertical line to the surface of the deformed area. Figure 10 shows a 3D scatter plot of the lower right region (the region of central angle θ = 90° to 180° in Figure 6) of the kiln shell 1, viewed in the +Z-axis direction, from the Z-axis coordinate position of 13.5 m to 15 m, along with the results of measuring the same region using a vertical line. Note that the data labeled "3DLS" on the right side of Figure 10 corresponds to the results obtained through steps #1 to #4 above.

[0086] 10, a comparison of the results of measurements using a conventional waterline and the results of the method of the above embodiment at four locations (four lines) on the figure confirmed that the magnitude of local deformation in both cases was roughly consistent. Measurements using waterlines are limited to lines. In contrast, the method of the above embodiment allows for a planar understanding of the shape of the wall surface of the kiln shell 1, allowing for detailed confirmation of the degree of local deformation of the kiln shell 1.

[0087] Step #4 corresponds to process (d).

[0088] (Step #5) Next, an approximation ellipse is calculated for each coordinate position in the Z-axis direction (axis coordinate position) based on the multiple divided region data df obtained in step #4. Any method can be used to calculate the approximation ellipse, for example, the least squares method. Furthermore, robust estimation processing may be used to reduce the influence of the presence of singular points. The calculation of this approximation ellipse is performed by the approximation ellipse generator 23. As an example, the "Robust Estimation Calculator Software" manufactured by Ideal Corporation can be used.

[0089] In step #5, the cross-sectional shape of the kiln shell 1 is approximated by an ellipse for each coordinate position in the Z-axis direction. Step #5 corresponds to process (e). Information about the obtained approximate ellipse is output from the output unit 29 as needed.

[0090] (Step #6) Next, the shape of the outer edge of the kiln shell 1 simulated by the multiple divided area data df obtained in step #4 (hereinafter sometimes referred to as the "measured outer edge shape") is compared with the shape of the outer edge of the approximate ellipse obtained in step #5 (hereinafter sometimes referred to as the "shape of the approximate ellipse"). One example of a comparison method is to calculate the difference between the value of r (the distance from the central axis of the cylinder) indicated by the divided area data df and the diameter of the approximate ellipse for each position in the Z axis direction, and then determine the distribution of these difference values. This step #6 is executed by the comparison processing unit 24.

[0091] Figure 11 is a drawing in which the shape of the measured outer edge obtained in step #4 and the shape of the approximate ellipse obtained in step #5 are superimposed at a certain coordinate position z1 in the Z-axis direction, and is shown in a manner that makes the deformation appear prominent, similar to Figure 9. Figure 11 confirms that at coordinate position z1, there are locations where the measured outer edge is located inside the approximate ellipse and, conversely, locations where the measured outer edge is located outside the approximate ellipse, depending on the position of the central angle θ (see Figure 6).

[0092] By performing color mapping for each coordinate position in the Z-axis direction according to the amount and direction of deviation from the approximate ellipse, it is possible to grasp the distribution of unevenness (local deformation) formed on the wall surface of the kiln shell 1 and the degree of the unevenness at any position. Figure 12 is a diagram in which color mapping according to the amount and direction of deviation from the approximate ellipse is performed on a 3D contour diagram based on the divided area data df obtained for the actual kiln shell 1 shown in Figure 8. In Figure 12, areas of the measurement outer edge located inside the approximate ellipse are displayed in red, and areas located outside the approximate ellipse are displayed in green.

[0093] It will be understood that the overall shape of the wall surface of the kiln shell 1 can be understood by referring to FIG.

[0094] When the shape of the kiln shell 1 of a rotary kiln is estimated, the kiln shell 1 is rotatable around an axis. However, as shown in FIG. 2, the first point cloud data dc1 is acquired using a 3D laser scanner while the kiln shell 1 is stationary. The approximate elliptical shape of the kiln shell 1 obtained through steps #2 to #5 is based on the stationary state of the kiln shell 1 when step #1 is performed. Therefore, the comparison result obtained in step #6 is also based on the stationary state of the kiln shell 1 when step #1 is performed.

[0095] In a large structure such as the kiln shell 1, the wall surface may temporarily deform due to some factor (gravity, external force, etc.). In this case, the location of the deformation is determined solely by the relative position within the kiln shell 1 at the time the first point cloud data dc1 is acquired. Specifically, assume that when the first point cloud data dc1 is acquired at a certain time T1, temporary deformation due to factors such as gravity occurs at a specific position corresponding to the top. If the first point cloud data dc1 is acquired at another time T2 with the kiln shell 1 rotated 90 degrees, temporary deformation occurs at the specific position corresponding to the top at this time T2. In this case, the top at time T1 and the top at time T2 are in different positions on the wall surface of the kiln shell 1. In other words, even if the axis coordinate position is the same, the shape of the approximate ellipse may change depending on the rotation angle of the structure.

[0096] In contrast, suppose that when the first point cloud data dc1 is acquired at a certain time T1, there is a permanent deformation at a specific position corresponding to the top. In this case, if the first point cloud data dc1 is acquired at another time T2 with the kiln shell 1 rotated 90°, there will be no deformation at the specific position corresponding to the top at time T2, but there will be deformation at a position 90° away from the top. In other words, if permanent deformation occurs, when the kiln shell 1 rotates, the deformed position will also follow this rotation. Therefore, regardless of the rotation angle of the structure, the shape of the approximate ellipse will be approximately the same at the same axis coordinate position.

[0097] Therefore, in order to determine whether the deformation is temporary or permanent, step #1 may be performed under two or more stationary states in which the kiln shell 1 is stopped at different angles. In this case, steps #2 to #6 are performed under each of the stationary states.

[0098] Figure 13A shows the distribution of deviations and deviation directions from the approximate ellipse obtained by performing step #1 under three different stop angles of the kiln shell 1 and then performing steps #2 to #6 under each stop angle, as shown by color mapping the 3D contour diagram in the same way as in Figure 12. Note that Figure 12 is the same as the "0°" diagram in Figure 13A.

[0099] 13A, for ease of comparison, the orientation of all 3D contour diagrams is standardized to the case where the stop angle is 0°. In other words, the illustration method has been adjusted so that areas corresponding to the same relative positions in each 3D contour diagram correspond to the same locations on the actual kiln shell 1.

[0100] 13A, when we focus on the position around the Z coordinate 27m (areas B0, B90, and B180), we see that regardless of the stop angle, there are green and red areas, and the color mapping pattern is the same. This suggests that there is a possibility that permanent deformation is occurring at the position of the kiln shell 1 around the Z coordinate 27m.

[0101] 13A shows that date Y1, when step #1 was performed with a stop angle of 0°, date Y2, when step #1 was performed with a stop angle of 180°, and date Y3, when step #1 was performed with a stop angle of 90°, are all different from one another, and date Y2 is about one year after date Y1, and date Y3 is about one year after date Y2. That is, Fig. 13A shows data obtained on date Y1 when step #1 was performed with a stop angle of 0° and through steps #2 to #6, data obtained on date Y2 when step #1 was performed with a stop angle of 180° and through steps #2 to #6, and data obtained on date Y3 when step #1 was performed with a stop angle of 90° and through steps #2 to #6. In Figure 13A, focusing on the position (areas A0, A90, A180) near the Z-axis coordinate 15m, area A180 at a stop angle of 180° and area A90 at a stop angle of 90° are shown in red (displaced inward from the approximate ellipse). In contrast, area A0 at a stop angle of 0° almost coincides with the approximate ellipse. This suggests that some kind of permanent deformation occurred in the kiln shell 1 at a position near the Z-axis coordinate 15m during the period from date Y1 to date Y2. The reason for this permanent deformation is likely to be, for example, excessive thermal load on the relevant area due to the operation of the rotary kiln during the period from date Y1 to date Y2.

[0102] Figure 13B shows the distribution of deviations and directions from the approximate ellipse for a kiln shell 1 at a different site from that shown in Figure 13A, using the same notation method as in Figure 13A. However, to obtain the data in Figure 13B, step #1 was performed on the same date with stop angles of 0°, 180°, and 90°. The distributions of deviations and directions from the approximate ellipse obtained by performing steps #2 to #6 with each stop angle were plotted in the same manner as in Figure 13A.

[0103] 13B, when we focus on the position around the Z coordinate 15 m (areas C0, C90, and C180), regardless of the stop angle, there are green and red areas, and the color mapping pattern is the same. This suggests that there is a possibility that permanent deformation is occurring at the position of the kiln shell 1 around the Z coordinate 15 m.

[0104] On the other hand, when we look at the position (area D) around 33 m to 35 m on the Z coordinate in Figure 13B, we can see that the position of the red area rotates according to the stop angle. This suggests that temporary deformation due to gravity or the like occurs around 33 m to 35 m on the Z coordinate in the kiln shell 1 shown in Figure 13B.

[0105] In this way, by performing steps #2 to #6 based on the first point cloud data obtained by performing step #1 under different stop angles and comparing the results obtained between different stop angles, it is possible to more accurately estimate the deformation pattern of the kiln shell 1.

[0106] Step #6 corresponds to step (f). In step (f), the difference between the shape of the measured outer edge and the shape of the approximate ellipse is compared for different stop angles. If a substantially similar difference exists at any stop angle, it is determined that a permanent local deformation has occurred. Information about the comparison result between the shape of the measured outer edge and the shape of the approximate ellipse is output from output unit 29 as needed.

[0107] (Step #7) Next, the flattening ratio of the approximate ellipse obtained in step #5 is calculated for each position in the Z-axis direction. The flattening ratio is calculated by finding the ratio between the major axis and the minor axis of each approximate ellipse obtained for each position in the Z-axis direction. This step #7 may also be performed by the approximate ellipse generation unit 23.

[0108] Figure 14 is a graph showing the change in the flattening ratio of the approximation ellipse for each position in the Z-axis direction, obtained by performing steps #1 to #5 on the actual kiln shell 1 shown in Figure 8. Similar to Figure 13A, Figure 14 also shows the change in the flattening ratio in the Z-axis direction for three different stop angles (0°, 90°, and 180°).

[0109] Figure 14 shows that the flattening of the ellipse is small near the fulcrum positions (near Z coordinate 6 m and Z coordinate 29 m) in the Z-axis direction, resulting in small elliptical deformation at these positions. This is presumably due to the large shell thickness at the fulcrum, which suppresses deformation. Furthermore, in the region between the fulcrums (Z coordinate 6 m to 29 m), the flattening of the ellipse for stop angles of 0° and 180° is almost the same, and is larger than the flattening for a stop angle of 90°. This result suggests that the stop angle may affect the ease of deformation of the overall shape of the kiln shell 1.

[0110] From this perspective, when estimating the shape of the wall surface of the kiln shell 1 based on the information on the flattening of the approximate ellipse, it is considered preferable to obtain first point cloud data corresponding to the shape of the wall surface of the kiln shell 1 under multiple stop conditions with different stop angles, calculate the flattening of the approximate ellipse for each stop angle, and comprehensively compare these data. In particular, if the initial stop angle is set to 0°, it is preferable to estimate the shape of the wall surface of the kiln shell 1 based on the flattening of the approximate ellipse obtained based on the first point cloud data for a stop angle of 0° and the flattening of the approximate ellipse obtained based on the first point cloud data for a stop angle of around 90° (e.g., 85° to 95°).

[0111] By performing the above process at the same stop angle periodically, for example once a year, and comparing the flattening ratio of the approximate ellipse at each time, it is possible to quickly grasp the change in shape of the kiln shell 1 over time. In this case, it is preferable to perform the above process at two or more stop angles, one near a stop angle of 0° and the other near a stop angle of 90°. The same applies to step #6.

[0112] Step #7 corresponds to step (g). Information relating to the flattening of the approximate ellipse according to the coordinate position in the Z-axis direction is output from the output unit 29 as needed.

[0113] The execution order of steps #6 and #7 may be reversed. Also, only the processing of either step #6 or step #7 may be executed. Furthermore, the execution of steps #1 to #5 may be completed, and the execution of steps #6 and #7 may be optional. This is also true for the second embodiment.

[0114] In the above embodiment, in step #7, the flattening of the approximate ellipse obtained in step #5 is calculated for each position in the Z-axis direction, but the overriding ratio may be calculated instead of the flattening (see Figure 15).

[0115] The overity ratio α is calculated as α = (a - b) / r, where r is the design radius of the kiln shell 1, a is the major axis of the approximate ellipse, and b is the minor axis of the approximate ellipse. The value of the design radius r is set based on the information at the time of designing the kiln shell 1 to be measured. Information regarding the design radius r is usually stored by the construction contractor or the contractor ordering the construction of the kiln shell 1, so this stored data can be read. Note that the value of the overity ratio may be calculated by multiplying the value of α obtained by the above formula by 100 and converting it into a percentage. By detecting the change in the overity ratio at each position in the Z-axis direction, the degree of deformation of the cross-sectional shape of the kiln shell 1 from a circular shape can be recognized, similar to detecting the change in the flattening ratio.

[0116] In step #7, both the flattening ratio and the overburden ratio may be calculated. This also applies to the following embodiments.

[0117] [Second Embodiment] A second embodiment of the method for evaluating the shape of a structure according to the present invention will be described, focusing on the differences from the first embodiment. Fig. 16 is a flowchart that schematically illustrates the procedure of the second embodiment of the method for evaluating the shape of a structure according to the present invention, following Fig. 1, and differs from the first embodiment in that steps #11 and #12 are executed after step #5.

[0118] Steps #1 to #5 are executed in the same manner as described above in the first embodiment, and the cross-sectional shape of the kiln shell 1 is approximated to an ellipse for each coordinate position in the Z-axis direction.

[0119] (Step #11) Next, the coordinates of the center of the approximate ellipse of the cross section of the kiln shell 1 are detected for each coordinate position in the Z-axis direction obtained in step #5.

[0120] (Step #12) Next, the displacement of the center coordinates of each of the approximate ellipses obtained in step #11 from a predetermined virtual reference line is recognized.

[0121] Here, the virtual reference line can be any straight line as long as it is substantially parallel to the longitudinal direction of the kiln shell 1. As an example, the virtual reference line may be the Z axis, which is the central axis of the cylinder fitted in step #2. As another example, the virtual reference line may be the Z axis of the reference coordinate system set in step #3.

[0122] As another example, the virtual reference line may be a straight line connecting the center coordinates of any two of the approximate ellipses detected in step #11. In this case, it is preferable that the two selected approximate ellipses are spaced apart to a certain extent in the longitudinal direction of the kiln shell 1. Here, the kiln shell 1 is typically supported by two or more supports and positioned at a slight incline. In this case, it is preferable that the two selected approximate ellipses are spaced apart by a distance equal to or greater than the minimum distance (minimum support distance) between adjacent supports in the longitudinal direction of the kiln shell 1.

[0123] 17A to 17C are graphs showing the results obtained by performing steps 1 to 5 and steps 11 to 12 on a kiln shell 1 at a different site from the kiln shell 1 for which the analysis results were shown in the first embodiment. Here, the virtual reference line BL is the Z axis, which is the central axis of the cylinder fitted in step 2.

[0124] 17A is a graph plotting the center positions of the respective approximate ellipses on a coordinate system having the virtual reference line BL as the origin and a plane perpendicular to the virtual reference line BL as a plane. The coordinate system shown in FIG. 17A corresponds to the "first evaluation coordinate system." The first evaluation coordinate system is a coordinate system in which the vertical axis is the vertical direction and the horizontal axis is the left-right direction when looking from one end of the kiln shell 1 to the other end (hereinafter referred to as the "reference measurement direction").

[0125] In the first evaluation coordinate system shown in FIG. 17A, if the center coordinates of each approximate ellipse are close to the origin, this means that the center of each approximate ellipse exists substantially on the virtual reference line BL.

[0126] In the first evaluation coordinate system shown in Fig. 17A, when the plot of the center coordinates of the approximate ellipse is located within the first quadrant p1, it means that the center of the approximate ellipse is displaced vertically upward and to the right with respect to the virtual reference line BL. The first quadrant p1 corresponds to the region in Fig. 17A where both the horizontal axis and the vertical axis have positive values.

[0127] In the first evaluation coordinate system shown in Fig. 17A, when the plot of the center coordinates of the approximate ellipse is located within the second quadrant p2, it means that the center of the approximate ellipse is displaced vertically upward and leftward with respect to the virtual reference line BL. The second quadrant p2 corresponds to the region in Fig. 17A where the horizontal axis has positive values ​​and the vertical axis has negative values.

[0128] In the first evaluation coordinate system shown in Fig. 17A, when the plot of the center coordinates of the approximate ellipse is located within the third quadrant p3, it means that the center of the approximate ellipse is displaced vertically downward and leftward with respect to the virtual reference line BL. The third quadrant p3 corresponds to the region in Fig. 17A where both the horizontal axis and the vertical axis have negative values.

[0129] In the first evaluation coordinate system shown in Fig. 17A , when the plot of the center coordinates of the approximate ellipse is located within the fourth quadrant p4, it means that the center of the approximate ellipse is displaced vertically downward and to the right with respect to the virtual reference line BL. The fourth quadrant p4 corresponds to the region in Fig. 17A where the horizontal axis has positive values ​​and the vertical axis has negative values.

[0130] 17A, the center coordinates of each approximate ellipse are frequently plotted in the fourth quadrant p4, which indicates that the kiln shell 1 being measured is deflected downward and to the right when viewed in the reference measurement direction.

[0131] Figure 17B is a graph plotting the center positions of the approximate ellipses on a coordinate system with the virtual reference line BL as the horizontal axis and the vertical axis as the vertical axis. In Figure 17B, the vertical direction is set to the positive direction on the vertical axis. The coordinate system shown in Figure 17B corresponds to the "second evaluation coordinate system." Figure 17B shows that the kiln shell 1 being measured is relatively deflected downward in the vertical direction. This result is consistent with the result shown in Figure 17A. Figure 17B also allows us to identify the position (longitudinal position) where the kiln shell 1 is distorted relatively significantly in the vertical direction.

[0132] FIG. 17C is a graph plotting the center positions of the approximate ellipses on a coordinate system with the virtual reference line BL as the horizontal axis and the left-right direction when the kiln shell 1 is viewed in the reference measurement direction as the vertical axis. In FIG. 17C, the right direction when the kiln shell 1 is viewed in the reference measurement direction is set as the positive direction of the vertical axis. The coordinate system shown in FIG. 17C corresponds to the "third evaluation coordinate system." According to FIG. 17C, it can be seen that the kiln shell 1 being measured is relatively deflected to the right when viewed in the reference measurement direction. This result is consistent with the result shown in FIG. 17A. Furthermore, according to FIG. 17C, it is possible to identify the position (longitudinal position) where the kiln shell 1 is distorted relatively significantly in the left-right direction.

[0133] When repairing a kiln shell 1, a portion to be repaired may be separated from the kiln shell 1, replaced with a new shell portion, and then joined together. When performing such repair work, by acquiring the results shown in Figures 17A to 17C in advance, the orientation of the shell portion can be adjusted according to the direction of deflection of the kiln shell 1 at that location when joining the replacement shell portion. This step #12 corresponds to process (i).

[0134] 17A to 17C, the number of approximate ellipses for which the center coordinates are plotted can be set arbitrarily. However, if the number of approximate ellipses is too small, the accuracy of the information regarding the deflection of the kiln shell 1 may become low. From this perspective, it is preferable to set the number of approximate ellipses for which the center coordinates are plotted to 100 or more, and more preferably 500 or more.

[0135] On the other hand, even if the number of approximate ellipses is set too large, the amount of calculation increases but in practice there is no significant difference in the accuracy of the information obtained. From this perspective, the number of approximate ellipses for which the center coordinates are plotted is preferably set to 5,000 or less, and more preferably set to 2,000 or less.

[0136] As in the first embodiment, in this embodiment, step #1 may be performed under two or more stationary states with different stop angles of the kiln shell 1. In this case, steps #2 to #5 and steps #11 and #12 are performed under the respective stop states.

[0137] In this embodiment, steps #6 and #7 described above in the first embodiment may be executed in parallel (see FIG. 18). Furthermore, in this case, as shown in FIG. 15, the overburden ratio may be calculated in step #7 instead of the flattening ratio.

[0138] [Third Embodiment] A second embodiment of the method for evaluating the shape of a structure according to the present invention will be described below, focusing on the differences from the first embodiment. Fig. 19 is a flowchart that schematically illustrates the procedure of the second embodiment of the method for evaluating the shape of a structure according to the present invention, following Fig. 1. This embodiment differs from the first embodiment in that step #8 is performed after step #3 is performed.

[0139] 20, the arithmetic processing device 20 that executes the method of this embodiment additionally includes a division processing unit 25, as compared to the arithmetic processing device 20 shown in Fig. 4. This division processing unit 25 is configured by software or dedicated hardware that can execute predetermined arithmetic processing, and is used to execute step #8, which will be described later.

[0140] (Step #8) The second point cloud data dc2 obtained in step #2 is divided into multiple sections along the reference direction (here, the Z-axis direction) set in step #3. For convenience, this divided point cloud data is referred to as "third point cloud data dc3."

[0141] The second point cloud data dc2 obtained in step #2 may be stored in the storage unit 27. In step #8, the division processing unit 25 may read information about the second point cloud data dc2 from the storage unit 27 and perform the processing described below.

[0142] The division processing unit 25 divides the second point cloud data dc2 into multiple sections based on the reference direction (i.e., the direction of the central axis 17) determined in steps #2 and #3 and a predetermined division number. As a result, the second point cloud data dc2 is converted into multiple third point cloud data dc3 divided into each section. In this division process, the widths of the sections may be uniform or may vary.

[0143] Next, the plurality of third point cloud data dc3 obtained in this manner are each subjected to fitting processing to a cylinder in the fitting processing unit 21. The fitting processing method is performed in the same manner as in steps #2 and #3. That is, by this processing, the central axis of the cylinder fitted to each third point cloud data dc3 is defined by an equation under the reference coordinate system of the second point cloud data dc2.

[0144] Depending on the kiln shell 1, some parts may bend due to their own weight over time after installation on site. For this reason, the first point cloud data dc1 based on the inner surface of the kiln shell 1 may differ from the shape of the fitted cylinder.

[0145] 21 is a diagram exaggerating the above point. The position of the central axis of the second point cloud data dc2 should change depending on the position in the Z-axis direction (the longitudinal direction of the kiln shell 1). In other words, the central axis 17 of the cylinder to which the second point cloud data dc2 is fitted, determined by the processing of step #2, may not accurately simulate the central axis of the internal space of the kiln shell 1.

[0146] In contrast, in step #8, a cylinder is fitted to third point cloud data dc3 obtained by dividing the second point cloud data dc2 into small sections in the Z direction in advance. That is, as shown in Fig. 22, multiple third point cloud data dc3 fitted to different cylindrical shapes at different positions in the Z direction are obtained.

[0147] As a result, even if the vertical position of the central axis of the kiln shell 1 changes depending on the Z coordinate as shown in Figure 21, it is possible to fit cylinders having different central axes to each small section (see Figure 23). Figure 23 shows a schematic diagram in which a certain third point cloud data dc3 is fitted to a cylinder 31a having a central axis 35a, another third point cloud data dc3 is fitted to a cylinder 31b having a central axis 35b, and yet another third point cloud data dc3 is fitted to a cylinder 31c having a central axis 35c.

[0148] (Step #4) The third point cloud data dc3 obtained in step #8 is divided into unit areas in the axial direction (Z direction) and the circumferential direction (direction of rotation based on the Z direction), thereby obtaining a plurality of divided area data df. That is, step #4 of this embodiment is different in that the object to be divided into unit areas is the third point cloud data dc3 obtained in step #8, rather than the second point cloud data dc2 obtained in step #2, but is otherwise the same as the first embodiment.

[0149] In step #4 in the first embodiment, the second point cloud data dc2, in which the intervals between adjacent points are varied, is converted into data for each equally spaced unit region 41. More specifically, each unit region 41 is converted into data on the distance from the central axis of the fitted cylinder.

[0150] In this embodiment, instead of the second point cloud data dc2, the third point cloud data dc3 obtained by fitting a cylinder to each of the subdivisions divided by the Z-axis coordinate is used. Therefore, the distance from the central axis of the fitted cylinder for each unit area 41 obtained in this embodiment takes into account the positional deviation of the central axis of the actual kiln shell 1, particularly the positional deviation in the vertical direction. Therefore, by performing steps #5 to #7 described in the first embodiment based on the divided area data df obtained by the method of this embodiment, it is possible to more accurately estimate the shape of the kiln shell 1.

[0151] This embodiment differs from the first embodiment only in that the central axis of the fitted cylinder may differ depending on the Z-axis coordinate, and the substantial processing is the same as in the first embodiment. Therefore, since the processing of steps #5 to #7 is the same as in the first embodiment, a description thereof will be omitted.

[0152] In this embodiment, the process of the second embodiment can be performed in parallel to detect the deflection of the kiln shell 1. In this case, as shown in Figure 24, after step #3 is performed, steps #4, #5, #11, and #12 may be performed sequentially as described above in the second embodiment.

[0153] As described above, in the process of this embodiment, step #8 is executed to divide the second point cloud data dc2 into multiple sections in a direction substantially parallel to the longitudinal direction of the kiln shell 1. Then, steps #4 and #5 are executed to fit each of the multiple point cloud data (third point cloud data dc3) obtained by the division process to a cylinder, and an approximate ellipse is obtained based on the central axis of each cylinder. In other words, according to this process, even if the central axis of the kiln shell 1 is misaligned, the central axis of the cylinder that serves as the reference for the approximate ellipse is set taking this misalignment into account. Therefore, an approximate ellipse of the kiln shell 1 for each axial position is obtained while taking the misalignment of the central axis of the kiln shell 1 into account.

[0154] On the other hand, steps 11 and 12 described in the second embodiment are primarily performed to detect the degree of deflection of the kiln shell 1, in other words, to detect the degree of displacement of the central axis of the kiln shell 1. In other words, these steps are performed to detect the direction and extent of displacement of the central axis of the kiln shell 1 from the reference line (virtual reference line BL). Therefore, it is not suitable to use the results of steps 8 and subsequent steps, in which the central axis of the cylinder to be fitted is set according to the longitudinal position of the kiln shell 1.

[0155] From this perspective, as shown in Figure 24, after step #3 is executed, the process may branch from step #8 onwards, and steps #4, #5, #11, and #12 may be executed sequentially. Note that steps #4 and #5 appear twice in Figure 24. These processes correspond to each other in terms of their content, but the data used in the processes differ. Therefore, in Figure 24, to distinguish between the process of step #4 that is executed without using the execution result of step #8 and the process of step #4 that is executed using the execution result of step #8, the former is denoted by the symbol #4a, and the latter is denoted by the symbol #4b.

[0156] For the same reason, in Figure 24, in order to distinguish between the processing of step #5 that is performed without using the execution results of step #8 and the processing of step #5 that is performed using the execution results of step #8, the former is marked with the symbol #5a and the latter is marked with the symbol #5b.

[0157] Another embodiment will now be described.

[0158] <1> In the above-described embodiment, in step #1, the entire longitudinal area of ​​the inner wall surface of the kiln shell 1 is scanned. In other words, the entire longitudinal area of ​​the inner wall surface of the kiln shell 1 is the target area. However, the present invention does not exclude the case where only a portion of the inner wall surface of the kiln shell 1 is the target area.

[0159] For example, a portion of the longitudinal area of ​​the kiln shell 1 may be designated as the target area, and the inner wall surface of the kiln shell 1 within this area may be scanned in step #1. In particular, a burner is typically installed at the front end of the kiln shell 1, of the two longitudinal ends. Therefore, during repair work, etc., it may be necessary to remove only the refractory bricks attached to the inner wall of the kiln shell 1 over a predetermined length from the front end. In this case, the target area may be the area of ​​the wall surface of the kiln shell 1 from which the refractory bricks have been removed.

[0160] <2> In the above embodiment, step #1 is a process of obtaining first point cloud data that reflects the shape of the inner wall of the kiln shell 1. However, step #1 may be a process of obtaining first point cloud data that reflects the shape of the outer wall of the kiln shell 1.

[0161] <3> In the above embodiment, the structure to be measured is a kiln shell 1. However, the method of the present invention is not limited to the kiln shell 1 and can also be applied to estimating the shape of the wall surface of other cylindrical structures. Specifically, it can be used to estimate the shape of structures such as road tunnels, water tunnels for hydroelectric power generation, and sewer pipes.

[0162] <4> The present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to provide a better understanding of the present invention, and the present invention is not necessarily limited to those having all of the configurations described. The scope of the present invention is defined by the claims, and it is intended to include all modifications within the meaning and scope of the claims.

[0163] 1: Kiln shell 11: 3D laser scanner 13: Reference ball 17: Central axis 18: End face 20: Processing device 21: Fitting processing unit 22: Divided area data conversion unit 23: Approximate ellipse generation unit 24: Comparison processing unit 25: Dividing processing unit 27: Memory unit 29: Output unit 31: Cylinder axis 31a, 31b, 31c: Cylinder 35a, 35b, 35c: Central axis dc1: First point cloud data dc2: Second point cloud data dc3: Third point cloud data df: Divided area data

Claims

1. A method for estimating the shape of a cylindrical structure, comprising: (a) a step of obtaining first point cloud data that conforms to the shape of a wall surface of a target area of ​​the structure over an entire target area of ​​the structure; (b) a step of fitting a cylinder to the first point cloud data to obtain second point cloud data; (c) a step of setting a reference coordinate system including a reference plane formed by a portion of the second point cloud data and a reference axis that is perpendicular to the reference plane and substantially parallel to the longitudinal direction of the structure; (d) a step of converting the second point cloud data or data based on the second point cloud data into a plurality of divided area data, which are data for each unit area defined by a reference length in a direction along a cylindrical axis that is the central axis of the fitted cylinder, and a reference angle in a circumferential direction around the cylindrical axis; and (e) a step of ellipsically approximating a cross section perpendicular to the reference axis for each axial coordinate position that is a coordinate position related to a direction along the reference axis based on the divided area data, to obtain an approximated ellipse for each axial coordinate position.

2. A method for estimating the shape of a structure as described in claim 1, characterized in that it comprises a step (f) of comparing the outer edge shape of the approximate ellipse with the outer edge shape indicated by the divided area data for each axis coordinate position.

3. The method for estimating the shape of a structure as described in claim 2, characterized in that step (f) includes a step of comparing, for each axial coordinate position, the diameter of the approximate ellipse with the distance between the divided area data and the cylinder axis.

4. A method for estimating the shape of a structure as described in any one of claims 1 to 3, characterized in that it comprises a step (g) of calculating the flattening or overriding rate of the approximate ellipse for each axis coordinate position.

5. A method for estimating the shape of a structure as described in any one of claims 1 to 3, characterized in that it comprises a step (i) of measuring the progress of displacement of the center position of the approximation ellipse for each axis coordinate position from a virtual reference line substantially parallel to the longitudinal direction of the structure.

6. The method for estimating the shape of a structure according to claim 5, wherein the virtual reference line is a straight line virtually connecting the center positions of the approximation ellipses at two different locations, or the reference axis.

7. The method for estimating the shape of a structure as set forth in claim 1, wherein the structure is configured to be rotatable around a central axis along the longitudinal direction of the structure, and step (a) includes a step of obtaining the first point cloud data for the structure in a first stopped state, and then rotating the structure to move the position of a top of the structure to a position different from the top, and obtaining the first point cloud data for the structure in a second stopped state after the structure is rotated, and steps (b) to (d) are executed based on the first point cloud data obtained in each of the first stopped state and the second stopped state, and step (e) includes a step of obtaining the approximate ellipse for each axial coordinate position in the first stopped state and the approximate ellipse for each axial coordinate position in the second stopped state, respectively.

8. The method for estimating the shape of a structure as described in claim 7, characterized in that it includes a process (f) of comparing, for each of the axis coordinate positions, an outer edge shape of the approximate ellipse with an outer edge shape indicated by the divided area data in each of the first and second stopping states, and a process (h) of estimating that a local deformation exists at the specific axis coordinate position of the structure when a discrepancy is confirmed between the outer edge shape of the approximate ellipse and the outer edge shape indicated by the divided area data at the specific axis coordinate position in both the first and second stopping states in the process (f).

9. A method for estimating the shape of a structure as described in claim 7 or 8, characterized in that it comprises a step (g) of calculating the flattening or overriding rate of the approximate ellipse for each of the axial coordinate positions for each of the first stop state and the second stop state.

10. The method for estimating the shape of a structure according to claim 1, 2, 3, 7 or 8, characterized in that the step (d) comprises: a step (d1) of dividing the second point cloud data into a plurality of sections along the reference axis and fitting a cylinder to each of the second point cloud data for each section to obtain third point cloud data; and a step (d2) of converting the third point cloud data into a plurality of divided area data defined by the reference length and the reference angle based on the cylindrical axis, which is the central axis of the cylinder fitted when obtaining the third point cloud data.

11. A method for estimating the shape of a structure as described in claim 1, 2, 3, 7 or 8, characterized in that step (a) includes a step of scanning an inner wall of the structure using a 3D laser scanner to obtain the first point cloud data.

12. The method for estimating the shape of a structure as described in claim 11, characterized in that the structure is a rotary kiln, and the target area is an area where the inner wall of the rotary kiln is exposed after refractory bricks have been removed.