Composite strain sensing network and tank deformation measuring method

Abstract

The invention discloses a composite strain sensing network which comprises FBG sensors and a flexible fabric strain sensor, and the multiple FBG sensors are located on an optical fiber. The optical fibers are transversely and longitudinally arranged on the surface of the tank body; each strain sensor is arranged next to the FBG sensor, and the pasting direction is consistent with the pasting direction of the FBG sensor. The invention further discloses a tank deformation measuring method. The method comprises the following steps: carrying out strain measurement on the tank; establishing a coordinate system, and calculating coordinate values of all sensor position points through a mathematical relationship between strain and deformation according to the positions of the measuring points of the composite strain sensing network; dividing the surface of the tank body into n sub-regions, establishing a quadric surface function for each sub-region, weighting according to the strain of each point, solving an equation by using a weighted least square method, and obtaining a curved surface function of each sub-region; and integrating the curved surface functions of all the sub-regions to obtain the overall deformation condition of the tank body. The method has the advantages of comprehensive and accurate measurement and the like.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of tank deformation monitoring, and particularly relates to a composite strain sensing network and a tank deformation measurement method. Background Art

[0002] Tank deformation monitoring is crucial for ensuring the structural safety, extending the service life, and improving the operation efficiency of large tanks. Large tanks such as oil storage tanks, chemical liquid storage tanks, and solid rocket barrels usually bear internal pressure, loads, and internal and external temperature changes. Deformation monitoring can detect potential safety hazards at an early stage, such as local expansion, cracks, or material fatigue, thus preventing structural failure, leakage, or explosion, and ensuring personnel safety and environmental protection. By monitoring the deformation in real time, the maintenance cycle can be optimized, sudden failures and unnecessary downtimes can be avoided, and at the same time, the operation efficiency can be improved and the maintenance cost can be reduced.

[0003] In addition, deformation monitoring helps to improve the design and management level of tanks. Long-term monitoring data can provide valuable feedback for engineers to help optimize the tank design and enhance its strength and safety. Deformation monitoring can also detect problems such as aging, fatigue, and corrosion of the tank, and timely take maintenance measures to extend the service life of the tank. For the requirements of meeting industry regulations and standards, deformation monitoring is also an effective means to ensure compliance and reduce legal risks.

[0004] In summary, tank deformation monitoring can not only ensure the safety of the tank during operation, but also help improve the operation efficiency, reduce the risk of environmental pollution, provide data support for equipment management and design improvement, and is an indispensable part of modern industrial management.

[0005] In the existing technical solutions, the methods for measuring the deformation of the tank body include single-point measurement methods, laser ranging, digital imaging technology, etc. Single-point measurement methods such as the girth measurement method, optical reference line method, total station, GPS method, etc. These methods rely on a small number of observation points to infer the deformation of the entire tank body. Due to the scarcity of measurement points, it is impossible to fully obtain the detailed information of local deformation and overall deformation, resulting in inaccurate inference of overall deformation. Especially in the case of a complex tank body structure and large local deformation, the health monitoring of the tank body loses accuracy and cannot effectively reflect the actual deformation situation. Although the accuracy of strain sensors (such as strain gauges) is relatively high, in actual applications, they need to be installed on the surface of the tank body and are affected by environmental factors (such as temperature, humidity, etc.), which will cause deviations in measurement data and thus affect the judgment of the health status of the tank body. Once multiple strain sensors are installed, the corresponding transmission cable layout will be very large and messy. Methods such as laser ranging and digital imaging technology provide relatively accurate non-contact measurements, but their implementation effects highly depend on the surface reflection characteristics of the measured object (such as surface smoothness, coating, etc.), and may be severely affected when used on the complex surface of the tank body. In addition, the calculation cost of these methods is relatively high. Especially in the case of large-scale monitoring, the calculation amount and processing complexity are also relatively large. The high calculation cost and technical threshold also make these methods not have wide application value in large-scale and real-time monitoring.

[0006] FBG sensors can measure tiny changes and are suitable for large-scale networking layout. They are not affected by external electromagnetic interference and do not require shielded cables. Only a few optical fibers can be used to connect hundreds of FBG sensors to form a large-scale sensing network, which is convenient for global monitoring of the tank body; the flexible fabric strain sensor has a larger strain measurement range and can adapt to complex surfaces and irregularly shaped objects, which can make up for the deficiency of the limited strain measurement range of FBG sensors and achieve wide-range measurement of the deformation of the tank body.

[0007] There are mainly the following problems in the existing use of FBG sensors to monitor the surface deformation of the tank body:

[0008] 1) The FBG sensors are wound in a single direction, making it difficult to measure the strain in the vertical direction and unable to comprehensively reflect the deformation of the tank body surface. They are arranged in a certain direction on the tank body surface;

[0009] 2) The strain measurement range of currently available FBG sensors on the market is very narrow, making it difficult to effectively monitor the situation of large deformation of the tank body;

[0010] 3) Currently, the early warning judgment of the tank body deformation is mainly made through the strain obtained by monitoring, but the specific deformation amount cannot be known. Summary of the Invention

[0011] In view of the technical problems existing in the prior art, the present invention provides a composite strain sensing network with high measurement accuracy and a method for measuring the deformation of a tank body.

[0012] To solve the above technical problems, the technical solution proposed by the present invention is as follows:

[0013] A composite strain sensing network includes FBG sensors and flexible fabric strain sensors. A plurality of FBG sensors are located on the optical fiber; each of the optical fibers is arranged horizontally and vertically on the surface of the tank body to form a mesh structure; each of the flexible fabric strain sensors is arranged closely adjacent to the FBG sensor, and the mounting direction is the same as the pasting direction of the FBG sensor.

[0014] Preferably, 2m FBG sensors are provided on each horizontally wound optical fiber; starting from one end of the optical fiber for numbering, the distance between each FBG sensor is equal to 1 / 2m of the circumferential length of the tank body, and the winding distance between each horizontal optical fiber is equal; 2n + 1 FBG sensors are provided on each vertically pasted optical fiber, all starting from one end of the optical fiber for numbering, and the distance between the FBG sensors on it is equal to the distance of the horizontally wound optical fiber pre-designed to be wound on the tank body.

[0015] Preferably, the starting positions of all horizontally wound optical fibers for pasting are the same, so that the FBG sensors with the same serial number in each horizontally wound optical fiber are on the same height on the side of the tank body; the pasting positions of all vertically pasted optical fibers on the tank body are determined according to the horizontally wound optical fibers, so as to ensure that the FBG sensors on each vertically pasted optical fiber accurately approach the FBG sensors with the same serial number on all horizontally wound optical fibers in sequence, and the FBG sensors with the same serial number on each vertically pasted optical fiber must be close to and staggered from the same horizontally wound optical fiber.

[0016] Preferably, the flexible fabric strain sensor is arranged closely adjacent to the FBG sensor at a preset position; the preset position is a position with large deformation or a key observation position of the tank body.

[0017] The present invention also discloses a method for measuring the deformation of a tank body based on the above-mentioned composite strain sensing network, including the steps of:

[0018] Measuring the strain of the tank body with the composite strain sensing network;

[0019] Taking the center of the bottom surface of the tank as the zero point to establish a coordinate system, and according to the positions of the measuring points of the composite strain sensing network, calculating the coordinate values (x, y, z) of all sensor position points through the mathematical relationship between strain and deformation;

[0020] Dividing the surface of the tank body into n sub-regions, establishing a quadratic surface function for each sub-region, weighting according to the strain magnitude of each point, and using the weighted least squares method to solve the equation to obtain the surface function of each sub-region;

[0021] Integrate the surface functions of all sub-regions to obtain the overall deformation of the tank body.

[0022] Preferably, when measuring the strain of the tank body with a composite strain sensing network, select according to the measured values of the FBG sensor and the flexible fabric strain sensor; first check its strain measurement value and determine whether it has exceeded the measurement range of the FBG sensor; if it has exceeded, directly select the strain measurement value of the flexible fabric strain sensor; if not, select the strain measurement value of the FBG sensor.

[0023] Preferably, the specific process of calculating the coordinate values (x, y, z) of all sensor position points through the mathematical relationship between strain and deformation amount is as follows:

[0024] Assume that the arc coordinates of the original coordinates are represented as (r0, θ0, H0), and the coordinates after deformation are (r, θ0, H). Then the coordinates after deformation are represented as:

[0025] r = r0 + Δr = r0 + ∈ 横 ·r0

[0026] H = H0 + ΔH = H0 + ∈ 纵 ·H0

[0027] Where Δr is the radius change caused by the transverse strain ∈ 横 ΔH is the radius change caused by the longitudinal strain ∈ 纵 The radius change caused by;

[0028] The coordinates (x, y, z) of this point after deformation are calculated by the following formula:

[0029] Abscissa x:

[0030] x = r·cos(θ0) = (r0 + ∈ 横 ·r0)cos(θ0)

[0031] Ordinate y:

[0032] y = r·sin(θ0) = (r0 + ∈ 横 ·r0)sin(θ0)

[0033] Height coordinate z:

[0034] z = H = H0 + ∈ 纵 ·H0

[0035] That is, the three-dimensional coordinate values (x, y, z) of this point position are obtained.

[0036] Preferably, the specific process of dividing the tank body surface into n sub-regions and establishing a quadratic surface function for each sub-region is as follows:

[0037] The surface of the tank body is divided into multiple "field"-shaped sub-regions through a composite strain sensing network. Each sub-region consists of 9 points, so the sub-region presents as a 2×2 grid; since there are 2n + 1 horizontally wound and 2m vertically pasted optical fibers, the entire surface of the tank body is divided into n×m sub-regions; according to the coordinate values (x, y, z), transverse strain, and longitudinal strain of the 9 position points in each sub-region, the surface function of each sub-region is fitted, and the form is as follows:

[0038] z = Ax 2 + By 2 + Cxy + Dx + Ey + F

[0039] A, B, C, D, E, and F are coefficients to be solved; during the fitting process, not only the coordinates (x, y, z) of each point are considered, but also the transverse strain ∈ 横 and the longitudinal strain ∈ 纵 are introduced to adjust the second derivative of the surface.

[0040] Preferably, the surface function is optimized by the weighted least squares method, and the objective function of the weighted least squares method is written as:

[0041]

[0042] The weight w i is set to be proportional to the strain value ∈ i :

[0043] w i ∝ ∈ i

[0044] The strain value of each position point considers the transverse strain ∈ 横 and the longitudinal strain ∈ 纵 , and the strain value used to calculate the weight of each position point takes the square root of the sum of the squares of these two:

[0045]

[0046] To sum up, the weight w i is a function of the strain value ∈ i , and the strain value is normalized to avoid the weight being overly biased due to the overly large strain value of some points:

[0047]

[0048] ∈ i is the strain value of the i-th point, is the sum of the strain values of all points.

[0049] Preferably, solve the quadratic surface coefficient equation: write the objective function in the form of vectors and matrices. First, define the error vector e and the coefficient vector p:

[0050] The error vector e = [e1, e2,..., e n T , where each e i is the error of the i-th point:

[0051] e i = z i -(Ax 2 + By 2 + Cxy + Dx + Ey + F)

[0052] The coefficient vector p = [A, B, C, D, E, F] T , which are the unknown parameters to be solved;

[0053] Construct a design matrix X and an observation value vector z, and transform the objective function into matrix form. The design matrix X is composed of (x i , y i ) of each point, and the form is as follows:

[0054]

[0055] X is an n×6 matrix, and each row contains the quadratic polynomial terms corresponding to the point (x i , y i );

[0056] The observation value vector z is the z value of all points:

[0057]

[0058] The objective function of the weighted least squares method is transformed into the following matrix form:

[0059] mine T We = (z - Xp) T W(z - Xp)

[0060] where W is a diagonal matrix containing weights w i :

[0061]

[0062] Take the derivative of the objective function with respect to the parameter p and set the derivative to zero; through matrix differentiation, the normal equation is obtained:

[0063] X T WXp = X T Wz(z - Xp)

[0064] X​T WX is a 6×6 matrix, X T Wz is a 6×1 vector. Finally, the coefficient vector p = [A, B, C, D, E, F] is obtained by solving this equation T .

[0065] Compared with the prior art, the advantages of the present invention are as follows:

[0066] In the composite strain sensing network layout used in the present invention, there are two types: horizontal layout and vertical layout, which can measure the strain values in multiple directions, more accurately restore the actual strain situation on the surface of the tank body, and the layout network ensures coverage of the entire surface of the tank body and is evenly distributed. The entire surface of the tank body can be evenly divided for analysis, providing more comprehensive strain measurement data. The composite strain sensor network is composed of FBG sensors and flexible fabric strain sensors, integrating the advantages of both sensors. The large-scale networking of FBG sensors ensures the monitoring range of the tank body surface, and flexible fabric strain sensors are pasted at key position points, avoiding the problem of too small measurement range of a single FBG sensor and ensuring that the deformations at all key positions of the tank body can be accurately monitored.

[0067] The quadratic surface function and weighted least squares method used in the present invention for fitting the surface of the tank body. The quadratic surface fitting not only ensures the accuracy and surface smoothness of the fitting, but also greatly reduces the computational complexity compared with cubic or higher-order surface fitting, optimizing the computational efficiency. The use of the weighted least squares method further improves the accuracy of the results, especially when dealing with the weighting of different measurement points, ensuring a more accurate fitting result.

[0068] The present invention solves the defects of the prior art in terms of measurement accuracy, environmental adaptability, computational efficiency, and maintenance cost through a composite strain sensing network, temperature compensation technology, and a quadratic surface fitting model. Compared with the prior art, the present invention can provide a more accurate, stable, and real-time tank deformation monitoring solution, significantly improving the effect and reliability of the tank health monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a schematic diagram of the installation position (dotted line position) of the FBG sensor of the present invention.

[0070] Figure 2 It is a schematic diagram of the horizontal and vertical mounting (partial) of the FBG sensor of the present invention.

[0071] Figure 3 It is a schematic diagram of the structure of the flexible fabric strain sensor used in the present invention.

[0072] Figure 4 It is a schematic diagram of the mounting of the FBG sensor and the flexible fabric strain sensor of the present invention.

[0073] Figure 5 Schematic diagram of mounting the FBG sensor and the flexible fabric strain sensor of the present invention; (a) is the horizontal installation; (b) is the vertical installation.

[0074] Figure 6 Schematic diagram of the tank body coordinates in the present invention; (a) is the overall view; (b) is the top view cross-section.

[0075] Figure 7 Flowchart of the method for measuring the deformation of the tank body by the composite strain sensing network of the present invention.

[0076] Figure 8 Embodiment diagram of the tank body deformation measurement system based on the composite strain sensing network of the present invention in specific applications.

[0077] Legend: 1. FBG sensor; 2. Flexible fabric strain sensor; 201. Signal output area; 202. Sensitive area; 203. Bonding area. Detailed implementation manners

[0078] The present invention will be further described below in conjunction with the specification drawings and specific embodiments.

[0079] As Figure 1 shown, the composite strain sensing network provided by the embodiment of the present invention includes an FBG sensor 1 and a flexible fabric strain sensor 2; wherein the FBG sensor 1 is covered in a net shape on the entire surface of the tank body, and the flexible fabric strain sensor 2 is installed in key areas, jointly forming a composite strain sensing network to realize the overall strain monitoring of the tank body and the wide-range strain measurement of key areas.

[0080] Specifically, the FBG sensors 1 are combined to form a sensing network, specifically as Figure 1 - Figure 2 shown: The optical fibers engraved with the FBG sensors 1 pasted on the surface of the tank body to be monitored are divided into two types: horizontal winding and vertical pasting. Each type has 2n + 1 roots (horizontal, generally not less than 9 roots) and 2m roots (vertical, generally not less than 8 roots). These two specifications of optical fibers need to be designed in advance according to the size of the tank body to be measured. Each horizontally wound optical fiber has 2m FBG sensors 1, and they are all numbered starting from one end of the optical fiber. The distance between each FBG sensor 1 is equal to 1 / 2m of the circumferential length of the tank body. The winding distance between each horizontal optical fiber shall not be greater than 1 / 8 of the height of the tank body to be measured, that is, 2n + 1 ≥ 9, and the distances are equal; each vertically pasted optical fiber is engraved with 2n + 1 FBG sensors 1, and they are all numbered starting from one end of the optical fiber, and the distance between the FBG sensors 1 on it is equal to the distance of the horizontally wound optical fiber wound on the tank body designed in advance.

[0081] Before pasting the optical fiber on the tank body, marking lines need to be made first. When each FBG sensor 1 is pasted, the surface smoothness of the pasting part needs to be ensured. The starting positions of all horizontally wound optical fibers for pasting are equal, and it is ensured that the FBG sensors 1 with the same serial number in each horizontally wound optical fiber are on the same height on the side of the tank body. The position of the longitudinally pasted optical fiber on the tank body is mainly determined according to the horizontally wound optical fibers, ensuring that the FBG sensors 1 on each longitudinally pasted optical fiber accurately approach the FBG sensors 1 with the same serial number on all horizontally wound optical fibers in sequence, and the FBG sensors 1 with the same serial number on each longitudinally pasted optical fiber must be infinitely close to the same horizontally wound optical fiber, with a slight stagger, generally with a distance less than 1 cm. Specifically as Figure 2 shown. The horizontally and longitudinally pasted FBG sensors 1 jointly measure the transverse strain and longitudinal strain at the same position point, enabling more accurate measurement of the strain situation at this point.

[0082] Since the FBG sensors 1 are affected by both temperature and strain, the FBG sensors 1 need to be temperature calibrated before mounting. During measurement, additional FBG sensors 1 need to be arranged on the tank body for temperature measurement. When calculating the strain, the temperature sensitivity coefficient of the pre-calibrated temperature sensor, as well as the temperature sensitivity coefficient and strain sensitivity coefficient of the strain sensor, should be substituted. After temperature compensation calculation, the discrete strain values of all FBG sensors 1 are obtained.

[0083] As Figure 3 shown, the flexible fabric strain sensor 2 includes a sensitive area 202, a bonding area 203, and a signal output area 201. The flexible fabric strain sensor can be customized in size according to the size of the object to be measured, ensuring both the measurement range and avoiding interference with other components; the strain measurement range of the flexible fabric strain sensor 2 is 0 - 40%, which is much larger than the strain measurement range of the FBG sensors 1, and can make up for the deficiency of the limited strain measurement range of the FBG sensors 1. Among them, when the flexible fabric strain sensor 2 is installed, it is close to the FBG sensors 1, and the mounting direction is the same as the pasting direction of the FBG sensors 1, and the distance between them is not greater than 1 cm, which is specifically related to the size of the measured tank body, ensuring that the FBG sensors 1 and the flexible fabric strain sensor 2 measure the strain at the same position point as much as possible. Specifically as Figure 4 and Figure 5 shown.

[0084] Considering that arranging too many flexible fabric strain sensors 2 will lead to redundant signal cables and make the overall system too bulky, the number of pasted flexible fabric strain sensors 2 is only 1 / 2 or even less than that of the FBG sensors 1. The flexible fabric strain sensors 2 are preferentially installed at positions where large deformations may occur or key observation positions of the tank body.

[0085] In the composite strain sensing network layout used in the present invention, there are two types: horizontal layout and vertical layout, which can measure strain values in multiple directions, more accurately restore the actual strain situation on the surface of the tank body, and the layout network ensures coverage of the entire surface of the tank body and is evenly distributed. The entire surface of the tank body can be evenly divided for analysis, providing more comprehensive strain measurement data. The composite strain sensor network consists of FBG sensor 1 and flexible fabric strain sensor 2. Combining the advantages of the two sensors, the large-scale networking of FBG sensor 1 ensures the monitoring range of the tank body surface, and flexible fabric strain sensor 2 is pasted at key position points, avoiding the problem of too small a measurement range of a single FBG sensor 1 and ensuring that the deformation of each key position of the tank body can be accurately monitored.

[0086] As Figure 7 shown, the embodiment of the present invention also provides a method for measuring the deformation of a tank body based on the above composite strain sensing network, specifically including the steps:

[0087] S1. Through the composite strain sensing network arranged as above, overall strain monitoring of the tank body and wide-range strain measurement of key areas are realized;

[0088] S2. Taking the center of the bottom surface of the tank bottom as the zero point to establish a coordinate system, according to the positions of the measuring points of the sensing network, and through the mathematical relationship between strain and deformation, calculate the coordinate values (x, y, z) of all sensor position points;

[0089] S3. Divide the surface of the tank body into n sub-regions, establish a quadratic surface function for each sub-region, weight according to the strain magnitude of each point, and use the weighted least squares method to solve the equation to obtain the surface function of each sub-region;

[0090] S4. Finally, integrate the surface functions of all regions to obtain the overall deformation situation of the tank body.

[0091] In step S1, since there are both FBG sensor 1 and flexible fabric strain sensor 2 at some key points, it is necessary to make a selection based on the measured values of the two sensors at these points. Since the strain measurement range of flexible fabric strain sensor 2 is much larger than that of FBG sensor 1, first check its strain measurement value and determine whether it has exceeded the measurement range of FBG sensor 1; if it has exceeded, directly select the strain measurement value of flexible fabric strain sensor 2; if not, select the strain measurement value of FBG sensor 1 to maintain data consistency as much as possible.

[0092] The specific process of step S2:

[0093] S201. Determine the relationship between strain and deformation

[0094] Assume that the shape of the tank body is cylindrical, and the original radius at a certain point on the tank body is r0. In a local area, the tank body is deformed, causing the radius to change. To estimate the coordinate value change (i.e., radius change) at a certain point on the tank body from the transverse strain and longitudinal strain, it is necessary to confirm the relationship between the strain and the deformation;

[0095] Strain sensors are arranged on the surface of the tank body to measure the change in the transverse strain ∈ 横 and the longitudinal strain ∈ 纵 of each point;

[0096] The transverse strain ∈ 横 is the strain along the circumferential direction (tangential direction), which describes the deformation of the point in the circumferential direction; the longitudinal strain ∈ 纵 is the strain along the axial direction of the tank body, which describes the deformation of the point along the height direction of the tank body.

[0097] S202. Determine the relationship between the coordinate system of the tank body and the deformation

[0098] As Figure 6 shown, assume that the original geometric shape of the tank body is a cylinder, the surface is circular, and the arc coordinate form of the original coordinates of a certain point is (r0, θ0), where r0 is the original radius of the point to the center of the circle, and θ0 is the angle of the point (i.e., the position on the circumference). The longitudinal strain will cause the deformation of the point along the axial direction of the tank body, affecting the change in the position of the point in the height direction. The transverse strain will cause the deformation of the point in the circumferential direction (i.e., tangential direction), affecting the position of the point on the circumference.

[0099] S203. Deduce the coordinate change from the strain

[0100] The longitudinal strain ∈ 纵 represents the deformation of the point along the axial direction (height direction of the tank body). Assume that the original height of the tank body is H0 and the deformed height is H. Then the longitudinal strain is defined as:

[0101]

[0102] Therefore, the height change amount ΔH caused by the longitudinal strain is:

[0103] ΔH = ∈ 纵 ·H0

[0104] However, the longitudinal strain does not directly affect the radius r of a certain point on the surface of the tank body, because the longitudinal strain mainly affects the height change of the tank body along the axial direction.

[0105] The transverse strain ∈ 横 refers to the strain along the circumferential direction. Assume that the deformed radius is r and the original radius is r0. The transverse strain is defined as:

[0106]

[0107] Therefore, the amount of radius change Δr caused by the lateral strain is as follows:

[0108] Δr = ∈ 横 ·r0

[0109] Obtain the effects of the two strains on a certain point on the tank surface:

[0110] 1. The longitudinal strain affects the change in the height direction of the point, but does not directly affect the radius.

[0111] 2. The lateral strain affects the change in the radius of the point, but does not directly affect the height direction.

[0112] Therefore, assuming that the arc coordinate form of the original coordinates is (r0, θ0, H0), and the deformed coordinates are (r, θ0, H), then the deformed coordinates can be expressed as:

[0113] r = r0 + Δr = r0 + ∈ 横 ·r0

[0114] H = H0 + ΔH = H0 + ∈ 纵 ·H0

[0115] S204. Deformation amount and coordinate value conversion

[0116] The coordinates (x, y, z) of the point after deformation can be calculated by the following formula:

[0117] Abscissa (x):

[0118] x = r·cos(θ0) = (r0 + ∈ 横 ·r0)cos(θ0)

[0119] Ordinate (y):

[0120] y = r·sin(θ0) = (r0 + ∈ 横 ·r0)sin(θ0)

[0121] Height coordinate (z):

[0122] z = H = H0 + ∈ 纵 ·H0

[0123] The three-dimensional coordinate values (x, y, z) of the point position can be obtained.

[0124] Step S3. Division of the tank surface and construction of sub-regions:

[0125] Step S301. Division of the tank surface area

[0126] The surface of the tank body is divided into multiple "field"-shaped sub-regions through a composite strain sensing network. Each sub-region consists of 9 points, so the sub-region presents as a 2×2 grid. Since there are 2n + 1 horizontally wound and 2m vertically pasted optical fibers, the entire surface of the tank body can be exactly divided into n×m sub-regions. Given the coordinate values (x, y, z), transverse strain, and longitudinal strain at 9 position points within each sub-region of the fiber Bragg grating sensor, the surface function of each sub-region can then be fitted based on these parameters, and finally these local surfaces are stitched together to obtain the shape of the entire tank body surface.

[0127] Step S302. Construction of the sub-region surface function

[0128] Considering the computational efficiency and the fact that there are only 9 point data in the currently selected sub-region and no more local data, a quadratic surface is selected as the fitting model. The fitting ability of the quadratic surface within each sub-region can effectively capture local variations and has good smoothness, which can better ensure that the surface after fitting has no sharp mutations or excessive fluctuations.

[0129] Fit a quadratic surface for each sub-region, in the form:

[0130] z = Ax 2 + By 2 + Cxy + Dx + Ey + F

[0131] Here, A, B, C, D, E, and F are the coefficients to be solved. During the fitting process, not only the (x, y, z) coordinates of each point are considered, but also the transverse strain ∈ 横 and the longitudinal strain ∈ 纵 are introduced to adjust the second derivative (i.e., curvature) of the surface.

[0132] Step S303. Optimization of the surface function by weighted least squares method

[0133] Use the weighted least squares method to solve. For each point (x i , y i , z i ), if the strain at this point is relatively large (indicating that this point changes faster or is more critical), a larger weight can be assigned, so that the error at this point is more important during the fitting process; conversely, if the strain at a certain point is smaller, a smaller weight can be assigned, so that the influence of this point during fitting is smaller.

[0134] The objective function of the weighted least squares method can be written as:

[0135]

[0136] The choice of weights can be based on the following principle:

[0137] 1) If the strain value of a point is large, it indicates that the change at that point is significant. To more accurately fit this point, a larger weight is assigned to this point.

[0138] 2) If the strain value of a point is small, it indicates that the change at that point is minor. The weight can be appropriately reduced to avoid this point overly influencing the fitting result.

[0139] To enable points with larger strain to have a greater impact on the fitting, the weight w i is set to be proportional to the strain value ∈ i :

[0140] w i ∝ ∈ i

[0141] The strain value of each position point should consider the lateral strain ∈ 横 and the longitudinal strain ∈ 纵 . The strain value used to calculate the weight for each position point is taken as the square root of the sum of the squares of these two:

[0142]

[0143] In summary, the weight w i can be a function of the strain value ∈ i , and the strain value is normalized to avoid the weight being overly biased due to the overly large strain value of some points:

[0144]

[0145] Here, ∈ i is the strain value of the i-th point, is the sum of the strain values of all points.

[0146] This normalization method ensures that the sum of all weights is 1.

[0147] Step S304. Solve the quadratic surface coefficient equation

[0148] Write the objective function in the form of vectors and matrices. First, define the error vector e and the coefficient vector p:

[0149] The error vector e = [e1, e2,..., e n T , where each e i is the error of the i-th point:

[0150] e i = z i - (Ax 2 + By 2 + Cxy + Dx + Ey + F)

[0151] ​The coefficient vector p = [A, B, C, D, E, F] T , which are the unknown parameters to be solved for.

[0152] To transform the objective function into matrix form, a design matrix X and an observation vector z need to be constructed:

[0153] The design matrix X is composed of (x i , y i ) for each point, and has the following form:

[0154]

[0155] Here, X is an n×6 matrix (n = 9), and each row contains the quadratic polynomial terms corresponding to the point (x i , y i ).

[0156] The observation vector z is the z-value of all points:

[0157]

[0158] The objective function of weighted least squares can be transformed into the following matrix form:

[0159] min e T We = (z - Xp) T W(z - Xp)

[0160] where W is a diagonal matrix containing the weights w i :

[0161]

[0162] To solve this problem, the objective function is differentiated with respect to the parameter p and the derivative is set to zero. Through matrix differentiation, the normal equation is obtained:

[0163] X T WXp = X T Wz(z - Xp)

[0164] Here, X T WX is a 6×6 matrix, and X T Wz is a 6×1 vector. Finally, the coefficient vector p = [A, B, C, D, E, F] T can be obtained by solving this equation.

[0165] Step S4. After completing the surface fitting of each sub-area, the surface functions of all sub-areas are spliced together to form the surface of the entire tank body, so that the overall deformation of the tank body can be known. The status detection is performed according to the set fixed period (such as every 1 minute or 1 hour), and compared with the previously set threshold. Once the threshold is exceeded, it is judged as abnormal and an alarm is issued, thereby realizing real-time monitoring of the tank body.

[0166] The present invention uses quadratic surface function and weighted least squares method to fit the tank surface. Quadratic surface fitting not only ensures the accuracy of fitting and surface smoothness, but also greatly reduces the computational complexity and optimizes the computational efficiency compared to cubic or higher-order surface fitting. The use of weighted least squares method further improves the accuracy of the results, especially when dealing with the weighting of different measurement points, ensuring a more accurate fitting result.

[0167] The present invention solves the defects of the prior art in terms of measurement accuracy, environmental adaptability, computational efficiency and maintenance cost by using a composite strain sensing network, temperature compensation technology and quadratic surface fitting model. Compared with the prior art, the present invention can provide a more accurate, stable and real-time tank deformation monitoring solution, significantly improving the effect and reliability of tank health monitoring.

[0168] like Figure 8 As shown, in specific applications, the fiber Bragg grating sensing network in the tank deformation monitoring system is connected to the fiber signal demodulation system, and the fiber signal demodulation system is composed of multiple fiber Bragg grating demodulators according to the number of optical fibers. The flexible fabric strain sensor 2 is connected to the flexible fabric strain signal acquisition system (such as an acquisition box) via the back-end transmission module.

[0169] Both ends of the optical fiber engraved with FBG sensor 1 are connected to the fiber Bragg grating demodulator by fusion-splice armored optical cable. The fiber Bragg grating sensor adopts double-ended form to avoid the problem that the entire chain signal cannot be collected due to damage to one end interface. If a single FBG sensor 1 of the optical fiber is damaged, it can be replaced with a fiber fusion splicer. The cables of the flexible fabric strain sensor 2 of each layer or each area will be bundled into a customized cable at a certain position on each layer, and then separated and connected to each port of the collection box near the collection box to avoid too many cables causing too much clutter on site.

[0170] The fiber optic signal demodulation system and the flexible fabric strain signal acquisition box are connected to the local signal analysis and monitoring platform through a signal transmission line. The local signal analysis and monitoring platform analyzes and processes the collected strain data to achieve the monitoring data analysis, early warning, visualization management and storage of the tank deformation, and can query the historical data. Through the local platform, the position of the tank deformation and the state change of the structure at this monitoring position can be directly displayed. Based on the deformation monitoring data collected by the composite strain sensing network, the reconstruction of the tank can be realized, enabling users to more quickly find the abnormal state points and master the operating state of the equipment. In addition, users can conduct correlation statistical analysis between different positions at the same time and different times at the same position through the local platform, enabling users to better understand the health status of the tank and ensure the safety of the tank. The local signal analysis and monitoring platform is transmitted to the remote monitoring platform in real time through a wired or wireless transmission network to achieve remote monitoring of the tank deformation.

[0171] The measurement method of the present invention is to form a composite strain sensing network by the FBG sensor 1 and the flexible fabric strain sensor 2. The two sensors are arranged on the tank according to a certain rule to detect the strain on the surface of the tank. The fiber optic grating temperature sensor is used for temperature compensation of the strain sensor. The surface function of the tank surface is fitted by an algorithm from the measured discrete strain data, and finally the deformation of the tank is obtained, which can accurately obtain the precise deformation details of the local or overall tank.

[0172] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions within the idea of the present invention belong to the protection scope of the present invention. It should be pointed out that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A composite strain sensing network, characterized in that, It includes FBG sensors (1) and flexible fabric strain sensors (2). A plurality of FBG sensors (1) are located on the optical fibers; each of the optical fibers is arranged horizontally and vertically on the surface of the tank body to form a mesh structure; each of the flexible fabric strain sensors (2) is arranged adjacent to the FBG sensor (1), and the mounting direction is the same as the pasting direction of the FBG sensor (1).

2. The composite strain sensing network according to claim 1, wherein There are 2m FBG sensors (1) provided on each horizontally wound optical fiber; starting from one end of the optical fiber for numbering, the spacing between each FBG sensor (1) is equal to 1 / 2m of the circumferential length of the tank body, and the winding spacing between each horizontal optical fiber is equal; there are 2n + 1 FBG sensors (1) provided on each vertically pasted optical fiber, all starting from one end of the optical fiber for numbering, and the spacing between the FBG sensors (1) on it is equal to the spacing of the horizontally wound optical fiber pre-designed on the tank body.

3. The composite strain sensing network according to claim 2, wherein The starting positions of all horizontally wound optical fibers for pasting are the same, so that the FBG sensors (1) with the same serial number in each horizontally wound optical fiber are on the same height on the side of the tank body; the pasting positions of all vertically pasted optical fibers on the tank body are determined according to the horizontally wound optical fibers, so as to ensure that the FBG sensors (1) on each vertically pasted optical fiber accurately approach the FBG sensors (1) with the same serial number on all horizontally wound optical fibers in sequence, and the FBG sensors (1) with the same serial number on each vertically pasted optical fiber must be close to and staggered from the same horizontally wound optical fiber.

4. The composite strain sensing network according to claim 1 or 2 or 3, characterized in that, The flexible fabric strain sensor (2) is arranged adjacent to the FBG sensor (1) at a preset position; the preset position is a position with large deformation or a key observation position of the tank body.

5. A method for measuring the deformation of a tank body based on the composite strain sensing network according to any one of claims 1-4, characterized in that, It includes steps: Using the composite strain sensing network to measure the strain of the tank body; Taking the center of the bottom surface of the tank bottom as the zero point to establish a coordinate system, and according to the positions of the measuring points of the composite strain sensing network, calculating the coordinate values (x, y, z) of all sensor position points through the mathematical relationship between strain and deformation amount; Dividing the surface of the tank body into n sub-regions, establishing a quadratic surface function for each sub-region, weighting according to the strain magnitude of each point, and using the weighted least squares method to solve the equation to obtain the surface function of each sub-region; Integrating the surface functions of all sub-regions to obtain the overall deformation situation of the tank body.

6. The method for measuring the deformation of the tank body according to claim 5, wherein When using the composite strain sensing network to measure the strain of the tank body, it is selected according to the measurement values of the FBG sensor (1) and the flexible fabric strain sensor (2); first, check whether the strain measurement value has exceeded the measurement range of the FBG sensor (1); If it has exceeded, directly select the strain measurement value of the flexible fabric strain sensor (2); if not, select the strain measurement value of the FBG sensor (1).

7. The method for measuring the deformation of the tank body according to claim 5, characterized in that, The specific process of calculating the coordinate values (x, y, z) of all sensor position points through the mathematical relationship between strain and deformation amount is as follows: Assuming that the arc coordinates of the original coordinates are expressed as (r0, θ0, H0), and the coordinates after deformation are (r, θ0, H), then the coordinates after deformation are expressed as: r = r0 + Δr = r0 + ∈ 横 ·r0 H = H0 + ΔH = H0 + ∈ 纵 ·H0 where Δr is the radius change caused by the transverse strain ∈ 横 ; ΔH is the radius change caused by the longitudinal strain ∈ 纵 . The coordinates (x, y, z) of this point after deformation are calculated by the following formula: Abscissa x: x = r·cos(θ0) = (r0 + ∈ 横 ·r0)cos(θ0) Ordinate y: y = r·sin(θ0) = (r0 + ∈ 横 ·r0)sin(θ0) Height coordinate z: z = H = H0 + ∈ 纵 ·H0 That is, the three-dimensional coordinate values (x, y, z) of this point position are obtained.

8. The method for measuring the deformation of the tank body according to claim 7, characterized in that, The specific process of dividing the tank body surface into n sub-regions and establishing a quadratic surface function for each sub-region is as follows: The tank body surface is divided into multiple "field"-shaped sub-regions through a composite strain sensing network. Each sub-region consists of 9 points, so the sub-region presents as a 2×2 grid; since there are 2n + 1 and 2m horizontally wound and vertically pasted optical fibers respectively, the entire tank body surface is divided into n×m sub-regions; according to the coordinate values (x, y, z), transverse strain, and longitudinal strain of the 9 position points in each sub-region, the surface function of each sub-region is fitted, and the form is as follows: z = Ax 2 + By 2 + Cxy + Dx + Ey + F A, B, C, D, E, and F are coefficients to be solved; during the fitting process, not only the coordinates (x, y, z) of each point are considered, but also the transverse strain ∈ 横 and the longitudinal strain ∈ 纵 are introduced to adjust the second derivative of the surface.

9. The method for measuring the deformation of the tank body according to claim 8, wherein, The surface function is optimized by the weighted least squares method, and the objective function of the weighted least squares method is written as: Set the weight w i to be proportional to the strain value ∈ i as follows: w i ∝ ∈ i The strain value at each position point considers the transverse strain ∈ 横 and the longitudinal strain ∈ 纵 , and the strain value used to calculate the weight at each position point is taken as the square root of the sum of the squares of these two: In summary, the weight w i is a function of the strain value ∈ i , and normalizes the strain value to avoid excessive weights caused by overly large strain values at certain points: ∈ i is the strain value of the i-th point, is the sum of the strain values of all points.

10. The method for measuring the deformation of the tank body according to claim 9, wherein, Solve the quadratic surface coefficient equation: Write the objective function in the form of vectors and matrices. First, define the error vector e and the coefficient vector p: Error vector e = [e1, e2, …, e n T , where each e i is the error of the i-th point:​ e i = z i -(Ax 2 + By 2 + Cxy + Dx + Ey + F) Coefficient vector p = [A, B, C, D, E, F] T , which are the unknown parameters to be solved Construct a design matrix X and an observation vector z, and transform the objective function into matrix form. The design matrix X is composed of (x i , y i ) for each point, in the following form: X is an n×6 matrix, where each row contains the quadratic polynomial terms corresponding to the points (x i , y i ); The observed value vector z is the z value of all points: The objective function of the weighted least squares method is transformed into the following matrix form: min e T We=(z - Xp) T W(z - Xp) where W is a diagonal matrix containing the weights w i : Take the derivative of the objective function with respect to the parameter p and set the derivative to zero; through matrix differentiation, the normal equation is obtained: X T WXp = X T Wz(z - Xp) X T WX is a 6×6 matrix, X T Wz is a 6×1 vector, and finally the coefficient vector p = [A, B, C, D, E, F] is obtained by solving this equation T .

Images