Method for compensating error of concrete constraint stress equipment caused by environment temperature

Abstract

The invention provides a method for compensating errors of concrete confined stress equipment caused by ambient temperature, which comprises the following steps of: building a template for pouring a concrete sample, and pouring the concrete sample; temperature sensors are arranged on a glass tube, an end metal block and metal frames on the two sides of the concrete confined stress equipment respectively, and the temperature history of the concrete test piece is measured; the temperature of the template for concrete sample pouring is constant at 14 DEG C so as to ensure that the concrete sample is not subjected to temperature deformation, and data measured by the displacement measuring device is only related to the environment temperature; constructing a temperature compensation neural network model; temperature values actually measured by the temperature sensors are substituted into the temperature compensation neural network model to calculate corresponding displacement, and displacement measurement data output by the displacement measurement device of the concrete constraint stress equipment are adjusted and compensated in real time. The accuracy and precision of a measurement result of the concrete constraint stress test equipment after compensation are greatly improved, and the concrete stress calculated according to test data is more accurate.

Description

Technical Field

[0001] The present invention relates to a method for compensating measurement errors of a concrete restraint stress device, specifically, a method for compensating measurement errors of a concrete restraint stress device caused by environmental temperature. Background Art

[0002] Since the cross-section of a mass concrete structure is relatively thick after pouring, the heat generated by the rapid hydration reaction inside it cannot be dissipated in time, resulting in an increase in its internal temperature. After the internal temperature rise reaches its peak, the internal temperature will slowly decrease; however, since the elastic modulus of the concrete is higher and its creep ability is weaker at this time, tensile stress is generated due to restraint in the late-age stage of the mass concrete structure. Once the tensile stress exceeds the tensile strength at the corresponding age, the concrete will generate temperature cracks.

[0003] The research on temperature crack resistance of mass concrete structures has always been the focus of attention in the engineering and academic fields. To study the relationship between the internal temperature - tensile stress - cracks of a mass concrete structure, a device used to simulate the cracking process of a mass concrete structure (such as a concrete dam), namely a concrete restraint stress device (also called a concrete restraint stress testing machine), is required. Its working principle is: by using a strain gauge to measure the deformation of a concrete specimen in real time, and based on the deformation of the concrete specimen under different temperature histories, the tensile stress generated by the concrete specimen due to restraint is simulated through the pushing and pulling actions of a motor, and complete or partial deformation restraint is imposed on the concrete specimen, thereby studying the relationship between the internal temperature - deformation - tensile stress - cracks of the mass concrete.

[0004] However, there is a major defect in the deformation measurement of the current concrete restraint stress test equipment used in laboratories: the influence of environmental temperature on test data is ignored, that is, the measured value is inaccurate due to the influence of environmental temperature, and the concrete restraint stress calculated based on this measured value is inaccurate. Since the cross-section of a mass concrete structure is very thick, up to dozens of meters, the temperature drop process of the concrete is very slow, mostly maintaining at 0.3 - 0.5 °C per day. Therefore, in the test process of simulating the deformation of a mass concrete structure due to temperature drop, it is necessary to accurately capture the deformation generated by the mass concrete structure during the above temperature drop process. Then, in the actual test process, due to the long measurement period, the change in the environmental temperature of the test equipment in the laboratory is often ignored, resulting in a large amount of deformation generated by the change in environmental temperature being coupled into the deformation measured by the deformation sensor. The data measured by the deformation sensor contains a large amount of errors, seriously affecting the test accuracy and the accuracy of test data! Summary of the Invention

[0005] In view of the above reasons, the purpose of the present invention is to provide a method for compensating measurement errors of a concrete restraint stress device caused by environmental temperature.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A method for compensating for errors caused by environmental temperature in a concrete restraint stress device, which includes the following steps:

[0007] S1. Build a formwork for casting a concrete specimen, and cast the concrete specimen; arrange temperature sensors on the glass tube, end metal block and two side metal frames of the concrete restraint stress device respectively to measure the temperature history of the concrete specimen.

[0008] S2. Keep the temperature of the formwork for casting the concrete specimen constant at 14 °C to ensure that the concrete specimen does not undergo temperature deformation, and the data measured by the displacement measuring device is only related to the environmental temperature.

[0009] S3. Construct a temperature compensation neural network model.

[0010] S3.1. Take temperature as the input layer and the corresponding displacement as the output layer, and determine the number of hidden layers by the empirical method or the trial method to construct the mapping relationship between variables based on the BP neural network.

[0011] S3.2. Obtain the temperature data of three temperature measurement points on the glass tube, end metal block and two side metal frames of the concrete restraint stress device to form a temperature data set {T 1 , T 2 , T 3 ...T n}, measure the displacement caused by temperature, and obtain a displacement data set {S 1 , S 2 , S 3 ...S n} as the input learning sample.

[0012] S3.3. Take temperature as the input quantity and displacement as the output quantity, and use the backpropagation algorithm to repeatedly train the weights and thresholds of the network to make the network output value infinitely close to the true value. When the objective function value meets the set requirements, the training is completed to obtain the temperature compensation neural network model.

[0013] The activation function in the temperature compensation neural network model of the present invention adopts the sigmoid function, and its expression is:

[0014]

[0015] The method for determining the number of hidden layers h is:

[0016]

[0017] In the formula, m and n are the numbers of nodes in the input layer and the output layer respectively, and a takes a constant between 1 and 10.

[0018] The objective function is set as the difference between the model output value and the actual monitored displacement value:

[0019]

[0020] Wherein, T is the temperature and S is the displacement value at this temperature;

[0021] The learning efficiency is set to 0.5, the error range is 10 -4 , and the maximum number of iterations is 2000 times;

[0022] S4. Substitute the temperature values measured by each temperature sensor into the temperature compensation neural network model determined in step S3 to calculate the corresponding displacement amount, and adjust the displacement measurement data output by the displacement measurement device of the compensation concrete restraint stress device in real time.

[0023] When training the temperature compensation neural network, it is necessary to perform linear normalization processing on the temperature and displacement amount data and map them to (-1, 1) to improve the accuracy of model training.

[0024] Since the present invention fully considers the influence of environmental temperature on the concrete restraint stress device and compensates it, eliminating the measurement error caused by the deformation of the displacement measurement device of the concrete restraint stress device due to the change of the ambient temperature, therefore, the accuracy and precision of the measurement result of the compensated concrete restraint stress test device of the present invention are greatly improved, and the concrete stress calculated based on the test data is more accurate! Description of the Drawings

[0025] Figure 1 It is a flowchart of the method for compensating the measurement error of the concrete restraint stress device caused by the environmental temperature in the present invention;

[0026] Figure 2 It is a schematic diagram of the installation position of the temperature sensor in the present invention;

[0027] Figure 3 It is a schematic diagram of the BP neural network structure in the present invention;

[0028] Figure 4 It is a flowchart of creating the temperature compensation neural network model in the present invention;

[0029] Figure 5 It is a graph of the convergence process of the temperature compensation neural network in the present invention;

[0030] Figure 6 It is a time history graph of the displacement of the concrete specimen measured after temperature compensation of the concrete restraint stress device in the present invention. Detailed Embodiment

[0031] The structure and features of the present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that various modifications can be made to the embodiments disclosed herein. Therefore, the embodiments disclosed in the specification should not be considered as limitations of the present invention, but only as examples of the embodiments, the purpose of which is to make the features of the present invention obvious.

[0032] As Figure 1 shown, the method for the present invention to compensate for the error caused by the environmental temperature in the concrete restraint stress device is as follows:

[0033] S1. Build a formwork for pouring concrete specimens and pour the concrete specimens; as Figure 2 shown, temperature sensors, namely temperature sensor 1, temperature sensor 2, and temperature sensor 3, are respectively arranged on the glass tube 1, the end metal block 2, and the two side metal frames 3 of the concrete restraint stress device to measure the temperature history of the concrete specimen.

[0034] The present invention selects to set temperature sensors at the glass tube, the end metal block, and the two side metal frames of the concrete restraint stress device because these positions are important components of the displacement measurement system of the concrete restraint stress device, and the change of their temperature will have a direct impact on the displacement of the concrete specimen. If there is a temperature change in the two side metal frames of the test equipment, the self-temperature deformation will have an indirect extrusion or tensile effect on the displacement measurement device, introducing errors. Arranging temperature sensors on the concrete restraint stress device in the present invention is different from existing devices and is a key step to eliminate the displacement error of the device.

[0035] S2. Keep the temperature of the formwork for pouring concrete specimens constant at 14 °C to ensure that the concrete specimen does not undergo temperature deformation, and the data measured by the displacement measurement device is only related to the environmental temperature.

[0036] S3. Build a temperature compensation neural network model.

[0037] S3.1. Use temperature as the input layer and the corresponding displacement as the output layer, and determine the number of hidden layers by the empirical method or the trial method to build the mapping relationship between variables based on the BP neural network.

[0038]

[0039] The weight, and a represents the node threshold.

[0040] The BP algorithm is the backpropagation algorithm, that is, the error function is gradually transmitted from the output end to the input end, and the weights and thresholds of the neural network are gradually corrected along the negative gradient direction of the error function.

[0041] The reverse optimization process is as follows:

[0042]

[0043] Among them, E is the error function, and o' is the theoretical value of the output.

[0044] The weight correction formula for the output layer is obtained from the above formula as:

[0045] V ji (t + 1) = V ji (t) - ηΔV ji (t) (2)

[0047] Among them, η is the learning step size.

[0048] Furthermore, the weight correction amount of the hidden layer is

[0049]

[0050] Furthermore, the weights of the output layer are obtained:

[0051] w ij (t + 1) = w ij (t) - ηΔw ij (t) (4)

[0052] S3.2. Obtain the temperature data of the three temperature measurement points of the glass tube, end metal block, and two-side metal frames of the concrete constraint stress device, form a temperature data group {T 1 , T 2 , T 3 ...T n}, measure the displacement caused by temperature, obtain a displacement data group {S 1 , S 2 , S 3 ...S n}, and use it as the input learning sample.

[0053] S3.3. Using temperature as the input quantity and displacement as the output quantity, use the backpropagation algorithm to repeatedly train the weights and thresholds of the network to make the network output value infinitely close to the true value. When the objective function value meets the set requirements, the training is completed to obtain the temperature compensation neural network model, as Figure 4 shown.

[0054] The activation function in the temperature compensation neural network model of the present invention adopts the sigmoid function, and its expression is

[0055]

[0056] The number of hidden layers is determined according to formula (6) and experience:

[0057]

[0058] Wherein, m and n are the numbers of nodes in the input layer and the output layer respectively, and a is a constant between 1 and 10.

[0059] The objective function is set as the difference between the model output value and the actual monitored displacement value, as shown in Equation (7):

[0060]

[0061] Wherein T is the temperature and S is the displacement value at this temperature.

[0062] The learning efficiency is set to 0.5 and the error range is 10 -4 , and the maximum number of iterations is 2000 times. The iterative convergence process of the neural network is shown in Figure 5 .

[0063] When training the temperature compensation neural network, it is necessary to perform linear normalization processing on the temperature and displacement data, and map them to the range of (-1, 1) to improve the training accuracy of the model.

[0064] S4. According to the temperature values measured by each actual temperature sensor, calculate the corresponding displacement through the temperature compensation neural network model, and adjust the displacement measurement data output by the compensation displacement measurement device in real time to eliminate the displacement measurement error introduced by the environmental temperature, so that the displacement measurement device only outputs the deformation of the concrete specimen itself.

[0065] Figure 6 is the time history diagram of the displacement of the concrete specimen after temperature compensation of the concrete constraint stress equipment by using the method disclosed in the present invention. It can be seen from the figure that during the test, the environmental temperature of the concrete stress testing machine is affected by the day-night temperature difference, and the temperature difference can reach 10 °C (that is, before compensation, the deformation generated by the displacement sensor is about 11 μm), while the deformation data increment measured by the corresponding displacement sensor after compensation is basically stable at about 1 μm. This shows that the deformation of the concrete stress equipment / testing machine caused by temperature basically disappears after compensation by Equation (7), with obvious effects, greatly improving the accuracy of the testing machine for temperature control of mass concrete, and making the test results more persuasive.

[0066] Finally, it should be noted that the above-mentioned embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for compensating the error of a concrete restraining stress device caused by ambient temperature, characterized in that: It includes the following steps: S1. Build a template for pouring concrete specimens and pour concrete specimens; arrange temperature sensors on the glass tube, the end metal block and the metal frames on both sides of the concrete restraint stress device to measure the temperature history of the concrete specimens; S2. Keep the temperature of the template used for pouring the concrete specimen constant at 14°C to ensure that the concrete specimen does not deform due to temperature, and the data measured by the displacement measuring device is only related to the ambient temperature; S3, constructing a temperature compensation neural network model; S3.1, using temperature as the input layer and the corresponding displacement as the output layer, determine the number of hidden layers through empirical methods or trial algorithms, and construct the mapping relationship between variables based on the BP neural network; S3.2, obtain the temperature data of the three temperature measuring points of the concrete restraint stress equipment glass tube, the end metal block, and the metal frames on both sides, and form a temperature data group {T1, T2, T3...T n }, measure the displacement caused by temperature, and obtain the displacement data set {S1, S2, S3...S n }, as input learning samples; S3.3, using temperature as input and displacement as output, the back propagation algorithm is used to repeatedly train the network weights and thresholds, so that the network output value is infinitely close to the true value. When the objective function value meets the set requirements, the training is completed and the temperature compensation neural network model is obtained; The activation function in the temperature compensation neural network model of the present invention adopts a sigmoid function, and its expression is: The method to determine the number of hidden layers h is: In the formula, m and n are the number of nodes in the input layer and output layer respectively, and a is a constant between 1 and 10; The objective function is set as the difference between the model output value and the actual monitored displacement value: In the formula, T is the temperature, and S is the displacement value at this temperature; The learning efficiency is set to 0.5 and the error margin is 10 -4 , the maximum number of iterations is 2000; S4, substituting the temperature values ​​measured by each temperature sensor into the temperature compensation neural network model determined in step S3 to calculate the corresponding displacement, and adjusting the displacement measurement data output by the displacement measurement device of the compensation concrete constraint stress equipment in real time.

2. The method for compensating the error of the concrete restraining stress device caused by the ambient temperature according to claim 1, characterized in that: When training the temperature compensation neural network, it is necessary to perform linear normalization processing on the temperature and displacement data and map them to (-1, 1) to improve the accuracy of model training.

Images