Real-time evaluation method of UV in-situ curing repair effect of drainage pipes based on machine learning

Abstract

The present invention discloses a real-time evaluation method for the UV in-situ curing repair effect of a drainage pipe based on machine learning, and the present invention relates to the technical field of repair effect evaluation. It comprises: collecting historical parameters, including surface temperature, surface humidity, curing time and UV irradiation intensity, and the corresponding hardness value of the cured material, as inputs of the prediction model for training, and establishing a curing hardness prediction model. The UV irradiation intensity during the repair process is obtained and corrected in combination with the ambient humidity. Real-time process parameters are collected, input into the model, and the real-time hardness value of the cured material is output. After the hardness value is corrected according to the uniformity of the light, a repair effect evaluation index is generated in combination with the parameters of the surface flatness, crack shape index and acoustic wave signal. By comparing with the evaluation threshold, different repair effect evaluation results are output. The method realizes intelligent real-time monitoring and evaluation of the curing repair effect of the drainage pipe, and improves the repair quality and efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of repair effect evaluation, and specifically to a real-time evaluation method for ultraviolet light in-situ curing repair effect of a drainage pipe based on machine learning. Background Art

[0002] In urban infrastructure, the maintenance and repair of drainage pipes are of vital importance. With the acceleration of urbanization, the pressure on drainage pipes is increasing. Pipe aging, corrosion and the influence of the external environment have led to frequent damage to the pipes. Traditional pipe repair methods, such as excavation and replacement, are usually time-consuming, costly, and have a great impact on the surrounding environment. Therefore, it is particularly important to develop a pipe repair technology that is efficient, economical and has less impact on the environment. In recent years, in-situ repair methods based on UV curing technology have gradually attracted attention. This method uses UV curing materials to repair the inside of the pipe, which can quickly restore the function of the pipe without damaging the surrounding environment.

[0003] However, despite the many advantages of UV curing technology, there are still some technical problems in practical application. First, environmental factors (such as temperature, humidity, etc.) and repair parameters (such as curing time, UV light intensity, etc.) during the curing process have a significant impact on the final hardness of the cured material. Accurately controlling and evaluating these parameters is essential to ensure the repair effect. Secondly, traditional methods for evaluating the hardness of cured materials often rely on manual sampling and laboratory testing, resulting in delayed and inaccurate evaluation results, making it difficult to reflect changes in the repair process in real time. Therefore, there is a lack of an effective method that can both monitor the repair process parameters in real time and evaluate the curing effect in a timely manner.

[0004] In order to solve the above technical problems, a real-time evaluation method based on machine learning has emerged. By collecting process parameters in the historical repair process, including surface temperature, humidity, curing time and UV light intensity, and combining them with the hardness value of the cured material, a rich data foundation can be provided for the machine learning model. As a deep learning model suitable for processing time series data, the long short-term memory network (LSTM) can effectively capture the nonlinear relationship in historical data and provide support for the prediction of the hardness of the cured material. In addition, by real-time monitoring of factors such as UV light intensity and ambient humidity, the accuracy of the curing process can be improved, further improving the evaluation accuracy of the repair effect.

[0005] In the prior art, the publication number CN116386789A discloses a method for real-time monitoring of the quality of ultraviolet in-situ curing repair of buried pipelines, and the specific steps are: first, collect the on-site in-situ curing data of UV-CIPP materials to construct a data set; then establish a genetic algorithm optimized support vector machine model for predicting the mechanical properties of UV-CIPP materials under the influence of multiple factors, referred to as a GA-SVM model, and train and verify the GA-SVM model through the data set to obtain the optimal GA-SVM model; secondly, according to the mechanical parameters of UV-CIPP materials required by the construction design and the optimal GA-SVM model, predict the optimal curing parameters and curing temperature; finally, input the predicted optimal curing parameters and curing temperature into the integrated control system of the buried pipeline ultraviolet in-situ curing repair trolley to start the pipeline repair construction; during the repair process, use the optimal GA-SVM model to monitor the quality of the buried pipeline ultraviolet in-situ curing repair in real time. However, in this scheme, when establishing the GA-SVM model, it may not be possible to take all factors affecting mechanical properties into consideration in the model. For example, factors such as soil type, degree of aging of pipelines, and chemical properties of the surrounding environment may have a significant impact on the results in actual repair, but they are not fully reflected in the model. At the same time, the GA-SVM evaluation model may rely too much on a single mechanical performance indicator and ignore other factors that may affect the repair effect. The environmental conditions at the construction site (such as temperature, humidity, light, etc.) may change in a short period of time. If the model fails to adapt to these dynamic changes, it may lead to inaccurate prediction results. Therefore, relying on only one evaluation model may reduce the real-time and effectiveness of the evaluation system.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The purpose of the present invention is to provide a real-time evaluation method for the UV in-situ curing repair effect of drainage pipes based on machine learning to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A real-time evaluation method for the repair effect of UV in-situ curing of drainage pipes based on machine learning, the specific steps include:

[0010] Collect the process parameters and corresponding hardness values ​​of the cured materials at different times during the UV in-situ curing repair process of the historical drainage pipes. The process parameters include the surface temperature, surface humidity, curing time and UV irradiation intensity of the drainage pipe repair site;

[0011] An LSTM prediction model is established, the process parameters obtained at different times are used as the input of the model, and the corresponding hardness value of the solidified material is used as the label. The LSTM prediction model is trained to obtain a solidification hardness prediction model with the input as the process parameter and the output as the corresponding hardness value of the solidified material.

[0012] The UV light intensity is calculated based on the output power of the UV light source during real-time repair, the distance between the UV light source and the pipeline repair location, and the effective emission area of ​​the light source. The UV light intensity is corrected by the ambient humidity to obtain the accurate UV light intensity.

[0013] Collect the real-time process parameters of the drainage pipe to be repaired during the repair process, and use the corresponding precise UV light intensity, surface temperature, surface humidity and curing time as input data, and input them into the trained curing hardness prediction model. The model outputs the real-time hardness value of the curing material.

[0014] Based on the real-time hardness value of the solidified material output by the model, the real-time hardness value of the solidified material is corrected by the uniformity of light illumination to obtain the average hardness value of the solidified material. The repair effect evaluation index is generated by combining the surface flatness, crack shape index and the resonance frequency and spectral entropy of the sound wave signal of the solidified material at the corresponding moment. The repair effect evaluation index is compared with the evaluation threshold, and different repair effect evaluation results are issued according to different comparison results.

[0015] Furthermore, a prediction model is established based on the LSTM long short-term memory network model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0016] ;

[0017] In the formula, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer;

[0018] At the same time, the hyperparameters of the LSTM model are set, and the hyperparameters of the LSTM model include: the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch processing number, and the number of hidden layer neurons;

[0019] The number of network layers is set to a four-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 64, the number of training times is set to 500, the batch size is set to 256, and the number of hidden layer neurons is 32.

[0020] Furthermore, the formula for calculating the UV irradiation intensity based on the output power of the UV light source during real-time repair, the distance between the UV light source and the pipeline repair location, and the effective emission area of ​​the light source is:

[0021] ;

[0022] In the formula, is the UV irradiation intensity, is the output power of the UV light source, is the effective emitting area of ​​the light source, is the shortest distance between the UV light source and the pipeline repair location;

[0023] The UV light intensity is corrected according to the ambient humidity to obtain the precise UV light intensity, wherein the formula for calculating the precise UV light intensity is:

[0024] ;

[0025] In the formula, To accurately measure the intensity of UV light, is the humidity attenuation coefficient, which is used to describe the attenuation effect of humidity on ultraviolet intensity. is the real-time ambient humidity, where the humidity attenuation coefficient Greater than 0.

[0026] Furthermore, based on the real-time hardness value of the solidified material output by the model, the real-time hardness value of the solidified material is corrected by the uniformity of the illumination to obtain the average hardness value of the solidified material, wherein the formula for calculating the average hardness value of the solidified material is:

[0027] ;

[0028] in, is the average hardness value of the solidified material at time t, Output real-time cured material hardness values ​​for the model, is the uniformity of illumination, where The calculation is based on the formula:

[0029] ;

[0030] In the formula, and Respectively represent the minimum and maximum values ​​of the precise UV light intensity, where the minimum and maximum values ​​of the precise UV light intensity and The angle between the incident direction of the light source and different sections of the pipe crack is calculated, and the specific formulas are:

[0031] ;

[0032] ;

[0033] In the formula, It is the angle between the incident direction of the light source and different sections of the pipe crack.

[0034] Furthermore, the logic for calculating the surface flatness at the corresponding moment is:

[0035] The logic for calculating the surface flatness at the corresponding moment is: obtain the absolute height difference between the bottom of the crack and the surface of the pipeline at the corresponding moment, and calculate the surface flatness of the pipeline based on the height difference. The formula is:

[0036] ;

[0037] In the formula, is the surface flatness of the pipeline at time t, is the crack length at time t, It represents the absolute difference between the height of the crack at the xth point and the pipe surface at time t, and x represents the position index of the crack point.

[0038] Furthermore, the method for obtaining the crack shape index is as follows: in real-time acquisition of image information of the repaired crack during the repair process, image enhancement preprocessing is performed on the acquired crack image, and crack contour is extracted by the Canny algorithm based on the enhanced crack image to obtain real-time contour feature information of the crack, and the geometric features of the crack are calculated according to the real-time contour feature information of the crack, wherein the geometric features of the crack include the area of ​​the crack and the perimeter of the crack, and the crack shape index is calculated according to the area of ​​the crack and the perimeter of the crack, wherein the specific formula for calculating the crack shape index is:

[0039] ;

[0040] In the formula, is the crack shape index at time t, is the area of ​​the crack at time t, is the crack perimeter at time t, where the crack area and crack perimeter The calculation is based on the number of crack pixels, and the specific formula is:

[0041] ;

[0042] ;

[0043] In the formula, and They represent the number of pixels in the crack area and the number of pixels on the edge of the crack contour at time t, respectively. Represents the actual area of ​​each pixel, Indicates the length of each edge pixel.

[0044] Furthermore, the specific acquisition logic of the resonance frequency and spectral entropy of the sound wave signal of the solidified material is: determining the maximum frequency of the emitted sound wave signal, and determining the sampling frequency based on the maximum frequency, wherein the specific formula for determining the sampling frequency is:

[0045] ;

[0046] In the formula, is the sampling frequency, is the maximum frequency of the emitted sound wave signal, where and is a positive integer;

[0047] By sending an acoustic wave signal to the curing material, using a sensor to obtain the acoustic wave signal transmitted by the curing material, performing a fast Fourier transform on the acoustic wave signal of the curing material, converting the time domain signal into a frequency domain signal, and obtaining the spectrum of the signal , based on the spectrum The power spectral density is calculated based on the following formula:

[0048] ;

[0049] In the formula, For frequency The power spectral density at Indicates the amplitude of the spectrum;

[0050] The resonant frequency is identified based on the power spectral density, and the specific formula is:

[0051] ;

[0052] In the formula, is the resonant frequency, represents the independent variable that makes the function reach the maximum value, that is, the power spectral density The frequency at which the maximum value is reached;

[0053] The specific formula for calculating the spectral entropy of the sound wave signal of the solidified material based on the power spectral density is:

[0054] ;

[0055] In the formula, is the spectral entropy of the acoustic wave signal of the solidifying material.

[0056] Furthermore, based on the average hardness value of the solidified material, combined with the surface flatness at the corresponding moment, the crack shape index, and the resonance frequency and spectral entropy of the solidified material sound wave signal, a repair effect evaluation index is generated, wherein the specific formula for generating the repair effect evaluation index is:

[0057] ;

[0058] In the formula, is the restoration effect evaluation index, is the average hardness value of the solidified material at time t, is the surface flatness of the pipeline at time t, is the crack shape index at time t, and are the weight indexes of the average cured material hardness value and the resonance frequency, respectively, where and and All are greater than 0;

[0059] According to the comparison between the restoration effect evaluation index and the evaluation threshold, different restoration effect evaluation results are obtained according to different comparison results. The logic is as follows:

[0060] when When the pipeline repair is judged to be in the early stage, an assessment result of weak repair effect is issued, and the repair work continues;

[0061] when When the pipeline is repaired, it is judged that the pipeline repair is in the process state, and the evaluation result of the repair effect is medium is issued, and the repair and improvement work of the pipeline cracks continues;

[0062] when When the pipeline repair is completed, it is judged that the pipeline repair is in the final state, and an evaluation result that the repair effect is excellent is issued, and the repair work is completed;

[0063] In the formula, is the evaluation threshold.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] First, the LSTM (Long Short-Term Memory Network) model is used to dynamically process time series data and analyze multiple parameters such as UV light intensity, ambient humidity, and temperature during the curing process in real time. This dynamic monitoring capability enables timely acquisition of real-time hardness predictions of curing materials, improving the reliability and consistency of the repair effect. Secondly, by calculating the UV light intensity in real time and correcting it according to the ambient humidity, the UV light conditions can be dynamically monitored and optimized during the repair process to ensure that the predicted hardness value is more accurate. Not only does it enhance the dynamic monitoring capability and improve the simulation effect of UV light intensity during the actual repair process, but real-time data analysis and UV light intensity correction can make more scientific decisions based on current environmental conditions and repair effects. Finally, the solution not only focuses on the hardness of the curing material, but also combines multi-dimensional indicators such as surface flatness, crack shape index, and acoustic wave signals to generate a comprehensive repair effect evaluation index. By integrating multiple evaluation indicators, it can provide a more comprehensive evaluation of the repair effect, making the evaluation of the repair effect more comprehensive and scientific. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a schematic diagram of the overall method flow of the present invention. DETAILED DESCRIPTION

[0067] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0068] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0069] Example:

[0070] See also Figure 1 , the present invention provides a technical solution:

[0071] A real-time evaluation method for the repair effect of UV in-situ curing of drainage pipes based on machine learning, the specific steps include:

[0072] Step 1: Collect the process parameters and corresponding hardness values ​​of the cured materials at different times during the historical UV in-situ curing repair process of the drainage pipe. The process parameters include the surface temperature, surface humidity, curing time and UV irradiation intensity of the drainage pipe repair site.

[0073] Surface temperature has a direct impact on the rate of chemical reactions. At higher temperatures, the curing reaction is generally accelerated, which may increase the hardness of the material. Temperature changes may also affect the physical properties of the cured material, such as toughness and brittleness, so monitoring temperature can help predict the performance of the material after curing.

[0074] Surface humidity affects the moisture content of the curing material, which in turn affects the curing process. Sometimes, too high humidity may lead to incomplete curing, causing internal defects in the material and reducing hardness. Changes in humidity may affect the chemical reaction mechanism of the curing agent, especially some curing agents are sensitive to moisture. Therefore, collecting humidity data can predict the curing effect.

[0075] Curing time is one of the important factors that affect the hardness of the cured material. Generally speaking, the longer the curing time, the higher the degree of cross-linking within the material, and thus the higher the hardness. By recording the hardness values ​​at different time points, the relationship between curing time and hardness can be established.

[0076] The intensity of UV light directly affects the rate of photopolymerization of the curing material. High-intensity UV light generally speeds up the curing process, allowing the material to reach the desired hardness more quickly. Uneven UV light intensity may result in different properties of the cured material in different areas, so monitoring the intensity can help predict the hardness value of the cured material.

[0077] Step 2: Establish an LSTM prediction model, use the process parameters obtained at different times as the input of the model, and use the corresponding hardness value of the solidified material as the label to train the LSTM prediction model to obtain a solidification hardness prediction model with the input as the process parameters and the output as the corresponding hardness value of the solidified material.

[0078] LSTM is specially designed to process and predict time series data. Since various parameters in the UV curing repair process of drainage pipes (such as temperature, humidity, UV light intensity, etc.) change over time, LSTM can effectively capture the temporal relationship and dynamic interaction between these parameters. This ability enables LSTM to perform well when processing time-related input data. Traditional recurrent neural networks (RNNs) face the problems of "gradient vanishing" and "gradient explosion" when processing longer sequences, making it difficult for the model to learn long-term dependencies. By introducing gating mechanisms (input gates, forget gates, and output gates), LSTM can effectively manage information flow, keep and forget past information, and thus avoid long-term dependency problems. This enables LSTM to better understand and predict complex time dependencies in the curing process.

[0079] A prediction model is established based on the LSTM long short-term memory network model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0080] ;

[0081] In the formula, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer;

[0082] At the same time, the hyperparameters of the LSTM model are set, and the hyperparameters of the LSTM model include: the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch processing number, and the number of hidden layer neurons;

[0083] The number of network layers is set to a four-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 64, the number of training times is set to 500, the batch size is set to 256, and the number of hidden layer neurons is 32.

[0084] Step 3: The UV light intensity is calculated based on the output power of the UV light source during real-time repair, the distance between the UV light source and the pipeline repair location, and the effective emission area of ​​the light source, and the UV light intensity is corrected by the ambient humidity to obtain the accurate UV light intensity.

[0085] The formula for calculating the UV irradiation intensity based on the output power of the UV light source during real-time repair, the distance between the UV light source and the pipeline repair location, and the effective emission area of ​​the light source is:

[0086] ;

[0087] In the formula, is the UV irradiation intensity, is the output power of the UV light source, is the effective emitting area of ​​the light source, is the shortest distance between the UV light source and the pipeline repair location;

[0088] The UV light intensity is corrected according to the ambient humidity to obtain the precise UV light intensity, wherein the formula for calculating the precise UV light intensity is:

[0089] ;

[0090] In the formula, To accurately measure the intensity of UV light, is the humidity attenuation coefficient, which is used to describe the attenuation effect of humidity on ultraviolet intensity. is the real-time ambient humidity, where the humidity attenuation coefficient Greater than 0.

[0091] The presence of water vapor will also affect the transmittance of ultraviolet rays. The real-time ambient humidity indicates the ratio of the actual content of water vapor in the air to the maximum content, expressed as a percentage. Changes in humidity will directly affect changes, thereby affecting the propagation of ultraviolet rays.

[0092] Humidity attenuation coefficient This coefficient is specifically used to describe the effect of humidity on UV intensity. It is used to quantify the effect of humidity on ultraviolet (UV) intensity. As the humidity of the environment increases, water vapor absorbs more UV rays, resulting in a decrease in UV intensity. The higher the humidity, the more obvious the absorption effect of water vapor on UV rays. Water vapor has absorption capacity in the UV wavelength range, especially in the short-wave ultraviolet (UV-C) and part of the medium-wave ultraviolet (UV-B) region. The higher the humidity, the greater the concentration of water vapor, which increases the absorption and scattering of UV rays, resulting in a significant attenuation of UV intensity. Humidity attenuation coefficient The specific method of obtaining is: set up a UV light source in a humidity-controlled environment, and use a light intensity meter to measure the light intensity under different humidity conditions. Under different relative humidity (for example, 0%, 25%, 50%, 75%, 100%), record the UV intensity and the UV intensity under humidity-free conditions, and perform data fitting based on the UV intensity and UV intensity data under humidity-free conditions to obtain the humidity attenuation coefficient.

[0093] Step 4: Collect the real-time process parameters of the repair process of the drainage pipe to be repaired, and use the corresponding precise UV irradiation intensity, surface temperature, surface humidity and curing time as input data, and input them into the trained curing hardness prediction model. The model outputs the real-time hardness value of the cured material;

[0094] Real-time process parameters include precise UV light intensity, surface temperature, surface humidity and curing time. Surface temperature and surface humidity can be obtained in the following ways: Surface temperature can be measured by a variety of methods and instruments. Infrared thermometer is a non-contact measurement tool that can quickly measure the temperature of the surface of an object. Its working principle is to estimate the temperature by detecting the infrared radiation emitted by the object. Thermistor (RTD) is also a common temperature sensor with high precision and stability. It has high accuracy and is suitable for long-term monitoring.

[0095] The measurement of surface humidity can generally be used: humidity sensors can directly measure the relative humidity in the air, but can also be used to measure the humidity on the surface of objects, especially for moisture monitoring during the curing process. Capacitive humidity sensors determine humidity by measuring changes in capacitance, and usually have higher accuracy and response speed. Suitable for rapidly changing environments.

[0096] Step 5: Based on the real-time hardness value of the solidified material output by the model, the real-time hardness value of the solidified material is corrected by the uniformity of light illumination to obtain the average hardness value of the solidified material. The repair effect evaluation index is generated by combining the surface flatness, crack shape index and the resonance frequency and spectral entropy of the sound wave signal of the solidified material at the corresponding moment. The repair effect evaluation index is compared with the evaluation threshold, and different repair effect evaluation results are issued according to different comparison results.

[0097] Based on the real-time hardness value of the solidified material output by the model, the real-time hardness value of the solidified material is corrected by the uniformity of the light to obtain the average hardness value of the solidified material. The formula for calculating the average hardness value of the solidified material is:

[0098] ;

[0099] in, is the average hardness value of the solidified material at time t, Output real-time cured material hardness values ​​for the model, The closer the value is to 1, the more uniform the lighting is. The calculation is based on the formula:

[0100] ;

[0101] In the formula, and Respectively represent the minimum and maximum values ​​of the precise UV light intensity, where the minimum and maximum values ​​of the precise UV light intensity and The angle between the incident direction of the light source and different sections of the pipe crack is calculated, and the specific formulas are:

[0102] ;

[0103] ;

[0104] In the formula, It is the angle between the incident direction of the light source and the different sections of the pipe crack. When the angle between the incident direction of the light source and the different sections of the pipe crack is 90 degrees, the intensity of ultraviolet light is the largest, that is, it is aimed at the pipe crack. Since the drainage pipe is generally circular, the further it deviates from the positive incident direction of the light source, the smaller the angle between the incident direction of the light source and the different sections of the pipe crack. To describe the UV intensity of the pipe crack that is not in the positive direction of the light source, the maximum and minimum precise UV intensity are found through the formula.

[0105] The logic for calculating the surface flatness at the corresponding moment is: obtain the absolute height difference between the bottom of the crack and the surface of the pipeline at the corresponding moment, and calculate the surface flatness of the pipeline based on the height difference. The formula is:

[0106] ;

[0107] In the formula, is the surface flatness of the pipeline at time t, is the crack length at time t, It represents the absolute difference between the height of the crack at the xth point and the pipe surface at time t, and x represents the position index of the crack point.

[0108] It represents the absolute difference in height between the crack at the xth point and the pipe surface at time t. It can be obtained by measuring using a surface roughness meter (such as a profilometer, laser scanner, etc.).

[0109] The method for obtaining the crack shape index is as follows: in real-time acquisition of image information of the repaired crack during the repair process, image enhancement preprocessing is performed on the collected crack image, crack contour is extracted by the Canny algorithm based on the enhanced crack image, and real-time contour feature information of the crack is obtained; geometric features of the crack are calculated according to the real-time contour feature information of the crack, wherein the geometric features of the crack include the area of ​​the crack and the perimeter of the crack; the crack shape index is calculated according to the area of ​​the crack and the perimeter of the crack, wherein the specific formula for calculating the crack shape index is:

[0110] ;

[0111] In the formula, is the crack shape index at time t, is the area of ​​the crack at time t, is the crack perimeter at time t, where the crack area and crack perimeter The calculation is based on the number of crack pixels, and the specific formula is:

[0112] ;

[0113] ;

[0114] In the formula, and They represent the number of pixels in the crack area and the number of pixels on the edge of the crack contour at time t, respectively. Represents the actual area of ​​each pixel, Represents the length of each edge pixel, where the actual area of ​​each pixel is consistent with the length of the pixel.

[0115] The specific acquisition logic of the resonance frequency and spectral entropy of the sound wave signal of the solidified material is: determine the maximum frequency of the emitted sound wave signal, and determine the sampling frequency based on the maximum frequency, wherein the specific formula for determining the sampling frequency is:

[0116] ;

[0117] In the formula, is the sampling frequency, is the maximum frequency of the emitted sound wave signal, where and is a positive integer;

[0118] By sending an acoustic wave signal to the curing material, using a sensor to obtain the acoustic wave signal transmitted by the curing material, performing a fast Fourier transform on the acoustic wave signal of the curing material, converting the time domain signal into a frequency domain signal, and obtaining the spectrum of the signal By analyzing the spectrum, we can find the frequency corresponding to the maximum value in the amplitude spectrum, which is the maximum frequency in the sound wave signal. Based on spectrum The power spectral density is calculated based on the following formula:

[0119] ;

[0120] In the formula, For frequency The power spectral density at Indicates the amplitude of the spectrum;

[0121] The resonant frequency is identified based on the power spectral density, and the specific formula is:

[0122] ;

[0123] In the formula, is the resonant frequency, represents the independent variable that makes the function reach the maximum value, that is, the power spectral density The frequency at which the maximum value is reached;

[0124] The specific formula for calculating the spectral entropy of the sound wave signal of the solidified material based on the power spectral density is:

[0125] ;

[0126] In the formula, is the spectral entropy of the acoustic wave signal of the solidifying material.

[0127] Based on the average hardness value of the solidified material, combined with the surface flatness, crack shape index and the resonance frequency and spectral entropy of the solidified material sound wave signal at the corresponding moment, the repair effect evaluation index is generated. The specific formula for generating the repair effect evaluation index is:

[0128] ;

[0129] In the formula, is the restoration effect evaluation index, is the average hardness value of the solidified material at time t, is the surface flatness of the pipeline at time t, is the crack shape index at time t, and are the average cured material hardness value and the weight index of the resonance frequency respectively.

[0130] Among them, higher hardness usually means that the material is stronger and can better withstand external pressure. The larger the value, the better the repair result. The larger the value, the better the repair effect. Therefore, the average hardness value of the cured material is closely related to the repair effect evaluation index. is a positive correlation. And the exponential function can reflect the gradual enhancement of the curing strength on the repair effect. For example, when When it reaches a certain value, its contribution to the repair effect will increase significantly, while below this value, its contribution may be relatively small.

[0131] Pipe surface flatness Surface flatness affects the contact and adhesion of materials. Better flatness means better repair effect, so surface flatness is closely related to the evaluation index of repair effect. The square root transformation can reduce the effect of surface flatness on the model, especially at high flatness values. At the same time, this transformation can increase the sensitivity of the model to small flatness values, reflecting that even small improvements can have a significant impact on the overall restoration effect.

[0132] Crack shape index The crack shape index affects the propagation characteristics of the crack and the difficulty of repair. Generally speaking, a lower crack shape index means that the crack is easier to repair. At the same time, as the repair progresses, the area and perimeter of the crack continue to decrease. The smaller the crack, the better the repair result. Therefore, the crack shape index is closely related to the repair effect evaluation index. is a negative correlation. Natural logarithmic transformation This transformation can reduce the impact of extreme values ​​(especially high fracture shape index values) on the model and make the data more concentrated, making it easier for the model to capture general trends. In addition, fracture shapes usually show exponential growth characteristics, so the use of logarithmic transformation can better handle this growth characteristic.

[0133] During the curing process, the acoustic signal will have a specific resonant frequency, which is closely related to the physical properties of the material. Fully cured materials usually show a higher resonant frequency because the elastic modulus of the material increases, so the resonant frequency is closely related to the repair effect evaluation index. It is a positive correlation.

[0134] The main frequency is the frequency with the highest energy in the spectrum of the acoustic signal, which usually represents the main vibration characteristics of the material. In the cured material, the main frequency will change due to the uniformity of the material and the increase of the overall stiffness, showing a higher main frequency and energy concentration. The concentration of the main frequency is described by spectral entropy. Spectral entropy is used to quantify the complexity and concentration of the signal. The lower the spectral entropy, the more concentrated the frequency components. Therefore, the spectral entropy repair effect evaluation index The spectral entropy is a measure of signal complexity. Generally, higher spectral entropy values ​​indicate increased signal complexity. Through exponential decay, high complexity can be mapped to low values, which can make the effect of spectral entropy on the repair effect smoother and also emphasize the positive effect of low complexity on the repair effect.

[0135] and are the weight indices of the average hardness of the solidified material and the resonance frequency, respectively. Since the influence of the average hardness of the solidified material on the repair result is greater than the resonance frequency, and and All are greater than 0;

[0136] According to the comparison between the restoration effect evaluation index and the evaluation threshold, different restoration effect evaluation results are obtained according to different comparison results. The logic is as follows:

[0137] when When the pipeline repair is judged to be in the early stage, an assessment result of weak repair effect is issued, and the repair work continues;

[0138] when When the pipeline is repaired, it is judged that the pipeline repair is in the process state, and the evaluation result of the repair effect is medium is issued, and the repair and improvement work of the pipeline cracks continues;

[0139] when When the pipeline repair is completed, it is judged that the pipeline repair is in the final state, and an evaluation result that the repair effect is excellent is issued, and the repair work is completed;

[0140] In the formula, is the evaluation threshold. The specific formula is:

[0141] ;

[0142] In the formula, is the evaluation threshold pre-set based on expert experience. is the pressure loss in the drainage pipe, where The calculation is based on the formula:

[0143] ;

[0144] In the formula, and are the inlet and outlet pressures, respectively. A smaller pressure loss value usually indicates a good repair effect. Therefore, during the repair process, the evaluation threshold It can be dynamically adjusted according to the pressure loss in the drainage pipe.

[0145] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0146] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0147] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0148] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A real-time evaluation method for the repair effect of UV-curing in situ of drainage pipes based on machine learning, characterized in that: The specific steps include: Collect the process parameters and corresponding hardness values ​​of the cured materials at different times during the UV in-situ curing repair process of the historical drainage pipes. The process parameters include the surface temperature, surface humidity, curing time and UV irradiation intensity of the drainage pipe repair site; An LSTM prediction model is established, the process parameters obtained at different times are used as the input of the model, and the corresponding hardness value of the solidified material is used as the label. The LSTM prediction model is trained to obtain a solidification hardness prediction model with the input as the process parameter and the output as the corresponding hardness value of the solidified material. The UV light intensity is calculated based on the output power of the UV light source during real-time repair, the distance between the UV light source and the pipeline repair location, and the effective emission area of ​​the light source. The UV light intensity is corrected by the ambient humidity to obtain the accurate UV light intensity. Collect the real-time process parameters of the drainage pipe to be repaired during the repair process, and use the corresponding precise UV light intensity, surface temperature, surface humidity and curing time as input data, and input them into the trained curing hardness prediction model. The model outputs the real-time hardness value of the curing material. Based on the real-time hardness value of the solidified material output by the model, the real-time hardness value of the solidified material is corrected by the uniformity of illumination to obtain the average hardness value of the solidified material. The repair effect evaluation index is generated by combining the surface flatness, crack shape index and the resonance frequency and spectral entropy of the sound wave signal of the solidified material at the corresponding moment. The repair effect evaluation index is compared with the evaluation threshold, and different repair effect evaluation results are issued according to different comparison results. Based on the real-time hardness value of the solidified material output by the model, the real-time hardness value of the solidified material is corrected by the uniformity of the light to obtain the average hardness value of the solidified material. The formula for calculating the average hardness value of the solidified material is: in, is the average hardness value of the solidified material at time t, Output real-time cured material hardness values ​​for the model, is the uniformity of illumination, where The calculation is based on the formula: In the formula, and Respectively represent the minimum and maximum values ​​of the precise UV light intensity, where the minimum and maximum values ​​of the precise UV light intensity and The angle between the incident direction of the light source and different sections of the pipe crack is calculated, and the specific formulas are: In the formula, is the angle between the incident direction of the light source and different sections of the pipe crack; The method for obtaining the crack shape index is as follows: in real-time acquisition of image information of the repaired crack during the repair process, image enhancement preprocessing is performed on the collected crack image, crack contour is extracted by the Canny algorithm based on the enhanced crack image, and real-time contour feature information of the crack is obtained; geometric features of the crack are calculated according to the real-time contour feature information of the crack, wherein the geometric features of the crack include the area of ​​the crack and the perimeter of the crack; the crack shape index is calculated according to the area of ​​the crack and the perimeter of the crack, wherein the specific formula for calculating the crack shape index is: In the formula, is the crack shape index at time t, is the area of ​​the crack at time t, is the crack perimeter at time t, where the crack area and crack perimeter The calculation is based on the number of crack pixels, and the specific formula is: In the formula, and They represent the number of pixels in the crack area and the number of pixels on the edge of the crack contour at time t, respectively. Represents the actual area of ​​the pixel, Indicates the length of edge pixels.

2. According to the method of claim 1, which is based on machine learning and is characterized by: A prediction model is established based on the LSTM long short-term memory network model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is: In the formula, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer; At the same time, the hyperparameters of the LSTM model are set, and the hyperparameters of the LSTM model include: the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch processing number, and the number of hidden layer neurons; The number of network layers is set to a four-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 64, the number of training times is set to 500, the batch size is set to 256, and the number of hidden layer neurons is 32.

3. The real-time evaluation method for the repair effect of UV-in-situ curing of drainage pipes based on machine learning according to claim 1 is characterized in that: The formula for calculating the UV irradiation intensity based on the output power of the UV light source during real-time repair, the distance between the UV light source and the pipeline repair location, and the effective emission area of ​​the light source is: In the formula, is the UV irradiation intensity, is the output power of the UV light source, is the effective emitting area of ​​the light source, is the shortest distance between the UV light source and the pipeline repair location; The UV light intensity is corrected according to the ambient humidity to obtain the precise UV light intensity, wherein the formula for calculating the precise UV light intensity is: In the formula, To accurately measure the intensity of UV light, is the humidity attenuation coefficient, which is used to describe the attenuation effect of humidity on ultraviolet intensity. is the real-time ambient humidity, where the humidity attenuation coefficient Greater than 0.

4. The method for real-time evaluation of the repair effect of UV-in-situ curing of drainage pipes based on machine learning according to claim 3 is characterized in that: The logic for calculating the surface flatness at the corresponding moment is: obtain the absolute height difference between the bottom of the crack and the surface of the pipeline at the corresponding moment, and calculate the surface flatness of the pipeline based on the height difference. The formula is: In the formula, is the surface flatness of the pipeline at time t, is the crack length at time t, It represents the absolute difference between the height of the crack at the xth point and the pipe surface at time t, and x represents the position index of the crack point.

5. The real-time evaluation method for UV in-situ curing repair effect of drainage pipes based on machine learning according to claim 1 is characterized in that: The specific acquisition logic of the resonance frequency and spectral entropy of the sound wave signal of the solidified material is: determine the maximum frequency of the emitted sound wave signal, and determine the sampling frequency based on the maximum frequency, wherein the specific formula for determining the sampling frequency is: In the formula, is the sampling frequency, is the maximum frequency of the emitted sound wave signal, where and is a positive integer; By sending an acoustic wave signal to the curing material, using a sensor to obtain the acoustic wave signal transmitted by the curing material, performing a fast Fourier transform on the acoustic wave signal of the curing material, converting the time domain signal into a frequency domain signal, and obtaining the spectrum of the signal , based on the spectrum The power spectral density is calculated based on the following formula: In the formula, For frequency The power spectral density at Indicates the amplitude of the spectrum; The resonant frequency is identified based on the power spectral density, and the specific formula is: In the formula, is the resonant frequency, represents the independent variable that makes the function reach the maximum value, that is, the power spectral density The frequency at which the maximum value is reached; The specific formula for calculating the spectral entropy of the sound wave signal of the solidified material based on the power spectral density is: In the formula, is the spectral entropy of the acoustic wave signal of the solidifying material.

6. The real-time evaluation method for the repair effect of UV-in-situ curing of drainage pipes based on machine learning according to claim 5 is characterized by: Based on the average hardness value of the solidified material, combined with the surface flatness, crack shape index and the resonance frequency and spectral entropy of the solidified material sound wave signal at the corresponding moment, the repair effect evaluation index is generated. The specific formula for generating the repair effect evaluation index is: In the formula, is the restoration effect evaluation index, is the average hardness value of the solidified material at time t, is the surface flatness of the pipeline at time t, is the crack shape index at time t, and are the weight indexes of the average cured material hardness value and the resonance frequency, respectively, where and and All are greater than 0; According to the comparison between the restoration effect evaluation index and the evaluation threshold, different restoration effect evaluation results are obtained according to different comparison results. The logic is as follows: when When the pipeline repair is judged to be in the early stage, an assessment result of weak repair effect is issued, and the repair work continues; when When the pipeline is repaired, it is judged that the pipeline repair is in the process state, and the evaluation result of the repair effect is medium is issued, and the repair and improvement work of the pipeline cracks continues; when When the pipeline repair is completed, it is judged that the pipeline repair is in the final state, and an evaluation result that the repair effect is excellent is issued, and the repair work is completed; In the formula, is the evaluation threshold.

Images