A method, medium and system for detecting construction quality of assembled internal partition walls

Abstract

The present invention provides a method, medium and system for detecting the construction quality of an assembled inner partition wall, which belongs to the technical field of building quality assessment, and comprises the following steps: collecting ultrasonic signals of three random points of each wallboard on the assembled inner partition wall, calculating the ultrasonic signal of the geometric center of each wallboard, and recording it as a wallboard receiving signal; establishing a wallboard vector; using Fourier transform to split the wallboard receiving signal of each wallboard to obtain a first receiving signal and a second receiving signal; calculating the attenuation of the ultrasonic signal between the wallboards, and recording it as a first attenuation set; establishing a finite element model of the assembled inner partition wall, and calculating in the finite element model, calculating the attenuation range of the ultrasonic signal between the wallboards, and recording it as a second attenuation range set; using a pre-trained signal attenuation-construction quality model, inputting the first attenuation set and the second attenuation range set, and obtaining a construction quality assessment vector of the assembled inner partition wall; and outputting the construction quality assessment vector of the assembled inner partition wall to an inspector.

Description

Technical Field

[0001] The present invention belongs to the technical field of building quality assessment, and in particular, relates to a method, medium and system for detecting the construction quality of assembled internal partition walls. Background Art

[0002] As the interior partition walls of buildings develop towards prefabricated ones, prefabricated interior partition walls are widely used due to their advantages such as fast construction and high material utilization rate. However, the overall stiffness of the wall is low, and the connection quality between the wall panels is prone to problems, which affects the overall construction quality of the prefabricated interior partition walls. Evaluating the construction quality of prefabricated interior partition walls is an important part of ensuring the quality of the project. Traditional methods for evaluating the construction quality of prefabricated interior partition walls mainly use manual observation and detection. Typical detection methods include: measuring the dimensional error of the wall body, checking the connection parts of the wall panels, and detecting hollowing of the wall panels by tapping. These methods have limited detection effects and cannot fully reflect the overall quality. Summary of the invention

[0003] In view of this, the present invention provides a method, medium and system for detecting the construction quality of prefabricated interior partition walls, which can solve the technical problem that the prior art has limited detection effect and cannot fully reflect the overall quality when detecting the construction quality of prefabricated interior partition walls.

[0004] The present invention is achieved in that:

[0005] A first aspect of the present invention provides a method for detecting the construction quality of an assembled interior partition wall, comprising the following steps:

[0006] S10, collecting ultrasonic signals of three random points of each wallboard on the assembled inner partition wall, the ultrasonic signals are continuously emitted by two ultrasonic transmitters arranged at diagonally opposite corners of the inner partition wall, and the ultrasonic signals emitted by the two ultrasonic transmitters have a frequency difference of more than 10 times; the ultrasonic signals are collected by an ultrasonic sensor, and the ultrasonic sensor is used to collect ultrasonic signals of three random points of each wallboard;

[0007] S20, calculating the ultrasonic signal of the geometric center of each wall panel according to the ultrasonic signals of three random points of each wall panel, and recording it as the wall panel receiving signal;

[0008] S30, establishing a wall panel vector, including a wall panel receiving signal, a straight-line distance between a wall panel geometric center and each ultrasonic transmitter, and a number of wall panel joints between the wall panel geometric center and each ultrasonic transmitter;

[0009] S40, using Fourier transform to split the wall panel receiving signal of each wall panel to obtain a first receiving signal and a second receiving signal;

[0010] S50, calculating the attenuation of ultrasonic signals between the wall panels, recorded as a first attenuation set;

[0011] S60, establishing a finite element model of the assembled interior partition wall, and calculating the attenuation range of the ultrasonic signal between the wall panels in the finite element model, which is recorded as a second attenuation range set;

[0012] S70, using a pre-trained signal attenuation-construction quality model, inputting the first attenuation degree set and the second attenuation range set, to obtain a construction quality assessment vector for the prefabricated interior partition wall;

[0013] S80, outputting the construction quality assessment vector of the prefabricated interior partition wall to inspection personnel.

[0014] On the basis of the above technical solution, the method for detecting the construction quality of an assembled internal partition wall of the present invention can also be improved as follows:

[0015] Among them, the step of calculating the ultrasonic signal of the geometric center of each wall panel based on the ultrasonic signals of three random points of each wall panel is specifically: calculating the coordinates of the geometric center of each wall panel; extracting the ultrasonic signals of three sampling points of each wall panel; constructing a distance vector from the sampling point to the geometric center; performing weighted averaging of the signals according to the distance vector to calculate the ultrasonic signal of the geometric center of the wall panel.

[0016] Among them, the signal attenuation-construction quality model training step includes: collecting a large amount of case data of interior partition walls with different construction qualities; setting the neural network model structure; determining the model training parameters; and training the model using the case data; wherein the case data is an interior partition wall construction quality assessment vector obtained by experimental detection and the corresponding first attenuation degree set and second attenuation range set obtained by using steps S10 to S60.

[0017] The step of calculating the attenuation of ultrasonic signals between wall panels specifically comprises: normalizing the two received signals to obtain energy; calculating the logarithmic energy attenuation rate of the signals between the wall panels; and obtaining a set of attenuation rates of all wall panel signals as a first attenuation set.

[0018] Among them, the steps of establishing a finite element model of the prefabricated interior partition wall and calculating the attenuation range of the ultrasonic signal between the wall panels in the finite element model are specifically: establishing a finite element model of the interior partition wall; setting material parameters; setting ultrasonic sources at both ends; simulating and calculating the signal attenuation between the wall panels; and obtaining the attenuation range set between the wall panels as the second attenuation range set.

[0019] The ultrasonic frequencies emitted by the two ultrasonic transmitters are 2 MHz and 20 MHz respectively.

[0020] Furthermore, the neural network model is a multi-layer feedforward neural network.

[0021] Furthermore, the number of samples used for model training is greater than 50.

[0022] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are executed, they are used to execute the above-mentioned method for detecting the construction quality of an assembled internal partition wall.

[0023] A third aspect of the present invention provides a system for detecting the construction quality of an assembled interior partition wall, which comprises the above-mentioned computer-readable storage medium.

[0024] Compared with the prior art, the beneficial effects of the method, medium and system for detecting the construction quality of assembled internal partition walls provided by the present invention are:

[0025] 1. A simplified dual-transmitter and single-receiver ultrasonic arrangement is adopted. Only two ultrasonic transmitters need to be set up. By continuously transmitting ultrasonic signals of different frequencies and using a single sensor to collect signals, the detection system is simplified and the implementation difficulty is reduced.

[0026] 2. A method is proposed to use the synthetic signal at the geometric center of the wall panel to represent the receiving characteristics of the entire wall panel, and the accuracy of signal acquisition is improved through the distance weighted average algorithm.

[0027] 3. A wall panel vector including ultrasonic signals, wall panel distance and number of joints was established, which enhanced the signal feature information and facilitated attenuation analysis.

[0028] 4. Using a neural network quality assessment model based on a large number of case training, intelligent mapping from ultrasonic signal attenuation data to internal partition wall quality grade is achieved, improving the accuracy of the assessment.

[0029] In summary, the technical solution of the present invention solves the technical problem that the prior art has limited detection effect and cannot fully reflect the overall quality when inspecting the construction quality of prefabricated internal partition walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0031] Figure 1 A flow chart of the method provided by the present invention; DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0033] like Figure 1 FIG. 1 is a flowchart of a method for detecting the construction quality of an assembled interior partition wall according to a first aspect of the present invention. The method comprises the following steps:

[0034] S10, collecting ultrasonic signals from three random points of each wall panel on the assembled inner partition wall, the ultrasonic signals are continuously emitted by two ultrasonic transmitters arranged at diagonally opposite corners of the inner partition wall, and the ultrasonic signals emitted by the two ultrasonic transmitters have a frequency difference of more than 10 times; the ultrasonic signals are collected by using an ultrasonic sensor, and the ultrasonic sensor is used to collect ultrasonic signals from three random points of each wall panel;

[0035] S20, calculating the ultrasonic signal of the geometric center of each wall panel according to the ultrasonic signals of three random points of each wall panel, and recording it as the wall panel receiving signal;

[0036] S30, establishing a wall panel vector, including a wall panel receiving signal, a straight-line distance between a wall panel geometric center and each ultrasonic transmitter, and a number of wall panel joints between the wall panel geometric center and each ultrasonic transmitter;

[0037] S40, using Fourier transform to split the wall panel receiving signal of each wall panel to obtain a first receiving signal and a second receiving signal;

[0038] S50, calculating the attenuation of ultrasonic signals between the wall panels, recorded as a first attenuation set;

[0039] S60, establishing a finite element model of the assembled interior partition wall, and calculating the attenuation range of the ultrasonic signal between the wall panels in the finite element model, which is recorded as a second attenuation range set;

[0040] S70, using a pre-trained signal attenuation-construction quality model, inputting a first attenuation degree set and a second attenuation range set, to obtain a construction quality assessment vector for the assembled interior partition wall;

[0041] S80. Output the construction quality assessment vector of the prefabricated interior partition wall to the inspectors.

[0042] The specific implementation methods of the above steps are described below:

[0043] The specific implementation of step S10 is:

[0044] First, determine the specific location of the assembled inner partition wall to be tested. The assembled inner partition wall is a rectangular structure with known length, width and height. Then, install two ultrasonic transmitters A and B on the diagonal lines of the inner partition wall. Their frequencies differ by more than 10 times, for example, the frequency of A is 2MHz and the frequency of B is 20MHz. The installation position of the two sensors is accurate to 0.1mm.

[0045] Before starting the test, the signal strength of the two ultrasonic transmitters needs to be calibrated. The method is: let the two transmitters A and B emit ultrasonic waves at the same time, and use standardized ultrasonic detection equipment to measure the strength of the two signals. If the difference between the two signal strengths exceeds 3dB, the transmitter's transmission power needs to be adjusted until the difference between the two signal strengths is within 3dB. The purpose of this step is to ensure that the strength of the two signals is close to facilitate subsequent analysis and processing.

[0046] Then, randomly select the wallboard to be tested on the inner partition wall, and randomly select 3 test points from the wallboard according to the simple random sampling method in statistics. When the two transmitters A and B transmit signals at the same time, use an ultrasonic detection sensor with a filter to collect ultrasonic signals at each test point. The filter will filter out signals with a frequency difference of more than 10 times. Each acquisition time is 0.5 seconds, and the sampling frequency is 10kHz. Repeat the signal acquisition 10 times for each test point, and take the average value as the test value of the point to reduce random errors.

[0047] The purpose of this step is to collect ultrasonic signals from several random points on the wall panel to provide raw data for subsequent analysis and calculation. Randomly selecting multiple points can improve the representativeness of the results, and repeated collection and averaging can reduce the impact of noise. Two ultrasonic transmitters with different frequencies can distinguish different ultrasonic signals.

[0048] The specific implementation of step S20 is:

[0049] Assume the number of wall panels is , wall panel number is .

[0050] Each wall panel The coordinates of the three random sampling points are .

[0051] Wall Panel The geometric center coordinates of The calculation formula is:

[0052]

[0053] Sampling point Euclidean distance to the geometric center The calculation formula is:

[0054]

[0055] Wall Panels The ultrasonic signals at the three sampling points are marked as .

[0056] Wall Panel Synthetic ultrasonic signal at the geometric center The calculation formula is:

[0057]

[0058] By repeating the above calculation, the expected signal values ​​in the two transmitter directions can be obtained respectively, and the average value of the two expected signal values ​​can be used to represent the ultrasonic signal intensity at the geometric center of the wall panel.

[0059] The purpose of this step is to estimate the ultrasonic signal at the geometric center point of the wall panel by extrapolation, laying the foundation for subsequent analysis and calculation.

[0060] The specific implementation of step S30 is:

[0061] The ultrasonic signal of the geometric center point of the wall panel obtained in step S20 is set as the wall panel receiving signal, which is recorded as A0.

[0062] According to the structural diagram of the inner partition wall, measure the straight-line distance from the geometric center of the wall panel to the two ultrasonic transmitters, and record them as L1 and L2 respectively. The accuracy is 0.1mm.

[0063] The number of wall panel joints in the direction connecting the geometric center point of the wall panel to the two transmitters is counted and recorded as N1 and N2 respectively.

[0064] The above data are organized into a vector [A0, L1, L2, N1, N2] as the feature vector of the wall panel.

[0065] This step extracts the position relationship characteristics of the wall panels and establishes a wall panel vector, laying the foundation for the subsequent analysis of the ultrasonic attenuation characteristics between the wall panels based on the signal propagation model.

[0066] The specific implementation of step S40 is:

[0067] The wall panel receiving signal A0 obtained in step S20 is subjected to Fourier transform to decompose it into signals of different frequency components.

[0068] According to the frequency characteristics of the two ultrasonic transmitters in step S10, it can be determined which frequency components in signal A0 correspond to the signal of transmitter A, denoted as A01; and which frequency components correspond to the signal of transmitter B, denoted as A02.

[0069] In order to improve the resolution accuracy, a filter can be used to remove the frequency components in A01 and A02 that deviate from the corresponding transmitter center frequency by more than 5%.

[0070] In this way, the ultrasonic signals A01 and A02 corresponding to the two transmitters at the geometric center of the wall panel are obtained, which can represent the receiving characteristics of the wall panel for the two transmitter signals.

[0071] The purpose of this step is to decompose the two frequency signals using signal processing technology, laying the foundation for the subsequent analysis of the ultrasonic propagation characteristics in the directions of the two transmitters.

[0072] For the specific implementation of step S50, the following sub-steps are taken:

[0073] First, according to step S40, two reception signals of all wall panels are obtained.

[0074] Then, the received signal is normalized to obtain the signal energy; specifically: Perform the two-norm normalization to obtain the energy :

[0075] ;

[0076] Next, calculate the logarithmic energy attenuation rate of the signal from transmitter A to wall panel B. Specifically, calculate the wall panel The logarithmic energy decay rate :

[0077] .

[0078] Afterwards, the above calculation is repeated to obtain the logarithmic energy attenuation rate of all wall panels i.

[0079] Finally, the signal attenuation rate set of all wall panels is taken as the first attenuation degree set.

[0080] By implementing this step, the attenuation rate of the ultrasonic signal between the wall panels is calculated, and the acquisition of the first attenuation degree set is completed.

[0081] The purpose of step S50 is to calculate the attenuation of the ultrasonic signal between the wall panels. This step is based on the comparison of the received signal strengths of the two transmitters and combined with the number of wall panel joints, which can effectively evaluate the propagation attenuation of the ultrasonic wave between the wall panels.

[0082] For the specific implementation of step S60, the following sub-steps are taken:

[0083] First, according to the layout design of the interior partition wall, a three-dimensional finite element model of the prefabricated interior partition wall is established. The model includes structural details such as wall panels and wall interfaces.

[0084] Then, set the material parameters of the finite element model, such as the density, elastic modulus, Poisson's ratio, etc. of the wall panels and interfaces. You can refer to the empirical parameters of the actual materials.

[0085] Next, ultrasonic emission sources are set at two diagonal positions of the model, with emission frequencies of fA and fB respectively.

[0086] Afterwards, finite element simulation was carried out to calculate the propagation and attenuation of ultrasonic waves in the model, and the signal attenuation results between every two wall panels were recorded.

[0087] Finally, the signal attenuation range results between the wall panels are extracted to construct the second attenuation range set. The results can be depicted as an attenuation matrix between the wall panels.

[0088] Specifically, this step includes:

[0089] 1) Establish a three-dimensional finite element model

[0090] Set the length of the inner partition wall ,Width ,high The model is constructed using four-node triangular acoustic units and hexahedral structural units. The total number of units is:

[0091]

[0092] in is the number of acoustic units, is the number of structural units.

[0093] 2) Set material parameters

[0094] Density of the wall material , elastic modulus Poisson's ratio .

[0095] Density of the interface material , elastic modulus , Poisson's ratio .

[0096] 3) Set the sound source

[0097] Point sound sources are set on the two diagonals of the model, with frequencies and .

[0098] The sound source intensities are and .

[0099] 4) Joint finite element analysis

[0100] Acoustic simulation uses the Helmholtz equation:

[0101]

[0102] Structural dynamics equations:

[0103]

[0104] middle, is the mass matrix, is the damping matrix, is the stiffness matrix.

[0105] Acoustic field obtained by acoustic junction coupling iteration .

[0106] 5) Extract results

[0107] Recording of sound pressure levels at wall panel interfaces .

[0108] Analyze the sound pressure attenuation coefficient between every two wall panels .

[0109]

[0110] Generate sound pressure attenuation coefficient matrix .

[0111] By implementing this step, the finite element model is used to simulate and obtain the signal propagation attenuation range data between the wall panels.

[0112] The purpose of step S60 is to obtain a second attenuation range set by establishing a finite element model to simulate and calculate the ultrasonic attenuation between the wall panels. This step provides theoretical model analysis support.

[0113] For the specific implementation of step S70, the following sub-steps are taken:

[0114] First, the first attenuation degree set obtained in step S50 is extracted.

[0115] Then, the second attenuation range set obtained in step S60 is extracted, for example, an attenuation matrix between wall panels.

[0116] Next, the above two sets of data are input into the pre-trained signal attenuation-construction quality model.

[0117] After that, the model accepts input and outputs the construction quality evaluation result vector of the prefabricated interior partition wall through transfer function calculation. The vector elements represent the quality grade of the interior partition wall, such as A, B, C, and D.

[0118] Finally, the quality assessment result vector is used as the construction quality assessment result of the prefabricated interior partition wall.

[0119] Through the implementation of this step, the pre-trained model is used to evaluate the construction quality of the internal partition walls and complete automated quality inspection.

[0120] The purpose of step S70 is to evaluate the construction quality of the interior partition wall using an intelligent model based on the signal attenuation data. This step realizes the intelligent transformation from signal analysis to quality evaluation.

[0121] The establishment and training of the signal attenuation-construction quality model takes the following steps:

[0122] First, actual cases of prefabricated interior partition walls with different construction quality levels are collected, and their ultrasonic signal attenuation data and quality assessment results are recorded. The number of cases is no less than 50, and the quality levels are set to four levels: A, B, C, and D. Data collection is obtained using a test method similar to steps S10-S40.

[0123] Then, the network structure of the neural network model is set, including the number of input layer, hidden layer and output layer units. In this embodiment, the number of input layer units is the number of wallboards n, the hidden layer is set to 10-100 units, and the number of output layer units is the quality level number 3.

[0124] Next, determine the training parameters of the network model, including the number of training times, batch size, learning rate, etc. Among them, the number of training times can be set to 1000 times, the batch size is set to 5-10, and the initial value of the learning rate is set to 0.01.

[0125] Afterwards, the collected wall panel signal attenuation data is used as input, and the corresponding quality assessment results are used as the expected output to train the established neural network model. The back propagation algorithm is used to update the network weight parameters. Cross-validation is set to prevent overfitting.

[0126] Finally, the training effect of the model is evaluated. The prediction accuracy of the model for the validation set is calculated. If the accuracy meets the requirement, for example, above 80%, the training is terminated to obtain the final signal attenuation-construction quality model.

[0127] Through the implementation of this step, a neural network was used to establish an evaluation model from ultrasonic signal attenuation data to the quality of interior partition wall construction. This model was obtained through a large number of case training and can realize automated quality evaluation.

[0128] The embodiment of the present invention takes the assembled inner partition wall of an office building as an example to carry out construction quality inspection. The inner partition wall is located on the first floor of the office building. It is a rectangular structure, 12 meters long, 8 centimeters thick, and 3 meters high. It is composed of 24 wall panels, each of which has a size of 1.2 meters × 3 meters and is made of gypsum board. The ambient temperature of the inspection site is 20°C and the humidity is 50%RH.

[0129] 1. Equipment layout

[0130] Select two ultrasonic transmitters with working frequencies of 30kHz and 300kHz as transmitters A and B, and install them at the intersection of the two diagonals of the inner partition wall to ensure that the ultrasonic wave can cover the entire range of the inner partition wall. Select a broadband ultrasonic sensor with a center frequency of 100kHz and a bandwidth of 5kHz-200kHz as the receiving sensor and install it in the center of the inner partition wall.

[0131] 2. Sampling point selection

[0132] Number the wall panels from 1 to 24 and randomly select 3 sampling points on each wall panel, for a total of 72 sampling points. Record the coordinates of each sampling point. For example, the coordinates of the 3 sampling points of wall panel 1 are (1.1 m, 0.8 m), (1.3 m, 1.5 m), and (0.9 m, 2.2 m).

[0133] 3. Signal acquisition

[0134] Turn on the ultrasonic transmitter and control it to continuously emit ultrasonic signals. At each sampling point, use the receiving sensor to collect 0.2 seconds of ultrasonic signals, repeat 5 times, and take the signal average as the ultrasonic signal of the sampling point. Complete signal collection for all 72 sampling points.

[0135] 4. Calculation of geometric center signal

[0136] For each wall panel, the coordinates of the geometric center of the wall panel are calculated based on the coordinates of its three sampling points. Then, the synthetic ultrasonic signal of the geometric center is calculated based on the distance weighted average algorithm. For example, the coordinates of the geometric center of wall panel 1 are (1.1 meters, 1.5 meters), and the Euclidean distances from the three sampling points to the geometric center are calculated:

[0137] ;

[0138] ;

[0139] ;

[0140] The calculation formula of the synthetic signal is:

[0141] ;

[0142] 5. Wall panel vector creation

[0143] Record the composite signal of all wall panels. Measure the distance from the geometric center of each wall panel to the two transmitters and count the number of wall panel joints on the path. Take wall panel 1 as an example, its vector is [S1, 2.5m, 3, 3.2m, 2].

[0144] 6. Signal decomposition

[0145] Perform a fast Fourier transform on the composite signal of each wall panel, set the frequency range to 20kHz to 400kHz, extract the components with a frequency range of ±5% around 294kHz and ±5% around 30kHz, and use them as the signals of the two transmitters, recorded as S1A and S1B.

[0146] 7. Signal attenuation calculation

[0147] Perform two-norm normalization on signals S1A and S1B to obtain energy values. Calculate the logarithmic energy attenuation rate of wall panel 1:

[0148] γ1 = ln(E1A / E1B) / (3 + 2)

[0149] Similar processing is performed on all wall panels to obtain a set of energy decay rates.

[0150] 8. Simulation calculation

[0151] Establish a three-dimensional finite element model of the interior partition wall. Set the wall panel material parameters: density 1200kg / m3, Young's modulus 3.5GPa, Poisson's ratio 0.25. Set sound sources with frequencies of 30kHz and 300kHz at the two diagonal positions of the model to simulate and calculate the sound pressure distribution. Extract the sound pressure attenuation matrix between the wall panels.

[0152] 9. Quality Assessment

[0153] The energy attenuation rate set and the simulated attenuation matrix are input into the pre-trained quality assessment neural network model. The model output results give the overall quality grade of the interior partition wall as "A".

[0154] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are executed, they are used to execute the above-mentioned method for detecting the construction quality of an assembled internal partition wall.

[0155] A third aspect of the present invention provides a system for detecting the construction quality of an assembled interior partition wall, which comprises the above-mentioned computer-readable storage medium.

[0156] Specifically, the principle of the present invention is:

[0157] 1) Different construction quality will lead to loose connections between wall panels, thus affecting the propagation and attenuation characteristics of ultrasonic waves through the wall panels.

[0158] 2) By analyzing the attenuation of ultrasonic signals between wall panels, the tightness of the wall panel connection can be evaluated, reflecting the overall construction quality of the interior partition wall.

[0159] 3) The dual-transmitter and single-receiver ultrasonic arrangement can obtain the attenuation characteristics of the wall panel interface from two directions, thereby improving reliability.

[0160] 4) The composite signal at the geometric center of the wall panel can effectively represent the ultrasonic reception of the entire wall panel. The distance and number of seams can describe the location-related information of signal attenuation.

[0161] 5) Fourier transform can decompose the mixed signal into independent signals in two directions, which is convenient for subsequent analysis and processing.

[0162] 6) The neural network model trained with a large number of samples can establish a nonlinear mapping relationship between the ultrasonic signal characteristics and the quality of the internal partition wall, and realize intelligent evaluation.

[0163] 7) Theoretical numerical simulation provides verification support for experimental results, and the combination of the two can improve the accuracy of the evaluation.

[0164] 8) This method integrates multiple technologies such as signal processing, analog computing, and intelligent algorithms, and uses the intrinsic connection between ultrasonic propagation and the quality of internal partition walls to achieve accurate detection of construction quality.

[0165] It should be noted that the variables used in the specific embodiments of the present invention are explained as shown in the following table:

[0166] Table 1 Variable explanation table

[0167]

[0168] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for detecting the construction quality of an assembled internal partition wall, characterized in that: The following steps are involved: S10, collecting ultrasonic signals of three random points of each wallboard on the assembled inner partition wall, the ultrasonic signals are continuously emitted by two ultrasonic transmitters arranged at diagonally opposite corners of the inner partition wall, and the ultrasonic signals emitted by the two ultrasonic transmitters have a frequency difference of more than 10 times; the ultrasonic signals are collected by an ultrasonic sensor, and the ultrasonic sensor is used to collect ultrasonic signals of three random points of each wallboard; S20, calculating the ultrasonic signal of the geometric center of each wall panel according to the ultrasonic signals of three random points of each wall panel, and recording it as the wall panel receiving signal; S30, establishing a wall panel vector, including a wall panel receiving signal, a straight-line distance between a wall panel geometric center and each ultrasonic transmitter, and a number of wall panel joints between the wall panel geometric center and each ultrasonic transmitter; S40, using Fourier transform to split the wall panel receiving signal of each wall panel to obtain a first receiving signal and a second receiving signal; S50, calculating the attenuation of ultrasonic signals between the wall panels, recorded as a first attenuation set; S60, establishing a finite element model of the assembled interior partition wall, and calculating the attenuation range of the ultrasonic signal between the wall panels in the finite element model, which is recorded as a second attenuation range set; S70, using a pre-trained signal attenuation-construction quality model, inputting the first attenuation degree set and the second attenuation range set, to obtain a construction quality assessment vector for the prefabricated interior partition wall; S80, outputting the construction quality assessment vector of the assembled inner partition wall to a testing personnel; The calculation of the geometric center is as follows: Assume the number of wall panels is , wall panel number is Each wall panel The coordinates of the three random sampling points are ; Wall Panel The geometric center coordinates of The calculation formula is: ； Wherein, step S20 specifically includes: Sampling point Euclidean distance to the geometric center The calculation formula is: ； Wall Panels The ultrasonic signals at the three sampling points are marked as ; Wall Panel Synthetic ultrasonic signal at the geometric center The calculation formula is: ； Repeat the above calculation to obtain the expected signal values ​​in the two transmitter directions respectively, and the average value of the two expected signal values ​​is used to represent the ultrasonic signal intensity at the geometric center of the wall panel; The signal attenuation-construction quality model training step includes: collecting a large amount of case data of interior partition walls with different construction qualities; setting a neural network model structure; determining model training parameters; and training the model using case data; wherein the case data is an interior partition wall construction quality assessment vector obtained by experimental detection and a corresponding first attenuation degree set and a second attenuation range set obtained by using steps S10 to S60; Among them, before starting the detection, it is necessary to calibrate the signal strength of the two ultrasonic transmitters. Suppose the two ultrasonic transmitters are A and B. The calibration method is: let the two transmitters A and B emit ultrasonic waves at the same time, and use standardized ultrasonic detection equipment to measure the strength of the two signals. If the difference between the two signal strengths exceeds 3dB, it is necessary to adjust the transmission power of the transmitter until the difference between the two signal strengths is within 3dB.

2. A method for detecting the construction quality of an assembled interior partition wall according to claim 1, characterized in that: The step of calculating the attenuation of ultrasonic signals between wall panels specifically comprises: normalizing the two received signals to obtain energy; calculating the logarithmic energy attenuation rate of the signals between the wall panels; and obtaining a set of attenuation rates of all wall panel signals as a first attenuation set.

3. A method for detecting the construction quality of an assembled interior partition wall according to claim 1, characterized in that: The steps of establishing a finite element model of the assembled interior partition wall and calculating the attenuation range of the ultrasonic signal between the wall panels in the finite element model are specifically: establishing a finite element model of the interior partition wall; setting material parameters; setting ultrasonic sources at both ends; simulating and calculating the signal attenuation between the wall panels; and obtaining an attenuation range set between the wall panels as a second attenuation range set.

4. A method for detecting the construction quality of an assembled interior partition wall according to claim 1, characterized in that: The ultrasonic frequencies emitted by the two ultrasonic transmitters are 2 MHz and 20 MHz respectively.

5. A method for detecting the construction quality of an assembled interior partition wall according to claim 3, characterized in that: The neural network model is a multi-layer feedforward neural network.

6. A method for detecting the construction quality of an assembled interior partition wall according to claim 3, characterized in that: The number of samples used for model training is greater than 50.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions, and when the program instructions are executed, they are used to execute the method for detecting the construction quality of an assembled interior partition wall according to any one of claims 1 to 6.

8. A prefabricated internal partition wall construction quality inspection system, characterized in that: A computer-readable storage medium comprising the method of claim 7.

Images