Material size automatic measurement method and device

Abstract

The invention discloses a material size automatic measurement method and device, and the method comprises the steps: predicting a sound wave emission parameter with an excellent sound wave reflection signal receiving effect of a product material, and obtaining a sound wave reflection signal and corresponding timestamp information; identifying a sound wave reflection source type of the sound wave reflection signal, and determining that the sound wave reflection source is a frequency spectrum characteristic of the sound wave reflection signal of the product material; identifying the reason of deviation of the sound wave reflection signal of the product material; the heterogeneity coefficient of the heterogeneous material is calculated, the propagation speed of the sound wave signal in the current heterogeneous material is corrected, and the thickness size of the single-layer product material and the thickness size of the multi-layer combined material are determined; and evaluating whether the technological parameters and the production parameters meet the production stability requirements after adjustment. The invention discloses a method and a device for automatically measuring the size of a material, which are used for enhancing the intelligence and automation level of product material detection and meeting the requirements of high-precision and nondestructive detection and high-efficiency production.

Description

Technical Field

[0001] The present invention relates to the field of automated measurement technology, and in particular to a material size automatic measurement method and device. Background Art

[0002] As the level of automation in industrial manufacturing continues to improve, the requirements for accuracy and efficiency of product quality control are also increasing. In modern production, especially in the production and monitoring of high-precision parts, how to achieve fast and accurate material size and thickness measurement has become a key issue. Traditional measurement methods mostly rely on contact tools such as calipers and micrometers. These contact measurement methods are not only time-consuming, but also susceptible to human errors and environmental factors, and it is difficult to meet the requirements of high efficiency and high precision. At present, ultrasonic measurement technology, as a non-contact measurement method, has shown great application potential in material thickness, quality inspection and defect identification. Ultrasonic waves measure through the propagation characteristics of sound waves in materials, which can effectively avoid friction and damage caused by contact methods. However, the existing ultrasonic measurement technology still faces some technical difficulties and challenges in practical applications. First, environmental factors have a great impact on ultrasonic signals. The propagation of ultrasonic signals is affected by environmental conditions such as temperature, humidity, and air pressure. These factors can cause the signal frequency and amplitude to drift, thereby affecting the accuracy of the measurement results. For example, in high or low temperature environments, the acoustic properties of materials may change, which in turn causes changes in the propagation speed and reflection characteristics of ultrasonic signals, making it difficult for traditional ultrasonic measurement methods to ensure consistency and reliability in different environments. Secondly, in the thickness measurement of existing heterogeneous materials, such as composites containing different minerals, fibers or particles, local density changes in materials will cause nonlinear changes in the propagation path of sound waves, which will cause measurement errors. Different regions of heterogeneous materials, such as regions with higher or lower density, will cause sound waves to pass through at different speeds during propagation, so that the propagation path of sound waves is no longer a simple linear function. Traditional sound wave propagation models assume the uniformity of materials, but in practical applications, especially when complex heterogeneous materials are involved, these assumptions are often not valid, making accurate measurements extremely difficult. In general, the existing ultrasonic measurement technology has many problems in terms of accuracy, adaptability, environmental interference suppression, and measurement of multi-layer composite materials, which limits its application in high-precision automated production lines and inspection lines. How to solve these technical problems and improve the adaptability of ultrasonic measurement technology in various complex environments has become a key issue that needs to be urgently solved in the current industrial measurement field. Summary of the invention

[0003] The present invention aims to solve the above problems in the prior art and provides a material size automatic measurement method and device.

[0004] An embodiment of the first aspect of the present invention provides a method for automatically measuring the size of a material, mainly including:

[0005] According to product materials of different specification standards, acoustic emission parameters, and the corresponding acoustic reflection signal reception effects, predict the acoustic emission parameters for which the acoustic reflection signal reception effect of the current product material is excellent, emit ultrasonic signals, and obtain the acoustic reflection signals and the corresponding timestamp information;

[0006] Separate the acoustic reflection signals through an independent component analysis algorithm, identify the types of acoustic reflection sources of the acoustic reflection signals, and determine the spectral characteristics of the acoustic reflection signals whose acoustic reflection sources are product materials;

[0007] According to the spectral characteristics of the acoustic reflection signals of the product material after calibration and the spectral characteristics of the reference acoustic reflection signals of the product material specification standard, identify the reasons for the deviation of the acoustic reflection signals of the product material, and adjust the process parameters of the production equipment;

[0008] Determine the inhomogeneity coefficient of the inhomogeneous material through the acoustic reflection signals of the inhomogeneous material and the acoustic reflection signals of the known homogeneous reference sample, correct the propagation speed of the acoustic signals in the current inhomogeneous material, and calculate the thickness dimensions of single-layer product materials and multi-layer composite materials;

[0009] Obtain the acoustic reflection signals of the product material after adjusting the process parameters and production parameters through an ultrasonic sensor, determine whether the production stability requirements have been met after adjustment, and continuously adjust the process parameters and production parameters.

[0010] Further, the step of predicting the acoustic emission parameters for which the acoustic reflection signal reception effect of the current product material is excellent, emitting ultrasonic signals, and obtaining the acoustic reflection signals and the corresponding timestamp information according to product materials of different specification standards, acoustic emission parameters, and the corresponding acoustic reflection signal reception effects includes:

[0011] The ultrasonic sensor is used to detect product materials of different specification standards and receive ultrasonic signals. The acoustic wave emission parameters of the ultrasonic sensor are recorded, and the receiving effects of the acoustic wave reflection signals are marked. The specification standards include material type, size, thickness dimension, number of layers, and density. The acoustic wave emission parameters include emission frequency, pulse width, and emission power. The receiving effects include excellent, good, and poor. According to the product materials of different specification standards, the acoustic wave emission parameters, and the corresponding receiving effects of the acoustic wave reflection signals, a cyclic neural network is used for model training to construct an acoustic wave emission parameter prediction model. The acoustic wave emission parameters with excellent receiving effects of the acoustic wave reflection signals for the current product materials are determined, and the acoustic wave emission parameters of the ultrasonic sensor are set. According to the layout of the detection line, the position of the ultrasonic sensor is adjusted, and the ultrasonic sensor is installed above the detection line. An image of the product materials on the production line is obtained through a camera above the detection line. If the product materials are directly below the ultrasonic sensor, the ultrasonic sensor is used to emit ultrasonic signals to the product materials on the detection line according to the set acoustic wave emission parameters and receive the acoustic wave reflection signals and the corresponding timestamp information.

[0012] Further, the acoustic wave reflection signals are separated by an independent component analysis algorithm, the types of acoustic wave reflection sources of the acoustic wave reflection signals are identified, and the spectral characteristics of the acoustic wave reflection signals with the acoustic wave reflection source being the product materials are determined, including:

[0013] According to the received acoustic wave reflection signals, the acoustic wave reflection signals are separated by an independent component analysis algorithm to obtain the acoustic wave reflection signals of different acoustic wave reflection sources. The short-time Fourier transform algorithm is used to convert the time-domain audio signals of the acoustic wave reflection signals into frequency-domain data, the spectral characteristics of the acoustic wave reflection signals are extracted and saved to the acoustic wave reflection monitoring database. The spectral characteristics include frequency, frequency peak, and amplitude. Through the acoustic wave reflection signal monitoring database, the spectral characteristics of the acoustic wave reflection signals of different acoustic wave reflection sources are obtained, the types of acoustic wave reflection sources are marked, and a random forest algorithm is used for model training to construct an acoustic wave reflection source identification model to identify the types of acoustic wave reflection sources of the acoustic wave reflection signals and determine the spectral characteristics of the acoustic wave reflection signals with the acoustic wave reflection source being the product materials. The detection environment data is obtained, and according to the detection environment data, the frequency drift amount and the amplitude drift amount are calculated to correct the frequency and amplitude of the acoustic wave reflection signals, and the spectral characteristics of the corrected acoustic wave reflection signals of different acoustic wave reflection sources are obtained and saved to the acoustic wave reflection signal monitoring database. The detection environment data includes temperature, humidity, and air pressure.

[0014] It also includes calculating the frequency drift amount and the amplitude drift amount according to the detection environment data to correct the frequency and amplitude of the acoustic wave reflection signals, specifically including:

[0015] The temperature, humidity, and air pressure data inside the detection line are obtained in real time through the temperature sensor, humidity sensor, and air pressure sensor installed on the detection line, and the frequency drift amount calculation formula is used Determine the frequency drift Δf in the current detection environment, and use the amplitude drift calculation formula ΔA = A1·(n1·sin(T) + n2·H - n3·log(P)) to determine the amplitude drift ΔA in the current detection environment, where T is the temperature, H is the humidity, P is the air pressure, f1 and A1 are the frequency value measured under standard conditions and the amplitude value measured under standard conditions respectively, l1, l2, and l3 are the influence coefficients of the detection environment on the frequency, and n1, n2, and n3 are the influence coefficients of the detection environment on the amplitude, which are obtained through experimental fitting of historical data; correct the frequency and amplitude of the acoustic wave reflection signal according to the spectral characteristics, frequency drift, and amplitude drift of the acoustic wave reflection signal.

[0016] Further, based on the spectral characteristics of the acoustic wave reflection signal of the corrected product material and the spectral characteristics of the reference acoustic wave reflection signal of the product material specification standard, identify the reasons for the deviation of the acoustic wave reflection signal of the product material, and adjust the process parameters of the production equipment, including:

[0017] Through the acoustic wave reflection signal monitoring database, obtain the spectral characteristics of the acoustic wave reflection signals of different acoustic wave reflection sources after correction, and calculate the similarity between the spectral characteristics of the obtained acoustic wave reflection signal of the product material and the spectral characteristics of the reference acoustic wave reflection signal of the product material specification standard by using the calculation method of cosine similarity; if the similarity is lower than the preset similarity threshold, then based on the obtained spectral characteristics of the acoustic wave reflection signal of the product material, and mark the reasons for the deviation of the acoustic wave reflection signal of the product material, use the random forest algorithm for model training to construct an identification model for the reasons for the deviation of the acoustic wave reflection signal, and identify the reasons for the deviation of the acoustic wave reflection signal of the product material. The reasons for the deviation of the acoustic wave reflection signal include but are not limited to product material density deviation and shape defects, and shape defects include but are not limited to cracks and pits; adjust the process parameters of the production equipment according to the reasons for the deviation of the acoustic wave reflection signal, and the process parameters include but are not limited to temperature, cooling rate, injection pressure, injection speed, and injection volume of the injection molding machine.

[0018] Further, determine the inhomogeneity coefficient of the inhomogeneous material through the acoustic wave reflection signal of the inhomogeneous material and the acoustic wave reflection signal of the known homogeneous reference sample, correct the propagation speed of the acoustic wave signal in the current inhomogeneous material, and calculate the thickness dimensions of the single-layer product material and the multi-layer composite material, including:

[0019] If the material with product material density deviation is a heterogeneous material, obtain the acoustic reflection signal of the corrected heterogeneous material and the acoustic reflection signal of the known homogeneous reference sample through the acoustic reflection signal monitoring database, and use the short-time Fourier transform algorithm to convert the time-domain audio signal of the acoustic reflection signal into frequency-domain data. The heterogeneous material includes, but is not limited to, foamed plastics, carbon fiber composite materials, and glass fiber composite materials; through the heterogeneity coefficient formula calculate the heterogeneity coefficient H of the heterogeneous material, where S t and S r are the acoustic reflection signal spectra of the heterogeneous material to be measured and the acoustic reflection signal spectra of the known homogeneous reference sample respectively, Cov(S t ,S r ) is the covariance of the spectra S t and S r , is the standard deviation of the spectrum S t , is the standard deviation of the spectrum S r ; based on the frequency-domain data of the acoustic reflection signal of the heterogeneous material and the acoustic reflection signal of the known homogeneous reference sample, extract the main frequency phase difference between the acoustic reflection signal spectrum of the heterogeneous material and the acoustic reflection signal spectrum of the known homogeneous reference sample, and based on the main frequency phase difference, judge the magnitude of the density of the heterogeneous material and the density of the known homogeneous reference sample, and determine the velocity correction direction factor D. If the density of the heterogeneous material is greater than the density of the known homogeneous reference sample, D is 1. If the density of the heterogeneous material is less than the density of the known homogeneous reference sample, D is -1; according to the heterogeneity coefficient and the velocity correction direction factor, use the velocity correction formula to correct the propagation velocity v a of the acoustic signal in the current heterogeneous material, where v0 is the propagation velocity of the acoustic signal in the known homogeneous reference sample; calculate the propagation time difference ΔT between the signal from the sensor to the upper surface and the lower surface through the timestamp information of the received ultrasonic signal; through the thickness dimension calculation formula Determine the thickness dimension h of the product material, where v0 is the propagation speed of the acoustic wave signal in a known homogeneous reference sample; based on the inhomogeneity coefficient of the inhomogeneous material, determine whether there is a local density direction change region in the inhomogeneous material, and based on the local inhomogeneity coefficient and the velocity correction direction factor of the inhomogeneous material, calculate the thickness dimensions of the local parts of the inhomogeneous material respectively, and determine the total thickness dimension of the inhomogeneous material; if the product material is a multi-layer composite material, then through the timestamp information of the received ultrasonic signal, calculate the propagation time differences between the signal from the sensor to the upper surfaces and the lower surfaces of each layer respectively, and through the thickness dimension calculation formula, calculate the thickness dimensions of each layer of the multi-layer composite material; calculate the dimension deviation between the product material dimension and the preset dimension, as well as the thickness dimension deviation between the product material thickness dimension and the preset thickness dimension. If the dimension deviation is greater than the preset dimension deviation threshold or the thickness dimension deviation exceeds the preset thickness dimension deviation threshold, then correct the thickness dimension of the product material by adjusting the production parameters in the production process, and the generated parameters include the accuracy of the production equipment, the gap of the mold, and the forming temperature.

[0020] It also includes determining whether there is a local density direction change region in the inhomogeneous material based on the inhomogeneity coefficient of the inhomogeneous material, and calculating the thickness dimensions of the local parts of the inhomogeneous material respectively and determining the total thickness dimension of the inhomogeneous material based on the local inhomogeneity coefficient and the velocity correction direction factor of the inhomogeneous material, specifically including:

[0021] If the inhomogeneity coefficient of the inhomogeneous material is greater than the preset coefficient threshold, then determine that there is a local density direction change region in the inhomogeneous material, segment the echo signal of the inhomogeneous material in time according to the preset number of segments, and determine the main frequency phase difference of each segment of the echo signal; calculate the inhomogeneity coefficient of the local part of the inhomogeneous material corresponding to each segment of the echo signal through the inhomogeneity coefficient formula; based on the main frequency phase difference of each segment of the echo signal, judge the size relationship between the density of the local part of the inhomogeneous material corresponding to each segment of the echo signal and the density of the known homogeneous reference sample, and determine the velocity correction direction factor D of the local part of the inhomogeneous material corresponding to each segment of the echo signal; according to the inhomogeneity coefficient and the velocity correction direction factor of the local part of the inhomogeneous material corresponding to each segment of the echo signal, use the velocity correction formula to correct the propagation speed of the echo signal in each local part of the inhomogeneous material respectively; through the thickness dimension calculation formula, determine the thickness dimensions of the local parts of the inhomogeneous material respectively, and calculate the total thickness dimension of the inhomogeneous material based on the thickness dimensions of the local parts of the inhomogeneous material.

[0022] Furthermore, obtaining the acoustic wave reflection signal of the product material after adjusting the process parameters and production parameters through the ultrasonic sensor, judging whether the production stability requirement has been met after adjustment, and continuously adjusting the process parameters and production parameters, including:

[0023] Obtain the acoustic wave reflection signal of the product material after adjusting the process parameters and production parameters through an ultrasonic sensor, and obtain the new spectral characteristics of the acoustic wave reflection signal and the corresponding timestamp information; based on the obtained new spectral characteristics of the acoustic wave reflection signal and the corresponding timestamp information, judge whether there is a deviation in the acoustic wave reflection signal, and whether the thickness dimension deviation between the thickness dimension of the product material and the preset thickness dimension exceeds the preset deviation threshold, and judge whether the production stability requirement has been met after adjustment; if the production stability requirement has not been met yet, continue to adjust the process parameters and production parameters until the production stability requirement is met.

[0024] The second aspect embodiment of the present invention provides a material size automatic measurement device, which mainly includes:

[0025] A processor and a memory; at least one instruction is stored in the memory, and the instruction is executed by the processor to enable the device to execute a material size automatic measurement method provided by the first aspect embodiment.

[0026] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0027] The present invention provides a material size automatic measurement method and device. By optimizing the acoustic wave emission parameters and real-time correcting the ultrasonic signal, the present invention can effectively improve the measurement accuracy and reliability. By analyzing the acoustic wave reflection signals of different materials and dynamically adjusting in combination with production environment data, the present invention can overcome the problem that traditional measurement methods are affected by environmental factors and ensure the stability of signals under different temperature, humidity and air pressure conditions. In addition, by real-time monitoring the changes in the production environment and performing adaptive correction, the present invention can effectively eliminate the measurement errors caused by process fluctuations and maintain high stability and accuracy during the production process. The present invention also has excellent non-uniform material size detection capabilities. By calculating the non-uniformity coefficient of non-uniform materials, correcting the propagation speed of acoustic waves in the current non-uniform materials, and based on the acoustic wave propagation time difference, accurately calculating the thickness dimension of non-uniform materials, and can also achieve accurate measurement of the thickness dimensions of single-layer and multi-layer composite materials. This function is particularly suitable for the automated detection of composite materials and complex structures, greatly improving the production efficiency and reducing the human measurement error. The present invention not only improves the accuracy and adaptability of the ultrasonic measurement system, but also enhances the intelligence and automation level of product material detection on the detection line, meeting the requirements of modern manufacturing for high precision, non-destructive detection and high-efficiency production. Description of the Drawings

[0028] Figure 1 It is a flowchart of a material size automatic measurement method and device of the present invention;

[0029] Figure 2 It is a schematic diagram of a material size automatic measurement method and device of the present invention;

[0030] Figure 3 Another schematic diagram of the automatic material size measurement method and device of the present invention. Detailed implementation manners

[0031] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] As Figures 1-3 , a specific automatic material size measurement method and device in this embodiment may specifically include:

[0033] Step S101, according to product materials of different specification standards, acoustic wave emission parameters, and corresponding acoustic wave reflection signal reception effects, predict the acoustic wave emission parameters with excellent acoustic wave reflection signal reception effects for the current product material, emit ultrasonic signals, and obtain the acoustic wave reflection signals and corresponding timestamp information.

[0034] Use an ultrasonic sensor to receive ultrasonic signals for product materials of different specification standards, record the acoustic wave emission parameters of the ultrasonic sensor, and label the acoustic wave reflection signal reception effects. The specification standards include material type, size, thickness dimension, number of layers, and density. The acoustic wave emission parameters include emission frequency, pulse width, and emission power. The reception effects include excellent, good, and poor. According to product materials of different specification standards, acoustic wave emission parameters, and corresponding acoustic wave reflection signal reception effects, use a recurrent neural network for model training, construct an acoustic wave emission parameter prediction model, determine the acoustic wave emission parameters with excellent acoustic wave reflection signal reception effects for the current product material, and set the acoustic wave emission parameters for the ultrasonic sensor. According to the detection line layout, debug the position of the ultrasonic sensor, and install the ultrasonic sensor above the detection line. Obtain an image of the product material on the production line through a camera above the detection line. If the product material is directly below the ultrasonic sensor, use the ultrasonic sensor to emit ultrasonic signals to the product material on the detection line according to the set acoustic wave emission parameters, and receive the acoustic wave reflection signals and corresponding timestamp information.

[0035] Exemplarily, on the inspection line of a certain manufacturing factory, if the product material is a foam plastic board and the product has different specification standards. Specifically, there are three specification standards for the foam plastic boards produced on the inspection line, including a polyurethane material type with dimensions of 300mm x 200mm, a thickness dimension of 10mm, a density of 0.3g / cm3, and 1 layer; a polyvinyl chloride material type with dimensions of 250mm x 200mm, a thickness dimension of 15mm, a density of 0.5g / cm3, and 2 layers; a polyethylene material type with dimensions of 500mm x 400mm, a thickness dimension of 20mm, a density of 0.4g / cm3, and 3 layers. To ensure the product quality and inspection effect, an ultrasonic sensor is used to detect the thickness dimension of these products and record the corresponding sound wave reflection signal reception effect. The sound wave emission parameters of the ultrasonic sensor include an emission frequency set to 5MHz, a pulse width of 2 microseconds, and an emission power of 30dB. During the experiment, according to the above specification standards and emission parameters, the ultrasonic sensor receives the reflection signals of different product materials respectively and labels their reception effects. For the first-specification polyurethane material, the reflection signal reception effect is rated as excellent; for the second-specification polyvinyl chloride material, the reception effect is good; for the third-specification polyethylene material, the reflection signal reception effect is rated as poor. Based on these labels and the specification standards of the product materials, a cyclic neural network is used for model training to construct a sound wave emission parameter prediction model to predict the influence of the sound wave emission parameters of materials with different specification standards on the reflection signal reception effect. After model training, the optimal sound wave emission parameters applicable to the inspection line are obtained, that is, an emission frequency of 5MHz, a pulse width of 2 microseconds, and an emission power of 30dB, which can ensure an excellent reflection signal reception effect. According to the layout of the inspection line, the ultrasonic sensor is debugged and precisely installed above the inspection line to ensure that it can correctly irradiate the foam plastic board being produced. According to the inspection line layout, the position of the ultrasonic sensor is debugged, and the ultrasonic sensor is installed above the inspection line. The position of the product is monitored in real time through a camera installed on the inspection line. When the camera detects that the product material is directly below the ultrasonic sensor, the ultrasonic sensor will emit signals according to the previously set sound wave emission parameters. The ultrasonic signals are reflected back by the product material, and the received sound wave reflection signals and timestamp information of the ultrasonic sensor are recorded.

[0036] Step S102, separate the sound wave reflection signals through an independent component analysis algorithm, identify the sound wave reflection source type of the sound wave reflection signals, and determine the spectral characteristics of the sound wave reflection signals whose sound wave reflection source is the product material.

[0037] According to the received acoustic wave reflection signal, the acoustic wave reflection signal is separated by the independent component analysis algorithm to obtain the acoustic wave reflection signals of different acoustic wave reflection sources. The time-domain audio signal of the acoustic wave reflection signal is converted into frequency-domain data by using the short-time Fourier transform algorithm, the spectral features of the acoustic wave reflection signal are extracted and saved to the acoustic wave reflection monitoring database, and the spectral features include frequency, frequency peak and amplitude. Through the acoustic wave reflection signal monitoring database, the spectral features of the acoustic wave reflection signals of different acoustic wave reflection sources are obtained, the types of acoustic wave reflection sources are labeled, and the random forest algorithm is used for model training to construct an acoustic wave reflection source recognition model to identify the types of acoustic wave reflection sources of the acoustic wave reflection signal and determine the spectral features of the acoustic wave reflection signal whose acoustic wave reflection source is the product material. The detection environment data are obtained, and according to the detection environment data, the frequency drift amount and amplitude drift amount are calculated to correct the frequency and amplitude of the acoustic wave reflection signal, and the spectral features of the corrected acoustic wave reflection signals of different acoustic wave reflection sources are obtained and saved to the acoustic wave reflection signal monitoring database. The detection environment data include temperature, humidity and air pressure.

[0038] Exemplarily, on a certain detection line, if the size of the currently produced foam plastic board is 500mm x 300mm, the thickness is 20mm, and the density is 0.4g / cm3. During the production process, the sensor receives acoustic wave reflection signals from different sources, including the reflection signals from the foam plastic board, the production equipment such as the metal fuselage or robotic arm, and the reflection signals from the workshop walls or other objects, and saves them to the acoustic wave reflection monitoring database. To effectively analyze and process these signals, the independent component analysis algorithm is first used to separate the received mixed acoustic wave signals, so as to extract the acoustic wave reflection signals from different reflection sources. The short-time Fourier transform is used to convert these acoustic wave reflection signals from the time domain to frequency domain data. By performing the Fourier transform on each separated signal, spectral features such as its frequency, frequency peak, and amplitude are extracted, obtaining three types of acoustic wave reflection signals, including a frequency of 5MHz, a frequency peak of 5.2MHz, and an amplitude of 0.8; a frequency of 6MHz, a frequency peak of 6.1MHz, and an amplitude of 1.0; and a frequency of 4.5MHz, a frequency peak of 4.7MHz, and an amplitude of 0.4. The spectral features of these three types of acoustic wave reflection signals are saved to the acoustic wave reflection monitoring database. Through the acoustic wave reflection signal monitoring database, the spectral features of the acoustic wave reflection signals from different acoustic wave reflection sources are obtained, and the types of acoustic wave reflection sources are labeled, including the reflection signals from the foam plastic board, the metal fuselage or robotic arm, and the workshop walls or other objects. The random forest algorithm is used for model training to construct an acoustic wave reflection source identification model to identify the types of acoustic wave reflection sources of the acoustic wave reflection signals, and it is determined that the spectral features of the acoustic wave reflection signals with the product material as the acoustic wave reflection source are a frequency of 5MHz, a frequency peak of 5.2MHz, and an amplitude of 0.8. In the production environment, environmental factors such as temperature, humidity, and air pressure may affect the propagation of ultrasonic signals, resulting in drift in the frequency and amplitude of the signals. For example, the current production environment has a temperature of 25°C, a humidity of 60%, and an air pressure of 1010hPa. Based on these production environment data, the frequency drift amount and amplitude drift amount of the signal can be calculated. Through the relevant environmental correction model, it is found that an increase in temperature will cause the frequency drift amount of the signal to be 0.2MHz and the amplitude drift amount to be 0.1. To ensure the accuracy of the measurement, these drift amounts are calculated and applied to correct each acoustic wave reflection signal. After correction, the frequency of the acoustic wave reflection signal of the product material becomes 4.8MHz and the amplitude is 0.7. Finally, these corrected spectral features are saved to the acoustic wave reflection monitoring database, providing a basis for subsequent production quality control and parameter adjustment.

[0039] Among them, according to the detection environment data, the frequency drift amount and amplitude drift amount are calculated to correct the frequency and amplitude of the acoustic wave reflection signal.

[0040] Temperature sensors, humidity sensors, and barometric pressure sensors installed on the detection line are used to obtain the temperature, humidity, and barometric pressure data inside the detection line in real time, and the frequency drift amount calculation formula is used. Determine the frequency drift amount Δf under the current detection environment, and use the amplitude drift amount calculation formula ΔA = A1·(n1·sin(T) + n2·H - n3·log(P)) to determine the amplitude drift amount ΔA of the current detection environment, where T is the temperature, H is the humidity, P is the barometric pressure, f1 and A1 are the frequency value measured under standard conditions and the amplitude value measured under standard conditions respectively, l1, l2, and l3 are the influence coefficients of the detection environment on the frequency, and n1, n2, and n3 are the influence coefficients of the detection environment on the amplitude, which are obtained through experimental fitting of historical data. According to the spectral characteristics, frequency drift amount, and amplitude drift amount of the acoustic wave reflection signal, correct the frequency and amplitude of the acoustic wave reflection signal.

[0041] Exemplarily, inside the detection line, the temperature sensor, humidity sensor, and barometric pressure sensor monitor the detection environment in real time, and the current temperature is obtained as 30 °C, the humidity is 65%, and the barometric pressure is 1012 hPa. According to these detection environment data and historical experimental data, use the frequency drift amount calculation formula and the amplitude drift amount calculation formula ΔA = A1·(n1·sin(T) + n2·H - n3·log(P)) to determine the frequency and amplitude drift amounts under the current detection environment. If the measured frequency is 5 MHz and the amplitude is 1.0 under the standard conditions of temperature T = 25 °C, humidity H = 50%, and barometric pressure P = 1010 hPa. According to the experimental fitting results, the influence coefficient of the detection environment on the frequency is l1 = 0.002, the influence coefficient on the humidity is l2 = 0.1, the influence coefficient on the barometric pressure is l3 = 0.01, and the coefficients of the amplitude drift amount are n1 = 0.05, n2 = 0.01, and n3 = 0.02. According to the frequency drift amount calculation formula, the calculated frequency drift amount is 0.3806 MHz. According to the amplitude drift amount calculation formula, the calculated amplitude drift amount is 0.6149. After obtaining the frequency drift amount and amplitude drift amount, the frequency and amplitude of the acoustic wave reflection signal can be corrected. If the original frequency of the acoustic wave reflection signal is 5 MHz and the amplitude is 1.0, through the above calculations, the corrected frequency and amplitude are 5.3806 MHz and 1.6149 respectively.

[0042] Step S103, according to the spectral characteristics of the corrected acoustic wave reflection signal of the product material and the spectral characteristics of the reference acoustic wave reflection signal of the product material specification standard, identify the reasons for the deviation of the acoustic wave reflection signal of the product material, and adjust the process parameters of the production equipment.

[0043] Monitor the database through the acoustic wave reflection signal, obtain the spectral characteristics of the acoustic wave reflection signals of different acoustic wave reflection sources after calibration, and calculate the similarity between the spectral characteristics of the obtained acoustic wave reflection signals of the product material and the spectral characteristics of the reference acoustic wave reflection signal of the product material specification standard through the calculation method of cosine similarity. If the similarity is lower than the preset similarity threshold, then based on the obtained spectral characteristics of the product material acoustic wave reflection signal, and mark the reasons for the deviation of the product material acoustic wave reflection signal, use the random forest algorithm for model training to construct an acoustic wave reflection signal deviation cause identification model to identify the reasons for the deviation of the product material acoustic wave reflection signal. The reasons for the deviation of the acoustic wave reflection signal include but are not limited to product material density deviation and shape defects. Shape defects include but are not limited to cracks, bubbles, and pits. According to the reasons for the deviation of the acoustic wave reflection signal, adjust the process parameters of the production equipment. The process parameters include but are not limited to temperature, cooling rate, injection pressure, injection speed, and injection volume of the injection molding machine.

[0044] Exemplarily, in an inspection line of a foam plastic board, the standard product specifications are: polyurethane material, size 300mm x 200mm, thickness dimension 10mm, density 0.3g / cm3. During the production process, the ultrasonic sensor monitors the spectral characteristics of the acoustic wave reflection signal of the product material in real-time through the acoustic wave reflection signal monitoring database. After calibration, the spectral characteristics of the acoustic wave reflection signal of the product material include: frequency of 5MHz, frequency peak of 5.2MHz, and amplitude of 0.8. Through the calculation method of cosine similarity, the spectral characteristics of the current product material are compared with the reference standard specification spectral characteristics, and the similarity value is obtained as 0.85, while the preset similarity threshold is 0.9. Since the calculated similarity value is lower than the preset 0.9, it indicates that there are certain deviations in the currently produced products. Based on the obtained spectral characteristics, after analysis, it is determined that the main reason for the deviation is the unevenness on the surface of the product material, and by marking this signal, the influence of this shape defect is confirmed. To further analyze the reasons for the deviation, use historical data and the reasons for the deviation of the acoustic wave reflection signal, and perform model training through the random forest algorithm to construct an acoustic wave reflection signal deviation cause identification model. According to the trained model, it is identified that the surface unevenness and possible material density deviation are the main reasons for the deviation of the acoustic wave reflection signal. To solve this problem, the process parameters of the inspection line are adjusted. First, the temperature control is fine-tuned to ensure that the mold temperature is stable at 185°C. Second, the cooling rate is adjusted to 1.2°C / min to ensure uniform cooling of the mold. At the same time, the injection pressure of the injection molding machine is increased from 90MPa to 95MPa, the injection speed is maintained at 18mm / s, and the injection volume is adjusted to 120g to better fill the mold and reduce the occurrence of surface defects.

[0045] Step S104: Determine the inhomogeneity coefficient of the inhomogeneous material based on the acoustic wave reflection signals of the inhomogeneous material and the acoustic wave reflection signals of a known homogeneous reference sample, correct the propagation speed of the acoustic wave signal in the current inhomogeneous material, and calculate the thickness dimensions of the single-layer product material and the multi-layer composite material.

[0046] If the material with a product material density deviation is an inhomogeneous material, obtain the corrected acoustic wave reflection signal of the inhomogeneous material and the acoustic wave reflection signal of a known homogeneous reference sample through the acoustic wave reflection signal monitoring database. Use the short-time Fourier transform algorithm to convert the time-domain audio signal of the acoustic wave reflection signal into frequency-domain data. Inhomogeneous materials include, but are not limited to, foam plastics, carbon fiber composite materials, and glass fiber composite materials. Through the inhomogeneity coefficient formula calculate the inhomogeneity coefficient H of the inhomogeneous material, where S t and S r are the frequency spectra of the acoustic wave reflection signals of the inhomogeneous material to be measured and the acoustic wave reflection signals of the known homogeneous reference sample, respectively. Cov(S t , S r ) is the covariance of the frequency spectra S t and S r , is the standard deviation of the frequency spectrum S t , is the standard deviation of the frequency spectrum S r . Based on the frequency-domain data of the acoustic wave reflection signals of the inhomogeneous material and the acoustic wave reflection signals of the known homogeneous reference sample, extract the main frequency phase difference between the frequency spectra of the acoustic wave reflection signals of the inhomogeneous material and the acoustic wave reflection signals of the known homogeneous reference sample. And based on the main frequency phase difference, judge the magnitude of the density of the inhomogeneous material and the density of the known homogeneous reference sample, and determine the velocity correction direction factor D. If the density of the inhomogeneous material is greater than the density of the known homogeneous reference sample, then D is 1. If the density of the inhomogeneous material is less than the density of the known homogeneous reference sample, then D is -1. According to the inhomogeneity coefficient and the velocity correction direction factor, use the velocity correction formula to correct the propagation speed v a of the acoustic wave signal in the current inhomogeneous material, where v0 is the propagation speed of the acoustic wave signal in the known homogeneous reference sample. Calculate the propagation time difference ΔT between the signal from the sensor to the upper surface and the lower surface through the time stamp information of the received ultrasonic signal. Through the thickness dimension calculation formula Determine the thickness dimension h of the product material, where v0 is the propagation speed of the acoustic wave signal in a known homogeneous reference sample. According to the inhomogeneity coefficient of the inhomogeneous material, determine whether there is a region of local density direction change in the inhomogeneous material. Based on the local inhomogeneity coefficient and the velocity correction direction factor of the inhomogeneous material, calculate the local thickness dimension of the inhomogeneous material respectively, and determine the total thickness dimension of the inhomogeneous material. If the product material is a multi-layer composite material, then through the timestamp information of the received ultrasonic signal, calculate the propagation time difference between the signal from the sensor to the upper surface and the lower surface of each layer respectively, and calculate the thickness dimension of each layer of the multi-layer composite material through the thickness dimension calculation formula. Calculate the dimension deviation between the product material dimension and the preset dimension, and the thickness dimension deviation between the product material thickness dimension and the preset thickness dimension. If the dimension deviation is greater than the preset dimension deviation threshold or the thickness dimension deviation exceeds the preset thickness dimension deviation threshold, then correct the thickness dimension of the product material by adjusting the production parameters in the production process, and the generated parameters include the production equipment accuracy, the gap of the mold, and the forming temperature.

[0047] Exemplarily, on the detection line, the thickness of a foam plastic material is being measured and needs to be compared with a homogeneous reference sample of a known foam plastic. Through the existing database, obtain the acoustic wave reflection signals of the foam plastic and the acoustic wave reflection signals of the known homogeneous reference sample. These signals are initially audio signals in the time domain. Use the short-time Fourier transform algorithm to convert these signals into frequency-domain data, and the spectral characteristics of the signals can be extracted from the frequency domain. If the spectra of the foam plastic and the reference sample are S t and S r respectively, calculate the covariance Cov(S t , S r ) between the spectra of the foam plastic and the reference sample, and the standard deviations and S t of the spectra of the foam plastic and the reference sample respectively. According to the covariance between the spectra of the foam plastic and the reference sample, and the standard deviations of the spectra of the foam plastic and the reference sample respectively, use the inhomogeneity coefficient formula Calculate the inhomogeneity coefficient H of the inhomogeneous material. This number can characterize the degree of inhomogeneity of the foam plastic material. If H = 1.2 is obtained through calculation, it indicates that the inhomogeneity of the foam plastic is relatively high. Calculate the phase difference of the main frequency based on the acoustic wave reflection signal spectrum of the foam plastic and the spectrum of the known homogeneous reference sample. The purpose of this operation is to infer the density difference between the foam plastic and the reference sample through the phase difference in the frequency domain. If the calculation result shows that the density of the foam plastic is less than that of the reference sample, the main frequency phase difference indicates D = -1, that is, the density of the foam plastic is low. If the calculation result shows that the density of the foam plastic is greater than that of the reference sample, the main frequency phase difference indicates D = 1, that is, the density of the foam plastic is high. Based on the inhomogeneity coefficient H = 1.2 and the velocity correction direction factor D = -1, use the velocity correction formula v a = v0×(1 + α·H·D) to correct the propagation velocity v a of the acoustic wave in the foam plastic, where v0 is the propagation velocity of the acoustic wave signal in the known homogeneous reference sample. If the acoustic wave propagation velocity v0 of the known homogeneous reference sample is 1500 m / s, the acoustic wave propagation velocity v a in the foam plastic is calculated to be 1450 m / s through the velocity correction formula, which indicates that the acoustic wave propagation velocity of the foam plastic is slightly slower than that of the reference sample. Use the timestamp information of the received ultrasonic signal to calculate the propagation time difference between the signal from the sensor to the upper and lower surfaces of the foam plastic. If the measured propagation time difference is Δt = 2 ms, through the thickness dimension calculation formula The thickness dimension of the foam plastic is calculated to be h=10mm. According to the heterogeneity coefficient of the heterogeneous material, it is determined whether there is a local density direction change area in the heterogeneous material. If the heterogeneity coefficient of the heterogeneous material 1.2 is greater than the preset coefficient threshold 1.0, then based on the local heterogeneity coefficient of the heterogeneous material and the speed correction direction factor, the local thickness dimension of the heterogeneous material is calculated respectively, and the total thickness dimension of the heterogeneous material is determined. If the foam plastic is a multi-layer composite material at this time, such as the outer layer is foam plastic and the inner layer is other materials, then the propagation time difference between the signal from the sensor to the upper and lower surfaces of each layer is calculated through the timestamp information of the received ultrasonic signal, and the thickness dimension of each layer is calculated respectively. If the thickness of the outer layer of foam plastic is calculated to be 3mm and the thickness of the inner layer material is 7mm, then the total thickness of the foam plastic material is 10mm. Finally, the deviation between the size of the foam plastic and the preset standard size is calculated. If the preset size is 10.5mm, the actual measured thickness size is 10mm, and the size deviation is calculated to be 0.5mm. If the deviation exceeds the preset dimensional deviation threshold of 0.3mm, or the thickness dimensional deviation exceeds the preset thickness dimensional deviation threshold of 0.2mm, the production parameters in the production process will be automatically adjusted. The adjusted parameters may include the accuracy of the production equipment, mold gap or molding temperature to ensure that the thickness of the foam plastic material meets the predetermined standards.

[0048] Among them, according to the heterogeneity coefficient of the heterogeneous material, it is judged whether there is a local density direction change area in the heterogeneous material, based on the local heterogeneity coefficient of the heterogeneous material and the velocity correction direction factor, the local thickness size of the heterogeneous material is calculated respectively, and the total thickness size of the heterogeneous material is determined.

[0049] If the heterogeneity coefficient of the heterogeneous material is greater than the preset coefficient threshold, it is determined that there is a local density direction change area in the heterogeneous material, and the echo signal of the heterogeneous material is time segmented according to the preset number of segments, and the main frequency phase difference of each echo signal is determined. The heterogeneity coefficient of each echo signal corresponding to the local heterogeneous material is calculated by the heterogeneity coefficient formula. According to the main frequency phase difference of each echo signal, the density of the local heterogeneous material corresponding to each echo signal and the density of the known homogeneous reference sample are judged, and the speed correction direction factor D of each echo signal corresponding to the local heterogeneous material is determined. According to the heterogeneity coefficient and speed correction direction factor of each echo signal corresponding to the local heterogeneous material, the propagation speed of the echo signal in each local heterogeneous material is corrected respectively using the speed correction formula. The thickness size of the local heterogeneous material is determined by the thickness size calculation formula, and the total thickness size of the heterogeneous material is calculated based on the thickness size of the local heterogeneous material.

[0050] Exemplarily, the thickness of a carbon fiber composite material with non-uniformity is being measured, and it is known that the density of the echo signals of this material varies at different positions. During the production process, it is found that the non-uniformity coefficient of this material exceeds the preset threshold 1, so further analysis and correction of the thickness measurement of this material are required. According to the ultrasonic signal data in the database, the calculated result of the non-uniformity coefficient of this carbon fiber composite material is 1.2, which is greater than the preset coefficient threshold of 1.0. This indicates that the non-uniformity of the carbon fiber composite material is relatively high, and there are regions with local density direction changes in the carbon fiber composite material. Therefore, the echo signals are processed by time segmentation. For example, the signals are divided into 5 segments, and each segment of the signal represents the acoustic wave reflection information at different positions of the material. Each signal segment is analyzed to extract its main frequency phase difference. The spectral data is obtained through short-time Fourier transform, and the main frequency phase difference of each segment of the signal is calculated. If the main frequency phase difference of the first segment of the signal is 0.01 radians, the second segment is 0.03 radians, the third segment is 0.02 radians, the fourth segment is 0.04 radians, and the fifth segment is 0.05 radians, this main frequency phase difference reflects the local acoustic wave propagation characteristics in each signal segment and judges the density difference of the material in these regions. The local non-uniformity coefficient corresponding to each segment of the echo signal is calculated using the non-uniformity coefficient formula. If the non-uniformity coefficient of the first segment is 1.5, the second segment is 2.0, the third segment is 1.8, the fourth segment is 2.2, and the fifth segment is 2.5, these non-uniformity coefficients indicate that as the echo signal progresses, the non-uniformity of the material gradually increases, especially in the latter part of the signal, where the non-uniformity coefficient increases significantly. According to the main frequency phase difference of each segment of the echo signal, it is further judged whether the density of the local non-uniform material corresponding to each segment of the signal is greater than the density of the known homogeneous reference sample. If the density of the reference sample is 1500 kg / m3. In the first segment of the signal, the main frequency phase difference is 0.01 radians, indicating that the density of this region is less than the density of the reference sample. Therefore, the velocity correction direction factor D for this segment is determined to be -1. In the second segment of the signal, the main frequency phase difference is 0.03 radians, and the density is greater than the density of the reference sample. Therefore, the velocity correction direction factor D for this segment is determined to be 1. The judgment results for the other segments are D = -1 for the third segment, D = 1 for the fourth segment, and D = 1 for the fifth segment. According to the non-uniformity coefficient and the velocity correction direction factor corresponding to each segment of the echo signal, the velocity correction formula is used to correct the propagation velocity of the echo signal in each local region. Given that the acoustic wave propagation velocity of the homogeneous reference sample is 1800 m / s, the propagation velocity of each segment of the signal is calculated through the correction formula. In the first segment of the signal, due to the relatively low non-uniformity coefficient and the velocity correction direction factor of -1, the corrected propagation velocity is 1700 m / s. In the second segment of the signal, the corrected propagation velocity is 1850 m / s, the third segment is 1750 m / s, the fourth segment is 1880 m / s, and the fifth segment is 1900 m / s.Using these corrected propagation speeds, the thickness dimensions are calculated. If the propagation time differences of the first echo signal are 1.2 ms, 1.5 ms for the second segment, 1.3 ms for the third segment, 1.6 ms for the fourth segment, and 1.7 ms for the fifth segment. Based on the propagation time differences and the corrected propagation speeds, the local thickness dimensions of each segment of the material can be calculated. For example, in the first segment, the local thickness calculated using the thickness dimension calculation formula is 2.04 mm, 2.78 mm in the second segment, 2.29 mm in the third segment, 3.01 mm in the fourth segment, and 3.23 mm in the fifth segment. Therefore, based on the local thickness dimensions of these five segments, the total thickness dimension of the inhomogeneous material is calculated to be 13.35 mm.

[0051] Step S105: Obtain the acoustic wave reflection signal of the product material after adjusting the process parameters and production parameters through an ultrasonic sensor, determine whether the production stability requirements have been met after the adjustment, and continuously adjust the process parameters and production parameters.

[0052] Obtain the acoustic wave reflection signal of the product material after adjusting the process parameters and production parameters through an ultrasonic sensor to obtain the new spectral characteristics of the acoustic wave reflection signal and the corresponding timestamp information. Based on the obtained new spectral characteristics of the acoustic wave reflection signal and the corresponding timestamp information, determine whether there is a deviation in the acoustic wave reflection signal, and whether the thickness dimension deviation between the thickness dimension of the product material and the preset thickness dimension exceeds the preset deviation threshold, and determine whether the production stability requirements have been met after the adjustment. If the production stability requirements have not been met yet, continue to adjust the process parameters and production parameters until the production stability requirements are met.

[0053] Exemplarily, in a certain inspection line, the standard specification of the produced foam plastic board is polyurethane material, with dimensions of 300mm x 200mm, a thickness dimension of 10mm, and a density of 0.3g / cm3. To ensure that the product meets the specifications, an ultrasonic sensor is used to monitor the acoustic reflection signal of the product in real time. When initial production started, the spectral characteristics of the measured acoustic reflection signal showed a frequency of 5MHz, a frequency peak of 5.2MHz, and an amplitude of 0.8. The propagation time difference of the ultrasonic signal was 0.012 seconds, and the calculated material thickness dimension was 9mm, with a deviation of 1mm. This deviation exceeded the preset threshold of 0.5mm, resulting in the production not meeting the stability requirements. Therefore, the production process and equipment parameters were adjusted, including improving the accuracy of the production equipment, adjusting the mold gap to a more precise 0.05mm, raising the forming temperature to 185°C, and adjusting the cooling rate to 1.2°C / min to ensure that the expansion and contraction of the material during forming were effectively controlled. After the adjustment, the acoustic reflection signal of the new product was continuously monitored through the ultrasonic sensor. The new spectral characteristics of the reflection signal showed a frequency of 5.05MHz, a frequency peak of 5.1MHz, and an amplitude of 0.85. The timestamp of the reflection signal showed a propagation time difference of 0.011 seconds, and the calculated product thickness dimension was 9.8mm. At this time, the thickness dimension deviation was 0.2mm, which was lower than the preset deviation threshold of 0.5mm. By comparing with the preset standard, the thickness dimension of the product material and the spectral characteristics of the acoustic reflection signal were already close to the preset requirements. After multiple rounds of adjustment and monitoring, the production process finally met the stability requirements, ensuring that the product quality met the standards and the acoustic reflection signal no longer showed deviations.

[0054] A material size automatic measurement device in this embodiment may specifically include:

[0055] A processor and a memory; at least one instruction is stored in the memory, and the instruction is executed by the processor to enable the device to execute the material size automatic measurement method described in the above method embodiment.

[0056] The above description is only a preferred embodiment of the present application and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by the specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the technical solutions formed by mutually replacing the above features with the (but not limited to) technical features with similar functions disclosed in the present application.

Claims

1. A method for automatically measuring material dimensions, characterized in that: The method comprises: According to the product materials and sound wave emission parameters of different specifications and standards, as well as the corresponding sound wave reflection signal receiving effects, the sound wave emission parameters with the best sound wave reflection signal receiving effect of the current product material are predicted, ultrasonic signals are emitted, and sound wave reflection signals and corresponding timestamp information are obtained; The sound wave reflection signal is separated by an independent component analysis algorithm, and the type of the sound wave reflection source of the sound wave reflection signal is identified, and the frequency spectrum characteristics of the sound wave reflection signal whose sound wave reflection source is the product material are determined; According to the spectral characteristics of the corrected product material acoustic wave reflection signal and the spectral characteristics of the reference acoustic wave reflection signal of the product material specification standard, the cause of the deviation of the product material acoustic wave reflection signal is identified, and used to guide the adjustment of the process parameters of the product production equipment; Determine the heterogeneity coefficient of the heterogeneous material through the acoustic wave reflection signal of the heterogeneous material and the acoustic wave reflection signal of the known homogeneous reference sample, correct the propagation speed of the acoustic wave signal in the current heterogeneous material, and calculate the thickness size of the single-layer product material and the multi-layer combination material; The ultrasonic sensor is used to obtain the acoustic reflection signal of the product material after the process parameters and production parameters are adjusted, to determine whether the production stability requirements have been met after adjustment, and to continuously adjust the process parameters and production parameters.

2. The method according to claim 1, wherein: The method predicts the acoustic wave emission parameters with the best acoustic wave reflection signal reception effect for the current product material according to the product materials and acoustic wave emission parameters of different specifications and standards, and the corresponding acoustic wave reflection signal reception effect, emits an ultrasonic signal and obtains the acoustic wave reflection signal and the corresponding timestamp information, including: Through the ultrasonic sensor, the product materials of different specifications and standards receive ultrasonic signals, record the ultrasonic wave emission parameters of the ultrasonic sensor, and mark the receiving effect of the sound wave reflection signal. The specifications and standards include material type, size, thickness, number of layers, and density. The sound wave emission parameters include emission frequency, pulse width and emission power. The receiving effects include excellent, good and poor. According to the product materials and sound wave emission parameters of different specifications and standards, as well as the corresponding receiving effects of sound wave reflection signals, a recurrent neural network is used for model training to build a sound wave emission parameter prediction model, determine the sound wave emission parameters with the best receiving effect of the sound wave reflection signal of the current product material, and set the sound wave emission parameters of the ultrasonic sensor. Debug the position of the ultrasonic sensor and install the ultrasonic sensor above the detection line. The camera above the detection line is used to obtain and generate the product material image on the line. If the product material is directly below the ultrasonic sensor, the ultrasonic sensor is used to emit ultrasonic signals to the product material on the detection line according to the set sound wave emission parameters and receive the sound wave reflection signal and the corresponding timestamp information.

3. The method according to claim 1, wherein: The method of separating the acoustic wave reflection signal by an independent component analysis algorithm, identifying the acoustic wave reflection source type of the acoustic wave reflection signal, and determining the frequency spectrum characteristics of the acoustic wave reflection signal whose acoustic wave reflection source is the product material includes: According to the received sound wave reflection signal, the sound wave reflection signal is separated by an independent component analysis algorithm to obtain the sound wave reflection signals of different sound wave reflection sources; the time domain audio signal of the sound wave reflection signal is converted into frequency domain data by a short-time Fourier transform algorithm, the spectrum characteristics of the sound wave reflection signal are extracted, and saved in the sound wave reflection monitoring database, the spectrum characteristics include frequency, frequency peak and amplitude; through the sound wave reflection signal monitoring database, the spectrum characteristics of the sound wave reflection signal of different sound wave reflection sources are obtained, and the type of the sound wave reflection source is marked, the random forest algorithm is used for model training, and a sound wave reflection source identification model is constructed to identify the type of the sound wave reflection source of the sound wave reflection signal, and determine that the sound wave reflection source is the spectrum characteristics of the sound wave reflection signal of the product material; the detection environment data is obtained, and the frequency drift and amplitude drift are calculated according to the detection environment data, the frequency and amplitude of the sound wave reflection signal are corrected, and the spectrum characteristics of the sound wave reflection signals of different sound wave reflection sources after correction are obtained, and saved in the sound wave reflection signal monitoring database, and the detection environment data include temperature, humidity and air pressure.

4. The method according to claim 3, wherein: The method of calculating the frequency drift and the amplitude drift according to the detection environment data and correcting the frequency and amplitude of the sound wave reflection signal comprises: The temperature, humidity and air pressure data in the detection line are obtained in real time through the temperature sensors, humidity sensors and air pressure sensors installed on the detection line. The frequency drift calculation formula is used Determine the frequency drift Δf under the current detection environment, and use the amplitude drift calculation formula ΔA=A1·(n1·sin(T)+n2·H-n3·log(P)) to determine the amplitude drift ΔA of the current detection environment, where T is temperature, H is humidity, P is air pressure, f1 and A1 are the frequency value measured under standard conditions and the amplitude value measured under standard conditions, respectively, l1, l2, and l3 are the influence coefficients of the detection environment on the frequency, and n1, n2, and n3 are the influence coefficients of the detection environment on the amplitude, which are obtained by experimental fitting of historical data; according to the spectral characteristics, frequency drift, and amplitude drift of the acoustic wave reflection signal, correct the frequency and amplitude of the acoustic wave reflection signal.

5. The method according to claim 1, wherein: The method of identifying the cause of the deviation of the acoustic wave reflection signal of the product material according to the spectral characteristics of the corrected acoustic wave reflection signal of the product material and guiding the adjustment of the process parameters of the production equipment includes: The spectral characteristics of the sound wave reflection signals of different sound wave reflection sources after correction are obtained through the sound wave reflection signal monitoring database, and the similarity between the spectral characteristics of the obtained product material sound wave reflection signals and the spectral characteristics of the benchmark sound wave reflection signals of the product material specification standard is calculated through the cosine similarity calculation method; if the similarity is lower than the preset similarity threshold, the spectral characteristics of the obtained product material sound wave reflection signals are obtained, and the reasons for the deviation of the product material sound wave reflection signals are marked, and the model is trained using the random forest algorithm to construct a sound wave reflection signal deviation cause identification model to identify the reasons for the deviation of the product material sound wave reflection signals, which include but are not limited to product material density deviation and shape defects, and shape defects include but are not limited to cracks and potholes; according to the reasons for the deviation of the sound wave reflection signals, the process parameters of the production equipment are adjusted, and the process parameters include but are not limited to temperature, cooling rate, injection pressure of the injection molding machine, injection speed and injection volume.

6. The method according to claim 1, wherein: The method determines the heterogeneity coefficient of the heterogeneous material by using the acoustic wave reflection signal of the heterogeneous material and the acoustic wave reflection signal of the known homogeneous reference sample, corrects the propagation speed of the acoustic wave signal in the current heterogeneous material, and calculates the thickness size of the single-layer product material and the multi-layer composite material, including: If the material with product material density deviation is a non-homogeneous material, the sound wave reflection signal of the corrected non-homogeneous material and the sound wave reflection signal of the known homogeneous reference sample are obtained through the sound wave reflection signal monitoring database, and the time domain audio signal of the sound wave reflection signal is converted into frequency domain data using the short-time Fourier transform algorithm. The non-homogeneous materials include but are not limited to foam plastics, carbon fiber composite materials and glass fiber composite materials; the non-homogeneous coefficient formula is used Calculate the heterogeneity coefficient H of heterogeneous materials, where S t and S r are the acoustic reflection signal spectrum of the inhomogeneous material to be tested and the acoustic reflection signal spectrum of the known homogeneous reference sample, Cov(S t ,S r ) is the spectrum S t and S r The covariance of The spectrum S t The standard deviation of The spectrum S r The standard deviation of the sound wave reflection signal of the inhomogeneous material and the sound wave reflection signal of the known homogeneous reference sample is calculated; the main frequency phase difference between the sound wave reflection signal spectrum of the inhomogeneous material and the sound wave reflection signal spectrum of the known homogeneous reference sample is extracted according to the frequency domain data of the sound wave reflection signal of the inhomogeneous material and the sound wave reflection signal spectrum of the known homogeneous reference sample, and the size of the density of the inhomogeneous material and the density of the known homogeneous reference sample is judged based on the main frequency phase difference, and the speed correction direction factor D is determined. If the density of the inhomogeneous material is greater than the density of the known homogeneous reference sample, D is 1, and if the density of the inhomogeneous material is less than the density of the known homogeneous reference sample, D is -1; according to the heterogeneity coefficient and the speed correction direction factor, the speed correction formula is used to correct the propagation speed v of the sound wave signal in the current inhomogeneous material a , where v0 is the propagation speed of the acoustic wave signal in a known homogeneous reference sample; the propagation time difference ΔT between the signal from the sensor to the upper surface and the lower surface is calculated using the timestamp information of the received ultrasonic signal; the thickness dimension is calculated using the formula Determine the thickness dimension h of the product material, where v0 is the propagation speed of the acoustic wave signal in a known homogeneous reference sample; determine whether there is a local density direction change area in the heterogeneous material based on the heterogeneity coefficient of the heterogeneous material, calculate the local thickness dimension of the heterogeneous material based on the local heterogeneity coefficient of the heterogeneous material and the velocity correction direction factor, and determine the total thickness dimension of the heterogeneous material; if the product material is a multi-layer composite material, calculate the propagation time difference between the signal from the sensor to the upper surface of each layer and the lower surface of each layer through the timestamp information of the received ultrasonic signal, and calculate the thickness dimension of each layer of the multi-layer composite material through the thickness dimension calculation formula; calculate the dimensional deviation between the product material dimension and the preset dimension, and the thickness dimension deviation between the product material thickness dimension and the preset thickness dimension. If the dimensional deviation is greater than the preset dimensional deviation threshold or the thickness dimension deviation exceeds the preset thickness dimension deviation threshold, the thickness dimension of the product material is corrected by adjusting the production parameters in the production process, and the generated parameters include the production equipment accuracy, the gap of the mold and the molding temperature.

7. The method according to claim 6, wherein: The method of judging whether there is a local density direction change region in the heterogeneous material according to the heterogeneity coefficient of the heterogeneous material, calculating the local thickness size of the heterogeneous material based on the local heterogeneity coefficient of the heterogeneous material and the speed correction direction factor, and determining the total thickness size of the heterogeneous material includes: If the heterogeneity coefficient of the heterogeneous material is greater than the preset coefficient threshold, it is determined that there is a local density directional change area in the heterogeneous material, and the echo signal of the heterogeneous material is time-segmented according to the preset number of segments, and the main frequency phase difference of each echo signal is determined; the heterogeneity coefficient of each echo signal corresponding to the local heterogeneous material is calculated by the heterogeneity coefficient formula; according to the main frequency phase difference of each echo signal, the density of the local heterogeneous material corresponding to each echo signal and the density of the known homogeneous reference sample are determined, and the speed correction direction factor D of the local heterogeneous material corresponding to each echo signal is determined; according to the heterogeneity coefficient and speed correction direction factor of the local heterogeneous material corresponding to each echo signal, the propagation speed of the echo signal in each local heterogeneous material is corrected respectively using the speed correction formula; the thickness size of the local heterogeneous material is determined respectively by the thickness size calculation formula, and the total thickness size of the heterogeneous material is calculated based on the thickness size of the local heterogeneous material.

8. The method according to claim 1, wherein: The method of obtaining the acoustic wave reflection signal of the product material after the process parameters and production parameters are adjusted by the ultrasonic sensor, judging whether the production stability requirements have been met after the adjustment, and continuously adjusting the process parameters and production parameters includes: Acquire the acoustic wave reflection signal of the product material after adjusting the process parameters and production parameters through the ultrasonic sensor, and obtain new spectrum characteristics of the acoustic wave reflection signal and corresponding timestamp information; based on the new spectrum characteristics of the acoustic wave reflection signal and the corresponding timestamp information, judge whether there is a deviation in the acoustic wave reflection signal, and whether the thickness dimension deviation between the thickness dimension of the product material and the preset thickness dimension exceeds the preset deviation threshold, and judge whether the production stability requirement has been met after adjustment; if the production stability requirement has not been met, continue to adjust the process parameters and production parameters until the production stability requirement is met.

9. An automatic material size measuring device, characterized in that: The device comprises: A processor and a memory; the memory stores at least one instruction, the instruction is executed by the processor, so that the device executes the automatic material size measurement method as described in any one of claims 1 to 7.

Images