Method and system for detecting waterproof performance of communication equipment

Abstract

The invention provides a method and system for detecting the waterproof performance of communication equipment, and the method comprises the steps: monitoring the motion track and aggregation state of liquid drops on the surface of the equipment in real time through employing a high-speed camera technology in an atomization process, and recognizing the aggregation region of the liquid drops through an image processing algorithm according to the image information obtained through high-speed camera shooting; dynamically adjusting the angle and the position of the atomizing nozzle according to the identified gathering area and the gathering degree thereof so as to avoid the situation that the water vapor gathering affects the waterproof performance detection result; a plurality of temperature and humidity sensors are arranged in the atomization cavity, temperature and humidity data of the atomization environment are collected in real time, and the influence of the data on the atomization effect is analyzed; a correlation model between the temperature and humidity data and the droplet evaporation rate is established through a machine learning algorithm, and the model considers and calculates a fluid dynamics simulation result and an image monitoring result at the same time.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method and system for detecting waterproof performance of communication equipment. Background Art

[0002] When testing the water resistance of communication equipment, it is necessary to use special spray equipment to atomize it to simulate the water vapor conditions that the equipment may encounter in the actual use environment. However, in the atomization process, how to ensure the accuracy of atomization is a key technical issue. Insufficient atomization accuracy may lead to the following problems: First, the spray coverage is uneven, some parts are greatly affected by water vapor, while other parts are rarely exposed to water vapor, which will lead to deviations in the test results. Second, the particle size distribution of atomized droplets is unreasonable. Droplets that are too large may form a water film on the surface of the equipment, causing local immersion, while droplets with too small a particle size are difficult to effectively simulate the actual environment. Third, improper spray angles of atomizing nozzles may cause water vapor to gather on the surface of the equipment to form water droplets, which cannot truly reflect the waterproof performance of the equipment in the atomized environment. Therefore, how to reasonably control the various parameters of the atomization equipment according to the characteristics of the equipment to be tested and improve the atomization accuracy is a technical problem that needs to be solved in the waterproof testing of communication equipment. Only by solving the problem of atomization accuracy can the authenticity and reliability of the waterproof test results be ensured, and strong data support can be provided for the waterproof design of communication equipment. Summary of the invention

[0003] The present invention provides a method for detecting the waterproof performance of a communication device, which mainly includes:

[0004] Obtain a three-dimensional model of the communication device to be tested, which includes the structural characteristics and material property information of the device. Based on the three-dimensional model, a computational fluid dynamics simulation method is used to simulate the distribution of water vapor on the device surface under different atomization parameters. Based on the computational fluid dynamics simulation results, the optimal atomization parameter combination is determined. The atomization parameters include atomization liquid flow rate, atomization nozzle diameter and injection pressure to ensure that the uniformity of atomization coverage and droplet size distribution meet the preset standards, where the droplet size distribution standards include average particle size and particle size distribution range. During the atomization process, high-speed camera technology is used to monitor the movement trajectory and aggregation state of droplets on the device surface in real time. Based on the image information obtained by the high-speed camera, the aggregation area of ​​the droplets is identified through an image processing algorithm. According to The identified aggregation areas and their aggregation degrees are used to dynamically adjust the angle and position of the atomizing nozzle to avoid the situation where water vapor aggregation affects the waterproof performance test results; multiple temperature and humidity sensors are arranged in the atomizing chamber to collect temperature and humidity data of the atomizing environment in real time, and analyze the impact of these data on the atomization effect; through machine learning algorithms, a correlation model between temperature and humidity data and droplet evaporation rate is established, which takes into account both computational fluid dynamics simulation results and image monitoring results; using the established correlation model, the evaporation process of droplets is dynamically predicted, and the atomization parameters are adjusted accordingly to ensure that the attachment time of droplets on the device surface matches the pre-defined standard usage environment, where the standard usage environment is determined based on the expected usage scenarios and conditions of the device.

[0005] The present invention provides a communication equipment waterproof performance detection system, which mainly includes: a three-dimensional model acquisition module for acquiring equipment structural characteristics and material property information; a computational fluid dynamics simulation module for simulating water vapor distribution and determining optimal atomization parameters; a high-speed camera monitoring module for real-time monitoring of droplet motion trajectory and aggregation state; a temperature and humidity data acquisition module for collecting temperature and humidity data of the atomization environment; a machine learning modeling module for establishing a temperature and humidity correlation model with droplet evaporation rate; a dynamic adjustment control module for predicting droplet evaporation and adjusting atomization parameters. The technical solution provided by the embodiment of the present invention may include the following beneficial effects:

[0006] The present invention discloses a method for detecting the waterproof performance of communication equipment. The method obtains a three-dimensional model of the equipment and optimizes the atomization parameters by computational fluid dynamics simulation to achieve uniform coverage. During the detection process, high-speed video is used to monitor the movement of droplets, and image processing is used to identify the aggregation area and dynamically adjust the position of the nozzle. At the same time, temperature and humidity sensors are arranged to collect environmental data, and a machine learning model is established to predict the evaporation process of droplets. The present invention combines simulation optimization, real-time monitoring and intelligent prediction to achieve precise control of waterproof performance detection, ensure that the detection environment matches the actual use scenario, and improve the reliability and representativeness of the detection results. This intelligent detection method not only improves the detection efficiency, but also provides strong support for the evaluation and improvement of the waterproof performance of communication equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 The present invention is a flow chart of a method for detecting waterproof performance of communication equipment.

[0008] Figure 2 A schematic diagram of a communication equipment waterproof performance detection method and system according to the present invention.

[0009] Figure 3 It is another schematic diagram of a communication equipment waterproof performance detection method and system of the present invention.

[0010] Figure 4 The present invention is a schematic structural diagram of a communication equipment waterproof performance detection system. DETAILED DESCRIPTION

[0011] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0012] like Figure 1-3 In this embodiment, a method for detecting waterproof performance of a communication device may specifically include:

[0013] Step S101, obtaining a three-dimensional model of the communication device to be tested, the three-dimensional model including the structural characteristics and material property information of the device; based on the three-dimensional model, a computational fluid dynamics simulation method is used to simulate the distribution of water vapor on the device surface under different atomization parameters.

[0014] Obtain the three-dimensional CAD model data containing the structural characteristics and material property information of the communication equipment to be tested, and convert it into STL format; import the converted STL format three-dimensional model data into ANSYSFluent computational fluid dynamics simulation software, set boundary conditions and solution parameters, including defining fluid properties, specifying inlets and outlets, and dividing the grid; perform multiple groups of numerical simulation calculations in ANSYSFluent software for different ambient temperatures and humidity atomization parameters to obtain water vapor concentration distribution results under different working conditions; extract the water vapor concentration numerical data on the equipment surface grid unit in each group of simulation results and save them as CSV format files ; Use Python's Matplotlib library to visualize the water vapor concentration data, generate a water vapor distribution cloud map on the device surface, and use a suitable color scheme to indicate the concentration; Use the K-means clustering algorithm to perform cluster analysis on the pixels in the water vapor distribution cloud map, and divide the pixels into different clustering areas based on the water vapor concentration value; Set a water vapor concentration threshold. If more than 20% of the pixel concentration values ​​in the clustered water vapor cluster area exceed the threshold, it is determined that the device has a condensation risk under the atomization parameters, and the device structure needs to be further optimized, including adding drainage holes or selecting materials with better surface hydrophobicity.

[0015] For example, first use SolidWorks software to open the 3D CAD model of the communication equipment and export the STL format file. Import the STL file in ANSYS Fluent software, define the air as an ideal gas, set the inlet temperature to 25°C, the humidity to 60%, the outlet pressure to 0 Pa, and use tetrahedral meshing, with the number of meshes being about 5 million. For 9 operating conditions with ambient temperatures of 15°C, 25°C, and 35°C and relative humidity of 40%, 60%, and 80%, numerical simulations are performed respectively. The solver uses the pressure-based implicit algorithm, and the turbulence model uses the k-epsilon model. Each operating condition is iterated for 1000 steps to reach convergence. The post-processing module is used to extract the water vapor concentration value of the grid unit on the surface of the equipment and export it to CSV format. Use Python's Matplotlib library to read the CSV file, and draw the water vapor distribution cloud map on the surface of the equipment in the form of a three-dimensional scatter plot. The color mapping uses the viridis scheme, and the concentration range is 0-25g / m³. K-means clustering is performed on the pixels in the cloud map, and the number of clusters k=3 and the maximum number of iterations max_iter=300 are set. The RGB values ​​of the pixels are used as feature vectors to obtain three clustering areas. The percentage of pixels with water vapor concentration exceeding 20g / m³ in each area is counted. If it exceeds 20%, it is considered that there is a risk of condensation under this condition. For example, under the condition of temperature 25°C and humidity 80%, the proportion of high-concentration pixels in the water vapor aggregation area reaches 35%. Therefore, it is necessary to add a drainage hole with a diameter of 3mm at the bottom of the device, and spray a hydrophobic nano-coating on the surface of the shell to reduce the surface tension. After optimizing the design, simulation verification is performed again until the risk of condensation is eliminated.

[0016] Step S102, based on the computational fluid dynamics simulation results, determine the optimal atomization parameter combination, the atomization parameters include atomization liquid flow rate, atomization nozzle diameter and injection pressure, to ensure that the atomization coverage uniformity and droplet size distribution meet the preset standards, wherein the droplet size distribution standard includes the average particle size and the particle size distribution range.

[0017] Obtain the physical properties and process requirements of the atomizing liquid to be optimized. The physical properties include viscosity, density and surface tension, and the process requirements include the atomizing liquid flow range and droplet size distribution requirements. Based on the physical properties and process requirements, establish a computational fluid dynamics simulation model of the atomization process in ANSYS Fluent software, and reasonably set boundary conditions and solution parameters. For the simulation model, set different combinations of atomizing liquid flow, nozzle diameter and injection pressure parameters, and carry out multiple groups of numerical simulation calculations. Extract droplet size distribution data and spatial coverage distribution data from each group of simulation results, compare the droplet size distribution data with the preset particle size standard, and calculate the root mean square value of the particle size deviation of the parameter combination. If the root mean square of the deviation is less than the preset threshold, the current parameter combination is output as the optimal solution for particle size optimization; otherwise, the particle swarm optimization algorithm is used to iteratively optimize the parameter combination with the goal of minimizing the root mean square of the deviation; on the basis of the optimal parameter combination for particle size, the atomization coverage distribution data is discretized, the variance in the coverage area is calculated, and the inverse of the variance is used to characterize the coverage uniformity; if the uniformity index is greater than the preset threshold, the current parameter combination is determined to be the comprehensive optimal solution; otherwise, the uniformity index is used as the optimization target, and the genetic algorithm is used to further optimize the parameter combination; based on the comprehensive optimal solution, the structural parameters of the atomization device are determined, and the three-dimensional structural design of the atomization device is completed in the CAD software to form a design scheme for the atomization system.

[0018] For example, a three-dimensional simulation model of atomized liquid flow was established in ANSYS Fluent software, and the atomized liquid was set to water, with a viscosity of 0.001 Pa·s, a density of 998 kg / m³, and a surface tension of 0.072 N / m. The nozzle diameter range is 2-0 mm, the atomized liquid flow range is 10-100 mL / min, and the injection pressure range is 5-0 MPa. The pressure-velocity coupling algorithm was adopted, the k-epsilon model was selected as the turbulence model, the pressure discretization format was PRESTO, and the momentum and turbulence discretization format was the second-order upwind format. The central composite design was adopted in the Design of Experiments module to generate 15 groups of parameter combinations. The droplet size distribution data in each group of simulation results was extracted and compared with the target particle size range of 20 μm±5 μm, and the root mean square of the deviation was calculated. The particle swarm optimization algorithm was adopted, the number of particles was set to 20, the learning factor c1=c2=0, the inertia weight w=8, the maximum number of iterations was 50, and the optimal parameter combination was output when the root mean square of the deviation was less than 0. On this basis, the uniformity of atomization coverage was evaluated, the droplet number density in the cross-sectional direction was discretized into a grid of 100×100, and the standard deviation of the grid droplet number density was calculated. The uniformity index was the inverse of the standard deviation. The genetic algorithm was used for further optimization, and the population number was set to 20, the crossover probability was 8, the mutation probability was 1, and the maximum evolutionary generation was 30. When the uniformity index was greater than 1000, the optimal parameter combination was determined to be: nozzle diameter 6 mm, atomization liquid flow rate 60 mL / min, and injection pressure 2 MPa. A three-dimensional structural model of the atomization device was established in SolidWorks, the nozzle aspect ratio was 20, and the diffusion angle was 60°, forming the final atomization system design.

[0019] Step S103, during the atomization process, high-speed camera technology is used to monitor the movement trajectory and aggregation state of droplets on the device surface in real time; based on the image information obtained by the high-speed camera, the aggregation area of ​​the droplets is identified through an image processing algorithm.

[0020] During the atomization process, a high-speed camera is used to monitor the motion trajectory and aggregation state of the droplets on the surface of the equipment in real time, and obtain image information of the droplet motion and aggregation; the obtained droplet image information is preprocessed, including using a median filter to denoise the image and using a histogram equalization algorithm to enhance the image, so as to improve the image quality and prepare for subsequent recognition; a droplet recognition model is constructed based on the characteristics of the droplet motion trajectory and aggregation state, and the model is trained using a support vector machine algorithm to obtain a trained droplet recognition model; the trained droplet recognition model is used to recognize the preprocessed droplet image to determine whether there is a droplet aggregation area in the image, and if so, the location of the aggregation area is determined. the position and range of the droplet aggregation area; according to the identified droplet aggregation area, the area of ​​the aggregation area, the number of droplets and the distribution density parameters are calculated to obtain a quantitative description of the droplet aggregation state; the quantitative description of the droplet aggregation state is compared with the preset threshold value, if it exceeds the threshold value, it is judged as abnormal aggregation, triggering an alarm signal and notifying relevant personnel to handle it; at the same time, according to the degree of deviation between the droplet aggregation state parameters and the preset threshold value, the working parameters of the atomization equipment are dynamically adjusted, including changing the atomization intensity by adjusting the vibration frequency of the atomizer, and controlling the droplet particle size by adjusting the droplet injection pressure, so that the droplet aggregation state is maintained within the normal range, avoiding the occurrence of abnormal aggregation, and ensuring the stability and uniformity of the atomization effect.

[0021] For example, in the monitoring of the atomization process, a high-speed camera is used to collect the motion trajectory and aggregation state of the droplets in real time. The camera frame rate is set to 1000 fps and the resolution is 1280×1024. The collected images are sequentially subjected to median filtering denoising and histogram equalization enhancement processing. The filter window size is 5×5 and the number of equalization levels is 256. The feature vector is constructed based on the roundness, area and movement speed of the droplet contour. The support vector machine model of the radial basis kernel function is used for training. The training set sample size is 500. The model parameters c=10 and g=01 are determined by cross-validation. The final model has an accuracy of 98% on the test set. The trained model is used to identify the image frame by frame, and the aggregation area is determined by contour analysis. The number of droplets in the area, the average particle size, the distribution density and other parameters are calculated. When the number of droplets exceeds 50 / cm², or the average particle size is less than 20μm, it is judged as abnormal aggregation and an alarm signal is sent to the control system. At the same time, according to the degree of deviation between the aggregation parameters and the preset threshold, the fuzzy PID control algorithm is used to dynamically adjust the atomizer vibration frequency and injection pressure, with a frequency adjustment range of 50-200 Hz and a pressure adjustment range of 1-5 MPa. Through real-time monitoring and feedback control, abnormal droplet aggregation is effectively avoided, and the uniformity of atomized liquid distribution is maintained at more than 95%, fully ensuring the stability of the atomization effect.

[0022] Step S104, dynamically adjusting the angle and position of the atomizing nozzle according to the identified aggregation area and its aggregation degree, so as to avoid the situation where water vapor aggregation affects the waterproof performance test result.

[0023] The image of the water vapor gathering area is obtained, and the image segmentation algorithm based on region growth is used to segment the image to obtain the water vapor gathering area. According to the segmented water vapor gathering area, the water vapor density is estimated by calculating the average gray value of the pixels in each area to obtain the degree of water vapor gathering. The angle and position parameters of the current atomizer nozzle are obtained. According to the water vapor gathering area, the degree of gathering and the current angle position of the atomizer nozzle, a model for adjusting the angle and position of the atomizer nozzle based on multivariate linear regression is established. The model is trained by the least squares method to obtain the optimized model parameters. The water vapor gathering area, the degree of gathering and the current angle position of the atomizer nozzle are input into the trained adjustment model, and the optimized angle and position parameters of the atomizer nozzle are obtained by model calculation. The current angle position of the atomizer nozzle is compared with the optimized angle position, and the Euclidean distance is calculated. If the distance is greater than the preset threshold, the atomizer nozzle motor is controlled to adjust the angle and position until it is adjusted in place. After the dynamic adjustment of the angle position of the atomizer nozzle is completed, the product surface image is obtained, and the Canny edge detection algorithm is used to detect the water vapor gathering on the product surface. The connectivity of the edge is analyzed to determine whether there is water vapor gathering. If there is no water vapor accumulation, it is considered that the adjustment effect meets the requirements and the subsequent waterproof performance test will continue; if there is water vapor accumulation, return to step 1 to re-acquire the water vapor accumulation area image and perform the next round of dynamic adjustment of the atomizing nozzle angle position until the water vapor accumulation is eliminated.

[0024] For example, during the atomization process, a high-speed camera is used to collect images of the water vapor gathering area. The camera frame rate is set to 500 fps and the resolution is 1024×768. The collected image is preprocessed, and the mean filter is used to remove noise. The filter window size is 3×3. Then the image is segmented using the regional growing algorithm. The growth criterion is that the grayscale difference is less than 10, and the number of growths is 5 times to obtain the water vapor gathering area. For the segmented water vapor gathering area, the average grayscale value of the pixel points in each area is calculated. The grayscale value is linearly related to the water vapor density. The larger the grayscale value, the higher the water vapor density. The degree of water vapor gathering is obtained accordingly. The pitch angle and horizontal position parameters of the current atomizing nozzle are obtained through the RS-485 bus, which are recorded as α and d respectively. Based on the water vapor gathering area, the degree of gathering S, and α and d, the angle and position adjustment model of the atomizer nozzle is established: Δα=k1S+k2α+b1, Δd=k3S+k4d+b2, where Δα and Δd are the angle and position adjustment values, and k1, k2, k3, k4 and b1, b2 are model parameters. The least squares method is used to train the 100 sets of collected data to obtain the optimized model parameters. The currently collected water vapor gathering area, degree of gathering, and α and d are input into the trained adjustment model to obtain the optimized angle α' and position d'. The Euclidean distance between the current angle position and the optimized angle position is calculated. If the distance is greater than 2° or 20 mm, the atomizer nozzle is adjusted by controlling the stepper motor with an adjustment accuracy of 1° and 1 mm until the optimized position is reached. After completing the dynamic adjustment, edge detection is performed on the product surface image using the Canny algorithm. The standard deviation of the Gaussian filter is 4, the high threshold of the dual threshold algorithm is 200, and the low threshold is 100. The 8-neighborhood connectivity analysis is used to determine whether there is water vapor accumulation. If not, it is considered that the adjustment meets the requirements. If so, it returns to readjust. After 3 iterative adjustments, the water vapor accumulation area is reduced by 85%, meeting the waterproof performance test requirements.

[0025] Step S105, arranging a plurality of temperature and humidity sensors in the atomization chamber, collecting temperature and humidity data of the atomization environment in real time, and analyzing the influence of these data on the atomization effect.

[0026] The temperature and humidity data of the atomization environment are obtained by collecting the temperature and humidity data through the temperature and humidity sensors arranged in the atomization chamber; the temperature and humidity data are preprocessed, and the moving average filtering algorithm is used to eliminate outliers and noise data; the ARIMA time series analysis model is used to model the preprocessed temperature and humidity data to obtain the time trend, periodicity and mutation point characteristics of the temperature and humidity data; the atomization effect evaluation index is obtained, and the atomization effect evaluation index includes the droplet size distribution and the atomization rate; the Pearson correlation coefficient is used to analyze the correlation between the temperature and humidity data and the atomization effect evaluation index, and a multivariate linear regression model of temperature and humidity and atomization effect is established; using A The RIMA time series analysis model makes short-term predictions on temperature and humidity. If the predicted value exceeds the preset threshold range, the alarm mechanism is triggered and the working parameters of the atomization equipment are adjusted according to the preset strategy. The temperature and humidity data under different seasons and weather conditions are obtained, and the temperature and humidity data are divided into different working conditions using the K-means clustering algorithm. For each working condition, a mapping table between temperature and humidity and atomization equipment parameters is established to optimize the control strategy of the atomization equipment. The support vector regression algorithm is used to establish a nonlinear prediction model of temperature and humidity data and atomization effect evaluation indicators. According to the prediction results of the nonlinear prediction model, early warning and dynamic optimization adjustment of the atomization effect are achieved.

[0027] Exemplarily, an SHT31 temperature and humidity sensor is arranged in the atomization chamber with a sampling frequency of 1 Hz, and the temperature and humidity data are collected in real time through the RS-485 bus. The collected temperature and humidity data are preprocessed, and a 5-point moving average filtering algorithm is used with a filter window size of 5, which effectively eliminates random noise and pulse interference. The filtered temperature and humidity time series data are modeled using the ARIMA (1,1,1) model, and the model order is determined by the AIC criterion to obtain the long-term trend, 24-hour periodicity and mutation point of temperature and humidity changes. The droplet size distribution is measured using a Malvern Spraytec laser particle size analyzer with a sampling time of 10 s and repeated 3 times to take the average value. The atomization rate is measured by weighing, and the atomization amount within 1 min is collected and repeated 3 times to take the average value. The Pearson correlation coefficients of temperature, humidity, particle size D50 and atomization rate were calculated. The temperature was negatively correlated with D50 (r=-0.85), and the humidity was positively correlated with the atomization rate (r=0.92). Based on this, a multivariate linear regression model of temperature, humidity and atomization effect was established. The ARIMA model was used to make a rolling forecast of temperature and humidity for the next 15 minutes. When the temperature forecast value exceeded 35℃ or the humidity forecast value exceeded 85%, an alarm was triggered and the atomizer frequency was adjusted from 100 kHz to 80 kHz. The temperature and humidity data of 30 consecutive days under typical working conditions in spring, summer, autumn and winter were obtained and clustered into 4 categories using the K-means algorithm. A mapping table of temperature, humidity, atomizer frequency and nozzle diameter was established based on the clustering results to achieve adaptive optimization of working conditions. The support vector regression algorithm was used to establish a nonlinear prediction model of temperature, humidity, particle size distribution and atomization rate. The Gaussian kernel function was selected, the penalty factor C=10, the training set sample size was 500, and the prediction mean square error was less than 5%. Based on the prediction results, when the predicted particle size D50 is less than 40 μm or the atomization rate is lower than 1 mL / min, an early warning is given 5 minutes in advance, and the nebulizer frequency and nozzle diameter are dynamically adjusted to continuously optimize the atomization effect.

[0028] Step S106, establishing a correlation model between temperature and humidity data and droplet evaporation rate through a machine learning algorithm, the model taking into account both computational fluid dynamics simulation results and image monitoring results.

[0029] Acquire temperature and humidity data and droplet image monitoring data during the droplet evaporation process, preprocess the data, extract key features, and construct a training data set; use computational fluid dynamics simulation software to simulate the droplet evaporation process and obtain theoretical predictions of droplet evaporation rate under different temperature and humidity conditions; compare the simulation prediction value with the actual droplet evaporation rate obtained by image monitoring, calculate the error, and use it as the training target of the machine learning model; use the support vector machine regression algorithm, with temperature and humidity features as input and evaporation rate as output, to construct a droplet evaporation rate prediction model; use grid search and cross-validation methods to optimize the kernel function type and kernel function parameters of the support vector machine regression algorithm; obtain droplet surface tension, droplet components and other factors affecting the droplet evaporation rate. The real-time data of influencing factors are introduced into the droplet evaporation rate prediction model; the feature selection and feature combination methods are used to screen out the feature subsets that have the most significant impact on the evaporation rate, and reduce the input dimension of the droplet evaporation rate prediction model; the optimized droplet evaporation rate prediction model is converted into a deployable program package, and the input and output interfaces of the prediction model are designed; the prediction model is integrated into the online droplet evaporation monitoring system, and the sensor data of temperature, humidity and other influencing factors are collected in real time. The prediction model is called to predict the evaporation rate, and the prediction results are displayed in real time on the human-computer interface; multiple droplet evaporation experiments are carried out to collect evaporation data under different environmental conditions and droplet properties for continuous training and optimization of the droplet evaporation rate prediction model.

[0030] For example, a high-precision temperature and humidity sensor SHT31 and a high-speed camera were used to collect temperature and humidity data and droplet images in real time during the droplet evaporation process, with a sampling frequency of 10 Hz. The temperature and humidity time series data were denoised using wavelet transform, and the Daubechies4 wavelet basis function was used to decompose the signal in 5 layers, and reconstructed after removing high-frequency noise. The droplet image was preprocessed, the Canny operator was used to extract the droplet edge, the droplet contour was fitted by Hough transform, and the morphological characteristics such as the droplet equivalent diameter and roundness were extracted. The droplet evaporation process was numerically simulated using ANSYS Fluent software, and the VOF method was used to track the gas-liquid interface. The Marangoni effect and thermocapillary flow were considered, the grid size was 10 μm, and the time step was 1 ms. The theoretical values ​​of the droplet evaporation rate under different temperature (20~50℃) and humidity (40%~90%) conditions were obtained through simulation. The simulation value was compared with the actual evaporation rate obtained by image analysis, and the relative error was calculated as the model training target. The prediction model was constructed by support vector machine regression (SVR) algorithm, and the Gaussian radial basis kernel function was selected. The penalty factor C=10, the kernel function parameter γ=1, and the training sample size were 500 groups. The SVR model parameters were optimized by grid search and 5-fold cross validation. The optimal parameter combination was C=15, γ=0.8, and the average validation error was 2%. The droplet surface tension, viscosity, density and other physical properties were introduced as additional input features. The three key features of temperature, humidity and surface tension were screened out by recursive feature elimination method to reduce the model input dimension. The optimized SVR model was converted into a C++ dynamic link library, and the model input and output interface was designed and integrated into the online droplet evaporation monitoring system. The system collected temperature, humidity and droplet image data every 1 s, called the prediction model to calculate the droplet evaporation rate, and displayed the real-time curve on the human-computer interface. A one-year droplet evaporation experiment was carried out, covering different seasons and day and night conditions, and more than 2000 sets of evaporation data were collected for continuous optimization of the prediction model. The model prediction accuracy can reach more than 95%, providing a reliable basis for the optimization control of the droplet evaporation process.

[0031] Step S107, using the established correlation model, dynamically predict the evaporation process of the droplets, and adjust the atomization parameters accordingly to ensure that the attachment time of the droplets on the device surface matches the pre-defined standard usage environment, where the standard usage environment is determined according to the expected usage scenarios and conditions of the device.

[0032] Obtain the expected usage scenarios and condition information of the equipment, and determine the temperature, humidity and air pressure parameters in the standard usage environment according to the instruction manual or the environmental parameter range provided by the customer;

[0033] By installing temperature and humidity sensors and air pressure sensors on the surface of the device, the temperature, humidity and air pressure data of the environment in which the device is located are collected in real time, and the collected data is uploaded to the cloud server through the wireless communication module;

[0034] In the cloud server, based on the historical data of the droplet evaporation process under different environmental conditions collected in advance, a regression model between the droplet evaporation time and the environmental temperature, humidity and air pressure is established using the multivariate linear regression method; using the regression model, the standard environmental parameters and the real-time collected environmental parameters are respectively input to calculate the predicted evaporation time of the droplets under the two conditions; it is determined whether the difference in the droplet evaporation time under the two environments exceeds the preset threshold. If it exceeds, it is determined that the droplet evaporation process in the current environment has a significant deviation, and a deviation value representing the degree of droplet evaporation deviation is obtained; the deviation value is input into the BP neural network model trained in advance using the droplet evaporation data under different environmental conditions, and the atomization frequency, droplet diameter and The adjustment value of the injection speed parameter is calculated and sent to the device end through the wireless communication module; after the control module on the device end receives the adjustment value of the atomization parameter, the piezoelectric ceramic vibration frequency, nozzle diameter and liquid pump pressure of the atomization device are controlled according to the adjustment value; after the adjustment is completed, droplet samples are collected on the surface of the device, the time from attachment to complete evaporation is measured, and the average evaporation time of the sample is calculated; the average evaporation time is compared with the droplet evaporation time under the standard environment, and the relative error between the two is calculated; if the relative error is within the preset acceptable range, it is determined that the adjusted atomization parameters meet the evaporation requirements of the standard environment; otherwise, the atomization parameter adjustment process is triggered again, and the next round of iterative optimization is performed until the relative error meets the requirements or the maximum number of iterations is reached.

[0035] For example, by consulting the manual of the equipment, it is determined that the standard ambient temperature is 25°C, the relative humidity is 50%, and the atmospheric pressure is 10325 kPa. Using the DHT11 temperature and humidity sensor and the BMP280 pressure sensor, connected to the NodeMCU development board through the I2C interface, environmental data is collected every 1 s, and uploaded to the Alibaba Cloud IoT platform through the MQTT protocol. In the cloud Python program, based on the existing 100 sets of droplet evaporation time data measured under different temperature (20~40°C), humidity (40%~80%) and air pressure (90~110 kPa) conditions, the LinearRegression class of the Scikit-learn library is used to construct a multivariate linear regression model to fit the relationship between the droplet evaporation time t and the temperature T, humidity H, and air pressure P: t=aT+bH+cP+d, where the coefficients a=-5, b=-2, c=1, d=20. Substituting the standard environmental parameters into the model, the expected evaporation time is 15s; substituting the real-time collected environmental parameters into the model, the current predicted evaporation time is 12s, the difference between the two is 3s, which exceeds the preset threshold of 2s, and it is determined that the evaporation process has deviated, and the deviation value is 20%. The deviation value is input into the three-layer BP neural network pre-trained based on 500 sets of evaporation data. The hidden layer uses the Sigmoid activation function, the output layer uses the linear activation function, the learning rate is 0.1, the number of iterations is 1000, and the calculation shows that the atomization frequency needs to be increased by 5kHz, the droplet diameter needs to be reduced by 10μm, and the injection speed needs to be increased by 2m / s. After receiving the adjustment value, the control module increases the PWM frequency of the piezoelectric ceramic from 120kHz to 125kHz, replaces the nozzle with an inner diameter reduced from 5mm to 4mm, and increases the liquid pump motor voltage from 5V to 6V. After adjustment, 5 10 μL droplet samples were added to the surface of the device, and the evaporation process was recorded by a high-speed camera. The Hough circle transform algorithm of the OpenCV library was used to track the change of the droplet contour. The average evaporation time of the sample was calculated to be 18 s, with a relative error of 3% from the standard value, which is less than the acceptable range of 5%. It was determined that the adjusted atomization parameters met the standard environmental requirements, and the optimization process was ended.

[0036] On the other hand, Figure 4 As shown, this embodiment also provides a communication equipment waterproof performance detection system, which specifically includes:

[0037] The three-dimensional model acquisition module is used to obtain the equipment structure characteristics and material property information; the computational fluid dynamics simulation module is used to simulate the water vapor distribution and determine the optimal atomization parameters; the high-speed camera monitoring module is used to monitor the droplet movement trajectory and aggregation state in real time; the temperature and humidity data acquisition module is used to collect the temperature and humidity data of the atomization environment; the machine learning modeling module is used to establish a correlation model between temperature and humidity and droplet evaporation rate; the dynamic adjustment control module is used to predict droplet evaporation and adjust the atomization parameters.

[0038] Exemplarily, the three-dimensional model acquisition module includes three-dimensional design software, such as SolidWorks, UG, Creo, etc.; the computational fluid dynamics simulation module includes fluid simulation software such as ANSYSFluent and PowerFLOW; the high-speed camera monitoring module is a high-speed camera, and the frame rate of the high-speed camera is not less than 1000 fps; the temperature and humidity data acquisition module is a sensor array composed of multiple SHT31 temperature and humidity sensors; the machine learning modeling module includes machine learning libraries such as scikit-learn and Keras; the dynamic adjustment control module includes a cloud server and a piezoelectric atomization device, including piezoelectric ceramics, nozzles, liquid pumps and other hardware facilities.

[0039] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. 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 a 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 above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) to form a technical solution.

Claims

1. A method for detecting the waterproof performance of communication equipment, characterized in that: The method comprises: Acquire a three-dimensional model of the communication device to be tested, the three-dimensional model including structural characteristics and material property information of the device, and simulate the distribution of water vapor on the surface of the device under different atomization parameters using a computational fluid dynamics simulation method based on the three-dimensional model; Based on the computational fluid dynamics simulation results, determine the optimal atomization parameter combination, wherein the atomization parameters include atomization liquid flow rate, atomization nozzle diameter and injection pressure, so as to ensure that the atomization coverage uniformity and droplet size distribution meet the preset standards, wherein the droplet size distribution standards include average particle size and particle size distribution range; During the atomization process, high-speed camera technology is used to monitor the movement trajectory and aggregation state of droplets on the surface of the device in real time, and the aggregation area of ​​droplets is identified through image processing algorithms based on the image information obtained by high-speed camera; According to the identified gathering area and its degree of gathering, the angle and position of the atomizing nozzle are dynamically adjusted to avoid the situation where water vapor gathering affects the waterproof performance test results; Arrange multiple temperature and humidity sensors in the atomization chamber to collect temperature and humidity data of the atomization environment in real time, and analyze the impact of these data on the atomization effect; Through machine learning algorithms, a correlation model between temperature and humidity data and droplet evaporation rate is established, which takes into account both computational fluid dynamics simulation results and image monitoring results; The established correlation model is used to dynamically predict the evaporation process of droplets, and the atomization parameters are adjusted accordingly to ensure that the attachment time of droplets on the surface of the device matches a predefined standard usage environment, wherein the standard usage environment is determined according to the expected usage scenarios and conditions of the device.

2. The method according to claim 1, characterized in that The method comprises: obtaining a three-dimensional model of the communication device to be tested, wherein the three-dimensional model includes structural characteristics and material property information of the device; and simulating the distribution of water vapor on the surface of the device under different atomization parameters by using a computational fluid dynamics simulation method according to the three-dimensional model, including: Obtain 3D CAD model data containing structural characteristics and material property information of the communication equipment to be tested, and convert it into STL format; Import the converted STL format 3D model data into ANSYS Fluent computational fluid dynamics simulation software, set boundary conditions and solution parameters, including defining fluid properties, specifying inlets and outlets, and dividing the grid; According to different ambient temperature and humidity atomization parameters, multiple sets of numerical simulation calculations were performed in ANSYS Fluent software to obtain the water vapor concentration distribution results under different working conditions; Extract the numerical data of water vapor concentration on the grid cells on the equipment surface in each group of simulation results and save them as CSV format files; Use Python's Matplotlib library to visualize water vapor concentration data, generate a cloud map of water vapor distribution on the device surface, and use a suitable color scheme to indicate high and low concentrations; The K-means clustering algorithm is used to cluster the pixels in the water vapor distribution cloud map, and the pixels are divided into different clustering areas based on the water vapor concentration value. Set a water vapor concentration threshold. If more than 20% of the pixel concentration values ​​in the clustered water vapor aggregation area exceed the threshold, it is determined that the device is at risk of condensation under the atomization parameters, and the device structure needs to be further optimized, including adding drainage holes or selecting materials with better surface hydrophobicity.

3. The method according to claim 1, characterized in that Based on the computational fluid dynamics simulation results, the optimal atomization parameter combination is determined, and the atomization parameters include atomization liquid flow rate, atomization nozzle diameter and injection pressure to ensure that the atomization coverage uniformity and droplet particle size distribution meet the preset standards, wherein the droplet particle size distribution standard includes the average particle size and the particle size distribution range, including: Obtaining physical property parameters and process requirements of the atomized liquid to be optimized, wherein the physical property parameters include viscosity, density and surface tension, and the process requirements include atomized liquid flow range and droplet size distribution requirements; Based on the physical properties and process requirements, a computational fluid dynamics simulation model of the atomization process is established in ANSYS Fluent software, and boundary conditions and solution parameters are reasonably set; For the simulation model, different combinations of atomizing liquid flow rate, nozzle diameter and injection pressure parameters are set to carry out multiple sets of numerical simulation calculations; Extract droplet size distribution data and spatial coverage distribution data from each set of simulation results, compare the droplet size distribution data with the preset particle size standard, and calculate the root mean square value of the particle size deviation of the parameter combination; If the root mean square of the deviation is less than a preset threshold, the current parameter combination is output as the optimal solution for particle size optimization; Otherwise, the particle swarm optimization algorithm is used to iteratively optimize the parameter combination with the minimum root mean square deviation as the goal; Based on the optimal combination of particle size parameters, the atomization coverage distribution data is discretized, the variance in the coverage area is calculated, and the inverse of the variance is used to characterize the coverage uniformity; If the uniformity index is greater than a preset threshold, the current parameter combination is determined to be a comprehensive optimal solution; Otherwise, the uniformity index is taken as the optimization target, and the genetic algorithm is used to further optimize the parameter combination; According to the comprehensive optimal solution, the structural parameters of the atomization device are determined, and the three-dimensional structural design of the atomization device is completed in the CAD software to form a design scheme of the atomization system.

4. The method according to claim 1, characterized in that: During the atomization process, the motion trajectory and aggregation state of the droplets on the surface of the device are monitored in real time by using high-speed camera technology; the aggregation area of ​​the droplets is identified by an image processing algorithm based on the image information obtained by the high-speed camera, including: During the atomization process, high-speed camera equipment is used to monitor the movement trajectory and aggregation state of droplets on the surface of the equipment in real time to obtain image information of droplet movement and aggregation; Preprocess the acquired droplet image information, including using median filtering to remove noise from the image and using histogram equalization algorithm to enhance the image, so as to improve the image quality and prepare for subsequent recognition; According to the characteristics of droplet motion trajectory and aggregation state, a droplet recognition model is constructed, and the model is trained using a support vector machine algorithm to obtain a trained droplet recognition model; The pre-processed droplet image is identified using the trained droplet recognition model to determine whether there is a droplet aggregation area in the image. If so, the location and range of the aggregation area are determined. According to the identified droplet aggregation area, the area of ​​the aggregation area, the number of droplets and the distribution density parameters are calculated to obtain a quantitative description of the droplet aggregation state; Compare the quantitative description of the droplet aggregation state with the preset threshold. If the threshold is exceeded, it is judged as abnormal aggregation, triggering an alarm signal and notifying relevant personnel to handle it; At the same time, according to the degree of deviation between the droplet aggregation state parameters and the preset threshold, the working parameters of the atomization equipment are dynamically adjusted, including changing the atomization intensity by adjusting the vibration frequency of the atomizer, and controlling the droplet particle size by adjusting the droplet injection pressure, so that the droplet aggregation state is maintained within the normal range, avoiding the occurrence of abnormal aggregation, and ensuring the stability and uniformity of the atomization effect.

5. The method according to claim 1, characterized in that The method of dynamically adjusting the angle and position of the atomizing nozzle according to the identified aggregation area and its aggregation degree to avoid the situation where water vapor aggregation affects the waterproof performance test result includes: S501, obtaining an image of a water vapor gathering area, and segmenting the image using an image segmentation algorithm based on region growing to obtain a water vapor gathering area; S502, estimating the water vapor density by calculating the average gray value of the pixels in each area according to the water vapor accumulation area obtained by segmentation, and obtaining the degree of water vapor accumulation; S503, obtaining the angle and position parameters of the current atomizing nozzle; S504, establishing an atomizing nozzle angle and position adjustment model based on multiple linear regression according to the water vapor gathering area, the degree of gathering and the current angle position of the atomizing nozzle; S505, training the model by the least square method to obtain optimized model parameters; S506, inputting the water vapor gathering area, the degree of gathering and the current angle position of the atomizing nozzle into the trained adjustment model, and obtaining the optimized angle and position parameters of the atomizing nozzle through model calculation; S507, comparing the current angle position of the atomizing nozzle with the optimized angle position, and calculating the Euclidean distance; if the distance is greater than a preset threshold, controlling the atomizing nozzle motor to adjust the angle and position until the angle and position are adjusted to the desired position; S508. After the dynamic adjustment of the angle position of the atomizing nozzle is completed, the product surface image is obtained, and the Canny edge detection algorithm is used to detect the water vapor accumulation on the product surface. By analyzing the connectivity of the edge, it is determined whether there is water vapor accumulation; if there is no water vapor accumulation, it is considered that the adjustment effect meets the requirements, and the subsequent waterproof performance test is continued; S509: If there is water vapor accumulation, return to step S501, reacquire the water vapor accumulation area image, and perform the next round of dynamic adjustment of the angle position of the atomizing nozzle until the water vapor accumulation is eliminated.

6. The method according to claim 1, characterized in that Arranging a plurality of temperature and humidity sensors in the atomization chamber to collect temperature and humidity data of the atomization environment in real time and analyzing the influence of these data on the atomization effect includes: Acquiring temperature and humidity data of the atomization environment, wherein the temperature and humidity data are collected by a temperature and humidity sensor arranged in the atomization cavity; Preprocessing the temperature and humidity data, and using a moving average filtering algorithm to remove outliers and noise data; The ARIMA time series analysis model is used to model the preprocessed temperature and humidity data to obtain the time trend, periodicity and mutation point characteristics of the temperature and humidity data; Obtaining an atomization effect evaluation index, wherein the atomization effect evaluation index includes droplet size distribution and atomization rate; The Pearson correlation coefficient was used to analyze the correlation between the temperature and humidity data and the atomization effect evaluation index, and a multivariate linear regression model of temperature and humidity and atomization effect was established; The ARIMA time series analysis model is used to make short-term predictions of temperature and humidity. If the predicted value exceeds the preset threshold range, an alarm mechanism is triggered, and the working parameters of the atomization equipment are adjusted according to the preset strategy; Acquire temperature and humidity data under different seasons and weather conditions, and use K-means clustering algorithm to divide the temperature and humidity data into different working conditions; For each of the working conditions, a mapping table of temperature, humidity and atomization equipment parameters is established to optimize the control strategy of the atomization equipment; A support vector regression algorithm is used to establish a nonlinear prediction model of the temperature and humidity data and the atomization effect evaluation index; According to the prediction results of the nonlinear prediction model, early warning and dynamic optimization adjustment of the atomization effect are achieved.

7. The method according to claim 1, characterized in that The correlation model between temperature and humidity data and droplet evaporation rate is established through a machine learning algorithm, and the model takes into account both computational fluid dynamics simulation results and image monitoring results, including: Acquire temperature and humidity data and droplet image monitoring data during droplet evaporation, preprocess the data, extract key features, and construct a training data set; The droplet evaporation process was simulated using computational fluid dynamics simulation software to obtain theoretical predictions of droplet evaporation rates under different temperature and humidity conditions; Comparing the simulation prediction value with the actual droplet evaporation rate obtained by image monitoring, and calculating the error as a training target for the machine learning model; The support vector machine regression algorithm is used to construct a droplet evaporation rate prediction model with temperature and humidity characteristics as input and evaporation rate as output. Optimizing the kernel function type and kernel function parameters of the support vector machine regression algorithm using grid search and cross validation methods; Acquire real-time data of droplet surface tension and other influencing factors of droplet components, and introduce them into the droplet evaporation rate prediction model; Using feature selection and feature combination methods, the feature subset that has the most significant impact on the evaporation rate is screened out, thereby reducing the input dimension of the droplet evaporation rate prediction model; Convert the optimized droplet evaporation rate prediction model into a deployable program package, and design the input and output interfaces of the prediction model; The prediction model is integrated into an online droplet evaporation monitoring system, sensor data of temperature, humidity and other influencing factors are collected in real time, the prediction model is called to predict the evaporation rate, and the prediction result is displayed in real time on a human-computer interface; Multiple droplet evaporation experiments were carried out to collect evaporation data under different environmental conditions and droplet properties for continuous training and optimization of the droplet evaporation rate prediction model.

8. The method according to claim 1, characterized in that The established correlation model is used to dynamically predict the evaporation process of the droplets, and the atomization parameters are adjusted accordingly to ensure that the attachment time of the droplets on the surface of the device matches the pre-defined standard use environment, wherein the standard use environment is determined according to the expected use scenario and conditions of the device, including: Obtain the expected usage scenarios and condition information of the equipment, and determine the temperature, humidity and air pressure parameters in the standard usage environment according to the instruction manual or the environmental parameter range provided by the customer; By installing temperature and humidity sensors and air pressure sensors on the surface of the device, the temperature, humidity and air pressure data of the environment in which the device is located are collected in real time, and the collected data is uploaded to the cloud server through the wireless communication module; In the cloud server, based on the historical data of the droplet evaporation process under different environmental conditions collected in advance, a regression model between the droplet evaporation time and the environmental temperature, humidity and air pressure is established using a multivariate linear regression method; Using the regression model, inputting standard environmental parameters and real-time collected environmental parameters respectively, and calculating the predicted evaporation time of the droplets in the two cases; Determine whether the difference in the evaporation time of the droplets under the two environments exceeds a preset threshold value. If so, determine that the evaporation process of the droplets under the current environment has a significant deviation, and obtain a deviation value indicating the degree of deviation of the evaporation of the droplets; The deviation value is input into a BP neural network model pre-trained using droplet evaporation data under different environmental conditions, and the adjustment values ​​of atomization frequency, droplet diameter and injection speed parameters are calculated by the BP neural network model, and the adjustment values ​​are sent to the device end through a wireless communication module; After receiving the atomization parameter adjustment value, the device-side control module controls the piezoelectric ceramic vibration frequency, nozzle diameter and liquid pump pressure of the atomization device according to the adjustment value; After the adjustment is completed, collect droplet samples on the device surface, measure the time from attachment to complete evaporation, and calculate the average evaporation time of the samples; Comparing the average evaporation time with the droplet evaporation time under a standard environment, and calculating the relative error between the two; If the relative error is within a preset acceptable range, it is determined that the adjusted atomization parameters meet the evaporation requirements of the standard environment; Otherwise, the atomization parameter adjustment process is triggered again to perform the next round of iterative optimization until the relative error meets the requirement or reaches the maximum number of iterations.

9. A communication equipment waterproof performance detection system, characterized in that: The system includes: a three-dimensional model acquisition module for acquiring equipment structural characteristics and material property information; a computational fluid dynamics simulation module for simulating water vapor distribution and determining optimal atomization parameters; a high-speed camera monitoring module for real-time monitoring of droplet motion trajectory and aggregation state; a temperature and humidity data acquisition module for collecting temperature and humidity data of the atomization environment; a machine learning modeling module for establishing a correlation model between temperature and humidity and droplet evaporation rate; and a dynamic adjustment control module for predicting droplet evaporation and adjusting atomization parameters.

Images