English
1. Most offline measurements on the market use manual contact eddy current film measuring instruments, with prices below 1000 yuan. Alternatively, it could be using the American Mage. The price of a standalone thickness gauge is around 450000 yuan - for an online spraying measurement control line. At a cost of several million or more. This scheme combines offline measurement of workpieces. Measure the spray thickness separately for finished workpieces. Introduce the use of non-contact photothermal sensors from Switzerland. Combining ultrasonic and infrared diffusion techniques to measure the thickness of sprayed surfaces. Has high cost-effectiveness. The advantage of simple operation. Can measure different colors. Different shapes, non-contact measurement of flat and curved sprayed workpieces. Real time access to data through ERP and browser. Cloud transmission of industrial control computers to display data and print in real-time on the computer desktop. Adopting a three-axis module structure. Servo motor controls XYZ direction movement. Place the workpiece on the platform. Set the probe measurement point. Automatically move the probe position. Non contact measurement of surface coating thickness on workpieces. Real time transmission to the industrial control screen. Display thickness value. Replaced manual contact eddy current thickness gauge. Improved work efficiency. Avoid scratches on the surface of non-destructive workpieces.
The second option is to use a 5-axis robot for measurement. Can measure the thickness of edge side spraying. Flexibility and maneuverability surpass the three-axis model structure.
2.
3.
Powder coating is an environmentally friendly alternative material for surface treatment processes, widely used in various fields such as automotive, construction, and household appliance industries. Apply powder coating under static electricity, in which the fluid carrying particles charges the particles through the corona inside the spray gun, and then spray the powder onto the grounding substrate. At this point, several physical phenomena occur: turbulent airflow (Re max ≈ 1.5e5, U max ≈ 20m/s), interacting polymer particles (2 μ m-180 μ m), superimposed electrostatic field (ψ max ≤ 1.2e5V), and gravity field. Due to the difficulty in studying the complex interactions involved through experiments and data, most research in the field of electrostatic spraying relies on past experience. Having economic foundation strength. It will be equipped with a fully automatic online measurement spray thickness control production line.
In order to quantify coating quality, attempting to use destructive measurement methods cannot obtain the distribution of particles on the substrate, either due to a lack of experimental comparisons with data studies, or indirectly deriving coating thickness based on velocity field distribution. However, it is necessary to obtain the distribution and thickness of the coating on the substrate in practical applications. Adopting a non-contact measurement method for data collection and post-processing quantification can not only optimize industrial spraying processes, but also contribute to the in-depth development of technology and the measurement and quality control of spray thickness uniformity.
2. Use photothermal method -3D non-contact measurement of coating thickness
The photothermal 3D online coating thickness measurement system can non-contact test the thickness of uncured coatings within 50cm. This system is based on advanced thermal optical technology, with computer-controlled light sources heating the tested coating in a pulsed manner. Then the built-in high-speed infrared detector records the temperature distribution in the time series at a certain distance. As the surface temperature decreases (which is related to the coating thickness and thermal properties of the coating material), the distribution of coating thickness within the detection range of the device is inferred as a spring magnetic field. The dynamic change process of coating surface temperature reduction has characteristics. He will be affected by the thickness of the coating and the thermal properties of the substrate material. The thinner the coating, the faster the surface temperature drops. The surface temperature provides a diffuse wavelength over time. The IN acquisition card calculates the coating thickness using a function formula by comparing temperature changes with wavelength voltage. Not affected by environment - color. The shape affects it.
Figure 3- Threshold filtering of raw data image. The above image is the original image, with pixels or regions marked as: A) a set of pixels that have reached the camera's maximum threshold by removing thin coatings through threshold filtering; B) Eliminating the pixel set of substrate outer coating thickness values through geometric filtering; C) Use appropriate filtering to remove the fixture area. The following image is the substrate image after threshold filtering, with A-type regions removed. Suitable for scaling the available data range.
2.3.1.2 Geometric filtering
Threshold filtering does not require removing all pixels that are detached from the substrate, as shown in region B in Figure 3. Since these pixels are usually smaller than the substrate, geometric filtering can be performed based on their geometric positions. Sort the remaining unfiltered data based on the position of rows and columns in the image. The coordinates of row (r) and column (c) are shown in Figure 5. Then, in the distribution of geometric positions, the coordinates of the eliminated pixel positions are lower or higher than the limit values defined in the percentages (low and high).
Figure 4- Geometric filtering based on geometric position distribution function.
Using the filtering function δ defined in formula (1) for mathematical representation, multiply this function by the pixel i associated with it, where D i is the coating thickness value.
The effect of the filtering process is shown in Figure 5, with limit values of 2.5 and 97.5% of the coordinates, respectively, to eliminate values containing non-zero coating thickness.
Figure 5- Geometric threshold between 2.5-97.5% of pixel coordinates. The upper image is the original image of the substrate, and the lower image has been geometrically filtered. Most of the B type pixels in Figure 3 have been eliminated, except for the pixels contained in the white box and a portion of the C area in Figure 3. Suitable for scaling the available data range.
In this case, it can be observed that the white box in Figure 5 does not eliminate all pixels located outside the substrate, most notably in the bottom pixels of the image. In addition, some fixture areas are still visible in the image. If thresholding is performed between 5-95% of the coordinate values, better filtering will be applied to all cases, as shown in Figure 6. Therefore, the optimal limit threshold can be empirically obtained.
Figure 6- Geometric threshold between 5-95% of pixel coordinates. The upper image is the original image of the substrate, and the lower image has been geometrically filtered. Most of the pixels in region B in Figure 3 and region C in Figure 3 have been eliminated. Suitable for scaling within the available data range.
2.3.1.3 Filter fixture area
The fixture used to fix the substrate during the coating process usually has a high coating thickness. Because it does not belong to the substrate, this area needs to be filtered out. In addition, it is necessary to eliminate coating pixels caused by fixture interference near the substrate. Eliminate the fixture area according to the schematic shown in Figure 7, where the corresponding area of the fixture is filled with blue. The identification of the substrate center and the range of row and column coordinates of the substrate play a key role in defining the fixture area.
Figure 7- Schematic diagram of fixture area
Identify the two ratios of the control area, as shown in equation (2).
If the row coordinates are less than the minimum row plus the height of the fixture, and the column coordinates are less than the coordinates of the substrate center minus the width w2 or greater than the coordinates of the substrate center plus w1, then the data will be eliminated as the fixture area. The filtering effect is shown in Figure 8.
Figure 8- Filter Fixture Area
2.3.2 Average Coating Thickness (ACT)
The primary performance parameter for characterizing coating quality is the average coating thickness. When comparing processes using the same amount of powder, it directly reflects efficiency, as the larger the average coating thickness of the substrate, the higher the proportion of deposited powder. Therefore, by using the filtering process described in the above sections and calculating the average coating thickness of the remaining pixels, ACT can be calculated, as shown in formula (4). The purpose of any painting process is to maximize efficiency, therefore it is necessary to improve ACT as much as possible.
In formula (4), an area scaling factor is specifically introduced for thinner coatings. In thinner coatings, the number of unfiltered pixels (N unfiltered) is much smaller than the number of pixels covering the entire substrate. In these cases, if the entire substrate area is not scaled and the average thickness of a single unfiltered pixel (Ai) area is directly calculated, it will result in incorrect results. The substrate area (A plate) is determined by the range of row and column coordinates. When measuring the range between different values and the substrate, the area of the substrate area range (A plate) is averaged across all standardized cases to eliminate the influence of variations, which are usually small for each case.
2.3.3 Center offset
The next performance parameter is the center offset. Quantifying this parameter requires identifying the highest coating thickness region (RHCT). This area contains a certain proportion of unfiltered pixels, reflecting the highest coating thickness. Then calculate the center offset based on the difference in row and column coordinates between the geometric center of the substrate (equation (5)) and the geometric center of the region (equation (6)), as shown in Figure 9.
In Figure 9, the geometric center of the substrate is represented in purple, and the geometric center of the region is represented in brown. It can be observed that the geometric center moves towards the bottom and right edge of the plate, indicating that the coating has some asymmetry. Therefore, the center offset will characterize the asymmetry of the coating.
Figure 9- Geometric center of the substrate relative to the center of RHCT.
The offset towards the bottom center indicates that the substrate is far away from the spray gun position, while the offset in the column coordinate direction usually indicates that the substrate is not necessarily perpendicular to the spray gun.
2.3.4 Non uniformity
The final performance parameter is non-uniformity. Quantifying non-uniformity is crucial whether it is to obtain a uniform coating as much as possible or to obtain a completely opposite coating based on the application. Quantify parameters based on the histogram of coating thickness, as shown in Figure 10.
Figure 10- Histogram of Coating Thickness
To generate a coating thickness histogram, all coating values were collected into a specified number of bins, as shown in Figure 10, using 20 bins. Therefore, each column represents the number of pixels containing thickness values within the bin range. Define non-uniformity using the average thickness values of the maximum count, N max, and corresponding bin D max in the histogram, as shown in equation (7).
Therefore, non-uniformity is a form of standardized weighted standard deviation, where the larger the deviation value, the more uneven the coating distribution. The weight of the count in bin i and N i (based on the maximum count) ensures that a small number of pixels with significant deviations from the maximum count thickness will not dominate the non-uniform values.
As shown in Figure 11, the relationship between ACT and voltage variation can be observed in the block diagram. The data in the block diagram consists of data from three substrates, each with three measurement values. Therefore, a total of 9 measurements were taken, as described in Section 2.1. The ACT value shows a slow increasing trend, and the spray gun is used at a constant voltage of about 29 kV.
Figure 11- Relationship between average coating thickness and voltage variation curve, as well as thickness contour lines. The bottom feature area is marked as A) thin band at the upper edge, B) thick band at the center, C) thick layer at the bottom, D) extremely thin corner, E) central diffusion band
Even though ACT tends to stabilize under high pressure, the contour contour contour contour of the median thickness (red line in the box plot) is significantly different. Comparing the 24 kV and 52 kV contour lines, it can be seen that in the former case, the coating thickness at the center band (B) and bottom edge band (C) is higher, while the coating thickness at the upper edge band (A) of the substrate is lower. In the center diffusion band (E), the lower coating thickness area is limited to the corner (D), resulting in a more uniform coating. This can also be quantitatively verified through the non-uniform plot in Figure 12. From the graph, it can be seen that at the lowest voltage of 10kV, the non-uniformity is at a low value, and at around 24kV, the non-uniformity reaches its peak, and then reaches its lowest value at the highest two voltages, which is consistent with the thickness profile described under the median condition.
Considering the histograms of coating thickness for three scenarios of 10kV, 24kV, and 52kV, it can explain the phenomenon of the lowest and highest voltage non-uniformity being at the same level, as shown in Figure 13.
圖 12- 電壓變化曲線和厚度等值線的不均勻性
Figure 13- Histogram of Coating Thickness Distribution at Specific Voltage
At a voltage of 10kV, there are more pixels in a single 'bin', indicating lower non-uniformity. In fact, the unevenness in this situation is lower than 52 kV. However, the coating thickness value associated with the highest "count" is relatively small, so the deviation between 10 kV and the highest "count" is relatively large. In the case of 52kV, this deviation is very small because the larger coating thickness is related to the highest "count", resulting in similar uneven values. For the medium voltage situation, it can also be seen from the histogram that the distribution of coating thickness values is large and the count is large, and the combination of the two leads to uneven peak values.
Figure 14 illustrates the center offset of row and column coordinates on the substrate. Negative column coordinates indicate that the center of RHCT is located to the left of the substrate center, while negative column coordinates indicate that it is located below the substrate center. It can be seen that the median values of the column coordinates are all positive, but the trend is not obvious. The median values of the row coordinates are all negative, and the higher the voltage, the closer it is to the center of the board.
The three selected cases of column coordinates, namely 18, 35, and 52 kV on the left side of Figure 14, align the column coordinates almost with the center of the substrate at 18 kV, and then offset more at 35 kV than at 52 kV. On the right side, it can also be seen that the row coordinates approach the center of the substrate as the voltage increases, with voltages ranging from 20 to 29 and 48 kV, which is reflected in the row coordinate center offset graph. It should be noted that, especially under medium voltage, there is usually a thick coating area at the bottom edge of the plate (see area C in Figure 11), which pushes the RHCT center towards the bottom, most prominently at 20 kV.
The center offset can also provide important information related to the coating process. In Figure 14, it is observed that the center offset of the column coordinates and most of the range box are within the positive range, which means that the offset systematically faces to the right of the substrate center. Otherwise, the range box will contain an offset value of 0.
Figure 14- The center offset of column and row coordinates varies with voltage.
The reason for this systemic effect may be ventilation in this situation. When considering the offset of row coordinates, except for the extremely low voltage of 10kV, all offsets are lower than the center of the substrate, and increasing the voltage brings them closer. Because the system effect here is gravity, its effect decreases as electrostatic force becomes dominant and higher voltage is used to push the center of RHCT towards the center of the substrate.
The above is a detailed introduction to the filtering process of collecting and processing experimental data on powder coating technology based on advanced thermal optical technology using measuring instruments to collect coating thickness data.
