WO2018103130A1 - Online quality inspection method for molded products - Google Patents

Abstract

Provided is an online quality inspection method of molded products: first, plastic parts (9) molded by an injection molding machine (1) are sent to a vibrating plate (2) through a conveyor belt; the vibrating plate (2) sends the plastic parts (9) one by one according to a certain frequency to the conveyor belt where manipulators (3) are located; the manipulator (3) grabs each of the plastic parts (9) into a detection black box (5) which detects appearance and color of the plastic parts (9) through a fill light (6) and an image collection device (7); and then plug detection with a standard accessory (10) is performed; a matching degree detection is conducted through changes of current signals caused by different damping; and then the collected information is sent to a control module (8); the qualified plastic parts (9) are placed back to the conveyor belt by the manipulator (3), and the unqualified plastic parts (9) are placed in a plastic part recycling box. This method is used to automatically monitor and classify plastic molded products with statues such as lack of plastic, unwanted plastic, and burrs or the like, and can send the data to a control system in time for statistics and quickly complete the system virtual warehousing.

Description

Online quality inspection method for molded products Technical field

The invention relates to an online quality detecting method for plastic molding products, and belongs to the technical field of plastic processing industry.

Background technique

At present, the quality inspection of most plastic parts processing and manufacturing industries is still in the manual inspection stage. The color detection is performed by observing the color of the rubber parts by the human eye, and then comparing with the standard color chart to judge whether the color is qualified. The detection of the shape is mainly measured by the vernier caliper, and then compared with the standard part size to see if the size of the rubber part is within the allowable error, thereby judging whether the shape is qualified. The matching of the mating connector is tested by hand and the standard parts, and the experience is used to judge whether the connector is qualified. For the local problems such as lack of glue, multi-adhesive, and flaking, the plastic part should be identified by the naked eye. The above detection methods all have the following problems: Whether the test is qualified depends on the subjective judgment and experience of the quality inspector. There is no specific numerical comparison and comparison, there is a large error, and the quality of the product is not guaranteed. Moreover, in the case of a large number of cases, the current manual detection method can only perform sampling detection, and it is impossible to detect each of the plastic parts, and the detection information cannot be transmitted in real time. Furthermore, artificial detection of prolonged work is prone to visual fatigue, work slack and other human factors, which will further increase the error of quality inspection.

Summary of the invention

The object of the present invention is to overcome the deficiencies in the prior art and provide an online quality detecting method for a molded product, which can be improved on the basis of the existing industrial control system, introduces intelligent analysis, automatically completes quality detection and virtual Storage.

According to the technical solution provided by the present invention, the online quality detecting method for the molded product comprises the following steps:

Step 1. The plastic parts produced by the injection molding machine according to the mold are sent to the vibrating plate through the conveyor belt, and the rubber parts are sent to the conveyor belt where the robot is located according to the set frequency by controlling the vibrating plate, and the debugged robot is transported from the conveyor belt. Setting the position to clamp the rubber to be tested;

Step 2: The robot grabs the rubber parts one by one to detect the image collection area set in the black box, and places the rubber parts according to the set angle, and collects the shape and color information of the rubber parts by the fill light and the image collecting device;

Step 3: The robot flips the rubber component to be tested horizontally by 180°, and collects the shape and color information of the rubber component again;

Step 4: Compare the collected shape and color information with the relevant information of the standard parts in the database. If it is qualified, proceed to the next inspection process. If it is not qualified, the robot will place the unqualified rubber parts into the plastic parts recycling box. And go to step 2;

Step 5. For the rubber parts with good shape and color inspection, the robot will control the rubber component and the standard component to check the connector matching degree, and judge the current magnitude change caused by the damping change during the insertion process. If the test is qualified, if the test is qualified, the glue is judged to be good, and the test information is transmitted to the control module. The robot puts the rubber into the good product box; if the test fails, it judges that the rubber is not Good, the robot puts it in the plastic parts recycling box and goes to step 2.

Specifically, the image capturing device takes a photo before and after the flipping of the rubber piece, and then performs image recognition, extracts the shape and color characteristics of the rubber component to be inspected, and transmits the shape and color of the standard component to the control module and the database. The color features are compared. If the error is within the range specified by the good product, it is determined that the shape and color of the component to be inspected are qualified.

Specifically, the method for detecting the connector matching degree is: installing a high-precision pressure sensor on the standard component, and outputting the signal as a current. When the rubber component to be tested is tested with the standard component, the damping between the two will be The change of the output current of the pressure sensor is caused, the difference between the measured current and the standard current recorded when the standard component is inserted is calculated, and it is judged whether the difference is within the allowable range to judge whether it is qualified.

Specifically, a plurality of high-precision pressure sensors can be installed at different positions of the standard parts, and the sum of the currents generated during the test is the measured current I _c , and the sum of the currents generated by the plurality of high-precision pressure sensors when the standard parts are interposed For the standard current I _b , calculate the difference between the measured current and the standard current to determine whether the error current is within the allowable range, as shown in the following equation:

(I _b -δ)<I _c <(I _b +δ)

Where δ is a given positive number, and its value is determined according to the connector specifications of different rubber parts; if the measured current I _{c is} within this error range, it is determined that the matching degree is qualified.

Specifically, the fill light adopts an LED array light source.

The invention has the advantages of automatically monitoring and classifying the good and bad products produced by the outflow machine for the occurrence of plastic molding products, such as lack of glue, multiple glues, and draping, and can timely send data into the control system for statistics and rapid. Complete system virtual storage.

DRAWINGS

Figure 1 is a flow chart of the present invention.

2 is a schematic structural view of an online quality detecting device.

detailed description

The invention will now be further described with reference to the accompanying drawings and embodiments.

As shown in FIG. 1 , the overall process of an online quality detecting method for a molded product according to the present invention is as follows:

Step 1. The plastic parts produced by the injection molding machine according to the mold are sent to the vibrating plate through the conveyor belt, and the rubber parts are sent to the conveyor belt where the robot is located according to the set frequency by controlling the vibrating plate, and the debugged robot is transported from the conveyor belt. Setting the position to clamp the rubber to be tested;

Step 2: The robot grabs the rubber parts one by one to detect the image collection area set in the black box, and places the rubber parts according to the set angle, and collects the shape and color information of the rubber parts by the fill light and the image collecting device;

Step 3: The robot flips the rubber piece to be tested horizontally by 180°, and collects the shape and color letter of the rubber piece again. interest;

Step 4: Compare the collected shape and color information with the relevant information of the standard parts in the database. If it is qualified, proceed to the next inspection process. If it is not qualified, the robot will place the unqualified rubber parts into the plastic parts recycling box. And go to step 2;

Step 5: For the rubber parts with the appearance and color inspection, the robot will control the rubber component and the standard component to perform the connector matching degree detection, and determine whether the connector matching degree is qualified by using the current magnitude change caused by the damping change during the insertion process; If the test is qualified, it is judged that the rubber piece is a good product, and the test information is transmitted to the control module, and the robot puts the glue piece into the good product frame; if the test fails, the glue piece is judged to be a defective product, and the robot puts it to the Glue the parts in the recycling box and go to step 2.

The invention can be improved on the existing industrial control system, and the image acquisition, the image recognition device, the glue image database and the glue parameter database are added on the basis of the control system of the conveyor belt and the robot.

The invention needs to establish a multi-angle image database of rubber parts, which can be provided by MES (manufacturing enterprise production process execution management system). It is also necessary to use the glue image recognition software to compare the physical object with the standard image, to make the product lack of glue, multi-glue and other analysis; and the data acquisition software to send the good and bad product information into the MES database.

As shown in Figure 2, a molded product online quality testing device is built, including: injection molding machine 1, vibrating plate 2, robot A (grabbing the rubber to be tested 9) 3, robot B (grabbing standard 10) 4, The LED array light source 6 (fill light) in the detection black box 5, the image acquisition device 7, and the control module 8 on the right side, may include an FPGA, a CPU, a DSP, an MES server, etc., and are designed according to the needs of field detection. The main functions realized by the device are: shape inspection of plastic parts, product color detection, and connector matching degree detection.

Robotic workflow:

1, the glue is grabbed. By the cooperation of the respective motors, the robot A 3 can grip the rubber member 9 to be detected from the conveyor belt connected to the vibration disk 2.

2. The glue is placed in the image collection area. After the glue member 9 is gripped, the robot A 3 sends the glue member 9 to the image acquisition area.

3, flip the plastic parts. During the image acquisition process, the rubber member 9 is turned over by the robot A 3 to realize the omnidirectional collection of the shape and color information of the rubber piece.

4. Push to the connector to check the matching degree. The control robot A 3 and the robot B 4 perform the rubber component matching degree test, and judge whether the current is qualified according to the current change caused by the different damping during the insertion process of the rubber component 9 to be tested and the standard component 10.

5. Return the qualified product to the conveyor belt, and return the non-conforming product to the rubber recycling basket.

The image capture device takes a picture before and after the flipping of the rubber piece, and then performs image recognition, extracts the shape and color characteristics of the rubber piece to be tested, and transmits it to the control module to compare with the shape and color characteristics of the standard parts in the database. If the error is within the range specified by the good product, it is determined that the shape and color of the component to be tested are qualified.

The position and angle of the rubber parts are specified, and a limiting member or a supporting member is arranged on the image collecting area, so that the rubber parts must be placed according to the specified position and angle, so that the image collecting device takes the picture. The photo of the rubber piece is a specific view surface, which greatly facilitates the image processing software to process and recognize and improve the recognition efficiency.

The method for detecting the connector matching degree is specifically: installing a high-precision pressure sensor on the standard component, and the output signal is a current. When the rubber component to be tested is connected to the standard component, the damping between the two causes the pressure sensor. The change of the output current, calculate the difference between the measured current and the standard current when the standard component is plugged, and determine whether the difference is within the allowable range to judge whether it is qualified.

In the embodiment, a plurality of high-precision pressure sensors are installed at different positions of the standard parts, and the sum of the currents generated during the test is the measured current I _c , and the sum of the currents generated by the plurality of high-precision pressure sensors when the standard parts are plugged together For the standard current I _b , the difference between the measured current and the standard current is calculated to determine whether the error current is within the allowable range. For example, four pressure sensors are arranged on the standard components, and the measured currents of the four sensors are I _c1 , I _c2 , I _c3 and I _c4 , and the standard currents of the four sensors are I _b1 , I _b2 , I _b3 and I _b4 .

Then measured current I _c =I _c1 +I _c2 +I _c3 +I _c4

Standard current I _b =I _b1 +I _b2 +I _b3 +I _b4

Determine whether the error current is within the allowable range by calculating the difference between the measured current and the standard current, as shown in the following equation:

(I _b -δ)<I _c <(I _b +δ)

Where δ is a given small positive number, its value is determined according to the connector specifications of different plastic parts, thus defining a standard range of current. If the measured current I _{c is} within this standard range, the matching test is qualified. During the movement of the manipulator, the measured current is greater than the standard range, indicating that the product shape is too large or the inner diameter is too small, and the measured current is less than the standard range, indicating that the product shape is too small or the inner diameter is too large.

The robot driver module of the present invention can use existing motor drivers and controllers, and can also integrate controller functions into the processor we added in the control module, omitting the original controller.

The MES server module can provide the standard picture library function, and accept the shape detection status and product color detection status of other products processed by other processors; it can also provide the machine action password and the receiving connector matching degree detection result. In this way, the good products and defective products produced by the machine are dynamically detected, and the data is sent to the system for statistics in time to quickly complete the virtual storage of the system.

It can be seen that the present invention can fully utilize the existing software and hardware facilities of the enterprise, and add a small-cost image acquisition and image processing analysis device, which can well control the cost of upgrading and transformation. The production of standard parts is not complicated and is very suitable for small and medium-sized enterprises.

Claims (5)

A method for online quality inspection of a molded product, characterized in that it comprises the following steps: Step 1. The plastic parts produced by the injection molding machine according to the mold are sent to the vibrating plate through the conveyor belt, and the rubber parts are sent to the conveyor belt where the robot is located according to the set frequency by controlling the vibrating plate, and the debugged robot is transported from the conveyor belt. Setting the position to clamp the rubber to be tested; Step 2: The robot grabs the rubber parts one by one to detect the image collection area set in the black box, and places the rubber parts according to the set angle, and collects the shape and color information of the rubber parts by the fill light and the image collecting device; Step 3: The robot flips the rubber component to be tested horizontally by 180°, and collects the shape and color information of the rubber component again; Step 4: Compare the collected shape and color information with the relevant information of the standard parts in the database. If it is qualified, proceed to the next inspection process. If it is not qualified, the robot will place the unqualified rubber parts into the plastic parts recycling box. And go to step 2; Step 5: For the rubber parts with the appearance and color inspection, the robot will control the rubber component and the standard component to perform the connector matching degree detection, and determine whether the connector matching degree is qualified by using the current magnitude change caused by the damping change during the insertion process; If the test is qualified, it is judged that the rubber piece is a good product, and the test information is transmitted to the control module, and the robot puts the glue piece into the good product frame; if the test fails, the glue piece is judged to be a defective product, and the robot puts it to the Glue the parts in the recycling box and go to step 2. The online quality detecting method for a molded product according to claim 1, wherein the image capturing device takes a photo before and after the flipping of the adhesive member, and then performs image recognition to extract the shape and color of the rubber component to be tested. The feature is transmitted to the control module to compare with the shape and color characteristics of the standard parts in the database. If the error is within the range specified by the good product, it is determined that the shape and color of the component to be tested are qualified. The online quality detecting method for a molded product according to claim 1, wherein the method for detecting the matching degree of the connector is: installing a high-precision pressure sensor on the standard component, and outputting a signal as a current, when the rubber component to be tested is When the standard part is tested for plugging, the damping between the two will cause the change of the output current of the pressure sensor, calculate the difference between the measured current and the standard current recorded during the connection between the standard parts, and determine whether the difference is within the allowable range. Internally judged whether it is qualified. The method for detecting on-line quality of a molded product according to claim 1, wherein a plurality of high-precision pressure sensors are mounted at different positions of the standard member, and the sum of the currents generated during the test is the measured current I _c , and the standard The sum of the currents generated by the multiple high-precision pressure sensors during the insertion is the standard current I _b , and the difference between the measured current and the standard current is calculated to determine whether the error current is within the allowable range, as shown in the following equation: (I _b -δ)<I _c <(I _b +δ) Where δ is a given positive number, and its value is determined according to the connector specifications of different rubber parts; if the measured current I _{c is} within this error range, it is determined that the matching degree is qualified. The method according to claim 1, wherein the fill light adopts an LED array light source.

Images