The present invention relates to an apparatus and a method for quality improvement in the automated machine-based casting method by means of identifying the cast parts by pattern recognition and structure recognition.
Metal production dates from the Copper Age, the time of transition from the Neolithic period to the Bronze Age. In antiquity, bronze was replaced by iron as the most important material, which in Europe was not capable of being cast until the Middle Ages.
During industrialization, cast iron became the most important construction material. By around 1900, series production parts were already being cast from aluminum for the automotive industry. In the 1970s, by the development of modern FEM simulation (method of finite elements) it became possible to simulate and optimize the casting process.
In respect of the prior art, reference is made at this point to Document DE 10 2015 102 308 A1. This is a method for labeling a cast part. According to the specifications in the description, the object of this method is to provide a method which makes it possible to produce cast parts that are permanently provided with readable information, in particular also in the state ready for use.
According to the specifications in patent claim 1, this object is achieved by a method for producing a cast part (G1, G2) provided with readable information (IG1, IG2), comprising the following working steps:
The methods for metal casting which are intended to place number parts or number stamps in the finished cast parts in order to identify them are very cost-intensive, susceptible to error and time-consuming, for which reason the object of the present application is to make do in all metal casting methods without manipulation of the model for the cast part identification, i.e. no part with a number or a continuous number stamping tool is installed or placed in the model.
This object is achieved by an apparatus for quality improvement in the automated machine-based casting method by means of identifying the cast parts by pattern recognition and structure recognition, having the following features:
The invention will be described by way of example with reference to an iron casting installation.
In detail:
The supply of the required amount of sand, the closing of the respective sand mold and the closing of the sand pressing mold are monitored by means of sensors and cameras. The molding sand mixture (mold material) is stocked in the molding sand store 2 (mold material), and the properties of the molding sand 7 (mold material) may be manipulated by additives before filling into the sand pressing mold.
The filling level and the throughput of molding sand 7 in the molding sand store 2 are monitored by video and sensors.
After the formation, the molds 12 are conveyed successively to the casting apparatus with a friction fit and/or form fit on a transport path 10. In this case, the front mold 12 forms the second part of the rear mold 12. The casting apparatus 3 consists of a container for the so-called melt? (liquid metal, for cast iron the melting temperature is about 1400 degrees) and a filling apparatus for filling the sand molds 12. The filling is monitored by cameras and sensors as well as laser scanners. These parts are not shown here for the sake of clarity. The liquid metal is manipulated by so-called seeding before filling a casting mold 12. This means that required additives are delivered through a lateral channel (not shown here) into the liquid metal in order to correspondingly influence the product characteristics.
After filling the molds 12 in the casting apparatus 3, the row of molds 11 is conveyed at the rate of the mold production 1 on the transport path 10 through the cooling device 4 (the transport path 10 and the cooling device 4 are one unit) in the direction of a shaking apparatus 8. Through the rate of the mold production 1 and by means of the monitoring by sensors and cameras, the number of molds 11 and the respective position of a particular mold 11 on the entire transport path 10 are known. The cast parts 24 in the molds 11 may therefore be assigned to each particular mold 11.
On the left side, the transport path 10 of
The cameras 15 and 16 located above track the cast parts 24 freed from the loose molding sand 7 and from the loose flashes as far as the conveyor belt 26. Thermal imaging cameras are preferably used for the cameras 15 and 16.
By means of the conveyor belt 26, the cast parts 24 are sent through a second cooling device 18, a camera 17 (CCD) arranged above tracking the route. Depending on requirements, more cooling installations may also be installed in the installation.
After passing through the cooling device 18, the residual flashes and larger parts of the feeder system are removed by means of a gripping apparatus in the form of a six-axis robot. This procedure may also be carried out manually or with the aid of manipulators. The flashes 27 are jointly captured by a video camera 19 in the further tracking of the cast parts 24, and the captured data are used to control the flash removal 20 of the residual flashes and feeder parts.
With the camera 19, the cast parts 24 are tracked further as far as the conveyor belt 25 which transports the parts to the inlet of the jet cleaning 22.
On the conveyor belt 25, the video camera 21 located above takes over the part tracking as far as the jet cleaning apparatus 22, or the transport apparatus 23 of the jet cleaning 22.
On the left side, we see the conveyor belt 25 which carries the cast parts 24 to the entry of the jet cleaning apparatus 22. The grid belt 23 of the jet cleaning apparatus 22 takes over the further transport through the jet cleaning apparatus 22.
During the jet cleaning, the cast parts 24 are cleaned of the residual contamination, for example incrustations of the molding sand, by means of bombardment with granules. This is done with the aid of granules which are accelerated by means of spinner wheels or compressed-air jet nozzles 28, which are respectively located above and below the conveyor grid belt 23 and are directed onto the cast parts.
Depending on the particle size, material and shape of the granules, this cleaning method leaves behind a structure on the surface of the cast parts 24, as may be seen in
Over the start of the conveyor belt 25, there is the camera 21 which concludes the pattern tracking of the cast parts before the jet cleaning parts 22. By the video pattern recognition and the video pattern tracking, the data of each cast part 24 in respect of the mold from which it comes, and which casting nest 43 it belongs to, are known as far as the entry into the cleaning installation 22.
The cast parts 24 are taken up by the grid belt 23 and delivered with an accurately defined speed through the jet cleaning apparatus 22. In this way, each cast part 24 is re-identified by the camera until leaving the cleaning installation 22.
The cleaning installation is followed by a structure scan device consisting of the camera 29 and the transport device 41. The camera 29 is a high-resolution stereo video camera and or a scanning apparatus equipped with a high-resolution graphene light sensor.
The transport device 41 is mounted vibration-free in order to improve the quality of the scan recording.
After the scanning with the camera 29 and the storing of the data of the structure scan face 47 (see
In the station 40, all the measurements take place in a continuous throughput testing method.
First, the cast parts 24 are checked for casting lattice defects by means of an eddy-current measuring device 30. In this case, the change in an applied electric field allows inference about the structure of the lattice of a cast part 24. The data obtained are analyzed and stored, and may later be assigned to the respective cast part 24.
Subsequently, a surface structure measurement 31 is carried out by means of lasers in order to check the surface of the cast parts for irregularities on the surface. Two lasers lying opposite one another are directed at a particular angle X onto the surface of the respective cast part at a point and synchronously scan the surface of the cast part 24. This creates a 3D profile of the respective surface, which is analyzed. The data of the measurement are assigned to the cast part 24 and stored. The data of the high-accuracy laser scan may likewise be used to check the scan data marking 47.
The cast part 24 is transported further and checked by means of the laser measuring device 32 in respect of the outer contours for planarity and curvature. The corresponding data are evaluated and stored for each cast part.
The cast part 24 is transported further on the belt 39 and checked by means of an ultrasound measuring apparatus 33 for cavity inclusions. The data for each cast part 24 are evaluated and stored.
The cast part 24 is transported further on the belt 39 and checked by means of a laser thickness measuring installation 42 in height for dimensional compliance. The lasers respectively scan the upper and lower edges of each cast part 24. The data obtained are evaluated and stored. The cast parts 24 are transported further on the conveyor belt 39 to the sorting installation 38.
Before entry into the sorting installation 38, the parts are captured with the camera 34 and identified using the structure scan face 47 by means of comparison of the stored structure data (reference pattern). The final check is carried out by comparison of the reference data with the measurement data which have been stored for each cast part 24.
In the sorting installation 38, the cast parts 24 are sorted in respect of various quality categories.
Exemplary category 35, cast parts with contour and thickness defects in category 36 and 37 with lattice and inclusion defects. These are cast parts which, for example, lie within the tolerance and are classified as good.
The outer edge 46 forms the lateral boundary of the sand casting mold. During the filling of the mold 11, the liquid casting iron or an alloy of different metals and additives flows through the main casting channel into the mold 11 and is distributed in the casting nests 43, which form the cavities for the future cast parts 24. Each casting mold is provided, for example, with eight casting nests. Each casting nest 43 of the sand casting mold (11, 12, 14) is provided with a casting nest number 44. By the casting nest number 44, the placement or location of the casting nest 43 and later of the molded cast part 24 in the respective sand casting mold (11, 12, 14) is known. This feature of the casting number 44 is an important detail in the pattern recognition and pattern tracking and in the analysis of casting defects. In this way, after the breaking 6 of the sand mold 12, all cleaned cast parts 24 can be assigned to the respective sand casting mold.
For example, after the breaking 6 of a mold 11, the placement of the cast part 23 from the casting nest 44 with the number 8 is shown on its transport route through the regions of the shaking screen 8, conveyor belt of the cooling installation 26, conveyor belt 25 to the jet cleaning installation. For reasons of clarity, we represent the pattern tracking, object tracking of only one cast part 24.
After the breaking of the mold 11 on the shaking apparatus 8, the cast parts are still connected to the flashes and feeder system parts 9. The cast parts 9 are captured by the video thermal image camera 5 and compared with the shape stored in the program and the placement parameters (classification). The technique of assigning the contents of digital images to a class of a classification system is an image analysis method. This may be subdivided into three subregions of segmentation, object recognition and image interpretation. The pattern recognition or object recognition is carried out by means of contour segmentation on the basis of edges or discontinuities. In this way, the identified cast parts 24 are observed and detected in the pattern tracking program by means of their classifier assigned by the program with the cameras 15, 16, 17, 21 in the regions of the shaking screen 8, conveyor belt of cooling installation 26, conveyor belt 25 to the jet cleaning installation as far as the conveyor belt 23. Here, the transport placement variation of the cast part with the casting nest number 8 on its way to the conveyor belt 23 of the jet cleaning installation 22 is shown.
Crucial for this are the size and shape and the hardness of the jet granule particles. The surfaces shown in the left column are generated with fine amorphous silicate particles. The surfaces shown in the middle column are generated with small round jet spheres. The surfaces shown in the right column are generated with coarse amorphous silicate particles. The size of the identifier face 47 in relation to the structure pattern of the surface is shown in the upper row of the structure recordings.
Category 1: is the cast part formation process with the following components: mold formation component consisting of the installation parts 1, 2, 13, 14, with the control module 49, the casting component 3 with the controller 50, the cooling components 4, 17 with the controller 50, the mold release component 8 with the controller 62, the residual flash removal 20 with the controller 54 and the cleaning component 22 with the controller 53 and the respective transport apparatuses 10, 26, 25, 23, 41, 39, with the controller 52. The sensors for the molding sand humidity, compressibility, mold material composition, the mold material temperature, the pressing pressure, (the adjustment parameters of the molding installation in general), the video sensor for the mold closure monitoring, the temperature sensor for the feeder temperature monitoring, the video sensor and the laser sensor for the mold filling monitoring.
Cooling temperature sensors, placement sensors of the transport apparatuses and the shaking apparatus 8, sound sensors, sensors of the residual flash removal 20, air pressure sensors and granule throughput sensors of the cleaning apparatus 22, rotational speed sensors of the transport apparatuses and other monitoring devices of the manufacturing process are not represented for reasons of clarity.
Category 2: the cast part measuring and testing process with the components: material lattice testing 30, 33, surface testing 31, contour testing 32, thickness testing 42 with the controller 55 and the stored data of the material composition of the chemical analysis, the previous treatment temperature and the origin of the melt.
Category 3: cast part recognition and tracking process with the components 5, 15, 16, 17, 19, 21 and the image processing module 60 and the controller 57.
Category 4: cast part marking and identification in the process by the components 29, 34, and the controller 59.
Category 5: cast part sorting process with the components: sorting installation 38 and the controller 56, the sensors for the transport placement monitoring, drive monitoring, scan sensors of the sorting function monitoring are not represented for reasons of clarity. The sensor data of Category 1 are transmitted to the data processing 61 via the data processings (shown here by dashes) and contain information relating to the actual state of the instantaneous manufacturing installation.
The sensor data of Category 2 are transmitted to the data processing 61 (dashed line) and contain information relating to the respective state of the cast part 24 being tested. The video data of the cameras of Category 3 are transmitted to the image processing 60, and the cast parts 24 are identified with the aid of a pattern recognition program based on contour segmentation, as described in
All data of Categories 1 to 5 are collected in the data processing 61, also referred to as big data, and delivered by a systematic data analysis program with an evaluation system as extracted data to the production data set in order to be used for active control and regulation and for interactive self-regulation by special programs of the overall manufacturing process.
Number | Date | Country | Kind |
---|---|---|---|
10 2019 007 188.3 | Oct 2019 | DE | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/DE2020/000238 | 10/12/2020 | WO |