Method and device for detecting and differentiating between contaminations and accepts as well as between different colors in solid particles

[0002] Product quality or color control is an essential quality criterion in many technical on-line processes. The production of plastic containers, in particular plastic bottles for the beverage industry, using granulated recycled material, requires separation and sorting of different color portions, different plastic types, e.g. polyethylene, polyamide, polyvynylchloride, the recognition and sorting of e.g. fractions contaminated with gasoline, diesel, benzene, toluene, xylene, to enable re-use e.g. in the food industry. To achieve this object, the reusable pure base materials must be differentiated and separated from the contaminated granulates and moreover, the fragments having different colors must be sorted to obtain portions of pure color.

[0003] Prior art methods and devices unsatisfactorily achieve this complex object to a limited extent only. Devices are known which permit recognition of color with subsequent color sorting by controlled air flow nozzles and using CCD color cameras. These systems are disadvantageous in that the measuring speed is unacceptably slow and the signal-to-noise ratios too low, and are therefore not suitable for use in the present technical process. Moreover, methods for identifying and separating different plastic materials have been developed. The methods used thereby are not suited for use with granulates or granulate-similar portions of recycled materials, since the intended identification and sorting of smaller or larger bottles, containers etc. requires 1000 times faster measuring speeds to cope with the minimum mass flow required for economic reasons. A further drawback consists in that the detection of colors and the differentiation of materials require separate systems using different physical principles, i.e. CCD cameras for detecting the color or NIR absorption for plastic classification, wherein the total cost per system is above the economically justifiable level. The largest drawback is the fact that none of the above-mentioned methods allows proof of contamination by foreign matter deposited in the base material. The latter is absolutely required e.g. within the scope of recycling plastic material used in connection with food, which is to be used a plurality of times as food containers.

[0004] It is therefore the underlying purpose and object of the present invention to propose a novel method or devices which not only enable identification of color and classification of material but also identification of contamination in the base material, wherein, for economic reasons, the three problems shall be solved simultaneously using only one single basic physical principle.

[0005] The present invention solves these three problems in that the recycling granulate, which is transported on a conveyor belt, is illuminated by a laser beam and the radiation re-emitted by the granulate is spectroscopically analyzed within a broad spectral range. The invention additionally provides means for extensive detection and evaluation of the spectra to determine the colors of the base materials and the contamination of the granulates, in real time. The invention is explained in more detail below in FIGS. 1 through 9.

[0006] FIG. 1 shows the basic construction of the overall system for on-line detection and sorting of contaminated granulates or tablets of different colors and different base materials;

[0007] FIG. 2 shows the optical arrangement for illumination of the measuring material and for detection of the radiation re-emitted by the measuring material using a polygonal mirror wheel and a toroidal mirror;

[0008] FIG. 3 shows the optical arrangement for illumination of the measuring material and for detection of the radiation re-emitted by the measuring material via a polygonal mirror wheel only;

[0009] FIG. 4 shows a side view of an essential part of the optical arrangement shown in FIG. 2;

[0010] FIG. 5 shows the construction of the toroidal mirror used for radiation deflection;

[0011] FIG. 6ashows a first selected arrangement of the sensor technology and signal processing for ultra-rapid evaluation of the optical signals;

[0012] FIG. 6bshows a second selected arrangement of the sensor technology and signal processing for ultra-rapid evaluation of the optical signals;

[0013] FIG. 6cshows a third selected arrangement of the sensor technology and signal processing for ultra-rapid evaluation of the optical signals determining contamination.

[0014] FIG. 7 shows an example of emission spectra for detecting and selecting the color;

[0015] FIG. 8 shows an example of emission spectra for determining the material;

[0016] FIG. 9 shows an example of emission spectra for determining contamination.

[0017] In accordance with FIG. 1, a laser (4) beam is guided over the measuring material (3) in a linear manner using an optical system (2) (further described in FIG. 2 or FIG. 3). The laser beam (4) scans the entire angular range (5) for detecting the entire width of the conveyor belt (6) which is supplied with granulate material (3) via a silo (7) or with tablets or other test material via a feed device (not shown in detail). Illumination of test samples (3) is effected point-by-point (9) through cycling of the laser light. Linear scanning is also feasible through use of a continuously radiating laser (1). The secondary light (10) generated through scattering, fluorescence, Raman laser radiation, reflection and other optical re-emission effects is detected by the optical system (2) in a wide angular region and supplied to a spectrometer (11) whose signals are processed by an evaluation unit (12), exemplarily illustrated in FIG. 6. Following classification of the test samples (3) into individual color classes, separate material classes and contaminants, they are sorted via a conventional separation system (e.g. via cycled nozzles) into different separate portions (13) which can be transported in this form to another production process or, if contaminated, to a disposal system.

[0018] FIG. 2 shows the laser (14) whose beam (15) is incident on a polygonal wheel (18) after collimation by a lens (16) and deflection via a reflector (17). The end surfaces (19) of the polygonal wheel (18) rotate at a high angular velocity and are formed as mirrors to azimuthally guide the laser beam (20), in a temporal saw tooth manner, over the three-dimensional concave toroidal mirror (21) to effect linear laser beam scanning (9) of the test material (3) on the conveyor belt (6) of FIG. 1. The divergent re-emitted light (10) (see FIG. 1) is also collected by the toroidal mirror (21), transformed into a converging bundle of rays (22) and reflected at the point (22a) (where the laser impinges) to gain access to the spectrometer (11), shown in the upper part of FIG. 1. An optical filter (23) is provided at the entrance of the spectrometer which removes stray light coming from laser radiation (14) which is reflected e.g. on the surfaces of optical components such as eg. the lens (16) and which would otherwise gain indirect access to the spectrometer. The optical filter (23) is designed to suppress the emission wavelengths of the laser (14) which are preferably at emission wavelengths of the YAG laser, i.e. at 1046 nm and their frequency-multiple wavelengths of 523 nm, 349 nm, and 262 nm. The light focused, by means of the lenses (25,26), on the gap (24) of the spectrometer is incident, via a first concave mirror (27), on the optical grid (28) which decomposes the spectrum of optical radiation into its wavelength components and projects same, in dependence on the wavelength and via a second concave mirror (29), onto the sensor system (30). The sensor system (30) consists e.g. of a linear CCD array. Depending on the spectral range to be examined, linear arrays of Si photodiodes or Si photoelements or corresponding arrangements of In Ga As can alternatively be used. FIG. 6 shows, in more detail, a specially formed sensor system arrangement using photomultiplier arrays.

[0019] FIG. 3 shows the laser (31), whose beam (32) is incident on the mirror (36) of the polygonal wheel (35) following collimation by a lens (33) and deflection via the reflectors (34). The end surfaces (36) of the polygonal wheel (35) rotate at a high angular velocity and are formed as mirrors to azimuthally guide the laser beam (37), in temporal saw-tooth motion, directly onto the measuring material (38). In accordance with FIG. 1, this produces a linear laser beam scan (9) of the test material (3) on the conveyor belt (6). The diverging re-emitted light (10) (FIG. 1) is gathered by the mirror (36), and reflected to reach, via a lens (41) and mirror (42), the spectrometer (11) (upper part of FIG. 1). An optical filter (43) is provided at the entrance to the spectrometer to block stray light coming from the laser (31) radiation which is e.g. reflected on the surfaces of optical components such as e.g. the lens (33) and which would otherwise gain indirect entrance to the spectrometer. The optical filter (43) is therefore designed such that it suppresses the emission wavelengths of the laser (31). These are preferably at emission wavelengths of the YAG laser, i.e. at 1046 nm and their frequency-multiple wavelengths of 523 nm, 349 nm, and 262 nm.

[0020] The light focused by the lens (44) on the gap (45) of the spectrometer is initially reflected by a first concave mirror (46) onto the optical grid (47) which decomposes the spectrum of the optical radiation into its wavelength components and which images same, in dependence on the wavelength and via a second concave mirror (48), onto the sensor system (49). The sensor system (49) consists e.g. of a linear CCD array. Depending on the spectral range to be examined, linear arrays of Si photodiodes or Si photoelements or corresponding arrangements of In Ga As can be alternatively used. A specially designed sensor system arrangement using photomultiplier arrays is shown in more detail in FIG. 6.

[0021] FIG. 4 shows a side view of an essential part of the optical arrangement of FIG. 2, once more illustrating the laser beam (51) and its point of impingement (52) on the reflector surfaces of the polygonal wheel (53), its reflection (54) on the toroidal mirror (55) and its point of impingement (56) on the conveyor belt (57). The path of the rays (58) of the radiation re-emitted by the granulate or the respective tablet, which extends coaxially and opposite to the laser beam (51) is also shown. This special inventive optical path for the exciting laser beam (51) and for the re-emitted beam (58) is achieved, in particular, through the use of the toroidal mirror (55), having different radii of curvature in the sectional planes of FIGS. 2 and 4. To detect the required angular range of re-emitted radiation, a toroidal mirror is required having dimensions of approximately 1.5 m×0.3 m. For manufacturing reasons, the mirror is made from individual segments (60,61) (see FIG. 5).

[0022] In accordance with FIG. 6a, the spectrally decomposed light (71) is detected by a multi-anode photomultiplier tube (72). The respective spectrum reaches, via signal preparation (73) and an analog/digital converter (74), a programmed logic (75) which effects material identification, color determination and detection of contamination in synchronization with the optical deflection unit (76) of FIGS. 2, 3 and 4, the laser (77) and a special synchronization unit (78) and in cooperation with mathematical algorithms (79) and externally calculated calibration vectors (80) and passes respective commands to the sorting unit (81).

[0023] FIG. 6bshows a further variant of the signal evaluation, wherein the programmed logic (75) in FIG. 6ais replaced by analog signal processing using mathematical models (82). The individual signal channels of the multi-anode photomultiplier tube (83) are processed in parallel via analog switching elements in accordance with a respectively predetermined mathematical model. A downstream summer (84) adds the signal contents of all parallel channels into one single resulting signal which is classified by a comparator (85) into the following categories contamination or alternatively into the respective material class, including color.

[0024] The third variant of the signal evaluation uses a processor (86) in accordance with FIG. 6cwhich compares the currently measured spectra with the reference spectra stored in a data memory (87), assesses them and passes corresponding commands on to the sorting unit (89). Alternatively or additionally, the respective current spectrum can be assessed and classified via mathematical models located in a memory (88).

[0025] FIG. 7 shows color detection according to the inventive method for differently colored PET materials in the form of different spectra for the colors red (90), brown (91), green (92) and transparent (93). One can see that clear color differentiation is possible.

[0026] FIG. 8 combines the spectra of different materials with the base materials PVC (94), PET (95), PEN (96) and PC (97) which also differ with clearly distinguished characteristic curve shapes.

[0027] The contamination spectra obtained according to the new method are shown in FIG. 9. Using the example of a cellulose lacquer diluent, the spectrum (98) shows a high contamination concentration, whereas the spectrum (99) shows a relatively low contamination concentration in the PET base material. In comparison therewith, the spectrum (100) shows the non-contaminated base material PET.