Seal Integrity X-ray Inspection
INNOVATION IN SEAL INTEGRITY ASSESSMENT
Innovation in seal integrity assessment methods is necessary for ensuring food quality and safety. New detection technology combined with new X-ray source technology, has facilitated that X-ray systems can be used at much lower energies than traditional X-ray systems used for foreign object detection. The usage of X-rays with an energy between 15 and 20 kV in on-line applications thus facilitates imaging of very small composition, density and/or thickness differences, down to the nano-scale.
This is the fundamental basis for making it possible to inspect and detect in real-time seal imperfections and leaking in e.g. plastic pouches. Deviations in the seal thickness down to fractions of a micron can thus be detected and the object rejected, at production speed up to 180 units per minute.
SEAL INTEGRITY OF FOOD POUCHES
So how does seal integrity assessment work? The results below summarizes the functionality for a HYMCIS solution for the on-line inspection of medium-sized food pouches. The pouches are conveyed through the system on the system built in conveyor, featuring a thin special conveyor tape. The speed is maximum 2 pouches per second at a belt speed maximum of 27 m/min. The pouches are fed into the system with the sealing PARALLEL to the conveyor motion direction.
Fig. 0 The food pouch considered.
The system will automatically inspect and classify the pouches according to the following:
1.Misaligned seal (out of tolerance – tolerances defined by Xx);
2. Missing sealing (minimum dimension/width of the missing seal is likely > 5 mm)
3. Product in seal
4. Foreign object in product
5. Product mass out of tolerance
The system will categorize as follows for pouches that cannot be correctly classified:
– Irregularly conveyed pouch (misaligned)
– Dirt (water etc) on pouch
– Dirt (water etc) on conveying belt or in other parts of the imaging “chain”
– Highly irregular pouches (out of shape)
Please note in particular, that the system is very sensitive and will detect even very thin contaminations in the seal. Therefore, water droplets etc will give false rejects.
Typical performance is as follows:
– (Seal) False positives: < 1 % – (Seal) False negatives: < 3 of 100
– (Foreign object) False positives: < 0.5% – (Foreign object) False negatives: < 3 of 100
HOW DO WE ASSESS SEAL INTEGRITY?
In order to ensure safe detection of defects and a low false reject rate, the samples must be imaged without obvious defects such as strong misalignment, package curvature etc.
As it is well known that the pouch seal area may fold or bend, the X-ray system may include a feature so that the seal is presented straight during the imaging, e.g. as described in Fig. 2, straightening out by a set of rollers that carry/conveys the seal across the X-ray imaging area.
X-RAY SEAL INTEGRITY - SYSTEM IMAGE CAPABILITY
If there are any seal integrity issues, such as defects in the product it will be shown in the X-ray image as dark zones. This will be shown later in this report. For the package from the image below the edge is wrinkled which is also shown as a dark zone in the X-ray image. When packages are wrinkled like this, the wrinkled area will not be possible to inspect, since a real defect will look quite similar on an X-ray image.
To avoid the packages being wrinkled, a set of rollers may be used just before scanning, which will flatten out the sealing. Figure 1 shows a package with wrinkled seal, which might be rejected, because it looks like contamination in seal. The sealing in this package will be flatten out with a roller, and a false reject will be avoided. In figure 1 we also see that the distance between the welded zones are too small, which indicates wrinkled sealing.
Fig. 1 X-ray image displaying wrinkles in the seal.
ROLLER SOLUTION FOR WRINKLED SEAL
Fig. 2 Drawing of a roller arrangement for straightening out possible folds in the seal area, and any other seal integrity irregularities.
ELEMENTS MISSING IN SEAL
Fig. 3 X-ray image displaying missing seal. Welding zone (crest/bead of the weld) are missing in the middle of the sealing (outer seal missing). This defect can easily be detected, since in that area, only one welding zone (the inner, defined by the crest/bead of the weld) can be detected. In the area with missing weld zone, the GLV is the same as outside of the weld sealing, which also can be seen from the profile plot.
FOLDS IN SEAL
Fig. 4 X-ray image displaying folds in seal. In this image the folds in the packages are very clear. Also by analyzing the sealing area with vertical profile lines, the folds shows up with very low GLVs compared to rest of the sealing.
PRODUCT FRAGMENT IN SEAL
Fig. 5 X-ray image displaying product fragment in seal. This particular image displays a fragment of meat stuck in the seal of the food pouch. Meat has a very high density (high attenuation) compared to the seal area and is easily detected both with horizontal and vertical profile lines. GLV for the meat is below 2100.
ELEMENTS MISSING IN SEAL
Fig. 6 Another X-ray image displaying missing elements in seal. Analysis and detection for missing seal is similar, but it is more well expressed in this sample. The width across the seal varies in this package with 66 pixels≈ 6,6 mm. This is detected using horizontal profiles and measuring the length in top and bottom of the package.
METAL FRAGMENT IN SEAL
Fig. 7 X-ray image displaying metal fragment in seal. Metal and foreign objects in general will appear very clear on an X-ray image, since they attenuates the X-rays much more than the product inside packages. This is seen in the profile, showing distinct dip coinciding with the position of the small metal piece.
The presentation above has shown the various types of defects in the seal of (food) pouches and described the methods for revelation/automatic detection of these. In general, the seal defects (contamination) are clear and consistent, but of course there is a lower detection capability when the contamination is very small.
Seal imperfections are clearly seen in the examples above, however the automatic correct classification is not considered likely if the seal imperfection is less than 5 mm in length.
Tolerances for the detection of misaligned seal must be provided on a case by case basis. The same is true for other defects – such as the determination of the mass of the product. The system capability is a tolerance measurement better than 1 mm.