Method for obtaining information from X-ray computed tomography data for optimizing the injection-moulding process of short-fibre-reinforced plastic parts
Method for obtaining information from short-fibre-reinforced plastic components sequentially produced by an X-ray computed tomography. A learning phase includes: generating CT data sets for a random sample of plastic components from a production process; extracting at least one defect-free region of the plastic components; determining characteristic feature(s) in the extracted regions, relevance of individual features, and regions which are characteristic of the plastic component type and production process thereof, over a predetermined period of the plastic components productions, which exhibit considerable characteristic differences between good parts and reject parts; and defining the feature(s) with its characteristic as trained classifier. An application phase includes: generating a CT data set of the plastic component for inspection; classifying the inspection part based on the trained classifier; examining the characteristic of the feature(s) for a negative trend; and automatically provide a negative trend alert and/or change process parameters to counteract the negative trend.
1. A method, with a learning phase and a subsequent application phase, for obtaining information from short-fibre-reinforced plastic components which are produced in series by means of an X-ray computed tomography method,
wherein the learning phase comprises the following steps:
a) generating CT data sets for a qualified random sample of plastic components produced by means of a production process;
b) extracting at least one defect-free region of the plastic components;
c) determining the characteristic of at least one feature in the extracted regions as well as the relevance of individual features and regions which are characteristic of the type of plastic component and the production process thereof and which, over the course of time of the production of the plastic components, exhibit considerable differences in their characteristic between good parts and reject parts;
d) defining this at least one feature together with its characteristic as trained classifier;
wherein the application phase comprises the following steps:
e) generating a CT data set of the plastic component to be inspected;
f) classifying the inspection part based on the trained classifier;
g) examining the characteristic of the at least one feature for the presence of a negative trend with respect to this feature in comparison with the plastic components in the defect-free region previously examined;
h) automatically changing the process parameters in a manner which counteracts the negative trend or giving a warning that this negative trend is present.
2. The method according to claim 1 , in which it is investigated whether the characteristic feature corresponds to the distribution of good parts and reject parts learned during the learning phase.
3. The method according to claim 1 , in which the at least one feature comes from the following group: distribution of the fibre orientation; local entropy, in particular Shannon entropy; local intensity distribution, in particular mean, variance or standard deviation; local wall thickness; geometric features, in particular lengths, diameters or distances.
4. The method according to claim 3 , in which the feature of the distribution of the fibre orientation is then examined as to whether it exhibits a change of direction which results in the formation of swirls.
5. The method according to claim 4 , in which the change of direction is determined in that, for any desired points in the plastic component, the orientation of the fibres in the space is determined and the frequency of each direction is calculated for an environment that can be predetermined in each case.
6. The method according to claim 1 , in which step g is carried out only at selected points in the CT data set and only in a volume that can be predetermined, in particular in a cube around the respective point as central point.
7. The method according to claim 6 , in which the edge length depends on the resolution of the CT data, in particular such that the relevant features can be imaged well.
8. The method according to claim 1 , in which, in addition to steps c and d, not only a defect-free region, but also at least one defect-prone region is extracted in the learning phase, and an examination of the characteristic features for the presence of a negative trend with respect to these characteristic features in comparison with the plastic components in the defect-prone region previously examined is effected.
9. The method according to claim 1 , in which the steps of the application phase are carried out either for every plastic component or only for a random sample of these.