Campaign analysis

Demo path

1. 1 Samples and conditions

   Read tube quality alongside the size of changes in operating conditions.

   Start with the selected samples

   The demo starts with 009, 011, 022 and 041. Columns compare tube results and machine and process parameters. Change samples opens the daily list: select up to six samples and inspect diameter, ovality, wall thickness, lumens and surface measurements in the sample details.

   Use one shared reference

   Click a sample ID in the matrix heading to make it the Reference. Values shows means in their original units; Differences shows how each sample differs from that reference. A + sign means an increase and a − sign means a decrease. The objective comparison uses this same reference.

   Associate each value with the right time

   An associated condition is not the reading at the instant the tube exits: it is the mean in an earlier source window, estimated using that parameter’s latency. Range is the maximum minus minimum of these sample means, not the height of a process spike. Warnings indicate incomplete coverage or configuration changes within the source window.

   Decide what best means

   In Which sample best meets the objective? choose lower ovality, lower roughness or higher minimum lumen circularity. Ranking includes only selected samples with the same materials, geometry, head and die as the reference. Changing the objective may change the winner; ties and insufficient data remain explicit. Associated conditions suggest what to test next, not a recipe to apply.

   Investigate a change

   Check the parameters you want to follow in the matrix and open Trends: the sample selection is preserved. You are not limited to pressure and speed. The Trend action on a condition associated with the best sample opens that signal in the context of that sample.

2. 2 Trends and recurrences

   Check whether a change is isolated or appears again with a similar shape.

   Read synchronised signals

   Add or remove parameters using the checkboxes. Charts share machine time and keep original units: two units use left and right axes, while more units use aligned lanes. To see whether pressure and speed change together, compare the same time, not just curve height. Bands identify sample source windows; gaps indicate unavailable readings.

   Isolate the shape to search for

   Set From and To just before and after the interesting shape, then run the search. Load the example spike sets a segment and two demonstration signals without changing selected samples. You can then change times and parameters. Constant signals cannot contribute a shape comparison; insufficient data requires a wider segment or different signals.

   Expand one dimension at a time

   Scope chooses which signals to search: the same parameters, or corresponding parameters on other extruders too. Extend to the whole day changes the time span instead: unchecked, search stays within sample source windows with ten minutes of context; checked, it searches the day excluding the starting segment.

   Inspect each candidate

   Similarity describes how closely shapes match, not the probability of a failure. The threshold separates accepted matches from less similar candidates; no significant recurrence is also a valid result. Click a candidate ID to inspect its actual signals. Back to the starting segment restores the original view without changing search criteria or samples.

3. 3 Temperatures and groups

   Check the intensity and duration of thermal excursions and which products share similar conditions.

   Look at excursions first

   For extruder 3 zone 4, read the minimum, maximum and its time, high reference, episode count and time above reference. Click an episode: the chart zooms in and the table identifies associated samples and groups through their source windows. A single associated group is selected automatically. Excursions have dashed bands; Chart data lets you inspect values, ten rows per page.

   Identify the products in each group

   The four groups use all 48 samples: diameter, ovality, minimum lumen circularity and the three extruders’ pressures. Temperature is examined afterwards; it does not create the groups. Click a member ID to open its measurements and highlight its latency-estimated source window in the chart, with a filled, solid-bordered band, ID and times, plus twelve minutes on either side. If the sample is associated with the active episode, both remain visible; otherwise you leave the episode detail while keeping the group. Show the whole day removes the zoom and sample band without changing the group. Sample selection and group membership do not change.

   Follow the initial group 3 example

   Initially the group with the most selected samples is highlighted: group 3 contains 009, 022, 035 and 041, three of them already selected. The 10:38–10:47 excursion is associated with sample 022 in group 3. The conditions below still compare the entire four-member group with the other 44 samples in groups 1, 2 and 4, not only exposed sample 022. Choosing a group not associated with the excursion returns to the daily overview.

   Interpret exposure and conditions

   A sample is exposed when its temperature source window overlaps an excursion. Samples in excursion lists the IDs associated with the selected episode: one episode can involve several samples and groups, so four excursions do not correspond to four groups. Mean of maxima summarises the maxima of individual sample windows, not the absolute daily maximum. A more frequent association is a lead to investigate, not proof of a cause.