Applying Western Electric Rules only to subsections of a chart

[Former Member]

Does anyone know of a method where the alarms in the alarms configuration tab of the iSPC chart could be applied to only a subsection of an iSPC chart, and not the entire dataset presented on an iSPC chart?

For example, if I have an xbar chart with sampled subgroups 1 through 25 displayed, would it be possible to only apply the alarm conditions to say, groups 15 through 25 on the xbar chart, and ignore the first 14 subgroups? Could I somehow do this in a transaction?

Thanks,

Justin

[jcgood25 · Accepted solution]

The chart calcs and alarming will always be based upon the entire dataset coming within your query results.

In BLS you also have the SPC/SQC action block which does all of the same behind the scenes efforts to produces the 'guts' of the iSPCChart, and will provide you with the same xml resultset as evidenced from the applet's right-click SPC Results detail information (summary stats, datapoints, alarms, etc).

In your Transaction you could provide the additional WHERE clause filter necessary to restrict the dataset to only the desired subset and use this to feed the SPC/SQC action block. Then you can find the necessary data in the relevant Rowset[#] and react accordingly.

[Former Member · Accepted solution]

Hi, Jeremy.

Just a reminder that the SPCChart does allow segmenting of the dataset based on date and control limit effectivity ranges. It should be do-able to refactor the alarm logic to reset accumulators whenever there is a control limit change..

Rick

[Former Member · Accepted solution]

Jeremy,

Is it possible to apply western electirc rule independently on the different subset of data on the same chart?

Thanks

SS

[former_member779911]

Hi,

I am new in SAP QM, can anyone please help me know how the western electric rule work & what configuration required for it.

Thanks & Regards

Salahuddin Shaikh

[former_member763929]

Thanks for coming to SAP Community for answers. The original question has received a helpful answer already, please post your question as a new question here:

Since you're new in asking questions here, check out our our tutorial about asking and answering questions (if you haven't already), as it provides tips for preparing questions more effectively, that draw responses from our members.

Please note, that your post here won't be answered.

[Former Member]

Thanks Jeremy for your response.

Your response got us thinking about what we might be able to accomplish to solve our problem. Here is a more detailed description what we are trying to solve.

We are using iSPC charts, and calculating our own control limits via a transaction. For our iSPC charts we've configured the alarm conditions in the iSPC chart display template. Here are two scenarios we are trying to solve:

1. We have a chart with 20 points, and SPC applied to only the last 10 points. We would like the SPC alarm conditions we set in the display template to begin starting at the first point with SPC applied (i.e. point 10 going forward).

One of our alarm conditions is set in the DT as alarming at seven consecutive points, with each point greater than the subsequent point. In our case, we would not want the alarm logic to consider any points for which there is no SPC (before the 10th point). Therefore, the earliest point where this alarm condition can show in our example would theoretically be point 17.

2. We are also plotting multiple phases of SPC on our charts (a single chart can have centerline value X from points 1 to 10 and centerline value Y on points 11 to 20). Is there a way to reset the alarm conditions setup in the DT between phases? Meaning that at the begining of the second SPC phase at point 11, no points prior would be considered in the alarm logic.

As an example, if points 5 through 10 were subsequently increasing, and point 11 (first point of new phase) was greater than point 10 (the seventh point increasing in a row), point 11 would not alarm because it is within a different SPC phase.

We are currently suppressing points in the chart that do not have SPC limits applied to them. This partially solves our problem, as now points without SPC do not exhibit alarm conditions. However we are trying to find out if suppressed points are considered in rule logic, where 7 points subsequently increasing causes an alarm, and if the first two points of the seven are suppressed, will the alarm still trip on the 7th point?

Thanks again for the guidance.