Manufacturing Intelligence: From Data to Decision Making
Most Manufacturing organizations today capture tons of data during the day to day transactions that are carried out on the shop floor. Thanks to cheaper disk space and IT departments eagerness to digitize all the data to be collected on the floor - large volumes of data are being collected. State of the art MES packages today enable a lot of data to be collected which complements the data that is captured by the ERP package that has been deployed. But inspite of so much of data the general feedback from shop floor supervisors is that they don't get the correct data at the correct time to help them take those crucial decisions on the floor. There are primarily two kinds of issues: the speed at which the required data can be recovered and the flexibility to slice and dice the available data accross different dimensions to help the shop floor manager take those critical decisions.
The challenges here can be summarized into the following buckets:
1) Integrate and Analyse data accross multiple systems
2) Ability to quickly generate these reports accross multiple dimensions
3) Ability to quicky identify and isolate problems
4) Ability to predict problems based on past data
Just to get an idea of how complex this can become is described below based on sample data collected from one such shop floor of a very large manufacturing organization.
1) Machine/Equipment
2) Data/ Time/ Shift
3) Employee/ Supervisor
4) Lot/Batch
5) Customer/ Supplier
6) Product/Product Family
7) Quality Codes/ Failure Codes
8) Job/Sales Order
The interesting thing to note here is all the above could potenatially come of multiple systems. The challenge would be to normalize the above the data onto a common baseline and help in the analysis.
Today companies are struggling to make real use of all this data accross multiple systems. They have all the data but to cull out the information or the trend from all the above discrete pieces of information is the key.
Comments
As far as real time data in shop floor is concerned, the simplicity of the presented data in the floor, is important from the supervisors' point of view. The supervisors who need to take real-time scheduling/dispatching/resourcing/loading decisions, need the data in a form that aids quick decison making. Along with the presentation of the data collected and collated, visual cues based on that data provide an extremely powerful tool to the supervisors in taking decisions on the shop floor. In addition to slicing the data along various dimensions as you have mentioned, data summarization levels targeted to meet decision making needs of the users of that data is also important in ensuring informed decisions at respective levels.
Posted by: Rupam Das | December 9, 2008 5:14 AM