The Manufacturing Operations Blog

Now that the case for Operational Intelligence is relatively well established, we need to discuss how to build a solution for it. Before we get there though, let’s recap the challenges we are up against

- Information is scattered all over the place. It’s in multiple source systems that came from different worlds of technology. Apart from ERP and MES, manufacturing data is in MES systems, Quality data bases, data historians, PLCs, SCADA systems, and whole variety of controllers, sensors and instruments.

- Manufacturing data is not organized, neither structured. If it did, you could just put a nice visualization tool over it and have your Operational Intelligence ready. Exactly the kind of pixie dust solution some vendors in this space offer. I won’t name them, as this blog does not believe in competitor bashing.

- Some of the manufacturing data is only on paper and in some cases not even on paper.

- It’s not just the data, even the manufacturing process model is scattered all over. No single system has a complete view of the manufacturing process.

- Manufacturing data comes in different levels of granularity often the context missing between two different levels of granularity. For example, you can get data from PLCs that’s in bits and bites but context that can relate that data to a manufacturing process or product is missing.

An effective Operational Intelligence solution must overcome the challenges. To overcome the challenges an OI solution would need all of the existing tools and technologies and some more.

- Data models that can help bring together data from multiple data sources with varying levels of granularity. (More about data models in a subsequent post…)

- Data integration tools i.e. traditional ETL tools to gather and bring together data from multiple data sources.

- Shop floor communication technology, OPC and all proprietary drivers that can help connect with real-time data sources such as equipments, PLCs and SCADA systems.

- Real-time data processing and transformation to convert raw data into information.

- Data warehousing and Business Intelligence techniques to make the information easy to report on.

- Visualization tools that can present the information to users in an easy to use manner.

- Exception and event management frameworks.

- Data mining and predictive analytics that help diagnose the problem proactively.

- Variety of data collection mechanisms including provisions for manual collection of data.

- Variety of data cleansing and data classification mechanisms including manual approaches.

If I made the solution seem impossible or the problems insurmountable, that’s unintentional. It may be difficult but it’s certainly not impossible. Most of the tools and technologies exist today.

Shop floor communication has been made a lot easier by vendors such as Kepware Technologies, Matrikon and ILS Technologies. For far too long, vendors like Kepware have been working behind the scenes and helping connect to variety of controllers and hardware equipments.

Data integration tools such as Oracle Warehouse Builder, Oracle Data Integrator (previously known as Sunopsys) and Informatica have made it easier to bring data together from multiple data sources.

Visualization and Business Intelligence technology has been made easier to use technologies such as from Hyperion, Siebel Business Analytics (now known as Oracle BI Enterprise Edition).

The time has come for a solution to weave it all together with a data model and by adding other missing components such as real-time data processing and events and exception management frameworks.

A credible solution for Operational Intelligence must acquire manufacturing data in real-time, establish the operating context for that data, and ensure that it is stored in a manner that makes it accessible at later times for a variety of uses.

The solution must facilitate the visibility of this real-time manufacturing data to users throughout the enterprise, and includes tools for enhanced decision support and role-based display of performance against operating metrics. Events and alarms should notify all who should be informed when important or unforeseen events take place.

An Operational Intelligence solution that meets these criterions can help companies achieve a higher level of manufacturing excellence by allowing them to better leverage their detailed manufacturing information.