Process intelligence helps users gain a specific understanding of their manufacturing processes and how to control their sources of variability.
For decades, companies have recognized the importance of understanding what is happening in their businesses at the enterprise level and using that information, known as business intelligence (BI), to make informed business decisions. This recognition led to the decision-support systems created in the 1970s and ultimately, in the 1990s, to the evolution of today’s more advanced BI systems. In the past decade, manufacturing intelligence (MI) systems, which access and aggregate manufacturing-related data from multiple sources, have gained significant momentum and increased their ability to focus on operational excellence.
Although MI gives companies valuable benefits, it does not provide the specific knowledge that all process manufacturers need. The key performance indicators (KPIs) that MI typically tracks and displays in dashboards are an important component of understanding what is taking place in the manufacturing organization. But once the enterprise level sees that yields are not in the expected range, for example, then the company must understand why, so that improvements can be made.
The missing piece
What MI is not delivering today is process intelligence, a critical subset of MI (see Figure 1). Process intelligence helps users gain a specific understanding of their manufacturing processes and how to control their sources of variability. Once this information is widely available, it can be used to achieve manufacturing-process excellence throughout the entire product life cycle using appropriate technological tools.
MI shows what is happening in the manufacturing process, and process intelligence explains why it is happening. Process intelligence solutions enable an understanding of the physics, biology, and chemistry of materials, recipes, and manufacturing processes. This understanding helps manufacturers improve process predictability, reduce start-up risks, improve product quality, decrease cycle times, and increase yields and compliance—especially in complex production environments regulated by good manufacturing practice (GMP) and the US Food and Drug Administration.
Process intelligence solutions improve control
Building process intelligence into overall MI and BI portfolios using a self-service data-access platform linked directly to the operational data stores of existing manufacturing-data infrastructure systems allows users to understand the sources of process variability. This technique enables them to achieve better process control from the process-design stage (i.e., in process development) throughout the entire product life cycle. Appropriate technology solutions for process intelligence use a hierarchy to help access and contextualize data for users across multiple levels of the ISA-95 model, offering a full view of manufacturing and process-development data.
For example, one global pharmaceutical manufacturer needed to reduce the burden on end users who wanted to examine their process data easily with standardized tools without waiting for the data and then massaging it for hours in spreadsheets before it could be analyzed. The manufacturer needed a single common system to access, contextualize, analyze, and report data and help interdisciplinary global teams compare batches manufactured at different plants.
The manufacturer planned to replace its proprietary system and provide self-service data access, contextualization, and analysis in a single, validated environment. Previously, every scientist used Excel spreadsheets to prepare reports, and these files would end up scattered around on various desktops. The company needed a robust system that could withstand a good manufacturing practice (GMP) audit.
The company looked for the following characteristics in a new system:
The IT department evaluated available commercial, off-the-shelf (COTS) systems according to these requirements and ultimately chose a COTS process intelligence platform for integrated self-service data access, contextualization, and analysis. The company realized that the data-aggregation, -contextualization, and -analytics offerings of commercially available data historians, enterprise resource planning systems, and statistics packages were not adequate to the task. The manufacturer needed the process-intelligence layer above these systems that made data available in a meaningful context to all levels of the enterprise.
After the company deployed the process-intelligence platform, it reduced the amount of time required to gather data for annual product reviews (APRs) by more than 90%. The high burden on IT for the support and development of in-house systems was eliminated.
In the end, process manufacturers usually see the following significant business benefits when adding the previously missing process intelligence piece: