From Raw Data to Meaningful Insight

The mass expansion of the industrial Internet of Things IIoT has brought with it a sudden and significant increase in the amount, complexity and accessibility of operational and equipment data in process manufacturing facilities. Combined with the emergence of artificial intelligence (AI) and machine learning (ML), this is providing the potential to uncover more meaningful insights than ever before. 

However, the journey from raw data to meaningful insight is still disjointed for many process manufacturers. The leading causes include limited data access and connectivity, a lack of time-series-specific analytical solutions, and collaboration difficulties, and addressing these issues is paramount to process optimization. 

Spreadsheet-caused challenges 

At most facilities, numerous data sources exist, creating equipment, process, quality, and inventory data, but this information is typically stored in a variety of different databases. Historically, spreadsheet-based analytics tools were used to aggregate, cleanse, and align all this data so insights could be extracted. However, this manual procedure was cumbersome and time-intensive for engineers and process experts, and it was certainly not the most efficient use of these resources’ skillsets. 

On top of these manual inefficiencies, the lack of live data connectivity left subject matter experts (SMEs) with perpetually out-of-date analyses. These challenges made it difficult for SMEs to wrangle data and prepare it for meaningful analysis. Furthermore, traditional solutions rendered sharing data and analyses across organizational teams and regions arduous or nearly impossible, limiting the ability for collaboration and knowledge transfer. 

Traditional spreadsheet-based workflows are still active in several facilities today, but this severely cripples organizations because these tools are decoupled from real-time data visualizations, making rapid and iterative data analysis prohibitive. Fortunately, better solutions are now widely available. 

Modern advanced analytics solutions 

To transition their analytics capabilities, increase operational efficiency, maximize profitability, and achieve ambitious corporate objectives—including digital transformation and sustainability metrics—process manufacturing organizations are rapidly implementing cloud-based advanced analytics solutions into their daily procedures. These software platforms are optimized to connect disparate data sources and immediately alleviate the challenges of live data connectivity. 

Additionally, these solutions provide native tools for data cleansing, time stamp alignment, and contextualization, empowering SMEs to quickly derive reliable insights referencing all available data. With live data connectivity right in the software, SMEs can apply their analyses to near-real-time data, whether it is stored in the cloud or on-premises. 

Because advanced analytics solutions are cloud-based, they are perfect for IIoT implementations, which gain direct access to the vast computing power and scalability of cloud software, advancing all types of Industry 4.0 projects, such as predictive maintenance programs and digital twins. 

Increased accessibility and collaboration yield innovative industry solutions 

Advanced analytics solutions also strengthen the collaboration between process, maintenance, and reliability teams, with built-in tools for sharing analyses and insights in easily digestible dashboards and reports. 

One petrochemical and refining company experienced a slew of reactor shutdowns, caused by the failure of a critical feed gas compressor on a polyethylene line, with the inability for immediate restart. On this line, an unplanned reactor shutdown creates a minimum of 4 hours of downtime, costing the plant upwards of $200k USD with every occurrence. These compressors were previously maintained on a preventative maintenance (PM) schedule, but this did not prevent unplanned shutdowns entirely. 
 
Following one compressor trip, machinery, controls, and electrical engineers worked together to identify the source. However, tracing electrical diagrams around the pump motor was time-intensive, and it failed to yield a root cause. 
 
Taking an alternative approach, a process engineer at the refinery opened a webpage in Seeq, an advanced analytics solution, to quickly locate the five most recent shutdowns and subsequent restarts (planned and unplanned) from decades of historical process data. Using “Capsules” and “Chain View” tools, they quickly focused on the shutdown and start-up time periods, overlaying the events, which led to identifying abnormalities in the discharge pressure profile of the two most recent start-ups (Figure 1).