Managing Data: Single- vs. Multi-Platform Approaches

There are advantages and disadvantages to both single-platform and multi-platform applications and software packages. Finding the right approach will help you put your data to work.

Eric Reisz, Panacea Technologies
Eric Reisz, Panacea Technologies

Big Data has become quite the buzzword in industry, but how do you know what data you should actually be collecting, and how best to put that data to work rather than just letting it take up space on your servers?

There are a wide range of solutions available to help. The ideal scenario would be to find one that meets every need, but the reality is that multiple applications and software packages are required to meet complex company-driven standards. So what exactly is the best approach for selecting software packages?

A number of platforms like to advertise that their solution is the “total” package, will fit all your needs, and is plug and play. In our experience, however, this isn’t always the case—and “plug and play” usually means that once it’s plugged in, you get to play with the system for a long time before it works.

When selecting a platform, you must keep in mind the end-user requirements. Let’s consider adding a historian component to an existing process area’s control strategy. If the intent is to use the historical data to support batch release and the scope is limited to process equipment controlled by a single distributed control system (DCS) or programmable logic controller (PLC) platform, then using the platform’s historian component with premier integration is a leading consideration.

Adding in various pieces of equipment or skids that function independently from the control system—for instance, lab equipment or filling skids—becomes trickier. In this scenario, adopting a historian platform that can coalesce the different data sources into a single data model becomes a more desirable approach.

An advantage to adopting a separate historian is that some historian platforms offer the additional capability of customizing data analysis. This is advantageous to some end users because it will allow them quick access to data and the additional ability of automatically processing it into a format usable for their purposes. In either case, it is important to have a reporting platform, which can automatically generate reports and greatly decrease the amount of paperwork that needs to be completed, freeing up more time for productive activities.

The concept of putting your data to work for you is gaining a lot of ground in the industry. After all, why collect data if it is just going to sit there? Put it to work for you! The various components of a plant work together to produce the final product, so a single break in the production chain can be devastating. Analyzing collected data to correlate events in a plant gives us the ability to predict when preventive maintenance is required before a catastrophic failure of a component happens. Making data work for you can mean dramatic drops in unscheduled downtime and maintenance costs.

For unique situations or requirements, another option would be to use multiple platforms to collect the data with direct or indirect redundancy. The reasons for using this method could include one oddball piece of equipment that will only talk to a particular application; or that a manufacturer never got around to migrating a system. It is not uncommon to run into instances where a specific application must be used because of equipment limitations.

Before looking for a software package, take stock of what is already available or being used by other groups to fulfill similar requirements. It is most likely that expanding the use of an existing solution will be more cost-effective and quicker to implement than adopting a new solution altogether.

In our experience, unless there are narrow end-user requirements, or the plant is unified on a single vendor, we don’t often see a single historian/reporting platform implemented. More commonly a multi-system solution that employs layered applications providing unique additional functionality can allow for high versatility and meet the needs of many different end-user groups. A multi-system solution can also have major business benefits by providing a uniform method of process control, data archival and analysis, and report generation from a multitude of data sources, forms and inputs.

A downfall of multi-system solutions is when there are multiple systems that are duplicating functionality. It is important to consider your end user, your plant environment, resource skill level and desired end state with data historization. Data is a utility and requires careful planning.

Eric Reisz is an automation validation engineer at Panacea Technologies, a certified member of the Control System Integrators Association (CSIA). For more information about Panacea, visit its profile on the Industrial Automation Exchange.

 

More in Data