Thursday, February 7, 2013

Big Data in Manufacturing

By Michael Schwarz, Invensys, Manager, Advanced Applications Product Marketing
The term “Big Data” describes the dramatically growing amounts of data being generated, transferred and stored. It is overlaid with the increasing use of information technology, global networking, and digital devices to manage, view, store, and control it.

Big Data is not just for consumers.  The same phenomenon is happening in business management, commerce, entertainment and social communications, as well as in manufacturing operations.  Big Data demands storage capacity and computing performance, driving new opportunities for enhanced intelligence and tools to offer users informed decision making.

Where Big Data does comes from?
The Big Data trend started in the 90’s, with R&D and commercial offerings responding to the growing requirements of managing the data streams of B2C portals, search engines and B2B, telecommunications and social networks. Today, with more than 1 billion people accessing the internet, large amounts of transactions and volumes of new data are generated every hour.  Just think of Twitter, with its massive volume of tweets being stored, transferred and posted by millions of global community members every minute.

The race to be at the forefront of Big Data in most public and private sectors is happening, and is served by its own large and rapid growing industry.   The commercial offerings for Big Data can be roughly classified into data storage, with significant increased performance (e.g. data in memory) and data management technologies and services (analytics applications, algorithms and architectures). 

 Big Data management capabilities have entered the mass market, and it is now possible to boost existing applications with faster storage technology at reasonable costs, reducing wait times and increasing usage of existing information systems to drive better decisions in day to day operations. 
New analytic reporting and viewing software tools combine algorithms to effectively analyze big data with innovative visualization techniques for collaborative use of information, providing self-service access to both “real time” manufacturing as well as the more latent business data.

Big Data in manufacturing
Some may see a mismatch between Big Data and their existing manufacturing data, but in reality, processes running 24/7 are already are generating, capturing, transferring and storing data in significant volume, sometimes in fractions of a second. Not every process operation leverages the data that is generated or available to their full benefit today, but the competitive pressures of the market for  cheaper  and better products and more efficient operations require companies to leverage all assets in the company that contribute to greater knowledge & control of the processes. Using the available Big Data from shop floor operations for higher effectiveness is a natural step.
The nature of Big Data is often described as technology which takes enterprise decision support, analytics and reporting to become information in real-time.  But for manufacturing companies, that Big Data is sometimes hidden in historian databases, MES databases, or even in the OEM equipment that is sitting on the shop floor.  These are traditional “manufacturing” databases holding time/date stamped, transactional data, and in some cases, may already have a reporting front end.   
And, it is typically not considered Big Data when it is distributed in multiple systems, such as Process Historians, MES or Quality, Asset, or Warehouse management systems. Big Data is different than a SQL Report.   Big Data takes the disparate bits of data from these real time databases and makes sense of it, by adding a layer of analytics to provide real cause and effect, or “what if” information, not just reporting alarms and events.  
 Information, in context, is invaluable.  Understanding how and why your plant is performing, against other plants in your enterprise and your competition is essential in today’s economic climate. This context is not complete without the combination of key measures, in near-real time, across maintenance, operations and supply chain performance.
Combining the so far separate data and information complements each other for improved planning and decision making with awareness of all operational activities taken into account.
Such source of information continued updated with the current status of operations execution enables the real time enterprise by analyzing and monitoring the cause and effect relationship between business control and process/production control.
Big Data is not just for manufacturing
Simulation software can also use large data sets to provide a more accurate understanding of operations and provide opportunities for process optimization. For the case of performance monitoring, large data sets of equipment performance history can be analyzed so that the behavior is reviewed for performance patterns. Reviewing performance trends enables engineers to predict equipment failure or determine optimal maintenance schedules.
In the case of optimization, the large data set won’t be historical; rather it will be in the form of data from many different pieces of equipment, market prices of utilities, raw materials, and product prices. The software can take all of those data points into account to optimize the process for profitability by making changes to the operating conditions.  Both cases rely on  technological and computing power advances to use large sets of data to gain significant information.
Big Data for Asset Management
Every decision made in a plant environment is dependent on quality and timely access to information. Most Enterprise Asset Management (EAM) systems are capable of capturing vast amounts of transactional data but users often find it difficult to navigate and interpret it.  This is due, sometimes, to the Maintenance department working independently, focusing only on specific areas of responsibility centered on asset availability and cost containment.  Without seeing the big picture and the impact on other departmental areas, the maintenance team can lack  the visibility to collaborate  and do their part to improve the overall operation. 
Technology advancements condition us to expect access to our  information, on our device of choice.  With self-service visualization of  key performance business measures, you are empowered with the ability to transform your Big Data into bite sized pieces you can act on.
Questions?  I’m happy to answer:

No comments:

Post a Comment