Safety Instrumented Systems / Systems Integration / Asset Management / Fieldbus / Yokogawa

Winds of Change Complicate Automation Lifecycle

Forecast Tumultuous, but "the Cloud" May Actually Help

By Jim Montague

Yokogawa Users Group 2012

Change wouldn't be so bad—if it wasn't happening so fast. But just like the winds in tornados and hurricanes can turn common objects into missiles, the sheer velocity of technical and organizational change can make it difficult and even dangerous to handle.

Fortunately, a little prioritized data and clear-eyed understanding can help process engineers and managers keep their hats—and heads—on when the winds of change are howling. Peter Reynolds, senior consultant, ARC Advisory Group, offered generous doses of both in his keynote presentation this morning to the 2012 Yokogawa User Group conference in New Orleans.

Reynolds reported that plant automation is beset by several main forces:

  • Most process facilities use commercial off-the-shelf (COTS) devices that make plant automation look like an enterprise data center;
  • Control complexity is high;
  • Cybersecurity is a worry;
  • Data integration is sub-optimal;
  • Automation roles aren't clearly defined;
  • Lack of interaction between plant automation and enterprise IT;
  • Not enough support staff; and
  • Process automation users view IT as not adding value.

"These situations increase risk, hinder optimization and prevent migration investment decisions," said Reynolds. "They're holding a lot of plants back."

Likewise, recent ARC research revealed that only 50% of about 50 end users and system integrators responding to a recent survey are actively managing their industrial control system (ICS) cybersecurity. Despite the demonstrated risks, "there's a lot of ‘set it and forget it' around cybersecurity," said Reynolds.    

Dealing with System Obsolescence

These challenges are compounded by the fact that process control systems remain focused on availability and uptime, while IT departments remain focused on confidentiality. Similarly, there's continuing friction over the fact that PC-based data processing and related COTS systems are updated and replaced every couple of years, while process control systems are supposed to last for a couple of decades. "This makes it difficult to justify changes and updates in control systems because it's hard to determine when many automation systems are worn out," said Reynolds. "It's simple to test individual components, but it's much harder to figure out when an overall automation system has reached the end of its life and when it's likely to fail."

To explore how obsolescence is managed, ARC surveyed 80 global manufacturing companies in January 2012 and found that obsolescence is less about technology and more about organizational processes and people, and defining all three helps their company's profitability. The respondents had many methods for dealing with lifecycle and obsolescence issues, including:

  • Managing supplier relationships with more specific, long-term key performance indicators (KPIs);
  • Managing spare parts more effectively;
  • Balancing whole migration with systematic migration, either horizontally or vertically;
  • Training and then retraining staff as needed;
  • Virtualizing data processing where applicable and then emulating old systems to retain useful look and feel;
  • Doing more strategic and long-range planning; and
  • Reducing risks of suppliers having service issues.

Consequently, the global manufacturers reported that the key factors in justifying automation were to:

  • Develop a financial assessment of the cost of NOT migrating;
  • Improve the rigor of control system reliability data to include actual failure events;
  • Develop a technology maturity model across each industry because some facilities are late adopters, while others keep up with new technology;
  • Benchmark against other companies that are running obsolete systems;
  • Drive value by creating long-term contracts and user-supplier relationships;
  • Standardize to help lower total cost of ownership; and take the tine to perform a solid market assessment before beginning a migration project.

"Manufacturers are also dealing with software version control, the need for smarter field instruments, security architectures and information about how to set up their information architecture and integrate basic data," Reynolds said. To improve automation sustainability, he recommended that end users:

  • Leverage standards and governance between automaton and enterprise IT groups;
  • Develop KPIs to align with company objectives;
  • Clarify roles, responsibilities and governance processes for technology groups;
  • Use tolls to main version control of software and hardware; 
  • Cross-train enterprise IT with manufacturing IT or process automation groups;
  • Ensure plant operations staff are trained and retained to use new software;
  • Building an internal knowledge base and train staff in it;
  • Develop supplier relationship management programs;
  • Provide mechanisms that align with company security architecture and allow remote support from suppliers; and
  • Develop a company-based cybersecurity center for excellence to create clarity around patch management, firewalls, password management and other issues.

The Cloud Is Here to Help

Reynolds added that one of the best ways for users and organizations to address this and other to-do lists is to employ virtualized computing and cloud-based computing and services. "The cloud is already happening, and it's coming to help with all the complexity in process applications too. Virtualization is beginning with detailed design projects and moving through to factory- and site-acceptance tests. Historians are some of the first process applications likely to go to cloud-based application provider services," he predicted.

In addition, private, on-site clouds that are higher cost and lower agility are migrating to infrastructure-as-a-service and platform-as-a-service models, but continuing on to end up in software-as-a-service settings that cost less and provide higher agility. "Distributed control systems are not going to the cloud, sensors are staying with their applications, and HMIs need to stay on premises. However, a lot of pain points related to their support services can be helped by going to the cloud," said Reynolds. "Many mainstream computing applications have been externalized, so engineers are asking, 'What about the plant?' "