Process Standards: Pursuing Best Practices
Most professionals know that ISA-84 is a safety-instrumented benchmark-standard. Interestingly, while the committee was chartered in 1984, it took until 1996 for first version to appear, largely due to the considerable, well-intentioned efforts aimed at specificity. Such extended thoroughness underscored the difficulties in reconciling the divergent aims participants, including both end users a...
Most professionals know that ISA-84 is a safety-instrumented benchmark-standard. Interestingly, while the committee was chartered in 1984, it took until 1996 for first version to appear, largely due to the considerable, well-intentioned efforts aimed at specificity. Such extended thoroughness underscored the difficulties in reconciling the divergent aims participants, including both end users and vendors that use standards and implement them in the automation, controls, and instrumentation they offer.
Much of that tendency to wrangle over minutiae has passed. Nevertheless, despite the best efforts of the standards generators, Edward Marszal, PE, CFSE, president of Kenexis Consulting Corp., says end-users often find applicable standards for equipment and process to be vague. And Marszal notes that there are circumstances where he spends entire days on legal aspects without addressing engineering.
Nevertheless, today most criteria are aimed at performance, avoiding disputes previously encountered among manufacturers on a standards committee. Subsequently, producing companies take a standard, such as ISA-84, and then go clause by clause echoing the verbiage in the corporate standard so it is specifically mandatory as to what is used, such as valve a, b, or c. So, anywhere there is a question about action or equipment employed in meeting the general standard, the corporate standard nails down specifics.
Corporate standards, similar to an ISA standard, typically have an expiration date; when the date approaches the corporate standards team has to review and revalidate it, or forge and reissue another version. This is particularly true of companies that maintain an ISO 9000 certification.
However, Marszal has seen a major refiner that wasn't going to follow industry-consensus standards at all, choosing performance-based internal standards. The company's view was that the external standards left excessive room for interpretation, allowing too much potential for widely divergent designs. The company took a proscriptive track, with great specificity on pumps and other elements. Marzal observed that, 'if a company is out there by itself—even if it's on an academically sound basis—being alone (compared to the bulk of the industry) carries a significant legal risk should something happen.'
Since adopting its original policy, that company has had a change of heart. It now allows more flexibility. In contrast, another refiner hired an outside firm to benchmark its peers on standards so as to be consistent with the industry mainstream.
Going forward, performance-based models are tailoring equipment and maintenance to avoid wasted resources from proscriptive-style standardization, such as over- or under-maintaining given equipment. And operating companies are increasingly implementing standards as a dictate from the top, instead of using the former site-specific approach. Currently there's a lot of capex being expended in compliance with a variety of standards and best practices. But Marszal wonders, 'will this effort evaporate if refinery (profit) margins dry up?'
Marszal says that it's encouraging that end-users are increasingly involved in writing standards. Previously there was a tendency to fight standards; now there's a recognition of the benefits of participation. These include:
Deferral of standards development costs;
Getting a wealth of expert knowledge; and
Avoiding the cost of retrofitting to an industry standard (by setting one's own).
Previously, most corporations had large staffs of automation engineers and designers who worked on projects of all sizes and also developed in-house standards. Independent consultant Alex Habib commented that, 'in the '80s-'90s, with corporate downsizing and decentralization, their central engineering staffs have shrunk, impacting the update of corporate standards.'
Knowledge of many
To keep up with the rapid development of the automation and control technologies, the need for standards is more important than ever. Standards ensure good product quality and plant safety, and also make sure that all pieces fit together during field construction.
When he was an end-user of automation equipment, working for multinational manufacturing companies in the food, pharmaceutical, and fine chemicals industries, Habib has found the following ISA standards to be most useful in implementing projects:
ISA-5 Control system documentation;
ISA-20 Instrument specification forms;
ISA-88 Batch controls; and
ISA-84 Safety instrumented systems (SIS).
Before the end-user has a workable process, manufacturers must provide the physical tools. Working in their own and customers' plants, device manufacturers are in a unique position, often helping in industry standards organizations and setting defacto standards for functionality, interoperability, and quality. Vendors of automation, via standards and best practices, aim to lower hurdles along the way, for themselves and for those using their tools.
Lee Neitzel, Emerson Electric senior technologist, observes that end-users formerly had to write a specialty communications driver that wouldn't have much longevity due to a change in either end of the system. With the multiplicity of functions and communications, lack of a standard was untenable. Consequently communications standards have been one of the most important and dynamic areas compared to other sectors.
Software has to be designed to reflect the processing standards. For example pharmaceutical clients are very interested in engineering their processes for conformance with the ISA-88 Batch standard. So Emerson says it's designed its software to reflect the hierarchies and structure of that standard.
Europe-based TUV develops safety standards; Emerson invites TUV's involvement from product conception through design, over the course of a couple years. The standards are engineered into the product, from electrical and communications through TUV's philosophy. So the product is designed around the standard before it is submitted to the standards organization for validation.
Human interface standards used to be a big concern. Today technology has addressed most of the issues, such as screen update and flicker rates.
Ergonomic issues, such as height of a monitor or keyboard layout, were factors. Standards in that area are now largely addressed by PC-component manufacturers, enabling Emerson to use many off-the-shelf components.
'It's important not to push to the limits of a standard in device design,' Neitzel advises. 'You're asking for trouble if you do. Some standards have a lot of options and flexibility built-in. Selling a product that pushes the envelope of a standard greatly increases the chance for interoperability problems, especially if the product is introduced into different markets because the standard matures around its core capabilities.'
Neitzel continues, 'Standards underpin manufacture of successful products. It's very important for a manufacturer to be very aware of standards and, indeed, to help develop them within standards committees. Then the company can be sure the standard is going in the right direction, is robust, and has all necessary features.
Standards exist to open up interfaces for a device or system, says Neitzel, and you measure the success of a standard by market acceptance. 'Adoption is driven by a need in the market and the ability of vendors to implement it. Wide acceptance is unlikely if a standard is complex. Moreover complexity breeds higher prices and it becomes more difficult for end-users to justify the additional cost,' he says.
So simplicity is key but flexibility and simplicity compete with each other all the time. 'Typically a lot of people in standards committees want flexibility, but when they reach implementation they recognize how difficult it is and wish they had made it simple. Two excellent examples of simplicity and resulting successful standards are Ethernet and SQL databases,' Neitzel says.
Standards also must be clear and technically correct, an example of a technically incorrect standard was the 'token bus,' IEEE 802.4—where the kinks were never quite eliminated. Similarly FDDI—at a 100 Megabit capacity when Ethernet was at 10 Megabits—was very forward thinking, but it was so complicated that it was too expensive to achieve any real degree of success. If the specification is not clear, you're going to see different sources implementing it in different ways—preventing interoperability and dooming the standard to failure.
Neitzel says, 'Companies involved in standards bodies have to be committed to accepting market needs. They have to be involved in the development, writing, and review of the technical specification—that may include implementation and prototyping. So there's potentially a tremendous amount of effort involved. Unsurprisingly there're plenty of companies that follow a standard—getting into the marketplace a lot later than the participating firms. And they don't have as good an understanding of the standard and while they are implementing, the innovating, standards-setting companies are already working on the next generation of standards and products. Best practices for a company? Be involved in all of the areas cited.'
Some manufacturers only want to market the cheapest possible product. But most device manufacturers face a ying and yang of:
Trying to differentiate their products from those of their competitors; and
Facing customers who are seeking a device filling a unique application or niche, while effectively buying the best value.
Neitzel adds, 'The foregoing is in opposition to standards, which effectively function as a leveling factor—everyone's stuff works the same. So (manufacturing) companies participating on a standards committee have effectively agreed to not compete in the technology governed by the standard. The (intercompany) differentiation is not at the interface level but rather at the device or system levels.'
Foundation for innovation
The leveling factor supplied by standards provides a firm foundation for innovation. With vendors and end users starting with the same standards-based requirements, higher quality solutions can be built faster. Bob LaPerriere, Rockwell Automation's director, Quality Assurance - Automation, Control and Information Group, says, 'Quality starts at the requirement analysis stage. It's a fine line between the nice-to-have versus must-have features. We're very conscious of the fact that we can overcomplicate products, so the focus on the customer's wants/needs is paramount.'
Once those baseline needs are met, vendors want to differentiate themselves. Over time, the improvements they generate can be considered for new standards. La Perriere considers Rockwell Automation's Highly Accelerated Life Testing (HALT) program, for example, to be a prime differentiator and a key factor in improving component mean time between failure (MTBF) ratings.
'By varying temperature, vibration, and voltage, testing is carried out on three levels: within specification, beyond (normal) operational limits, and to hard failure. The fundamental cause of failure is evaluated and, in a cost-effective manner, the weakest component is redesigned,' he says.
Keith Seibert, manager, Quality Assurance, Rockwell Automation Twinsburg (OH) Production says, 'Another differentiator is quality data. Everyone has it, but many companies don't have the ability to (evaluate and) react to the data in a timely manner. Quality engineers can end up spending 80% of their time mining the data and 20% acting on it. In contrast, all of our data is automated, flipping the 80:20 ratio around. In effect, quality engineers are working on a true exception-oriented basis. The oversight and corrective-action process is continuous.'
Feeding the lessons learned back to engineering is key in assuring higher reliability in the next generation of controller or other product.
Says Seibert, 'Everyone does ISO (standards), but another differentiator is Rockwell's holistic approach to standards. All standards, including environmental, are integrated because, realistically, they cannot be separated. Often manufacturers can end up not seeing the proverbial forest for the trees if they use a silo approach in implementing standards in design engineering.'
In the race for the highest quality solution, standards lower the hurdles so companies can achieve more. Individual success and best practices may vary, but innovation lives on.
5 tips to simply apply standards as best practices
Augment and clarify standards where needed to ensure corporate consistency across sites.
Get involved in standards most important to your organization's success.
Ensure hardware and software incorporate a standards-based approach.
Structure software and data structures modularly, so 80% or more can be automated.
Don't overdo... keep standards application simple and consistent, in line with the bigger picture!