A 'new' way of thinking about problem solving derived from the Russian Navy has potential to improve safe designs of process control systems.

Presently accepted tools, such as Hazardous Analysis and Operability (HAZOP) studies, What-if reviews, and event trees are used to conduct formal safety audits of process designs to ensure the consequences of incidents have been eliminated or reduced to acceptable levels.

Formalized safety audits bring together knowledgeable disciplines where information is presented, discussed, analyzed, and recorded. The assumption however, is most hazardous conditions result from defects in design, material, workmanship, or human error.

Progressing through the P&IDs, each line, vessel, and element is analyzed to ensure causes of risk have been identified and mitigation methods provided.

Using line by line, single item jeopardy analysis methods, like HAZOP reviews, miss asking questions related to system performance, process interactions, and initiating events.

During the 40's, Genrich Altshuller, an engineer working in the Russian navy-patent department, began what became a life long search for a systematic procedure to guide inventors to promising technical areas. Altshuller's work attracted enthusiasts who carried on what was later named 'Teoriya Resheniya Izobretatelskikh Zadatch' (TRIZ), which translates to 'Theory of the Solution of Inventive Problems.'

TRIZ's methodology helps eliminate contradictions and makes it an effective methodology for creative technical problem solving.

When confronted with a complex technical problem, most people seek the solution by drawing on previous experience and knowledge of solutions applied to similar problems. Depending on our experience, solutions are sought in one of four quadrants (See Solution space diagram). If the first solution concept is unsuccessful, another is sought, but almost always our solution exploration remains in the quadrant of our knowledge.

Finding solutions to complex technical problems often requires 'thinking outside out
knowledge-based box.' Anticipatory failure determination is a methodology to develop
creative solutions to complex technical problems and is effective in conducting system failure analysis.

Anticipatory failure determination (AFD) is a methodology developed using TRIZ concepts. AFD is more generalized than HAZOP or What-if list and provides a tool for analyzing systems and functions.

When AFE and HAZOP are used concurrently, analysis includes how to cause an event to happen (AFD), and what might happen as a result of the event (HAZOP).

When entire control systems were supplied by a single vendor, most users assumed the manufacturer performed a system risk assessment and had implemented methods to 'gracefully' accommodate hardware and/or software faults.

Deployment of 'open' system solutions shifts the burden to conduct system wide risk assessments from manufacturers to system integrators and users. Control systems assembled using 'open' hardware, software, and networks should be analyzed to determine system performance during periods of 'heavy use' and/or device faults. Where findings indicate system performance could be jeopardized, measures must be incorporated to protect against catastrophic system failures.

Using AFD methodologies provides an effective means of conducting system analysis.

For more information on TRIZ and AFD, visit www.controleng.com/freeinfo:

CONTACT INFORMATION: Ideation International Inc. 21800 West Ten Mile Road Suite 119 Southfield, MI 48075 P: 248/353-1313 F: 248/353-5495 e-mail: support@ideationtriz.com web: www.ideationtriz.com