Piping systems exist in every plant and facility.  In nuclear facilities, these systems are critical and require detailed evaluation.  In many other facilities, such detailed evaluations are not required, and the design of piping support systems is typically done by means of cookbook methods.  But cook-book design is almost always conservative; it almost always results in over-designs, with more supports or larger supports than are actually needed.  This can add significantly to total project cost and schedule.

Two Bechtel engineers in the San Francisco Office, Milton Dong and Dr. Hong-Ming Lee, have been awarded a technical grant to drive down these costs by putting piping design on a more automated basis.  Their approach will be to develop a computer program, using state-of-the-art nonlinear program algorithms, that automatically determines the location and minimum number of hangers for typical non-critical piping systems.  The program, called HALOS (hanger auto located and optimized system), will be able to select required supports from feasible locations using prescribed rules and optimize the number of supports by iteration through a complete stress and cost analysis.

When implemented, HALOS will (1) lower the design cost by decreasing the number of engineering job-hours required; (2) cut fabrication, stocking, warehousing, and installation costs by reducing the number of pipe supports as well as structural member size and weld requirements; (3) shorten both the design and construction schedule; (4) improve engineering quality; and (5) capture and use expert knowledge in pipe stress and pipe support design. HALOS will be particularly useful in piping layout design, piping stress analysis, and pipe support design for petrochemical plants, cogeneration plants, and pipelines. Says Dong, “I believe this program will cover 80 percent of all projects requiring minimum analysis and design evaluation and will reduce the number of supports in a typical cogeneration plant by 20 percent.”

The HALOS project will consist of three phases.  Phase 1, the phase covered by the technical grant, will be devoted to gathering data, selecting hangers, and optimizing the number of hangers. This phase will comprise five tasks, each involving the development of a specific module which directly interfaces with the company’s pipe stress program, MEl01.  In Task 1, an input module will be developed and an expert rule will be generated based on the largest support span. In Task 2, a hanger locator module will be developed to create nodes with feasible locations for supporting or hanging pipe and components; and data can be extracted from 3D structural steel files for feasible locations.  In Task 3, an optimizer module, with forward or backward iteration, will be developed to select support types and trial locations.  In Task 4, an evaluator module will be developed to compare and check allowables from project expert and report files.  Finally, in Task 5, a summary module will be developed to report results and various iteration procedures.

The resulting HALOS program will be rule-based, with the rules initially being set up in accordance with petroleum and cogeneration applications via B31.1 and B31.3 hanger span tables. These rules will also serve as a guide for evaluating code stress, nozzle load, clearance check, available space, and feasible hanger type. 

Dong will lead the project. He will design the overall working process, build the expert knowledge base, help work out the optimization algorithm, develop criteria to select hanger location and type, and verify the feasible solution from the HALOS program. Lee will develop a module to interface with ME101, along with other HALOS modules, and will also help work out the optimization algorithm. “Nowadays, the tendency is to outsource such technology,” says Lee, “but I think this can be shortsighted. In-house R&D of this kind is a must for us if we are to stay competitive.”

In Phase 2, the knowledge base for supporting pipe will be expanded with various criteria and allowables.  The special skills of several types of designers and analysts will be captured. In Phase 3, the system will be expanded to interface with or incorporate the full hanger sketch and hanger design program. The final product will include full signature-ready reports, hanger design drawings, and piping isometrics. 

BECHTEL ENGINEERING BULLETIN

Page 11, Year 1994