skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Optical Design Using an Expert System

Abstract

We present, as a different perspective on optimization, an expert system for optimization of optical systems that can be used in conjunction with damped least squared methods to find minima for specific design forms. Expert system optimization differs from global optimization in that it preserves the basic structure of the optical system and limits its search for a minima to a relatively small portion of the design space. In general, the high density of local minima obscures the general trend of the merit function in the region of interest for systems with a large number of variables and constraints. Surprisingly, there may be a potential decrease of an order a magnitude in the merit function for a region of solution space. While global optimization is well-suited to identifying design forms of interest, expert system optimization can be used for in-depth optimization of such forms. An expert system based upon such techniques was used to obtain the winning entry for the 2002 IODC lens design problem. The expert system used is discussed along with other design examples.

Authors:
Publication Date:
Research Org.:
Lawrence Livermore National Lab., CA (US)
Sponsoring Org.:
US Department of Energy (US)
OSTI Identifier:
15005176
Report Number(s):
UCRL-JC-154627
TRN: US0401581
DOE Contract Number:
W-7405-ENG-48
Resource Type:
Conference
Resource Relation:
Conference: 2003 Annual Society of Photo Optical Instrumentation Engineers Conference, San Diego, CA (US), 08/03/2003--08/08/2003; Other Information: PBD: 1 Aug 2003
Country of Publication:
United States
Language:
English
Subject:
70 PLASMA PHYSICS AND FUSION TECHNOLOGY; 99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; DESIGN; ENGINEERS; EXPERT SYSTEMS; OPTICAL SYSTEMS; OPTIMIZATION

Citation Formats

Lerner, S A. Optical Design Using an Expert System. United States: N. p., 2003. Web.
Lerner, S A. Optical Design Using an Expert System. United States.
Lerner, S A. Fri . "Optical Design Using an Expert System". United States. doi:. https://www.osti.gov/servlets/purl/15005176.
@article{osti_15005176,
title = {Optical Design Using an Expert System},
author = {Lerner, S A},
abstractNote = {We present, as a different perspective on optimization, an expert system for optimization of optical systems that can be used in conjunction with damped least squared methods to find minima for specific design forms. Expert system optimization differs from global optimization in that it preserves the basic structure of the optical system and limits its search for a minima to a relatively small portion of the design space. In general, the high density of local minima obscures the general trend of the merit function in the region of interest for systems with a large number of variables and constraints. Surprisingly, there may be a potential decrease of an order a magnitude in the merit function for a region of solution space. While global optimization is well-suited to identifying design forms of interest, expert system optimization can be used for in-depth optimization of such forms. An expert system based upon such techniques was used to obtain the winning entry for the 2002 IODC lens design problem. The expert system used is discussed along with other design examples.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Fri Aug 01 00:00:00 EDT 2003},
month = {Fri Aug 01 00:00:00 EDT 2003}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • The rejection of nitrogen from low-quality natural gas has historically been accomplished through low-temperature cryogenic systems and generally applied to large volumes of gas. Until recently, the treatment of small gas volumes has presented difficulty in meeting economic targets. The Gas Research Institute (GRI) has indicated that in 1988, total nonassociated raw gas reserves in the lower 48 states was 144 TCF. Of these reserves, an estimated 19.2 TCF (13.3%) contain nitrogen greater than or equal to 4% and that only 1.7 TCF of this high nitrogen gas is in production. Nitrogen rejection from associated gas is also an expandingmore » market and includes the applications of coal seam gas and GOB gas processing. UOP has developed and commercialized its NITREX process for these markets. This process separates nitrogen from methane through the cycling of multiple fixed beds of adsorbent. The NITREX system requires little pretreatment and is a shop-fabricated system with maximized skid mounting of components. The economic attractiveness of upgrading low-quality natural gas has been extended to low gas volumes through the use of this technology. POLYBED{reg_sign}PSA systems were first commercialized for the separation of light gases in 1966. This technology operates by using a fixed bed of adsorbent to adsorb gases of higher boiling points and molecular weights from lighter gases. Commercially attractive separations include the purification of hydrogen from light gases, the upgrading of raw helium for liquefaction and the manufacture of ultra pure methane for chemical plant feedstock.« less
  • The SP-100 shield subsystem design process has been modified to utilize the GE Corporate Reserch and Development program, ENGINEOUS (Tong 1990). ENGINEOUS is a software system that automates the use of Computer Aided Engineering (CAE) analysis programs in the engineering design process. The shield subsystem design process incorporates a nuclear subsystems design and performance code, a two-dimensional neutral particle transport code, several input processors and two general purpose neutronic output processors. Coupling these programs within ENGINEOUS provides automatic transition paths between applications, with no source code modifications. ENGINEOUS captures human design knowledge, as well as information about the specific CAEmore » applications and stores this information in knowledge base files. The knowledge base information is used by the ENGINEOUS expert system to drive knowledge directed and knowledge supplemented search modules to find an optimum shield design for a given reactor definition, ensuring that specified constraints are satisfied. Alternate designs, not accommodated in the optimization design rules, can readily be explored through the use of a parametric study capability.« less
  • ZEX is an expert system, which is designed to support operation of the ZEUS experiment. Experience gained in the OO analysis and design of the system is summarized in the paper.
  • In this paper an attempt is made to develop a new algorithm for distribution system planning. Generalized algorithm is developed for obtaining the optimal feeder path and optimal location of substation based on minimum loss criteria. The proposed method can handle all the heuristic rules. Also presented are a new load flow algorithm for solving radial distribution feeders and a branch conductor optimization algorithm. The load flow and branch conductor optimization techniques are used as subroutine in the generalized distribution planning algorithm. Several problems have been successfully tested using the proposed algorithm and results are highly satisfactory.
  • An integrated expert system was developed to facilitate the design of window glass for structural strength, hydrostatic loads, sound attenuation, and solar control. The integrated software consists of a text-based interface, a rule-based expert system, and two neural networks. The text of a glazing design guide is linked by related topics and concepts. The software's design feature lets the user enter design parameters for the window choice via an interactive consultation in to a rule-based expert system that critiques the design. The technical aspects of the glass's structural strength are based primarily on linear methods published by the American Architecturalmore » Manufacturers Association Statistical correlations for the new nonlinear failure prediction for glass strength are used to automatically design for the minimum glass thickness required to withstand a specified load. Neural networks estimate values for sound attenuation and solar transmission characteristics from laboratory test data on selected glass constructions.« less