{"@type": "dcat:Dataset", "accessLevel": "public", "accrualPeriodicity": "irregular", "bureauCode": ["026:00"], "contactPoint": {"@type": "vcard:Contact", "fn": "Ole Mengshoel", "hasEmail": "mailto:ole.j.mengshoel@nasa.gov"}, "description": "We present in this article a case study of the probabilistic\r\napproach to model-based diagnosis. Here, the diagnosed\r\nsystem is a real-world electrical power system, namely the\r\nAdvanced Diagnostic and Prognostic Testbed (ADAPT) located\r\nat the NASA Ames Research Center. Our probabilistic approach\r\nis formally well-founded, and based on Bayesian networks and\r\narithmetic circuits. We pay special attention to meeting two of the\r\nmain challenges \u0097 model development and real-time reasoning\r\n\u0097 often associated with real-world application of model-based\r\ndiagnosis technologies. To address the challenge of model development,\r\nwe develop a systematic approach to representing electrical\r\npower systems as Bayesian networks, supported by an easy-touse\r\nspeci\u0002cation language. To address the real-time reasoning\r\nchallenge, we compile Bayesian networks into arithmetic circuits.\r\nArithmetic circuit evaluation supports real-time diagnosis by\r\nbeing predictable and fast. In experiments with the ADAPT\r\nBayesian network, which contains 503 discrete nodes and 579\r\nedges and produces accurate results, the time taken to compute\r\nthe most probable explanation using arithmetic circuits has a\r\nmean of 0.2625 milliseconds and a standard deviation of 0.2028\r\nmilliseconds. In comparative experiments, we found that while\r\nthe variable elimination and join tree propagation algorithms\r\nalso perform very well in the ADAPT setting, arithmetic circuit\r\nevaluation was an order of magnitude or more faster.\r\n\r\n**Reference:**\r\n\r\nO. J. Mengshoel, M. Chavira, K. Cascio, S. Poll, A. Darwiche,\r\nand S. Uckun. \"Probabilistic Model-Based Diagnosis: An Electrical\r\nPower System Case Study\u201d. Accepted to IEEE Transactions on\r\nSystems, Man, and Cybernetics, Part A, 2009.", "distribution": [{"@type": "dcat:Distribution", "description": "Article", "downloadURL": "https://c3.nasa.gov/dashlink/static/media/publication/SMCA08-04-0141-Paper.pdf", "format": "PDF", "mediaType": "application/pdf", "title": "SMCA08-04-0141-Paper.pdf"}], "identifier": "DASHLINK_67", "issued": "2010-09-10", "keyword": ["ames", "dashlink", "nasa"], "landingPage": "https://c3.nasa.gov/dashlink/resources/67/", "modified": "2025-03-31", "programCode": ["026:029"], "publisher": {"@type": "org:Organization", "name": "Dashlink"}, "title": "Probabilistic Model-Based Diagnosis for Electrical Power Systems"}