Fault diagnosis is the problem of determining the root causes of process upsets. This paper presents a very efficient method of identifying the possible causes of process disturbances using the signed directed graph (digraph) representation of process interactions. The analysis is based on forming logical statements (rules) derived from the process digraph; these are evaluated using on-line data to yield the diagnosis. Evaluation of rule antecedents is more efficient than the previous algorithmic approach of Shiozaki et al. In the rule-based approach, the diagnostic criteria are represented explicitly, not hidden by a complex algorithmic procedure. This allows the diagnostic rules to be tailored to reflect the best available knowledge of plant behavior. The rules generated by this technique can be integrated with other rules on plant operations using an expert systems framework. M. A. Kramer, B. L. Palowitch, Jr. Department of Chemical Engineering Massachusetts Institute of TechnologyCambridge, MA 02139 IntroductionIn most plants, diagnosis of process upsets is left to the abilities of process operators. When a process alarm is activated, the operator determines the seriousness of the situation and initiates appropriate action. The diagnostic decision is based on the type of alarm, the values of related process variables, and the operator's background, training, and mental model of the plant.This method of diagnosis has certain disadvantages. The availability of process experts may depend on work shift, employee turnover, vacations, and the like. Operators may be well trained in standard procedures but ill-equipped to handle unusual events. Stress associated with alarm situations can compound the difficulty of decision making. The operator's mental model of the process may be incorrect. These factors make computer-based process monitoring and diagnosis desirable.The problem of fault diagnosis has been addressed by many authors and is the subject of books by Himmelblau (1978) and Pau (198 1). Different process representations distinguish various approaches to this problem. Quantitative approaches involving filtering and estimation have been reviewed by Isermann (1 984). Qualitative approaches involving fault trees and related diagrams have been reviewed by Lees (1983). Another qualitative approach, involving the signed directed graph (SDG), has been developed by O'Shima and coworkers (Iri et aI., 1979;Tsuge et al., 1985;Shiozaki et al., 1985). Digraph-based meth- ods are attractive because relatively little information is needed to set up the digraph and perform the diagnosis. The approach of O'Shima and coworkers is afgorithmic in nature, and involves tracing the possible sources of disturbances using the information stored in the SDG.The SDG represents pathways of causality in the fault-free process. The nodes of the SDG correspond to state variables, alarm conditions, or failure origins, and the edges (branches) represent the causal influences between the nodes. The direction of deviation of the nodes is repres...

Background: Biomarker measurements improve the phenotyping of patients with severe uncontrolled asthma (SUA) and predict therapeutic responses. The use of biomarkers in asthma, however, remains underused.Objective: To test the hypothesis that biomarker measurements of patients with SUA remain markedly underused and contributes to asthma morbidity and oral corticosteroid use.Methods: Leveraging claims data linked to electronic health record data, we calculated biomarker use by providers treatingpatients with SUA from January 2017 to August 2020.Results: From 3.6 million clients, 3817 had a primary diagnosis of asthma; most were between 50 and 60 years old. Also, 63.2% were female patients; those under ages 10 years were primarily boys. Of the 728 patients who reported race, 69.9% were white and 21.8% were African American. Of the 840 who reported ethnicity, 14% were Latinx. A predetermined definition of SUA identified 348 patients with SUA. In a nested sample of 151 patients with SUA, 43% were managed by primary care physicians (PCP), 4% by specialists, and 49.7% by both. Of this sample, 61.5% had a measurement of serum eosinophils, 9.9% total immunoglobulin E values, and 9.3% radioallergosorbent skin tests; 38% received no tests, whereas 9.9% had more than one. Specialists ordered a biomarker test 4.6 times more often than did PCPs, whereas PCPs ordered 70% of the prednisone prescriptions for recurrent asthma exacerbations.Conclusion: Specialists were more likely to order biomarkers than were PCPs. Patients managed exclusively by PCPs were more likely prescribed oral prednisone. Real-world evidence shows that biomarkers are infrequently used to characterize patients with SUA, especially among patients exclusively managed by PCPs. Programs that encouraged biomarker use may improve SUA management and oral corticosteroid burden.

No abstract

No abstract

No abstract