Marc Buchner

Leak detection in an experimental heat exchanger process: a multiple model approach

An application of a multiple model nonlinear filtering technique is discussed for the detection of leaks in a heat-exchanger process. The results of simulation study and an experimental study on a laboratory heat-exchanger process are presented. A simple decision logic scheme (threshold crossing) is used to process the outputs of the leak-detection algorithm to indicate the status of the process, i.e. whether or not there is a leak. The results of the study provide preliminary indication of the feasibility of the proposed diagnostic technique to detect small leaks in an experimental heat-exchanger process. The study also indicates that the technique is robust in the midst of significant flow disturbances on the inlet water flow, using only simple models for processing of the outlet temperature data, but is sensitive to the aggregate heat transfer parameter used in the detection algorithm. >

Detection and visualization of tandem repeats in DNA sequences

One conspicuous feature of DNA is the extent to which nucleotide subsequences repeat in the genome. Several strongly repetitive tandem (or contiguous) repeats are known to be associated with genetic diseases, while weaker repetitive structures are thought to be representative of historical events associated with sequence repetition. Thus, it is important to develop sensitive and rapid automation of the detection and identification of repeat sequences. A new algorithm for examining periodic patterns in DNA sequences is developed. The algorithm uses the short-time periodicity transform to compute the closest periodic sequence of fixed length at each nucleotide position in a given sequence to be analyzed. Each such subsequence is then compared to its closest periodic sequence to provide a quantitative measure of the amount of repetition within the sequence. In addition to being used to detect the presence of repeat subsequences in DNA, the periodicity explorer algorithm provides a potentially useful visualization of periodic patterns in a DNA sequence through a graphical display of the relative energy in the optimal periodic projections of the analyzed sequences, i.e., the DNA periodogram. Computationally, the algorithm is linear in the length of the analyzed sequence.

Avatar-based depression self-management technology: promising approach to improve depressive symptoms among young adults.

Major depressive disorder is prevalent among American young adults and predisposes young adults to serious impairments in psychosocial functioning. Without intervention, young adults with depressive symptoms are at high risk for worsening of depressive symptoms and developing major depressive disorder. Young adults are not routinely taught effective depression self management skills to reduce depressive symptoms and preempt future illness. This study reports initial results of a randomized controlled trial among young adults (18-25 years of age) with depressive symptoms who were exposed to an avatar-based depression self-management intervention, eSMART-MH. Participants completed self-report measures of depressive symptoms at baseline and at 4, 8, and 12 weeks follow-up. Participants who received eSMART-MH had a significant reduction in depressive symptoms over 3 months, while individuals in the attention-control condition had no change in symptoms. In this study, eSMART-MH demonstrated initial efficacy and is a promising developmentally appropriate depression self-management intervention for young adults.

Application of an Innovative Process Diagnostics Algorithm to Tube Leak Detection in a Heat Exchanger

Abstract This paper describes an experimental investigation of a multiple model nonlinear filtering approach to leak detection in a laboratory heat exchanger. Although the work is still in its embryonic stage the results obtained are promising.