Edward M. Marszal, PE, ISA84 Expert

The proper placement of gas detectors can be difficult and complex to determine. Our last blog post discussed how gas detectors were placed in the past and current technologies for detector placement, including geographic coverage mapping. This post will address where the technology is headed in the future. The next level of analysis, which is defined in the ISA 84.00.07 technical report, is scenario coverage. While geographic coverage is a big step over grid placement because it allows for multiple different critical cloud sizes for different hazards, it still suffers from the lack of a more sophisticated mechanism for considering where leaks are coming from. Also lacking is information on the dimensions of the leak created clouds as this will be considering factors such as wind direction. The mechanism that can better deal with these issues is scenario coverage.

RAI.GasDetectionMapIn scenario coverage, the covered area is not the desired output, but instead the fraction of area that is covered, the intention is to calculate the fraction of release scenarios that are covered. In this method, cloud dimensions from a release of process equipment (as determined through dispersion modeling) are determined from each piece of equipment in a zone. These releases are then simulated in the process area and compared against the detector locations to determine if a release gas cloud overlaps the location of a detector. If so, that release scenario is determined to be “covered.” For each release source in a zone, the released gas cloud is distributed in many directions (usually around 720 directions), and the effect of wind direction is considered in this distribution process. As such, if a typical zone is considered, there could be hundreds of thousands of release scenarios that are considered. The software then determines what fraction of the release scenarios are detected and compares that against a quantitative target. In addition to the quantified coverage target, colorized maps can be drawn that show, by the overlaid color, the frequency at which an undetected gas cloud is expected to exist in a certain geographic location. This method provides a quantum leap in richness of knowledge about the degree of effectiveness of a gas detection array, but it comes at a price of increased analysis time. We have seen, though, that many end users find enough value in the additional insight provided by scenario coverage to go through the additional effort.

If this level of analysis is not enough, more can be done, but at this point, the actual commercial applications are limited due to the significantly increased analysis cost and effort. Loss prevention specialists know all too well the limitation of the dispersion modeling that is often employed in scenario coverage. Dispersion models are quite weak in their ability to analyze “near field” effects of obstructions, and they are all but useless for indoor releases where the size and shape of a released gas cloud is almost entirely driven by the action of the building’s HVAC system.

RAI.CFDModelingIn cases where dispersion modeling is less than effective, a more rigorous method for determining the size and shape of a released gas cloud is through Computational Fluid Dynamics (CFD). Gas detector placement using CFD models is a type of scenario coverage modeling, but the results are much harder to develop and to interpret. Essentially, the analyst builds a three-dimensional model of a facility, places releases, and then determines the gas concentration at the various detector locations after a release has occurred to determine if the release is “covered.”

While the richness of data results from a CFD modeling run is orders of magnitude richer than for dispersion modeling-based scenario coverage, it comes at a high price. CFD modeling consumes large amounts of analyst time for the development and interpretation of the models, and also the sheer computational effort often makes it prohibitive. Whereas a dispersion modeling-based scenario coverage assesses hundreds of thousands or even millions of scenarios, CFD assessment of a zone will typically only include 6-10 scenarios. The reason for the paucity of scenarios in CFD assessment is mainly attributed to computational resources. In order to run millions of scenarios that would be used in scenario coverage, CFD would require literally dozens of years of computing time. This degree of effort is simply not feasible. As a result, analysts need to use judgment to select a small subset of scenarios that would be representative of the complete data set. While this type of judgment works well for indoor facilities, where release patterns can be more easily predicted, it is difficult to reasonably represent the full range of outdoor scenarios and weather conditions. In any case, CFD-based analysis is already being used successfully for indoor scenarios and, with advancements in methods and computing power, could soon be feasible for outdoor analysis too.

15 Jan, 2015  |  Written by  |  under Education

By Patrick Naillon, Global Blended Learning Developer

Many of today’s plants are facing a labor shortage issue as experienced technicians and operators are retiring and new employees require additional training to ensure consistent productivity. One way many plants are dealing with this issue is by re-evaluating their training programs to ensure they are meeting their current needs.

Computer-based learning, or eLearning, is dramatically improving employee training and productivity. It is estimated that more than 40% of Fortune 500 companies use some form of eLearning technology, while companies that use eLearning tools have a potential to boost productivity by 50%. Companies that offer best practice eLearning generate around 26% more revenue per employee.

Clearly, eLearning is the way to go. But how does that impact the world of O2 sensors, pH calibration, and electronic analyzers in your plant?

Emerson offers Rosemont Analytical online, self-paced training courses, featuring animated lessons, professional narration, 3D modeling, knowledge checks, and a final assessment to test your staff. These courses can be taken any time, from any internet-connected pc, laptop, or tablet. Even a smart phone can manage many eLearning courses. Courses can also be ‘paused’ allowing your users to return to the instruction without needing to start over again – a valuable tool if time is limited to view instructional materials.

Self-scoring courses allow a company to assign a number of I&E techs, sales reps, plant managers, and others, to take a required training course, and have the final scores for each person routed to your HR department for confirmation of course completion and score. A series of courses may be created for a particular job classification in your company – for example: pH Theory, Electronic Analyzers, pH Sensors, Conductivity Theory – allowing companies to create an ‘online university’ of course material and track each employee and their career goals.

RAI.pullquoteAnd this increase in training actually comes at a lower cost – no more scheduling employees for training that eats into working hours or brings your company to a halt; no more hiring expensive trainers, or flying trainers (or employees) long distances; no more hotel and meal bills. Plus, you realize cost savings from elimination of prep time for training (booking meeting rooms, printing handout materials, providing meals and coffee service).

Best of all – eLearning has the best record of retention of all training types. By allowing your employees to work at their own pace, and allowing lessons to be revisited prior to the final assessment, eLearning provides a better environment for any student. Many types of eLearning are meant to be accessed at any time – for example, tutorials showing how to use a 1056 Analyzer could include lessons showing how to run a sensor calibration. Your employees could at any time access this tutorial, for a refresher or to learn a new skill for their job. The tutorial functions like a hardware guru available 24/7 to everyone in your organization.

Online learning can help your plant improve employee productivity at all levels to ensure operational efficiency; address the looming labor shortage issue and increase production quality.