It’s All a Matter of Misdetection and False Alarm Rates

Almost all chemical detectors begin life as general purpose instruments, and only over time, are adapted to specific applications including personal exposure, area monitoring, fence line monitoring, or even environmental screening.  Each application represents a niche and has a particular set of technical challenges. Automatic emergency system trigger applications assume a very accurate as well as fast response to a catastrophic chemical release.  False alarms are a major problem in emergency trigger applications utilizing current technology for two very good reasons:

1. false alarms may damage a product or process by turning off machinery (refrigerators, etc.) or initiating a damaging remediation system ( emergency water system) or,
2. false alarms cause workers to ignore the alarm or turn it off thus eliminating the last defense against a catastrophic event (implicated in recent mining and oil well disasters)

The Shur-Shot is a great example of a detector specifically designed to shut down equipment or activate a remediation system such as a water curtain in the event of a significant release with an almost zero false alarm rate.  Accurate information from a reliable chemical detector allows the plant or equipment operator to make a high stakes decision: should I shut down a process or even damage a plant in order to save it.  False alarm rate is a paramount concern in activating remediation systems and extreme sensitivity is usually not as important.

There are two sources of false alarms in chemical detectors: 1) random events in the environment that cause glitches in the detection system but have nothing to do with co-existing chemicals, and 2) chemicals in the environment that confuse the detector and cannot be distinguished from the target chemical.  Problem (1) is an engineering problem: How can the detector’s vulnerability be engineered out of the device?  Shock, vibration, electrical noise, and temperature sensitivity are all common culprits.  Problem (2) requires a better detection strategy which usually means better chemistry combined with better detection software.  Before you can fix the problem you have to understand it.  RMC’s Sensor Companion™ pictured on the right is a miniature data processing system specifically designed to collect false alarm data and document false alarm situations with its integral camera.  It can do this in the laboratory or in the field.  You can then take the Sensor Companion™ portable unit back to a convenient computer and analyze the situation using the Sensor Companion™ support software.  You can even set up new alarm conditions from your computer and take them back to the field for further evaluation.

Finding a Solution to a False Alarm Problem

There are two ways of reducing troublesome false alarms caused by chemical interferences in a chemical detection system: 1) changing the detection mechanism, for example, finding a very specific chemical reaction between a detector coating and the target chemical, and 2) using software to improve the detection process by enhancing the chemical signal and attenuating chemical noise. Each of these approaches has its merits and drawbacks.  Detector developers often see these strategies as mutually exclusive, which they are not, and greatly favor one approach simply based on their background and experience.  We have the physical chemists and the software guys.  One usually wins out in the end.  To quote one of my favorite bosses: “if my physical chemists did their job I wouldn’t need you”.  A great example from my own experience follows.

A Tale of Two Detectors – The XM22 ACADA Program

Two contractors, each representing the opposite approaches above, competed for the early development work on the Army’s XM-22 ACADA.  Both used Ion Mobility Spectrometers (IMS), the Army’s preferred technology, for portable chemical warfare detectors. Based on an understanding of the proton and electron affinity of the target chemicals one company was able to find a dopant material that when added to the sample stream eliminated the detection of most nuisance chemicals in the background.  They then used a relatively simple combination of analog electronics and a microcomputer to detect the signal and present it to the operator.  The other contractor  used no dopant material and relied only on digital signal processing and pattern recognition software to differentiate between the target chemicals and background.  The first contractor won the day, but only with the addition of a complex permeation tube system to bleed in the dopant chemical.  Had the second contractor been successful the resulting hardware would have been the simplest and best solution.  A hybrid system that combined dopant and digital signal processing software would have provided the customer with the ultimate performance however.

Before You Fix It You Have To Understand It

In order to improve a chemical detector’s performance you have to understand the misdetection and false alarm problems.  RMC’s Sensor Companion™ is designed to help you collect laboratory and field data to document false alarm problems, develop new detection algorithms using digital signal processing, and use actual detector data to verify improved performance by simulating alarm conditions and false alarms.  You can program the Sensor Companion™ internal camera to take a picture when an alarm occurs so changes in the environment that might cause a false alarm can be documented.