Bringing Real-Time Detection from Vision to Reality
By Ed Rutkowski
Phil Smith, PhD, CIH, FAIHA, provided an overview of the past, present, and future of real-time detection systems during a live virtual session held June 2 as part of AIHce EXP 2020. Smith, the director of OSHA’s Industrial Hygiene Chemistry Division, suggested that the most urgent use of real-time detection systems would be to understand intra-work shift variability in exposures to chemicals. In some circumstances, workers may experience dangerously high instantaneous exposures that are not captured by traditional sampling equipment.
Real-time detection systems have come a long way since their earliest applications, which involved the use of colorimetric indicators. Later, analog systems offered improvements but were weakened by low-voltage signals and limited by the finite lengths of cables; networking wasn’t possible. Digital systems overcame these problems through Bluetooth and Wi-Fi technologies and the ability to network with existing infrastructure. In the near future, Smith said, real-time detection will use artificial intelligence to make predictions based on instantaneously collected data.
Most IH and OHS professionals use real-time exposure data for four purposes, Smith said. The most critical use of real-time data is for activating alarms to warn workers when exposures are approaching conditions that are immediately dangerous to life and health. Alarms are often used when workers may be exposed to fast-acting health stressors, such as hydrogen sulfide in the oil and gas industry or low-oxygen atmospheres. A second purpose for real-time data collection is to sample exposures in a worker’s breathing zone, through data-logging sensors clipped to the worker’s lapel. Other current uses of real-time exposure data include conducting workplace surveys and employing fixed sensors at work sites for alarm purposes.
For the coming generation of digitized and networked systems, Smith identified three priorities: they must keep employees safe from immediate harm by identifying and predicting IDLH conditions; they must prevent future harm by identifying out-of-control conditions in routine processes; and they must provide a clearer picture of workers’ exposures. Current data-logging systems allow IHs to identify instantaneous exposures, but only after the fact, because the IH must first download the data. To be most useful for IHs, real-time systems would allow IHs to see spikes in exposures as they happen—or better yet, predict them before they happen.
Networking will be crucial for realizing the potential of IH sensors, Smith said. Workplace systems are already using sensors to track the locations and movements of workers and vehicles. These systems tap into the networking possibilities of the Internet of Things to provide immediate site-level awareness of certain hazards, and it’s only a matter of time before they incorporate IH instruments, too. “The industrial hygiene community should be involved in designing these [instruments] so we get what we need out of” these systems, Smith said.
In the near future, networked systems will allow IHs to see the peaks and valleys of exposure. No longer will IHs have to download the data; it will automatically go to a centralized location and will be accessible from anywhere. Networked data, combined with AI, will allow the creation of algorithms for low-level decisions.
Although real-time detection systems will change the nature of sampling, IHs will still have an important role to play. The role of the IH, Smith said, will be to determine whether a given sensor is appropriate and will provide reliable data, to conduct initial and periodic exposure assessments, to assign the correct sensors to workers, and to use their expertise to assess exposures. They will also decide whether the network should expand to new locations or to accommodate additional sensors.
Perhaps the most significant benefit of future real-time detection systems is they will allow IHs to be more effective with the same amount of resources. “We don’t need more IH support,” Smith said. “We just need to redirect what we have.”
Related: For more information about sensors, read “Predictive Purposes: Will Big Data Change Industrial Hygiene?” in the March 2018 Synergist.
Ed Rutkowski is editor-in-chief of The Synergist.