Where's the Leak?

To effectively respond to any chemical odor complaint, one first needs to know the source of the chemical release.

A chemical emitted from an individual site and the effect of that release — on not only the emitting site, but also on neighboring sites and nearby communities — are of concern to plant personnel, community leaders, and regulators. The issue frequently first manifests itself in terms of someone complaining about an odor. This may be an employee at the site, someone at a nearby site, or a resident of a nearby town. Sometimes the odor may be easily identified in terms of its substance; in other cases, not. Often, though, especially in a situation where a number of industrial facilities are in close proximity to one another, it may be difficult to determine which site is the source of the offending emission. This results in periods of uncertainly as to who is responsible. It may lead to ongoing complaints and sometimes to good, old-fashioned fingerpointing.

In the afternoon of May 28, 2009, an unusual number of workers in a petrochemical plant were rushed to the infirmary complaining about shortness of breath, headaches, dizziness, and nausea. (The plant is located in a rural area and fl anked by several other industrial sites.) Most of the complaints came from a unit that was shut down for routine annual maintenance. Workers reported smelling a strong odor, but their description of the odor characteristics could not establish a link to a specific chemical.

The plant environmental health and safety manager was informed. Contact was made with the managers of different units within the facility to check for a possible chemical release during the time of the reported smell, which was between 4 and 5 p.m. Results of the EH&S manager's inquiry were negative, and a team was assembled to identify the source of the chemical leak.

The plant safety people provided questionnaires to those who reported sick. The questionnaire's main queries were: Where was each individual located when he or she first detected an unusual odor? What time did he or she smell the odor? What chemical odor characteristics were present (e.g., pungent, rotten egg, etc.)? What was the intensity of odor on a number scale of 1 (low), 2 (medium), and 3 (high)?

From the information gathered, a grid was superimposed onto a map of the plant site and each individual's location at the time the odor was first detected was marked. (The grid is helpful because people could specify only the general area where they started to feel sick, rather than an exact location.) Five meteorological towers were then identified: two inside and three outside the plant's fenceline.

The petrochemical plant also had a network of sensors for measuring volatile organic compounds. A copy of archived hourly average concentration of measured volatile organic compounds was obtained. These data could help to identify chemical(s) with concentrations beyond he background baseline values.

Data Analysis

It was assumed all of the sickness was from one source and all of the claims were legitimate. For each location within the grid using the data gathered via the questionnaires, the team strove to identify individual locations and the number of people at each location. Meteorological data for the month of May were obtained, from which the data for the required time window were extracted. The wind direction for the period of interest was mainly from east to south/southeast.

The VOC measurement did not identify any chemical with concentration above background concentration. The odor description by individuals was inconclusive, so the team could not identify the type of chemical via this method. The focus shifted into finding the leak source and, most importantly, whether the source was within the plant.

Source Location

The minimum required information to run a dispersion scenario is the identity of chemical, release location, release rate, and weather data. Meteorological information was the only data available in this instance. Armed with only these data, the team could at best locate the general area of the source leak but could not quantify the magnitude of the release. Wind direction and atmospheric stability were used to create a reverse corridor for identifying possible leak source locations. Reverse corridor is the opposite of a Backtrajectory method, which is commonly used in air quality studies to examine the likely path of a chemical plume released from a point source. The reverse corridor tries to narrow the search zone for possible chemical release locations.

A reverse corridor is constructed from the position of each impacted individual. The corridor is a wedge drawn from the position of each individual using the opposite wind direction. So if the wind is blowing from the south, the reverse corridor is drawn from the north.

The wind was swung from east to south/ southeast, and the reverse corridors were drawn, with the hashed area depicting the possible source locations identified on the maps. Finally, all of the areas obtained from different wind directions were combined to create the overall area for the possible locations of the odorous chemical release. It clearly shows the source of the obnoxious odor was located beyond the petrochemical plant's property line.

This was both a relief and a concern for the plant's EH&S personnel. At least they now knew with a high degree of certainty that they had not caused the workers' sickness, and they were relatively sure who likely had. The next step would be to contact the facility that was located within the source location area.

This article originally appeared in the September 2009 issue of Occupational Health & Safety.

Featured

Artificial Intelligence