Modelling a runaway chemical reaction using EFFECTS to reproduce the conditions of the T2 Laboratories incident

In the complex and hazard-prone world of chemical processing, the importance of predicting and understanding the severity and extent of potential incidents cannot be overstated, as it enables safety measures to be taken to prevent or minimise the consequences of devastating events.

One of the stark reminders of the risk associated with chemical processing was the catastrophic incident at T2 Laboratories, a chemical manufacturer in Jacksonville, Florida, USA, where a runaway chemical reaction occurred, resulting in an explosion that caused fatalities and injuries.

A runaway chemical reaction is a dangerous scenario where a thermally unstable reaction system exhibits an uncontrolled accelerating rate of reaction leading to a rapid increase in temperature and pressure¹. This can be caused by the failure of the cooling or control systems, inappropriate reaction scaling, or accidental introduction of impurities or incompatible substances. The implications for safety are significant — such reactions can lead to explosions, toxic releases, or fires, posing severe risks to health, the environment, and property. 

Figure 1. Self-sustaining chain of events that can lead to a thermal runaway².

This blog explores the critical importance of consequence modelling in understanding the impact of a runaway chemical reaction. Drawing on the process conditions from the T2 Laboratories incident, we will demonstrate the capabilities of EFFECTS, Gexcon’s advanced and user-friendly integral consequence modelling software, to simulate a runaway chemical reaction and quantify the subsequent effects of an explosion. Additionally, this blog emphasises the broader significance of consequence modelling as one of the proactive steps in supporting the prevention of similar catastrophic events in the chemical industry.

T2 Laboratories incident summary

On the afternoon of December 19^th, 2007, T2 Laboratories was the scene of a severe explosion disaster. The event claimed the lives of 4 T2 employees and caused additional injuries to 28 members of the public within the surrounding area, damaging the facility, and causing collateral damage to surrounding businesses.

Figure 2. Aerial photograph of T2 Laboratories taken December 20^th, 2007³.

The US Chemical Safety Board’s (CSB) investigation found that the catastrophe at T2 Laboratories was due to an uncontrollable exothermic reaction occurring specifically in the metalation stage within a batch reactor, during the production of a gasoline additive—methylcyclopentadienyl manganese tricarbonyl (MMT). 

A malfunction in the reactor’s cooling system initiated an undesired secondary exothermic reaction, which is believed to have set off the runaway reaction. This led to a rapid and unmanageable rise in both pressure and temperature. In mere minutes, the escalating pressure breached the reactor’s limits, causing a detonation that ignited its volatile contents and resulted in an explosion. 

Modelling the T2 Laboratories’ runaway chemical reaction scenario in EFFECTS

Approach

The EFFECTS model “Rupture of Vessels” is an explosion model which supports the definition of multiple causes of a vessel rupture, including runaway reaction and predicts overpressure effects and propelling debris range from catastrophic ruptures of process equipment. Therefore, the “Rupture of Vessels” model is deemed to be the most suitable to model the runaway reaction that occurred at T2 Laboratories.  

Using the “Runaway reaction, non-ideal gases/vapours” modelling approach removes the need to define the composition of the mixture at the moment of the explosion by defining the liberated energy during the chemical reaction. However, the mixture composition at the moment of the explosion could be defined in EFFECTS as the software allows the addition of new chemicals or user-defined mixtures in EFFECTS’s DIPPR chemical database. 

Figure 3. Selection of the cause of vessel rupture in EFFECTS’s “Rupture of Vessels” explosion model.

The CSB report mentioned that the mixture ignited at the moment of the explosion. The testing of the runaway reaction by the CSB’s laboratories determined the burst pressure and peak temperature to be 123.3 bar and 661.2 °C, respectively, at which conditions the reaction mixture is expected to be in a pressurised liquefied gas state.

As the mixture was released instantaneously, the “BLEVE Fireball” model in EFFECTS could be used for predicting the heat radiation effect of the explosion. In addition to assessing heat radiation effects, this model can also calculate overpressure effects. However, the “BLEVE Fireball” model lacks the assessment of the projected fragment reach, which is why the “Rupture of Vessels” model is selected to assess overpressure effects for this specific case.

Input parameters

The experimentally determined burst pressure and peak temperature of 123.3 bar and 661.2 °C, respectively, are used to specify the process conditions in the “Rupture of Vessels” model. 

The investigation report provides information about the process equipment such as the volume of the reactor (9.27 m³) and wall thickness (7.62 mm). The dimensions of the reactor and the mass of the empty reactor can be calculated based on the volume of the mixture contained within the batch reactor. The width and height of the reactor are calculated to be 1.97 m and 3.72 m, respectively. The weight of the empty reactor is 16,391 kg.  

The environmental conditions are set to the average of December meteorological data in Jacksonville. The figure below shows the input parameters used in EFFECTS’s “Rupture of Vessels” model. 

Figure 4. Input parameters used in the EFFECTS’s “Rupture of Vessels” model to simulate the runaway reaction at T2 Laboratories.

The definition of vulnerable areas in EFFECTS allows the estimation of damage to the surrounding buildings and people inside and outside those buildings. The population at the time of the accident was distributed inside and outside the surrounding buildings as seen in Figure 5. 

Figure 5. Injuries and business locations as reported to CSB³.

Based on the information provided in Figure 5 regarding the number of injured people during the accident, we can define vulnerable areas on the EFFECTS map as population polygons. The population polygons enable the definition of population both outside and inside buildings. This is achieved by specifying the fraction of the population inside buildings incorporated into the approximation of population distribution in the affected areas, as presented in Table 1.

	Number of people	Inside fraction of people (-)
Building A	16	0.9
Building B	6	1
Building C	10	0.4
Building D	8	1
Building E	3	1
Building F	10	1
Building G	12	0.93
Outside area of the reactor	9	0
T2 building	4	0.25

Table 1. Definition of vulnerable areas in EFFECTS based on Figure 3 presenting the injured and killed people as dots in the map.

Figure 6 displays the input parameter configuration for a vulnerable area labelled as the “Outside area of the reactor” (an area situated within the white circle shown in Figure 5). The presence of nine individuals in this area, as noted in Table 1, is reflected in the “number of people” field in Figure 6. The model employs a 300 mbar overpressure level, considered as the level at which total destruction occurs, to assess the fatalities resulting from exposure to this overpressure level.

Figure 6. Definition of the vulnerable area labelled as “Outside area of the reactor” in EFFECTS.

Modelling results and implications 

Experts from the CSB assessed the damages to buildings, people, and the reach of the flying fragments. That is why the modelling approach used is aimed at predicting the reported damages caused by the explosion. 

Based on Figure 5, which illustrates the fatalities and injuries at various business locations within T2 Laboratories, it is evident that the reported fatalities occurred exclusively in the area surrounding the reactor (shown as a white circle in Figure 5). This area is defined as a vulnerable area called the “Outside area of the reactor” in EFFECTS, as highlighted in Figure 6. 

Two overpressure limits have been assessed in EFFECTS: 300 mbar and 100 mbar. At 300 mbar, buildings collapse on top of the people that remained indoors, while those outdoors suffer from total body displacement and death. Meanwhile, at 100 mbar, window glass shatters, resulting in minor injuries for people inside buildings due to glass fragments. 

Figure 7 illustrates the EFFECTS simulation of the runaway reaction. In green, the area surrounding the reactor is displayed as a vulnerable area (where 9 people were present). In dark-shaded red, the 300 mbar overpressure contour is displayed (within which fatalities occur due to total destruction).

According to the EFFECTS simulation, only a part of the outside area of the reactor is reached by the 300 mbar overpressure contour. The 100 mbar overpressure contour, which corresponds to a level where glass in windows shatter injuring the people inside the building, reaches the outside area of the reactor and the T2 building. This level of overpressure will only result in injuries.

Figure 7. Overpressure contours resulting from the reaction runaway calculated in EFFECTS.

Figure 8 shows an EFFECTS report summarising the consequences of the runaway reaction on the defined vulnerable areas. The report specifies the wind angle for the worst-case scenario, distance from the explosion centre, overpressure effects and number of lethalities per vulnerable area. EFFECTS reports a total number of 4 fatalities out of 9 people that were present in the area around the reactor.

This is the exact same number of fatalities reported to CSB (see Figure 5). No other fatalities were reported from the accident. However, the buildings were exposed to significant overpressure. These overpressure effects are calculated in EFFECTS, and the corresponding results can be seen in Figure 8.

Figure 8. EFFECTS report summarising the explosion results in the vulnerable areas.

According to the CSB report, the rupture of the reactor resulted in the projection of five big fragments. Notably, two of the reactor’s large steel support columns flew up to a distance of 305 m from the reactor. The reactor head damaged a building at a distance of 121 m from the reactor. Additionally, the agitator shaft was thrown 106 m away, and the piping flew several tens of meters onto the other businesses and wooded areas surrounding T2. 

The maximum distance of the heaviest fragment of the reactor calculated by EFFECTS is 163 m (see the pink contour “maximum range fragments” in Figure 9), which corresponds to the majority of reported flying fragments.  

Figure 9. EFFECTS report summarising the explosion results in the vulnerable areas.

Through simulating this runaway reaction scenario using EFFECTS, we can understand the possible outcomes of such a scenario, including the impacts of any subsequent explosion. This information is valuable in informing safety strategies. For example, preventative measures, such as designing proper emergency pressure relief systems and incorporating essential safety features such as cooling water systems that are aligned with the process conditions and substances used in the process, can be developed based on these insights. Furthermore, enhancing the competence of employees in understanding the potential of runaway reactions, along with conducting thorough inspections, can be considered as other safeguard approaches.

Conclusion

To conclude, the modelling of a runaway chemical reaction scenario in EFFECTS, drawing on data from the T2 Laboratories incident, has provided crucial insights into the consequences of such events. This example showcases the value of advanced consequence modelling tools for understanding hazard severity, supporting risk reduction, and preventing similar events from occurring.

Through accurate simulation of the scenario, EFFECTS generates valuable information that assists in making informed decisions for safety in chemical facilities to safeguard staff, assets, and the surrounding environment.

Any questions about modelling runaway chemical reaction scenarios in EFFECTS? 

Reach us at effects@gexcon.com.

References

1. American Institute of Chemical Engineers (no date) CCPS Process Safety Glossary: Runaway reactions. Available at: https://www.aiche.org/dippr 

2. Dutta, S. and Bellair, R.J. (2022) Losing Your Heat Balance: Insights into Thermal Hazard Assessments, AICHE. Available at: https://www.aiche.org/resources/publications/cep/2022/march/losing-your-heat-balance-insights-thermal-hazard-assessments 

3. US Chemical Safety and Hazard Investigation Board (2009) Investigation Report: T2 Laboratories, Inc. Runaway Reaction (Four Killed, 32 Injured). Available at: https://www.csb.gov/t2-laboratories-inc-reactive-chemical-explosion/ 

4. American Institute of Chemical Engineers (no date) DIPPR: Design Institute for Physical Properties. Available at: https://www.aiche.org/dippr  

Writers

---