On May 22, 2026, roughly 50,000 residents of Garden Grove, California were ordered to evacuate their homes while emergency crews spent four days trying to cool an overheating tank of methyl methacrylate at a nearby facility. The tank bulged, cracked, and came close to a boiling liquid expanding vapor explosion (a BLEVE) before pressure finally dropped enough to stand down the worst-case scenario.
GKN Aerospace is a well-established manufacturer of components for commercial and military aircraft, operating facilities around the world. This incident is not a story about a negligent operator. It is a reminder that reactive chemical storage hazards can develop at any facility, and that the gap between a well-run operation and a serious incident is often a question of whether the right engineering analysis was in place and kept current.
Methyl methacrylate (MMA) is a well-characterized reactive monomer with a well-documented hazard profile. The conditions that can drive an uninhibited polymerization runaway in MMA storage are understood, the reaction kinetics have been studied extensively, and the engineering controls required to prevent this scenario are not experimental.
What happened in Garden Grove is an opportunity for the process engineering community to discuss current practices around reactive monomer storage and how process simulation and hazard analysis can help anticipate events such as this one.
What the Chemistry Tells You
MMA undergoes exothermic free-radical polymerization. In storage, the reaction is suppressed by a chemical inhibitor, typically monomethyl ether hydroquinone (MEHQ), which requires dissolved oxygen to function. That dependency creates two simultaneous failure conditions that can compound each other: inhibitor depletion and oxygen-depleted vapor space. Either condition alone increases polymerization risk. Together, they create the conditions for a runaway.
The reaction is autocatalytic. Once polymerization initiates in a bulk storage environment, heat release accelerates the reaction rate, which generates more heat, which accelerates the reaction further. Adiabatic temperature rise in uninhibited MMA can drive vapor pressure well beyond vessel design limits in a timeframe that does not allow for manual intervention.
The valve seizure reported in the Garden Grove incident is consistent with localized polymer formation on valve internals. MMA polymerizes readily on metal surfaces, particularly in stagnant zones near isolation valves where inhibitor concentration and oxygen availability may be locally lower than in the bulk liquid. This is a recognized failure mode in MMA handling systems, and one that process engineers working with reactive monomers should have on their hazard register.
Reporting from the scene indicated that by the time responders arrived, the tank temperature was rising approximately one degree per hour. A degree per hour is not a sudden event. It is a trend line. That rate of change, applied to a reactive monomer storage vessel with known polymerization kinetics, raises the question of whether simulation-informed alarm setpoints were in place to surface that signal earlier in the progression. It is worth asking, as an industry, how we instrument and monitor reactive chemical storage against the scenarios our models define.
What a Validated Process Model Defines
A CHEMCAD simulation of an MMA storage system is not a theoretical exercise. It produces specific, actionable engineering parameters that define the boundary between safe operation and runaway initiation.
A complete model of this system would include reaction kinetics for uninhibited and partially inhibited MMA polymerization, adiabatic temperature rise calculations as a function of initial inhibitor concentration and temperature, pressure buildup rates under runaway conditions at varying fill levels, heat input from ambient conditions and any process utilities, and sensitivity analysis across the parameter space most likely to degrade in service: inhibitor concentration, dissolved oxygen, and storage temperature.
From that model, you can establish a safe operating envelope with quantitative boundaries. Not rules of thumb. Not general guidance from a safety data sheet. Specific temperature limits, minimum inhibitor concentrations, and vapor space oxygen requirements that, if maintained, prevent the reaction from initiating. Those numbers then translate directly into DCS alarm setpoints, inspection intervals for inhibitor sampling, and operating procedures for vapor space management.
Relief system adequacy is the other critical output of this analysis. The question is specific: is the pressure relief device sized for the uninhibited polymerization case, or only for the scenarios considered during initial facility design? A relief device sized for a fire case or utility failure may be undersized for a full polymerization runaway. FERST, Fauske's emergency relief sizing tool powered by CHEMCAD, is designed to answer that question rigorously. FERST applies DIERS methodology to size relief devices for tempered, gassy, and hybrid reactive systems, and supports dynamic simulation of a relief event using measured adiabatic calorimetry data as the source term. If your facility stores MMA or any reactive monomer and that relief sizing analysis has not been run against the worst credible runaway scenario, it should be.
The MOC Connection
Incidents like this one are also a reminder of how important it is to keep process models current. Any modification to equipment, operating conditions, or chemical handling procedures at a reactive monomer storage facility should trigger a review of the existing simulation against the new configuration. MOC is not only a regulatory requirement — it is the mechanism by which a facility's process model stays connected to its physical reality. When equipment changes and the simulation is not updated, the safe operating envelope the model defines may no longer reflect what is actually in the plant.
A living process model, maintained through MOC discipline, is a forcing function for engineering rigor. If your simulation requires updated inputs every time something changes, engineers remain in the loop on whether new configurations have been evaluated against the hazard scenarios that matter. That discipline is not unique to any one facility or industry segment. It applies wherever reactive chemicals are stored at scale.
What Rigorous Simulation Requires
To model an MMA storage system to the standard of care, the simulation work should include, at minimum, a validated reaction kinetics model for MMA polymerization calibrated against published data, a dynamic scenario for inhibitor depletion as a function of storage time and temperature, a pressure-temperature profile for the runaway case at maximum fill volume, confirmation that relief device sizing accounts for the maximum credible runaway rate, and sensitivity analysis showing the consequence of operating at the margin of the safe envelope.
None of this is computationally intensive by modern standards. It is time-intensive to do correctly, which is precisely why it needs to be treated as an engineering deliverable rather than an informal estimate. A properly documented simulation package for an MMA storage system is auditable, defensible, and updateable. It is also the foundation on which every downstream safety decision about that system should rest.
The Broader Point
Our thoughts are with the residents of Garden Grove and the emergency responders who worked around the clock to manage this situation. Incidents like this one carry a real human cost, and that should not be lost in the technical discussion.
The process engineering community has the methods and the tools to define the hazard envelope for reactive chemical storage with precision. The Garden Grove incident is a useful prompt for any facility handling reactive monomers to ask whether their current simulation work fully covers the worst credible storage scenarios, whether those models are current, and whether the alarm setpoints on the plant floor reflect what the simulation says.
When the model does not exist, the hazard does not disappear. It just goes unquantified until the plant tells you what it is.
Talk to a CHEMCAD engineer to learn how CHEMCAD's dynamic simulation, or FERST software powered by CHEMCAD, handles reactive monomer runaway scenarios.
