Protective Measures

Protective measures to ensure safe operation have either to deal with, or to mitigate, the consequences of the runaway reaction. Protective measures include emergency venting or relief systems, inhibition, and containment.

In common with using prevention as a basis of safety, it is essential that a full evaluation of the hazards of the process is carried out, before the type of protective measure is chosen and designed. The identification and definition of the "worst case" scenario is particularly important as, in contrast to prevention, any protective measure has to be able to cope with the worst case runaway reaction.

Protective systems are rarely used on their own and some preventive measurements are usually included to reduce the demand on the protective systems. It should be recognized that it is not always possible to design a protective system to cope with the consequence of a runaway reaction. Protective systems such as crash cooling, drown out and reaction inhibition (see below) involve the detection of the onset of the runaway reaction and subsequent corrective action to prevent it occurring. In addition to suitable detection methods and the availability of process compatible systems, these techniques need time to act and are of limited use when the runaway is caused by a sudden event. In such cases, if preventive measures are also unsuitable to ensure safe operation then the process may need to be radically re-designed or even abandoned.

Emergency Venting and/or Relief Systems

Venting is often the most practical system for the relief of runaway reactors, and regardless of other safety systems, a vent will normally be present on a reactor, directing any flow to a known location rather than resulting in an exploding reactor. Over-pressure venting or pressure relief is fitted to chemical reactors to cope with one or more of the following:

The vented material may be:

The venting of such materials direct to the atmosphere is undesirable, generating hazards in the vicinity of the reactor, where the bulk of the liquid will fall out from the atmosphere and be deposited on local plant. A flammable mixture may result in an explosive atmosphere, and toxic, corrosive or very hot liquids are a danger to personnel. An expensive clean-up procedure can be expected where the liquid deposits on process plant, particularly when solidification is likely.

Thus, simply having a vent on a reactor is not enough, and some consideration must be given to treatment of the discharge or at least to the direction and location of the end of the vent line. More often, a pressure relief valve is installed instead of atmospheric venting.

Few chemical reactors are operated without any pressure relief system, but relief systems sized to cope with service fluids or fire engulfment are rarely adequate to protect against the effects of a runaway reaction.

The discharge may be purely gaseous or vapor or it may be a vapor-liquid mixture. 2-phase discharge is common among foaming systems. The presence of 2 phases in non-foaming system is caused by the phenomenon of "liquid swell" as follows.

When a runaway reaction occurs and a vent/relief operates to release pressure, bubbles of gas or vapor form in the reaction mixture and rise to the liquid surface where they disengage. While the bubbles are still in the liquid, however, their volume increases that of the liquid and the liquid level rises. If the liquid level reaches the vent/relief, liquid as well as vapor is discharged. A 2-phase mixture usually requires a larger vent/relief area than a single-phase gas or vapor discharge.

Chemical reaction systems can also contain solid materials such as catalysts. The relief mixture in such cases will often be 3-phases. It is currently believed that the presence of the third solid phase has little effect on the required relief area, though particular care should be taken to ensure that the relief and associated downstream equipment do not become blocked by the presence of solids.

It is important that a suitable system is fitted for the batch being processed. Many reactors in the pharmaceutical and specialty chemical industries are multi-purpose, and different mitigation systems may be relevant in each batch. Relief and disposal systems are difficult to modify once installed. It may be necessary to have more than one system available in these cases.

Back on Top

 

Inhibition

Inhibition involves injecting small quantities (in ppm) of inhibitor to the reactor at an early stage of the runaway. The inhibitor used can either halt the reaction completely or reduce the reaction rate to delay further runaway for a time period, allowing time for absent ancillary services or agitation to become available.

The suitability of a system for inhibition is dependent on the reaction mechanism occurring in the reactor. Inhibition is ideally suited to free-radical-initiated reaction systems (such as some polymerization reactions), pH-dependent reactions and systems where the catalyst can be removed by the inhibitor.

Where possible, inhibition should be used at least as an initial protection against thermal runaway. The advantages are the passive nature and complete containment within the reactor without the need for extreme pressure design of the reactor. This is particularly important where the pressure generated during runaway is due to the vapor pressure of the system. Such reaction systems require very high pressure ratings for complete containment. Since no venting is involved, inhibition is applicable to foaming and highly viscous systems.

In spite of their apparent ideal suitability, inhibition systems are rarely used in industry. Disadvantages of inhibition include the lack of published information regarding the design of injection systems, uncertainties over the mixing efficiency and distribution of the inhibitor, and the fact that reaction may not be altogether halted and may continue at a reduced rate before a second runaway occurs. The injector system must be designed to give good distribution of the inhibitor within the reactor to provide sufficient mixing, and provide the mixing in as short a time as possible. Problems exist with the mixing of such small quantities within the bulk mass in the reactor, as good mixing is necessary to prevent hot spots, i.e. pockets of reactant remaining uninhibited and generating the temperatures and pressures that would occur if the reaction was allowed to continue.

The time taken to inject and distribute the inhibitor is also an important factor. Aids to mixing within a non-agitated system are desirable to reduce the number of injectors required on a large system. Consideration should also be given to the effect of viscosity on the mixing, as many systems, particularly polymerization reactions, will become more viscous as the reaction proceeds. The increased viscosity will impede mixing, and the system should be designed for the worst case, i.e. when the reaction is nearing completion.

Passive injection, i.e. systems requiring no external power, should be preferred. Thermal runaway is likely to be caused by loss of power resulting in loss of control system, agitation or cooling. It is therefore likely that stirring of the reactor, to distribute the inhibitor, is not possible in the event of power failure. Any agitation available will be due to residual stirring (for a short time after power loss), convection currents or separate agitation provided by the injector.

Additional considerations must be given to the methods for early detection of the onset of runaway and actuation of the injector, in order to optimize the timing of the injection in relation to the runaway process.

Detecting the onset of a runaway for actuation of the injector can be done by use of several parameters. The actuation can be linked to the temperature or pressure in the reactor, the failure of agitation or the loss of cooling to the system. A pressure-control system would normally be used for normal operation, and loss of control can happen at any point during the runaway. Pressure is often a better choice of parameter for runaway detection, as over-pressurization of the reactor is the main concern. Using temperature as a trigger may not always be effective, as some systems, particularly those generating non-condensable gases, can give high pressure rises at small temperature differentials.

Linking the injector to the loss of cooling or agitation is also practical, although the injector may be actuated before runaway has commenced. Mechanically rather than electrically controlled actuators should be used in this case, as the cause of failure of services is likely to cause failure of the control system.

Back on Top

 

Containment

Containment within the reactor should always be considered where practicable. As with inhibition, the system is passive and no venting occurs, making it suitable for foaming and highly viscous systems.

A reactor system to be used for total containment should be designed to withstand the maximum pressure produced by the runaway reaction. In the case of vapor-generating reactions, where the pressure increases exponentially with temperature, this may not be a practical solution. Impractical wall thicknesses (in terms of economic cost as well as construction) may result. Clean-up following an incident may be difficult if the reactants are allowed to solidify.

Encasement of the reactor in concrete or in a steel or concrete bunker may provide an alternative. Where a vapor is present, the cooling effect of the expansion due to reduced pressure may be used to slow the reaction significantly such that it can be contained by the system.

 

Quench and Dumping

Quenching and dumping are more usual methods of inhibiting a runaway reaction. In both cases a quantity of cold inert diluent is added to the reaction mixture and stops the reaction by cooling it.

The ideal substance is water, when applicable, which is cheap, readily available and has a high specific heat. In some cases water reacts exothermically with the reaction mixture and an alternative diluent must be used. An example is the runaway decomposition of a sulphonation reaction, where concentrated sulfuric acid is used as diluent.

The actual addition of diluent is best carried out by quenching, where the diluent is added quickly to the reactor from a storage vessel mounted above it. When the runaway is detected (e.g. by temperature rise) a valve opens automatically and the quench liquid runs rapidly into the reactor under gravity. When there is not enough free space in the reactor to introduce an adequate quantity of diluent, dumping can be used. In this case the reactor contents are run off (dump) into another vessel containing the quench liquid.

Quench systems are suited to most reaction systems: "gassy", foaming and viscous mixtures can be quenched with a suitable quench liquid. It is possible to have complete containment of the chemical system using quenching, preventing toxic or hazardous vapors or gases being passed to the atmosphere.

If the quench liquid is carefully chosen, the quench liquid may itself react with the discharge and inhibit or halt the reaction. This active quenching is an ideal substitute to reactor inhibition where larger quantities of the inhibiting medium are required. In passive quenching, the two-phase vented mixture is sparged through a dedicated liquid in a quench tank, which cools the reaction preventing further runaway, condenses vapors and allows only non-condensable gases to escape or pass on to a vent stack or for further treatment.

The major uncertainty in quench systems is the design of the sparger for optimum mixing and subsequent heat transfer from the quench fluid to the two-phase mixture. The presence of large vapor bubbles may prevent some of the vapors from contacting the cold liquid interface. The size and number of holes need to be properly defined.

Care should also be taken in ensuring that the reacting mass does not solidify in the sparger when the discharge first hits the cold quench fluid in the vent line. Reacting liquids that will become solid at ambient temperatures are one of the few situations where quench tanks should not be used. A different form of quenching, non-sparged, may be suitable. Discharging the reactor to the void space above the quench fluid and letting the reacting mass hit the liquid will quench the reaction and at least give some benefit. Mixing of the quench and discharge is the main problem in this case.

The following should be considered when using a quench system:

 

Back on Top