Thermal relief sizing

Consider Problem 9-9, part a. This time use alcohol as a liquid medium with a thermal expansion coefficient of 1.12 X 10 3/°C. The heat capacity of the alcohol is 0.58 kcal/kg °C, and its density is 791 kg/m3. Determine the relief size required. 

One example of a discontinuity is a drop in reaction rate due to depletion of reactants as the reaction nears completion. In this case, the step-wise method can still be used, but will tend to oversize. This is because heat losses in the full-scale reactor will cause the reaction to reach completion at a lower temperature/ pressure than was measured in an adiabatic small-scale test, corrected for thermal inertia (see Annex 2). It is not recommended to attempt to take account of heat losses from the full-scale reactor in sizing the relief system (e.g. by modelling them within a computer simulation). This is because a slight overestimation of the rate of heat loss could cause a large underestimation of the relief size required. 

A reactor has a volume of 2 m3. The worst case runaway reaction has been identified and the data from a suitable adiabatic, low thermal inertia test, with a thermal inertia ( ) of 1.05, is given in Figure 6.4. Under these conditions, the reactor would contain 793 kg of reactants. The reacting system is a vapour pressure system. It is desired to relieve the runaway via a safety valve, if possible, with a set pressure of 0.91 barg (relief pressure of 1.0 barg = 2.0 bara). Evaluate the required relief size for an overpressure of 30% of the absolute relief pressure, which gives a maximum pressure of 2.6 bara = 1.6 barg. 

The same example problem as used in Chapter 7 and A5.9.2 above will be used. A reactor of volume 3.5 m3 has a design pressure of 14 barg (maximum accumulated pressure 16.41 bara). A worst case relief scenario has been identified in which a gassy decomposition reaction occurs. The mass of reactants in the reactor would be 2500 kg. An open cell test has been performed in a DIERS bench-scale apparatus, in which the volume of the gas space in the apparatus was 3800 ml, and the mass of the sample was 44.8 g. The peak rate of pressure rise was 2263 N/m2s at a temperature of 246°C, and the corresponding rate of temperature rise was 144°C/minute. These have been corrected for thermal inertia. The pressure in the containment vessel corresponding to the peak rate was 20.2 bara. The liquid density at 246°C is estimated as 820 kg/m3. The gas generated by the runaway has a Cp/Cv value of 1.3. The problem is to evaluate the relief size required. 

The flows required for thermal relief are very small, and there are special thermal relief valves on the market that accommodate this specific application. Oversizing a thermal relief valve is never a good idea, and orifice sizes preferably below API orifice D are recommended. 

For liquid-packed vessels, thermal relief valves are generally characterized by the relatively small size of valve necessary to provide protection from excess pressure caused by thermal expansion. In this case, a small valve is adequate because most liquids are nearly incompressible, and so a relatively small amount of fluid discharged through the relief valve will produce a substantial reduction in pressure. 

The ASME code requires every pressure vessel that can be blocked in to have a relief valve to alleviate pressure build up due to thermal expan sion of trapped gases or liquids. In addition, the American Petroleum Institute Recommended Practice (API RP) 14C, Analysis, Design, Installation and Testing of Basic Surface Safety Systems on Offshore Production Platforms, recommends that relief valves be installed at vari ous locations in the production system and API RP 520, Design and Installation of Pressure Relieving Systems in Refineries, recommends various conditions for sizing relief valves. 

A reaction was believed to be thermally neutral, as no rise in temperature was observed in the laboratory. No cooling was provided on the pilot plant, and the first batch developed a runaway. Fortunately the relief valve was able to handle it. Subsequent research showed that the reaction developed 2 watts/kg/°C. Laboratory glassware has a heat loss of 3-6 watts/kg/°C, so no rise in temperature occurred. On the 2.5-m3 pilot plant reactor, the heat loss w as only 0.5 watt/kg/°C [21]. Reference 22 lists heat losses and cooling rates for vessels of various sizes. 

Equation 9-45 describes the fluid expansion only at the beginning of heat transfer, when the fluid is initially exposed to the external temperature Ta. The heat transfer will increase the temperature of the liquid, changing the value of T. However, it is apparent that Equation 9-45 provides the maximum thermal expansion rate, sufficient for sizing a relief device. 

The primary application of the VSP is to obtain data necessary to calculate the vent design (size and relief setting) for emergency venting of nonvolatile and reactive runaway systems. The calculated vent design is to limit the maximum pressures at the point of emergency venting to acceptable levels. A secondary application is to provide thermal stability data for reactive systems. 

The RSST calorimeter (see Annex 2) is a pseudo-adiabatic, low thermal inertia calorimeter, intended for screening purposes. It can identify the system type and measure adiabatic rate of temperature-rise and rate of gas generation by the reacting mixture. It is therefore well-suited to the task of selecting the overall worst case scenario from a small number of candidates. Alternatively, a calorimeter designed to obtain relief system sizing data may be used for this purpose (see Annex 2). 

Thermal expansion and fire cases are not required to be checked, if the existing equipment is re-used, with the same service and also the same level control setting. Overpressure relief requirements due to each utility failure, fire cases and any other combination scenarios need to be estimated. API 521 (2014) has a comprehensive list of effects for utilities failure. All the PRV manifolds shall be checked to estimate back pressures at the PRVs. PRD overpressure calculations for equipment shall be documented as shown in Table 3.4. Vacuum relief (if the vessel/s is/are not designed to withstand full vacuum) shall also be documented. All the flare scenarios and flare network shall be properly documented. An example of PRV sizing calculations for the system shown in Figure 3.5 is presented in Table 3.4. 

We plan to adopt some design improvements. They are simplification of reactor internal pump power supply system, change of FMCRD motor from stepping motor to induction motor, change of process computer from special ordered computer to work station computer, down sizing of radioactive dispersal system, change of flammability control system (FCS) from thermal recombiner to passive autocatalytic recombiner, size up safety relief valve and so on. 

The pressure relief system was not provided primarily to meet code requirements but rather Is an Integral part of the over-all protection system which includes the reactor power setback and scram systems Strict adherence to the code is impractical because the maximum thermal energy output of the reactor cannot be defined. Further at the hifpier conceivable powers fuel melting would probably occur in any event. Prevention of a fuel melting accident with consequent release of fission products depends primarily on the nuclear control and cool ant supply system and only Incidentally on the pressure relief systems. 3 Shutoff valves are not provided at the end of each instrument take-off connection, and the minimum size of instrument take-off connections is not necessarily l/2 or 3/if Inch as required by the ASA Code. 

---