Scaling performance issues

Managing Scale and Scope of Large-Scale Process Operations discusses the performance issues associated with data analysis and interpretation that occur as the scale of the problem increases (such as in complex process operations). 

Coherence and maintainability of the system degrade rapidly if different developers choose their own schemes to solve these technology issues. Large-scale performance and scalability are determined primarily by how the components are partitioned and distributed across machines and by the nature of communication across machines. 

The best data to use for determining whether an incompatibility exists will obviously be from testing the actual scenarios and conditions that are identified. However, this is often not practical or possible. Small-scale tests can be performed in a laboratory that can give an indication whether a reaction is expected. However, be wary of concluding that since no reaction is seen on a small scale, no effects will be realized in an industrial facility. Heat transfer effects and scale-up issues are especially important to be careful of when extrapolating small-scale results. Differences in heat transfer, mixing and other scale-up effects can cause a significant and potentially... 

If the desired yield target cannot be achieved then an antisolvent system can be selected using the same techniques. The antisolvent should be fully miscible with the primary solvent and have a low solubility for the solute. It should be noted that the addition of an antisolvent to reduce solubility and generate supersaturation may introduce scale up issues, caused by the differences in micro-mixing performance between the laboratory and manufacturing plant. 

While this route worked well on multigram scale, significcint issues were identified during the scale-up process. First, reproducibility of the alkylation was problematic. Sometimes acid 8 (Scheme 4) would form a gel when dissolved in DMF. Therefore, the intermediate sodium salt 15 was isolated and used as the alkylation precursor. This was done with NaOH/IPA however, tedious azetropic drying of the IPA solution was required to obtain reasonable yields. Also, product isolation was difficult due to the hygroscopic nature of sodium salt 15. In order to successfully perform the subsequent alkylation, sodium salt 15 was dissolved in DMF, which also needed to be azeotroped again in order to provide a dry stream for the alkylation. 

The above is a partial list of frequent scale-up issues that arise in bulk drug processing with consequences of lower chemical yields, or worse yet, loss of control over the impurity profile, as well as slower processing, excessive damage to microbial cells and crystalline solids, undesirable particle size distributions and any from a wide range of assorted shortfalls in process performance. 

Fiorentino and Newton (1998) x-ray Identifying of scale-up issues for predicting large-scale BP reactor performance... 

As concerns the second assumption, direct access to the gas pore network is not a stringent requirement for keeping the Pt surface active. The condensed phases of water and ionomer possess finite oxygen permeabilities, which will render oxygen concentrations finite at electrolyte-covered Pt particles. Again, this is an issue of nonuniform reaction rate distributions that will be dealt with in the section Hierarchical Model of CCL Operation. In that section, a two-scale performance model will be presented, which couples transport and distribution of protons and oxygen at the agglomerate level and at the macroscopic scale. 

In a prophetic vision of our physical world, Richard Feynman issued in 1959 his famous statement "There is plenty of room at the bottom". He was essentially suggesting experiments and technologies which could be performed at extremely small scales. We know nowadays that Feynman s expectations were beyond what has been achieved since he made his prediction. 

Use Scalable Heat Transfer. The feed flow rate scales as S and a cold feed stream removes heat from the reaction in direct proportion to the flow rate. If the energy needed to heat the feed from to Tout can absorb the reaction exotherm, the heat balance for the reactor can be scaled indefinitely. Cooling costs may be an issue, but there are large-volume industrial processes that have Tin —40°C and Tout 200°C. Obviously, cold feed to a PFR will not work since the reaction will not start at low temperatures. Injection of cold reactants at intermediate points along the reactor is a possibility. In the limiting case of many injections, this will degrade reactor performance toward that of a CSTR. See Section 3.3 on transpired-wall reactors. 

---