By-pass streams

The exposure of sensors in a by-pass stream (which can be valved off), is an alternative way of collecting monitoring data although correlation is required between the main-stream and the by-pass The use of a side-stream taken either side of a choke in the main-stream can provide a useful monitoring point. Traps where product streams can be condensed can offer alternative sampling systems. 

Concentration of polymer of chain length j Concentration of polymer of chain length J Concentration of active polymer of chain length j Concentration of polymer of chain length j in the by-pass stream of a by-pass CFSTR... 

Material balance calculations on processes with by-pass streams are similar to those involving recycle, except that the stream is fed forward instead of backward. This usually makes the calculations easier than with recycle. 

Remark 1 Note that given the process units, inputs, and outputs such a graph representation includes all possible alterative interconnections among the units, the inputs, and the outputs. Note also that the one-way arcs directly from the inputs to the outputs correspond to by-pass streams of the process units. Arcs originating from a unit and returning to the same unit without going through any other node have not been incorporated for simplicity of the presentation. 

Remark 2 An important consequence of allowing heat flow across the pinch if the economic versus operating trade-offs suggest it, is that the optimal HEN structures that are obtained in Ciric and Floudas (1990) are simple and do not feature by-pass streams. As a result, these structures may be more interesting from the practical application point of view. 

The topology of the HEN shown in Figure 8.28 features a split of Cl into three streams. In the top two branches there are two heat exchangers in series, while in the third branch there is one heat exchanger. Note that there is also a by-pass stream directed to the inlet of the HI — Cl exchanger. 

Note that the recoveries in the optimal solution shown in Figure 9.6 are at their lower bounds for r 2 and r fc2, and at the upper bound for. A comparison between sharp and nonsharp sequences Aggarwal and Floudas (1990) shows that significant savings can result from optimizing the degree of nonsharp separation. In particular, in example 2 of their paper, which corresponds to the same optimization model presented in the illustrative example but with different product compositions, the optimal solution consists of a single column with a by-pass stream. This corresponds to 70% cost reduction versus the sharp separation case. 

In Appendix A of Aggarwal and Floudas (1990), a procedure is presented that calculates upper bounds on the flow rates of the overall by-pass streams. These bounds are useful in restricting the feasible domain of the remaining flow rates. 

The set of constraints for the superstructure of Figure 10.2 of the illustrative example is as follows (note that we omit the by-pass streams around the reactor units for simplicity of the presentation) ... 

Once optimum performance of the plant has been reached, tracer experiments may be carried out to indicate deviations from optimum conditions. Often the reasons for malfunction are found, like unwanted by-pass streams, or obstruction of vessels and pipes which can cause changes in flow-rate or the appearance of dead zones. 

The flow configuration comprising of two perfectly-mixed reactors of volumes Vi and V2 is demonstrated in Fig.4.3-2. A tracer in a form of a pulse input is introduced into reactor 1 and is transferred by the flow Qi into reactor 2 Q - Ql is the by-pass stream. The tracer is assumed to accumulate in reactor 2, while flow Ql is leaving the reactor. This configuration simulates a situation demonstrated in ref.[81], designated as "short-circuit". It should be noted that the behavior in reactor 1 is independent of what happens in reactor 2. 

Cases a and b in Fig.4.4-3a demonstrate the effect of q. By increasing q, i.e. the by-pass stream, the mean residence time in the tubular reactor tp is increased from 0.1 to 0.5 time units. Case a and c demonstrate the effect of Xi = X2- By increasing this quantity from 10 to 500, the mean residence time in the perfectly mixed reactor is decreased, and the response becomes instantaneous. In the computations. At = 0.002 for q = 1 and 0.01 for q = 5. 

The first problem is the key issue. The immediate approach is to see the reactor synthesis as the inverse problem of the decomposition of a real reactor in compartments of ideal mixing patterns. Thus, the chemical reaction network would consists of a combination of ideal models, CSTR s and PFR with connections and by-pass streams. 

Example. Given Shell exit temperature, Tout = 100 F tube fluid temperature, t, at reversal of tubes = 71° F overall heat transfer coefficient, C/, multiplied by the surface area of the exchanger per baffle section, Ag (sq.ft.) or UAg = 2,000 shell flow rate, W, (lb./hr.) multiplied by the specific heat of the sheU fluid, C, (Btu/lb. °F) or WC = 10,000. Ratio of stream heat capacities, JR, equals the tube fluid flow rate, w (Ibs./hr.) times the specific heat of the tube fluid, c, (Btu/lb. °F) divided by WC equals one, or i = wcjWC = 1 magnitude of the by-pass stream expressed as a fraction of total flow, I = 0.6. 

---