Principales retos en la implementación y operación de tecnologías RFID

1.7.1

Introduction & Importance

Measurement of the solids RTD in CFB risers and downers is very useful for several purposes. In non-catalytic gas / solid reactions, such as biomass pyrolysis, the main applications of the solids RTD are to ensure adequate gas / solids contact times and to assess axial and radial heat and mass transfer (Huang, Qian, Zhang, & Wei, 2006). In other situations, the solids RTD is used to control solids mixing and residence time, to characterize reactor hotspots, and to identify unreacted material (Lackermeier & Werther, 2002).

Many different experimental techniques have been employed to directly measure the solids phase RTD, all of which involve the use of a solid tracer material that is assumed to accurately represent the flow patterns and characteristics (e.g. density, size, shape) of the actual solids phase. These various experimental methods and results are reviewed in the following section.

1.7.2

Experimental Methods

In short, the experimental procedure for measuring the solids phase RTD in a CFB riser / downer involves:

i. Pulsed or step injection of a small amount of tracer particles, or introduction of a single tracer particle into the surrounding flow,

ii. Measurement of tracer concentration with time as the tracer flows past the detector(s).

Figure 1.7 shows a typical arrangement for a tracer injection / detection system in a CFB riser. Depending on the locations and methods of tracer injection and detection, the RTD may be measured between any two locations.

Figure 1.7 - Schematic diagram of a tracer injector / detector system in a CFB riser (from Ambler et al., 1990)

Many different tracers and tracer injection techniques have been used to measure the solids phase RTD in CFB reactors. These various methods are grouped by tracer particle property: radioisotope, ferromagnetic, salt, chemical, luminescent (i.e. fluorescent and phosphorescent), and other miscellaneous types. An excellent summary of previous experimental investigations into the solids RTD in CFB risers is provided in Harris et al. (2003a) and in Gao et al. (2012), including information about: tracer type, riser geometry (e.g. height H and diameter D), and operating conditions (e.g. superficial gas velocity, solids flux, reactor temperature, particle properties). Huang et al. (2006) also provide a summary of previous solids phase RTD studies. Their critical reviews are omitted here for conciseness.

Solids phase RTD studies in downers were reported in Roques et al. (1993), Wei et al. (1994), and Huang et al. (2006), in which all used phosphorescent tracer particles. The use of phosphorescent tracer particles seems to be the most reasonable compromise between RTD accuracy and ease of implementation. Phosphorescent particles used in RTD studies were activated using strong visible light flashes to create very close

approximations to true pulse injections. However, preferential activation of particles near the wall versus particles near the column centerline may be a problem if the flash brightness is too low or when operating at high solids mass flux. Excessive tracer may also be activated if the light flash is not collimated. Radioactive tracers can give accurate results but require long down times between experiments as the tracer is recovered and the reactor is decontaminated from residual tracer. Use of salt tracer is easy to implement but results in large errors in the injection and detection / sampling stages. Organic chemical tracer is moderately difficult to implement and has problems with re-adsorption on the solid particles.

The response time of solids tracer detection and sampling is one of the greatest limiting factors of RTD measurement. For example, in Bader et al. (1988), Rhodes et al. (1991), and Zheng et al. (1992), the temporal resolution of salt sampling was on the order of 0.5 s, which is far too long for accurate RTD measurement according to Harris et al. (2003a). Scintillation counters used with radioactive tracers (e.g. Ambler et al., 1990, Patience et al., 1990) have extremely short response times and are generally quite accurate, but they may erroneously detect radiation too early or too late if the emitted radiation is not collimated into the detector. Nearby particles at rest (e.g. solids collected in a cyclone dipleg) can also be detected erroneously if proper shielding is not used. Gas chromatography (GC) is often used to detect organic tracer desorbed from solid particles. GC also has adequate response times but may give erroneous RTD results if the organic tracer re-adsorbs onto the solid particles and is not captured by the detector. Photomultiplier tubes and photocells are typically used with phosphorescent and fluorescent tracer particles. They have good temporal resolution but their sensitivity may be reduced at heavy solids loadings by non-phosphorescent particle shielding (Yan et al., 2009). Like radiation detectors, these sensors may also detect tracer too early or too late if the emitted light is not collimated into the detector.

Harris et al. (2002a) listed the following numerous advantages of the phosphorescent tracer technique:

• instantaneous, non-intrusive activation of tracer by a light pulse, • simple, non-intrusive online detection of the tracer by a light detector,

• the tracer is identical to the rest of the bed material (when not blended with the actual bed material used in the reaction),

• no extra particles or gas are added to disturb the flow,

• no tracer accumulation in the bed since the bed particles deactivate, and • low cost compared with radioactive tracer studies.

However, the method also has a few minor disadvantages:

• excessive or improper activation of tracer if the light flash is not plane collimated, • early or late detection of activated tracer if the detected light is not collimated at

the detector, and

• preferential activation and / or detection of particles near the wall versus particles near the column centerline if the flash brightness is too low or when operating at high solids mass flux.

Detection of tracer outside of the measurement plane would tend to skew the measured RTD to be narrower and earlier than the true RTD due to detector bias toward the slightly brighter solids upstream of the detector. Overall, the phosphorescent technique seems to be the most reasonable compromise between RTD accuracy, ease of implementation, and cost after taking precautions for proper tracer activation and detection.

1.7.3

Experimental Results

Experimentally determined solids phase RTDs in CFB risers / downers are used to: • estimate the gas-solids contact time;

• quantify the extent of backmixing in the reactor; • identify problematic flow regions;

• apply the findings to tune hydrodynamics and reaction kinetics models.

Solids backmixing can have a strong negative impact on conversion and is typically to be avoided. Experimental RTDs have also given evidence for the existence of a core- annulus flow structure in risers (e.g. Rhodes et al., 1991; Harris et al., 2003a).

The effect of superficial gas velocity and solids circulation rate on the measured RTD in risers / downers has been studied extensively. On the whole, the results were fairly consistent. Most authors concluded that increasing the superficial gas velocity led to a

decrease in the mean residence time, decreased axial dispersion, and a tendency toward plug flow in both the gas and solids phases (e.g. Ambler et al., 1990; Rhodes et al., 1991; Smolders & Baeyens, 2000b; Harris et al., 2003a; Yan et al., 2009). Harris et al. (2003a) demonstrated that increasing gas velocity decreased the RTD signal’s variance (i.e. a measure of peak spread) and breakthrough time (i.e. time required for tracer to be initially detected).

The solids loading has a less clear influence on the solids RTD. In general, increasing solids flux has been found to cause increased mean residence time, increased axial dispersion, shrinking peak height, increased tail length, and in some cases the appearance of a second peak in the RTD (e.g. Harris et al., 2003a; Ambler et al., 1990; Smolders & Baeyens, 2000b). However, conflicting results were reported in Rhodes et al. (1991) and Yan et al. They found that axial dispersion actually decreased slightly with increasing solids flux. They acknowledged that this relationship was quite weak and claimed that the effect of the superficial gas velocity was much more influential on mixing.

The riser diameter has also been shown to have an effect on the solids RTD. Conflicting conclusions are given in Rhodes et al., where it was found that increasing the riser diameter decreased backmixing, and Pugsley et al. (1997), who found an opposite trend. An interesting result in downer reactors was obtained by Huang et al. (2006), where it was found that there was essentially no significant scale-up effect on backmixing when increasing the reactor diameter. They also claimed that the radial solids mixing was more intense in the larger diameter reactor, which they surmised to be advantageous for downer scale-up.

Figure 1.8 compares the solids axial Péclet number versus superficial gas velocity in a similar sized riser versus downer as reported by Harris et al. (2003) and Wei et al. (1994), respectively. The dimensionless Péclet number, Pe = LUp / Dax, where L is the column length, Up is the bulk particle velocity, Dax is the axial dispersion coefficient, is the ratio of convective transport to diffusion-like or dispersion-like transport. Higher Péclet numbers indicate less backmixing approaching plug flow behavior. The results showed that the axial Péclet number was roughly five times greater in the downer versus the riser.

The solids flux was similar between the two studies. However, it must be noted that there were significant differences between the column inlet and outlet configurations in the two studies; therefore, the differences in the reported Péclet numbers should not be considered absolute or authoritative. Furthermore, Wei and Zhu (1996) showed that dispersed particles in risers have axial Péclet similar to the solids in a downer, indicating near plug flow behavior for the dispersed particles.

Figure 1.8 – Comparison of typical solids axial Péclet number observed in a CFB downer (from Wei et al., 1994) versus a CFB riser (from Harris et al., 2003a)

1.7.4

Motivations

Since axial dispersion in downers has been shown to be very low compared to risers, the corresponding relative impact of the gas-solids separator on the overall solids RTD may be significant, but is unknown. As shown, the solids RTD has not previously been measured in the separator of a CFB downer. Moreover, the impact of solids RTD on the separator performance has also not been investigated. Therefore, there is a need to measure the solids RTD in ICFAR’s cone-shaped gas-solids separator and to determine the impact of the RTD on separator performance. Furthermore, there is a need to integrate

both the gas and solids RTD results into a realistic flow model, which has not been done in a downer gas-solids separator.