Читать книгу Continuous Emission Monitoring - James A. Jahnke - Страница 82

A sample probe (sometimes called a “stinger”) can be made by merely inserting an open metal tube into the stack or duct, or multipoint probes may be used to sample from several locations when the flue gas is stratified (Shahin 2016). This may be adequate in sampling situations where no particulate matter is present. However, flue gases free of particulate matter do not often occur in those sources that are subject to CEM regulation. An open tube can be easily plugged when particulate matter is present, especially if the flue gas has high moisture content. Also, water may condense and combine with the particulate matter to produce an agglomerated material that can plug the probe more readily. To minimize such problems, a filter can be placed at the end of the probe (Figure 3‐4a–c).

Filters made of sintered stainless steel and porous ceramic materials are commonly used to prevent particles from entering the sample tube. Sintered metal is made by compressing micrometer‐sized metal granules under high pressure and elevated temperatures. The metal fuses and acquires porosity depending on the compression pressure. Sintered stainless steel filters that are capable of filtering out particles of 5 to 50 μm have been used as probe filters. Some systems use filters that exclude particles greater than 1–2 μm in size, but the finer the filter, the more difficult it will be to draw the sample gas through the filter, and pump capacity will need to be increased.

Sintered filters or ceramic filters can become plugged by the particles impacting on and penetrating into the porous material. To minimize plugging, a baffle plate can be attached to the filter to deflect particles from the filter surface (Figure 3‐4b). Particles will then follow streamlines formed around the plate, whereas pollutant gases will still diffuse into the probe. Another way to minimize plugging is to attach a cylindrical sheath around the filter (Figure 3‐4c). Gas will still diffuse into the annular space between the filter and baffle, but most particles will not be able to make the 90° change in the direction to enter the space. If the end of such a sheath is partially closed off with a porous plate to provide for gas diffusion, the external filter probe can undergo a probe calibration check. Calibration gas can be injected into the sheath to flood out the stack gas during the calibration intervals.

In the probe designs shown in Figure 3‐4a–c, the coarse particulate filter is attached at the end of the probe. In another probe design, this filter is mounted externally, in a housing mounted outside the stack (Figure 3‐5).

In filter assemblies mounted outside of the stack, a heater either can be fitted into the assembly or placed around the outside of the filter holder. This allows the hot flue gas sample to remain hot as it is drawn through the filter and passed through to the heated sample line and the remainder of the conditioning system. The principal advantage of this configuration is that the coarse filter can be easily unclamped and inspected. The entire probe assembly does not need to be unbolted and removed from the stack to replace the filter, and if the probe becomes plugged with particulate matter, the plug can be pushed out with a rod. Also, if the probe is mounted at an angle (Figure 3‐5), water or acid condensed in the probe can roll back into the stack.

Figure 3‐4 (a) A simple probe filter. (b) Sintered filter with a baffle plate deflector. (c) Sintered filter with a deflector sheath.

Extractive systems designed with external filter housings can readily undergo a calibration verification at the filter housing. Calibration gases can be injected into the annular space between the filter and the housing to provide a check of system integrity from the filter back to the analyzer. This is difficult to do with in‐stack filters because the filter is constantly subject to the flue gas flow and excessive amounts of calibration gas may be needed to continually flood the area around the filter during a calibration verification.

Figure 3‐5 A course filter assembly mounted outside of the stack.

Other variations of the designs as shown in Figure 3‐5 are also used. One variation uses a coarse filter of 10–50 μm porosity at the probe tip, but incorporates a 1‐μm‐pore‐size external fine filter at the flange assembly. Another variant uses a bellows valve to close off the external filter from the probe to reduce the amount of gas necessary to perform a probe calibration check.

A major problem associated with the probes illustrated in Figures 3‐4a–c and 3‐5 is that the sintered filters still can plug with particulate matter. Another system, designed to minimize this problem, utilizes an inertial filter (Figure 3‐6).

In the inertial filter system, a pump (usually an eductor pump) pulls the flue gas through an internal cylindrical filter. The internal filter can be made of sintered stainless steel or porous ceramic, usually having a pore size of 2–5 μm. As the flue gas moves down the tube, a sample is pulled off from the filter at a direction perpendicular to the gas flow (radially) using another sampling pump. The gas velocity through the tube can vary from 70 to 100 ft/s, whereas the radial velocity of the gas pulled from the filter may be only 0.005 ft/s. Larger particles in the flue gas are swept through the tube because of their inertia in the gas stream and are exhausted back into the stack. Because of the low radial sample velocity, larger particles are also less likely to break from their streamlines and enter the filter. The 70–100 ft/s flow also aids in sweeping the particles off of the filter surface and back into the gas stream. An inertial filter can be incorporated as part of the probe inside the stack or in a system external to the probe, such as the one shown in Figure 3‐7.

Figure 3‐6 The inertial filter.

Conceptually, this system may appear ideal, but actually, submicron sized particles (<1 μm diameter) can follow the radial sample flow and enter the tubular filter. These embedded particles can further assist in the filtering action to reduce the filter pore size and remove particles down to 0.5 μm diameter from the sample stream. However, this also means that the filter can eventually become plugged.

Filter plugging can be a problem with any fully extractive system. Plugging can be minimized by “blowing back” on the filter using high‐pressure gas, plant air, or steam – air at pressures from 60 to 100 pounds per square inch (psi) is blown back through the filter, opposite to the normal direction of gas flow. The blowback can be pulsed by first pressurizing a surge tank and suddenly releasing the pressure to shock the particulate matter out of the pores of the filter. Depending upon particle characteristics and concentration, filters are blown back at periods of 15 minutes to 8 hours for durations of 5–10 s; 15‐minute blowback cycles are common. Care must be taken in the blowback system so that the blowback gas does not cool the probe to the extent that acids or other gases condense.