Exhaust Emissions from In-Use General Aviation Aircraft (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

74 Exhaust Emissions from In-Use General Aviation Aircraft wall is less than the gas, which is often the case when engine exhaust is sampled. Particle loss due to turbulent diffusion is the largest size-dependent loss where the loss is highest for the smallest particles. Other size-dependent losses are due to inertia, gravity, bending of the tubing, and elec- trostatics. Particle sizes of combustion-generated particles as they exit the device are generally less than 500nm; therefore, losses due to thermophoresis and diffusion dominate. In recognition of the importance of particle loss in sampling systems, in 2008 a spreadsheet model was developed at United Technologies Research Center (UTRC) to predict particle trans- port as a function of particle size. This model could then be used to assess the performance of various sample line configurations (Liscinsky et al. 2010). The resulting Excel-based tool assumes steady-state flow and calculates particle losses using standard equations taken from Yook and Pui (Yook and Pui 2005) and Willeke and Baron (Baron and Willeke 2001). Although Baron created a very powerful and widely used spreadsheet tool called Aerocalc (Baron 2001) that contains many of the same particle transport calculations, Aerocalc treats each loss mechanism as a separate calculation. The UTRC tool simplified the analysis of a sample line by integrating the effect of five different particle loss mechanisms over ten different sample line sections. The UTRC tool predicts transport efficiency for particles over a range of sizes, based on characteristics of the flow, the transport line, and ambient conditions. The UTRC model was validated in laboratory testing during development. Subsequently, it has been found that experimental data taken on practical sampling systems at campaigns sponsored by NASA APEX and AFFEX and EPA VARIAnT have agreed better than expected with the modeling predictions. Given that the measurement of particle loss is tedious and prone to error, the use of a predictive tool for line loss has evolved to become a recommended practice. SAE E-31 is develop- ing an Aerospace Information Report, AIR6504, which details the entire theory of line loss in the standard sampling system used to measure nvPM from aircraft engines as described in AIR6241 (SAE International). When published, the tool will expand on the UTRC tool, include losses in the VPR, and account for CPC counting efficiency. Figure I-1 is a schematic of the particle sampling system used during the first campaign (October 2014). The total sample flow rate was 18 SLPM, with a line length of 40ft from the tripod collection probe to the instrument trailer interface. The setup for the second campaign (June 2015) has a similar instrument setup but the third campaign (October 2015) used a longer line length (140ft) from the tripod collection probe to the trailer interface. Figure I-2 shows the losses as a function of particle size predicted by the UTRC line loss tool as a function of loss mechanism. The plots show that the shorter line had higher transport as expected; however, the smallest particles have the highest losses and below 20 nm ~50% of the particles are not transported in the shorter line compared to ~30% in the longer line. Also the losses increase dramatically when the particles are less than 20nm and, in the longer sample line, the losses of 10nm particles are 95%. Given that the measured particle size distributions indicate that most of the particles are less than 20nm, comparison of data among the different sampling systems requires a correction for line loss. Furthermore, to use the particle measurements as input to models of particulate emissions requires a correction for particle number that is at least a factor of 2.

Figure I-2. Predicted Particle Losses for Campaign 1 and 2 (top plot) and 3 (bottom plot). Figure I-1. Layout of the Particle Sampling System for Campaign 1.

76 Many of the aircraft measured for ACRP Project 02-54 engine tests had no fuel flow gauge. Furthermore, aircraft with analog fuel flow gauges as opposed to digital gauges often do not register fuel flow at idle (dial is below the lowest mark). In these cases, other methods must be used to estimate a fuel flow for the engine state of interest. Aircraft engine operating manuals typically display some sort of plot or a combination of plots that allow the pilot to relate engine RPM to fuel flow. These plots typically start at 50% power and above and tend to be for a mixture full-rich setting. For aircraft with constant speed propellers (i.e., variable pitch propellers), the manifold pressure for a given RPM is required to estimate a fuel flow. In these cases, a limited set of engine states, propeller RPM, and manifold pressures were chosen based on the pilot’s operation of the test aircraft and the manual’s descrip- tion of sample operating conditions. For example, many aircraft manuals state the fuel flow for a representative cruise state with 24 inches of manifold pressure and 2400 RPM. None of the operating manuals investigated for this report mention fuel flow at taxi and/or idle. A data point for fuel flow at low power states is important, however, in anchoring the fuel flow estimate. Thus, when available, the manual fuel flows were supplemented with FOCA data for the taxi state, which is a measured value. FOCA defines the taxi state as whatever the operating manual states. The RPM for the taxi state is set at 1000 RPM for most aircraft, based on pilot’s actual use. When no FOCA data was available for the aircraft in question, the closest engine type was chosen, taking into account the maximum fuel flow, the engine horsepower, and the com- pression ratio. In some cases, the engine manual specifies enough operational points that a fuel flow at taxi is not required to anchor a fit of these data points. Combining the data described above produces plots that relate the fractional fuel flow (fuel flow/max fuel flow) to the fractional engine RPM (RPM/max RPM). An exponential fit to this data is found to be more appropriate than multiple polynomial fits because an exponential fit allows fitting the entire RPM space with a single function. This fit then allows for the fuel flow for any given engine RPM to be estimated. A different plot is generated for distinct engine types, with some engine subtypes grouped as in the engine manuals (e.g., Lycoming O-320-A, -E are grouped separately from Lycoming O-320-B, -D). Figure J-1 plots the equations used to estimate fuel flows when no appropriate cockpit data was available. Fractional fuel flows for a great variety of engine types follow a similarly shaped curve when the data is put in these relative terms. 95% confidence limits for the average of these curve fits are shown as the shaded grey region. Uncertainties are greatest at low engine states (0.2 – 0.6 fractional engine RPM). At these fractional engine RPMs, the fractional fuel flow 95% confidence limits are ± 0.05 (or 5%). This method of determining fuel flow can be verified with engines for which fuel gauges are installed [e.g., aircraft with a Lycoming O-360-A4M engine (Figure J-2)]. The measured fuel A P P E N D I X J Estimating Fuel Flows for Piston Engines