QTR_4.06
Cruise Performance Monitoring
Download This Article (PDF - 2.1 MB)

CASE STUDY 2:
Investigation of a cruise fuel mileage shortfall

An airline expressed concern to Boeing that its new 737-800/CFM56-7B airplanes equipped with Aviation Partners Boeing (APB ) blended winglets were exhibiting fuel mileage performance more than 2 percent worse than the Boeing database level, while its older, nonwinglet 737-800s (all approximately two years old) displayed fuel mileage performance similar to the database level. The airline, which collects cruise fuel mileage data on an ongoing basis, based its analysis on ACMS-collected cruise fuel mileage data analyzed using the Boeing Airplane Performance Monitoring (APM) program.

In initial discussions between Boeing and the airline, it was explained that the database being used by the airline to represent the 737-800 with blended winglets was based on the original winglet flight test results completed in early 2000. This is the same database used in the Flight Crew Operations Manual, the FMC, and the operational flight planning database. Additional flight tests had led to Boeing’s latest, best assessment of the delivered performance of the winglets, which showed slightly less improvement than the original testing. This revised database, based on several additional flight test programs conducted in 2000 and 2001, includes a different winglet drag increment (relative to a nonwinglet 737-800) and an aeroelastic correction absent in the earlier database.

At Mach 0.79, the difference between the two databases varies from 0.2 percent to 2.3 percent, depending on the exact conditions flown, with the airline’s database predicting a better fuel mileage increment because of the winglets in all cases.

Boeing’s analysis of the data using the revised database concluded that two of the airplanes appeared to display fuel mileage performance about 3 percent below the latest Boeing assessed winglet level — even more than the 2 percent originally suggested by the airline.Upon request, the airline provided Boeing with ACMS data for two of its 737-800 airplanes with blended winglets. Boeing analyzed the data for each airplane using both the operational database and the revised database.

While the data was, on average, about 0.5 percent closer to the newer database level than the operational database level, Boeing’s analysis did not agree with the airline’s analysis.

Boeing’s analysis of the data using the revised database concluded that two of the airplanes appeared to display fuel mileage performance about 3 percent below the latest Boeing-assessed winglet level — even more than the 2 percent originally suggested by the airline.

Further discussions with the airline revealed that it had been using passenger weight allowances of 70 kg per passenger, including carry-on baggage, and 13 kg per checked bag for all of its flights. As of June 1, 2002, the airline changed to the higher passenger weight allowances recommended in the Joint Aviation Requirements – Operations (JAR-OPS) 1. Checked baggage would be weighed whenever possible; otherwise, JAR-OPS 1 checked baggage weight allowances would be used. The average passenger weight allowances are significantly higher than the 70 kg per passenger the airline had been using. Because the data sent to Boeing for the two winglet-equipped airplanes was collected prior to June 2002, it was basedon the lighter weight allowance of 70 kg per passenger. The airline’s analysis was based on data using a combination of the weight allowances.

As the investigation continued, the airline sent additional data to Boeing for the same two winglet equipped airplanes — but only for conditions recorded after June 1, 2002, based on the higher JAR-OPS weight allowances. The airline also included data for one more winglet-equipped airplane, as well as for three nonwinglet airplanes. The data for the two winglet-equipped airplanes showed an immediate fuel mileage improvement of about 2.4 percent for each airplane, based on analyzing only the data from JAR-OPS weight allowances collected after June 1, 2002. This result led quickly to the belief that the previous 70 kg per passenger weight allowance was too light.

Although both the fuel mileage and thrust required changed significantly between data based on 70 kg per passenger and data based on JAR-OPS passenger weight allowances, the thrust-specific fuel consumption (TSFC) hardly changed. Errors in the estimated weight of an airplane present themselves as high or low drag but do not affect the fuel flow (i.e., TSFC) deviations calculated by APM.

Although a significant improvement was observed for both of the winglet-equipped airplanes originally analyzed with data recorded before June 1, 2002, the results for all six airplanes were still not as good as what Boeing experience indicated for this model. These results supported the conclusion that, for this airline, the JAR-OPS passenger weight allowances for scheduled flights more accurately reflect the true weight of the passengers plus carry-on baggage.Further investigation determined that this airline operates its fleet of 737-800s in a mix of both scheduled and holiday charter flights, using the specific JAR-OPS weight allowances called out for each. The data sent to Boeing for the six airplanes included a mixture of data from both these types of flights. The average passenger weight allowance recommended for scheduled service is 84 kg per passenger and 76 kg per passenger for charter service (both are higher than the 70 kg per passenger originally used by the airline). At Boeing’s request, the airline separated all of the post June 1, 2002, data into two groups: charter service and scheduled service. The data for each group was reanalyzed separately (see figs. 1 and 2).

The analysis revealed a significant discrepancy in demonstrated fuel mileage and thrust-required levels between the charter and scheduled flights. If airplane weight is underestimated, perceived airplane performance will be poorer than expected. Weight that is unaccounted for shows up as increased airplane drag and decreased fuel mileage. In this analysis, the TSFC deviations remained consistent between both sets of data, but the thrust-required (drag) deviations increased significantly for the charter flights — a strong indication of unaccounted-for airplane weight.

In this situation, Boeing proposed that the JAR-OPS passenger weight allowances as recommended for holiday charter flights were underestimating the airplane weight for this particular airline’s charter operations. Although the airline was receptive to the possibility that the JAR-OPS passenger weight allowances might be too light for its holiday charter flights, it was not fully convinced. The airline believed that the JAR-OPS weight allowances for scheduled flights could just as easily be incorrect, in which case their airplanes were performing as poorly as the charter flight data indicated.

To determine which weight allowances were correct, the airline and Boeing agreed to collect delivery flight performance data on the airline’s next new airplane delivery, a 737-800 with production blended winglets installed.

The advantages of collecting delivery flight data as opposed to in-service data are:

After collecting cruise performance data on the delivery flight, the airline would continue with its standard in-service data collection on both scheduled and charter flights. Comparing the results from the delivery flight with the results obtained in-service would help determine which JAR-OPS passenger weight allowances gave the airline more accurate airplane gross weights. If the weight allowances were too heavy (the airplane was actually lighter than estimated), then the in-service performance would appear to be better than the delivery flight performance. If the weight allowances were too light (the airplane was heavier than estimated), then the in-service performance would appear to be worse than the delivery flight level.

The airline provided Boeing with the first 10 weeks of ACMS in-service data for the airplane following delivery, separating the data for charter flights and scheduled-service flights. For this analysis, the data was analyzed relative to the most recent 737-800 with winglets database. Although the delivery flight results showed the airplane to be slightly better than the demonstrated database level, the early in-service charter flight results show the airplane with an average perceived fuel mileage 3.3 percent worse than the demonstrated level (see fig. 3). Unaccounted-for weight shows up as airplane drag (thrust required). According to the charter flight data, the airplane experienced a 4.4 percent increase in thrust required on entering service, partially offset by a 0.7 percent drop in engine TSFC, for a 3.6 percent drop in fuel mileage from the delivery flight level. When the same airplane’s scheduled service data for the same time period was analyzed, the fuel mileage was much closer to the delivery flight level. Average in-service fuel mileage for the first 10 weeks of operation deviated from the delivery level by only 0.8 percent (only 0.5 percent below the demonstrated level), which is within the ACMS’s ability to determine fuel mileage over a given time period.

These results supported the conclusion that, for this airline, the JAR-OPS passenger weight allowances for scheduled flights more accurately reflect the true weight of the passengers plus carry-on baggage than the weight allowances recommended for the charter flights.

In addition, both are much more representative than the original 70 kg per passenger the airline had been using. The weight allowances for scheduled flights of 84 kg per passenger produce a more accurate zero fuel weight buildup and a truer representation of the actual performance of the airplane, with or without winglets.

The results of this case study identified a fleetwide airplane weight buildup issue for this particular airline. Boeing suggested that the JAR-OPS holiday charter passenger weight allowances appeared to be too light for this airline’s operations, with unaccounted-for weight showing up as excess airplane thrust required (drag). Using the JAR-OPS-recommended passenger weight allowances for scheduled flights, the fuel mileage performance for its 737-800s — with and without winglets — is close to predicted and reflects Boeing expectations based on numerous flight tests and delivery flight results.


CHARTER SERVICE DATA ONLY
JAR-OPS 1 PASSENGER WEIGHT ALLOWANCES

=
Airplane with APB winglets relative to winglet database

=
Nonwinglet relative to nonwinglet database



SCHEDULED SERVICE DATA ONLY
JAR-OPS 1 PASSENGER WEIGHT ALLOWANCES

=
Airplane with APB winglets relative to winglet database

=
Nonwinglet relative to nonwinglet database



PERFORMANCE SUMMARY
DELIVERY FLIGHT VERSUS IN-SERVICE
(ALL DEVIATIONS ARE RELATIVE TO THE WINGLET DATABASE)

=
Delivery flight

=
Post delivery charter service

=
Post delivery schedule service




Download This Article (PDF - 2.1 MB)