Bombardier CL-600-2B16 N823KD

The airplane was turning toward the final approach course about 5 miles northeast of the destination airport when a “Master Warning” light illuminated on the glareshield and, 1 second later, a corresponding red message was displayed on the engine indicating and crew alerting system (EICAS), with an “engine oil” voice advisory. Twenty-three seconds later, while the airplane was about 1,000 ft pressure altitude and 122 kts, on a shallow intercept angle for the final approach course, the crew announced to the airport air traffic control tower, “…lost both engines… emergency… (I’m/um) making an emergency landing.” The tower controller acknowledged the transmission and cleared the airplane to land. Shortly after, a flight crewmember replied, “eh we’re clear to land but we’re not gonna make the runway uh we’ve lost both engines.” The airplane touched down on a highway while in a slight left bank. It then veered right and travelled off the highway. The airplane’s right wing struck a non-frangible highway sign; the airplane then veered further to the right and impacted a concrete sound barrier wall. A postcrash fire ensued and the cabin attendant and two passengers were able to egress through the baggage compartment door in the tail section of the airplane. The two flight crewmembers were fatally injured and one ground occupant sustained a minor injury. Analysis of data from the flight data recorder (FDR) indicated that during the approach both engines began a commanded decrease in power, comparison of this deceleration to prior flights showed that the engine deceleration during the accident flight was consistent with previous flights and not consistent with a fuel cutoff event, combustor blowout, or engine flameout event. About 1 second after reaching the lowest engine core (N2) speeds of 62.8% (No. 1 engine) and 63.3% (No. 2 engine), N2 briefly increased to 65.0% (No. 1 engine) and 64.6% (No. 2 engine) consistent with the throttle command increasing. At that point, N2 rolled back on both engines and decreased to a sub-idle state, and interturbine temperature (ITT) increased for the rest of the recording. This behavior was consistent with both engine compressors operating in an unrecoverable rotating stall. Examination of both engines revealed no evidence of catastrophic internal mechanical failure. Fuel samples from various engine components, fuel supply lines, fuel tanks and the auxiliary power unit (APU) were collected and sent to two separate facilities for evaluation. The sampled fuel was consistent with normal Jet A fuel and no anomalies were noted. Operational testing of each main fuel control (MFC) unit indicated they were typical of an in-service MFC; no anomalies were noted that would have precluded normal operation. Both engines were sent to the manufacturer for further examination and disassembly, and a series of variable geometry (VG) tests were completed to assess the VG actuators’ total travel, actuation pressures, and rotational forces, and the VG system’s OPENED and CLOSED positions and drag torques. The examination revealed the same results for both engines: corrosion was observed in the high-pressure compressor (HPC) case flow path area, with the most significant corrosion found in the VG stage 5 area. Extensive corrosion was observed in the HPC case VG stage 5 stator vane spindle bores. Additionally, the VG stage 5 stator vanes were unable to travel fully (that is, the distance from fully OPENED to fully CLOSED) when tested using the specified maintenance procedures, and higher than normal actuation pressures were required to move the VG hardware through its full range when compared to other engines without corrosion on the HPC spindle bores, with a slower than normal VG system response when tested with pressurized air. This condition can have a significant negative impact on compressor stability during startup, which can lead to hung engine starts. At low power conditions, as was the case at the time of the accident, it can lead to sub-idle rotating stalls. It is likely the corrosion limited the VG hardware travel as the flight crew reduced the power for landing, resulting in near-simultaneous, sub-idle rotating compressor stalls and a subsequent loss of thrust in both engines, which was unrecoverable at the low altitude. Chemical analysis of the corrosion collected from the compressor case and VG system hardware revealed corroded steel and elements commonly found in a sea salt environment. The corrosion buildup likely occurred over time as the airplane was continually exposed to salt air associated with marine climates. Since its manufacture, the airplane was primarily based at airports located in close proximity to the ocean (first with the previous operator based in Barbados, and then with the current operator based in Fort Lauderdale, Florida). Twenty-five days before the accident, a hung start occurred on both of the accident airplane’s engines while the pilots were preparing for taxi. The operator consulted with the engine manufacturer to troubleshoot the issue, using a fault isolation logic flowchart with 27 logic blocks requiring a “YES” or “NO” response. Block 21 of the flowchart required a pressure check of the VG system (titled Maintenance Practice [MP] 68). During the troubleshooting of the hung start events, MP 68 was not performed because the engines were started and no further anomalies were noted, allowing discontinuing of troubleshooting in accordance with the flowchart. With the concurrence of the engine manufacturer, the airplane was returned to service and flew 33 uneventful flights (excluding the accident flight) over the next 25 days, accruing 57 hours of flight time until the accident. According to the engine manufacturer, a hung start may be an indicator of corrosion buildup in the engine and will result in poor engine starting and operating performance. (In addition to the hung starts twenty-five days before the accident, the operator experienced 7 additional hung start events in the previous 10 years.) One way corrosion could have been identified in the engine, and specifically of the VG system components, was through the MP 68 pressure check. However, because this step was so late in the fault isolation hung start guidance, and it was not a required maintenance check, the airplane was returned to service after successful engine start and no other subsequent engine start issues. Thus, the corrosion of the VG system components continued to go undetected and eventually led to the sub-idle compressor stall during the accident flight. As a result of the accident investigation, the engine manufacturer published an updated version of the fault isolation hung start guidance to give precedence to the VG system testing by making it step 2 in the troubleshooting logic tree.

Quoted verbatim from the NTSB record.