(posted with map and spreadsheet this time) A number of friends have asked what think about MH370 as a pilot and Ive been thinking about this a bit. This is lengthy, but for good reasons. Here is why I think the search for MH370 is probably being conducted in the wrong area due to serious mistakes and bad assumptions used in the satellite ping analysis. The highest probability scenario is that MH370 crashed in the zone depicted on the attached map. That zone is about 50 to 100 miles wide by 500 miles long between the Maldives and the east coast of Africa, near the area of Somalian piracy. This zone is along a segment of a straight line between the position of MH370 when it turned westerly and a point near Mogadishu, Somalia. That line first passes directly over Palau Langkawi the probable emergency airport with a 13,500 foot sea level runway, best able to handle a fully fueled and loaded 777-300 Heavy aircraft. That line continues to the point in the Strait of Malacca where radar contact was lost if the plan had descended to 3,000 ot 5,000 feet to attempt a landing but continued at that altitude it would quickly slip below radar coverage due to the curvature of the earth. That line then crosses over the Maldives at Kuda Hudavoo located just south of Male, where multiple fisherman and other people reported seeing a Malaysian Airlines jet at very low altitude at the approximate time it would have reached this point. If we Ignore the erroneous directional information based on erroneous ping analysis, but we use the same flight duration estimates based on the cessation of pings, and fuel exhaustion calculations, but we use this directional assumption, the end of flight MH370 would be in the zone depicted zone between the Maldives, off the east coast of Africa. Some allowance should be made for north or south drift, or a late turn due to one engine shutting down before the other. For reasons that are unclear and very frustrating, this high probability area is not being searched. The answers to two key questions bring me to the above conclusions, which frustratingly have been difficult to communicate to anyone who matters: 1. Is there a reasonable, high probability scenario that fits the facts and assumption sets above? Yes. Per 777 pilot Goodfellow MH370 was a “heavy” aircraft with a takeoff weight that would greatly exceed it safe landing weight until significant fuel was burned off or dumped. An underinflated nose wheel could cause overheating of the nose wheel on takeoff, resulting in a smoldering, slow burning fire that could produce a large volume of smoke and toxic fumes in the wheel well that is located directly below the electronics bay that is located directly below the cockpit. Such a fire would not trigger deployment of emergency oxygen masks in the cabin and the crew would be reluctant to turn on and use their cockpit oxygen intended to keep them conscious during depressurization emergencies, not fire survival. Oxygen and fires mix explosively, and pilots are reluctant chance adding oxygen to an aircraft fire. The use of the autopilot to help manage inflight emergencies is common practice amongst seasoned pilots, who program in a series of emergency airports along their route of flight, and appropriate safe minimum altitudes the aircraft could descent to, in order to have the autopilot ready to quickly assist in an emergency. This practice is important in modern aircraft with two man crews instead of three. In the event of an inflight emergency, they switch to the appropriate autopilot program and let the aircraft automatically turn, navigate and descend towards the airport, freeing the pilots to deal with other aspects of managing the emergency. Pilots are trained in any emergency to: first aviate (fly the plane), second navigate (have a plan and fly there) and third communicate (but only after the first two are taken care of and stable.) An aircraft fire could have many sources, and could be electrical, fuel or other. The first response would be to disable all nonessential systems that can be disabled from the cockpit by pulling the main circuit breaker busses. This would turn off the avionics including the voice transmitters and the transponder without rotating the knob on these devices. They would then turn off all the individual breakers and turn the main busses back on. They would attempt to bring up systems one at a time to try to identify the source of the fire. Either damage from a fire below or in the electronics bay, or these emergency actions by the crew, could reasonably account for loss of transponder, loss of voice communications, loss of other systems, etc. without the need to assume nefarious intent. Certain essential equipment cannot be turned off from the cockpit or has its own secondary or backup power sources, and these would continue to function. Assuming the pilots had programmed into the first autopilot emergency program, a minimum descent altitude of 3,000 to 5,000 feet the aircraft would turn towards the programmed course to the emergency destination, and then descend to the minimum safe altitude programmed. This would be done to prevent the aircraft from flying into the ground (or water) while the crew is distracted. The intent would be to allow the plane to reach that emergency airport with as little intervention by the crew as needed, at which point the crew would then attempt to land the aircraft. A nose wheel fire could cause sufficient damage to prevent the electronics from functioning or turning back on, or the smoke and fumes could overwhelm the crew and passengers as they are attempting to manage the problem. In this scenario, without any further crew intervention, unless the pilots had actually programmed in an actual approach sequence and enabled Boeings “auto land” system, which us doubtful, the aircraft would probably continue to fly the same course, overflying the airport at the programmed minimum descent altitude and fly that course and altitude until fuel exhaustion. The plane would not be within in cell coverage range, assuming the passengers were still conscious. This is a high probability scenario that requires no wild speculation, mal-intent, or low probability assumptions. 2. Is the satellite ping analysis reliable? I think it is not reliable. Consider GPS. For reasons discussed further below, satellite based positioning (GPS, or Global Positioning Satellites) require a constellation of about 25 satellites in geostationary orbits transmitting very precisely timed signals and data to GPS receivers. To achieve reasonable accuracy most receivers employ 12 channels so they can receive and analyze signals from up to 12 satellites at once. It takes at least 3 to get any kind of reliable position in 3 dimensional space. Consider radar. Radar transmits a signal then turns off the transmitter and listens to receive a reflection of its signal off an object (target). Dividing the time from transmission to reception in half accounts for the round trip, and using the speed of light, radar can determine the distance between the transmitter antenna and the target. It is critical to understand that the received signal is a reflection of the original transmitted signal, and there no latent delay of that signal before it is returned from the target. The transmitted radar pulses also trigger a transponder in the aircraft telling it to transmit a uniquely assigned code and altitude that are transmitted back separately and linked to that radar target. ACARS system is not radar, does not operate like radar, is subject to distortions, does not operate like GPS nor contain its features. The method the satellite engineers are using to estimate distance the way radar does by assuming that the reply delay when an aircraft is pinged is precisely consistent form ping to ping, and that that delay or latency is known or knowable. Neither is true, and the magnitude of the errors in these assumptions result in errors of distance estimates that are wildly unreliable and very large. Only one point at one ping is actually known, and all the others are projections from that, meaning the errors compound with each projection based on the next ping. In the ACARS system the satellite transmits a ping message essentially saying “are you there.” The aircraft does not reflect this signal and the satellite is not equipped to receive a reflected signal. Instead, the aircraft receiver detects this ping and then sends it to the computer system on the aircraft whose program determines what will happen next. Once the computer processes this ping and does whatever else it is programmed to do, and manages whatever other tasks are burdening the computer, it transmits a response ping back to the satellite essentially saying “yes, I am here.” If the satellite responds to that, the two systems will negotiate how and when the aircraft should send its data. If the satellite does not reply nothing further happens until the next ping. In order to use this exchange of transmissions to estimate distance, two things must be known. How long is the latency between the instant the aircraft receives the signal and the time it transmits its reply; and how consistence is that latency from ping to ping? Radio travels at the speed of light, so very small errors in the estimates of this latency and its consistency can produce huge errors in the distance estimates. How big are those errors? We’re talking about milliseconds equating to hundreds or thousands of kilometers. Per the spreadsheet attached, a latency estimate error of 1 second produces a distance estimate error of 299,792 kilometers. An estimated latency error of 1/10th of a second (100 milliseconds) produces an estimated distance error of 29,379 kilometers. A latency estimate error of 1/100th of a second (10 milliseconds) produces an estimated distance error of 2,938 kilometers. The problem is that if a precise latency is not known and subtracted, more or less time is assumed for the radio signal travel. At the speed of light the signal covers a long distance during this error. Keep in mind that many things are going on in the aircrafts systems that could affect the actual length of time, including multiple inoperative systems from which the ACRS system is attempting to collect data. Does it lock up momentarily, run diagnostics, attempt to reconnect to the non-functioning devices, or does its CPU become overloaded with tasks such that it cannot process and reply as quickly? Considering that only 10 milliseconds produces a 2,938 kilometer error, these details matter greatly. In my opinion, in order to rely on this in any way, a full or double blind test is warranted. The engineers should be given only the actual starting location of several actual aircraft, then nothing more than the latency times for a series of 6 or 7 hourly pings. If they can, with nothing more than they have for MH370, consistently predict the actual flight paths of those aircraft, then I might be convinced. But my IT friends tell me that latency delays in this situation and these kinds of systems probably vary in the 3 to 5 millisecond range under normal circumstances when a system is not at its capacity. That is a pretty big distance error. If used to chart the first two points of a projected route, and then that result is used to predict the third, and so on, the compounded errors not only wildly affect the predicted distance, but also wildly affect the prediction of the direction of the flight. What about using Doppler to estimate the general direction of flight towards or away from the satellite? Can the Doppler shift in the aircraft’s transmitted frequency be measured and be used reliably? I do not think it can be. To illustrate why, let’s assume this system operates at a frequency of 1,000 mhz. NASA, ARRL and a great deal of other research tells us that a radio signal in this frequency range (400 mhz to 2,000 mhz) passing through both the atmosphere and ionosphere will experience a Doppler frequency shift of 1 to 2 hertz per mhz based soley on natural conditions, and subject to many variables having nothing to do with motion. As illustrated on the spreadsheet a 1,000 mhz radio signal will experience an inconsistent and variable Doppler shift up or down from the original frequency of 1,000 to 2,000 Hertz. What is the significance of this? In other words, how large is that shift compared to the shift that would be caused by the motion of an aircraft? If the aircraft were flying directly towards or directly away from the satellite at 450 knots, my calculations show it would produce the maximum possible motion generated Doppler shift and that would be about 772 Hertz. So the natural Doppler shift caused by nature is far larger than the information we are looking for. To measure a shift of 772 Hertz against a background of a possible up or down frequency shifts of 2,000 hertz requires that the engineers know exactly how much Doppler effect they are seeing at each ping is due to motion of the aircraft, and how much is due to these environmental factors. It is also important to know at what angle the radio signal traveled through the atmosphere and ionosphere on its way to the satellite. The environmental Doppler shift and signal bending are both affected by this angle. If the signal was bent, it also traveled a longer distance and therefor took a longer time to travel to the satellite than the straight line. That further complicates the distance estimate. And the Doppler shift is also affected by solar flare conditions at the time, as well as the daily effects of solar radiation energizing the ionosphere when sunlight strikes the particles in the different layers of the ionosphere, changing its density and height of the ionosphere above sea level. To know the angle of the signal the engineers must know where the aircraft is. That is what we don’t know, thus even more errors. All of this is why radar works like it does, and why GPS requires lots of satellites and environmental monitors to make needed corrections in real time, in order for satellites to determine position. We can’t know much about MH370’s distance, direction or speed because we don’t’ know what Doppler shift is caused by natural environmental conditions and what is caused by motion, or what the actual latency delays were. The significance of my questioning the analysis is that this untested, unproven, unreliable method and its highly speculative results have been given greater importance than the observations, known or highly probable eyewitness information, and the most probable real world event scenario to determine where to search. I have no way of knowing if Boeing or Rolls Royce actually did receive real data transmissions that they are not telling us about. If so, they might have a basis other than this faulty ping analysis. But, if this questionable ping analysis really is all we have for ACARS, then the teams should regroup, return to old reliable methods, and search in the zone illustrated on this map that has a higher probability of locating MH370.
Posted on: Sun, 06 Apr 2014 06:06:59 +0000
Trending Topics
Recently Viewed Topics
© 2015