Aiming at the failure of the tail-wing deicing system in the process of power-on inspection of a certain type of aircraft, the failure phenomena were checked on the aircraft.
Based on the fault analysis method, the causes of the failure were analyzed, the accurate fault location was obtained, the finished product failure was solved, and the safety and quality of aircraft production were effectively guaranteed.
In the course of flight, with the increase of flight altitude, the external humidity of the aircraft increases gradually, and the temperature decreases gradually, reaching below the freezing point in a certain range, which will lead to the phenomenon of icing outside the aircraft. The icing on the wing and tail will increase the frictional force of the airflow on the wing surface, hinder the airflow and reduce the airplane. Lift, which will seriously affect the safety of aircraft flight. According to statistics, during 1967-2015, the major flight safety accidents caused by aircraft icing accounted for 14.9% of the total flight accidents worldwide, which shows that the anti-icing system plays an extremely important role in the safe flight of aircraft. In the process of power-on inspection of a certain type of aircraft, it is found that the position switching can not be realized when the function switching of the left and right sides of the aircraft is simulated under icing condition. In view of the extreme importance of the anti-icing system, it is necessary to carefully study the mechanism of the failure and effective solutions. During the power-on inspection of a certain type of aircraft, it was found that the tail-wing deicing controller could not switch normally during the analog operation, and the abnormal operation of the optocoupler could lead to abnormal left-right switching signal of the product, which resulted in wrong judgment of CPLD programmable logic device and DSP digital signal processor, and abnormal switching of the distribution box. According to the working principle of the system, we can get its fault tree as shown in Figure 1. There are five failure modes in fault tree: A: Deicing Controller Fault B: Distribution Box Fault C: Electric Heating Component Fault D: External Fault E: Software Fault. There are four modes causing A fault: F: monitoring circuit fault, X1: interface circuit fault, G: communication fault, X2: deicing start signal abnormal. Among them, F corresponds to two failure modes: X11: short-circuit monitoring circuit fault, X12: phase-out monitoring circuit fault, and G corresponds to two failure modes: X13: CAN communication fault and X14: 429 communication fault. There are three modes causing B fault: H: sensor circuit fault, X3: main switch fault of distribution box, X4: regional switch fault of distribution box. H corresponds to two failure modes: X15: short-circuit sensor fault and X16: phase-missing sensor fault. There are two modes causing C failure: X5: short circuit or open circuit of heating module, and X6: local resistance of heating module increases. There are two modes of causing D fault: X7: connecting cable fault, thermostatic element X8: power supply fault. There are two modes of causing E failure: X9: software logic defects, X10: software coding errors. In the process of troubleshooting, the performance of the system is normal after replacing the distribution box. Therefore, B, C, D and E fault modes can be eliminated, and the fault can be located as A fault mode. The distribution box returned to the laboratory to detect that there was no trigger signal for its switching function.
After further inspection, it was found that the input isolation optocoupler had no output. Then the optocoupler was returned to the manufacturer for testing and analysis. According to the conclusion of quality inspection, it was found that the bonding wire of the light-emitting chip at the input end of channel 4, 5 and 6 of the optocoupler had been broken, and the high current at the input end burned out obviously. Traces. In the sub-system connected with it, the output of the other controller’s 1, 2 and 3 channels are not damaged (output terminal finite current resistance 3K), so the high current signal will not occur due to the switch-on of the interface circuit. The fault of high current input can only occur between the Backup 1, Switch 1, Master 1 signals and the 28VGND signals. Potential difference (forward or reverse overvoltage) occurs as shown in Figure 2. The photocoupler input terminal is light emitting diode. When the forward overvoltage signal is input, the peak current of the forward conduction of the diode should be less than 40 mA, otherwise it may cause large current damage.
When the reverse overvoltage signal is input, the reverse voltage tolerance value of the diode is 5V. When the voltage value exceeds 5V, the light emitting diode will be broken down and the instantaneous resistance value is infinite.
Small, but also can cause large current damage.
Two deicing controllers were tested, and the input of the optocoupler was monitored.
When the power was started, it was found that the reverse voltage of 7, 9 and 11 legs exceeded 5V voltage value. The voltage signal is generated during the charging process of the power supply circuit. After the controller is powered on, the capacitor is charged instantaneously, and the negative charge accumulates rapidly to the negative end of the capacitor, which results in a positive voltage at the 28VGND signal end, while the backup 1, Switch 1 and Master 1 pins of the optocoupler input end are blocked by the inductance of the filter circuit, and the potential remains unchanged, resulting in a reverse voltage at the optocoupler input end. When the voltage is repeatedly loaded on the optocoupler input, it can cause the optocoupler input to be reversely broken down. Because of the infinite current resistance of the optocoupler input, the phenomenon of over-current damage occurs. According to the analysis, due to the over-current of optocoupler, the left-right switching signal of the product is abnormal, which results in the wrong judgment of CPLD programmable logic device and DSP digital signal processor, and makes the distribution box abnormal switching. To this end, the protection of the optocoupler is realized by adding a current limiting resistance of 1K at the input end of the optocoupler. After improvement, a single deicing controller is tested to monitor the deicing start-up and on-off process for many times, and the product is fault-free. The improved product is installed for testing, and the failure phenomenon has not recurred. Therefore, it can be determined that the circuit improvement can effectively avoid the recurrence of the system failure phenomenon, and also improve the reliability and safety of the product.