Analysis of Engine Vibration Data using an Alternative Vibration Sensor Position and Acquisition Equipment

Engine vibrations are among the most important parameters used to evaluate engine operating service limits. Several case studies showed that these vibration levels could be affected by the malfunction of the internal acceleration sensors, causing a false positive error. If vibration levels overcome the ones described by the Original Equipment Manufacturer (OEM) the engine must be dismounted from the aircraft, causing delays in the normal operating modes of the airline company. Lufthansa Technik AG (LHT) is interested in solving the problem. One of the first approaches is to use an external sensor to validate the information obtained from the internal ones. The goal of the project is to find correlation curves between the different sensors and to use the data produced by the external one as a reference for the main signals. The validation is performed in the test cell of LHT and this is the first thesis that deals with this problem and it is used as a bridge to provide LHT with the missing know-how about the topic. After an initial literature research, it has been proven that the easiest way to obtain a better understanding of the dynamical behavior of the engine was to perform Experimental Modal Analysis (EMA), in the form of an impact hammer test. LHT doesn’t possess the required equipment to perform such tests and for this reason, an external engineer was hired and asked to find the correlation between the sensors. Up to now, only one engine has been tested and this thesis focuses on the results related to this day of testing with the main emphasis on the impact hammer test. Three tests were performed on the engine: the first test is a vibration survey, consisting of a run down and a run up, that has been performed to analyze the response of the sensors in normal operating conditions: in this test, the engine rotational speed is varied following a ramp profile. The second and third tests are impact hammer tests performed in the test cell of LHT and the preparing hall of the same building. The data from the two hammer tests have been analyzed and compared looking for similarities and changes in the results. The data obtained from the vibration survey showed that a visible analogy is present between the shapes of the plots derived from the different sensors. Results have also revealed that the actual response of the system is strongly dependent on the speed profile used during the phases of run up or run down. The data obtained from both the hammer impact tests have shown that no significant resonant frequencies are present in the frequency range of interest of the engine. Moreover, the comparison of the two impact tests showed that it is possible to use data from the test done in the preparing hall as representative of the whole system. This thesis has been especially focused on the analysis of the results obtained from the impact tests, identifying the main experimental mistakes and suggesting several signal processing techniques to perform the proper evaluation of the data. The comparison of the vibration survey data has proved that a correlation is present. To overcome the dependence on the speed profile, a new test has been proposed. This new test is based on a step profile of the rotational speed: the length of each step is chosen to be long enough so that all system transient responses can reach a steady level condition. This test will allow creating characteristic curves between the steady-state response of all the sensors.