A method for measuring Crank Angle resolved Thrust Loads

• Main Author: Thiyagarajan, Janakiraman
• Format: Others
• Language: English
• Published: KTH, Energiteknik 2016
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-191490

This thesis report elucidates the experimental and analytical methods of determining the axial thrust loads of a twin scroll turbocharger mounted on a 6-cylinder diesel engine, tested at Scania CV AB, Sodertalje. Literature survey revealed several methods to experimentally determine the thrust loads but the concept of using strain gauges was nally selected. The key objectives of this thesis are to determine an experimental method to measure axial thrust loads, develop an analytical model to predict axial loads and to perform a root cause analysis. It was noted that axial load peaks are of immense interest here than the average values of the axial loads. The method adopted to measure axial loads consists of a weakened bearing that was capable of withstanding an estimated axial load and at the same time undergo strains that could be sensed by strain gauges. The test operating points consisted of a series of engine speeds and loads. Root cause analysis results suggested that there was a minimal drift of 0.25 N force per 10 degree rise in temperature on the bearings. Also, the main reason for the drift in strain gauge measurements was the plastic deformation of the axial bearing that experienced higher axial loads due to transient testing. A 1D model based on the Control Volume approach including both the pressure and impulse contributions predicting the thrust loads was modelled. The factored axial force measurements during most of the operating conditions were negative (forces acting from the turbine towards the compressor) with the highest value recorded as -90 N at 1400 rpm and 80 % engine load. The highest factored axial force (negative) predicted by the model during the same operating point was -100 N. The di erence between the predicted and measured max load varied between 10 % to 15% among all the operating conditions considered for correlation. It could be concluded that the experimental method established could be used to measure axial loads. Further, future work in terms of non-invasive measurements, CFD methods and usage of fast pressure sensors on compressor side to predict transient conditions would improve the current method. === Denna avhandling rapport belyser de experimentella och analytiska metoder for att bestamma de axiella tryckbelastningar av en Twin Scroll turboaggregat monterat paen 6-cylindrig dieselmotor, testas vid Scania CV AB, Sodertalje. Litteraturstudie visade era metoder for att experimentellt bestamma axialbelastningar men begreppet att anvanda tradtojningsgivare slutligen valts. De viktigaste mlen for denna avhandling ar att bestamma en experimentell metod for att mata axiella tryckbelastningar, utveckla en analytisk modell for att forutsaga axialbelastningar och utfora en orsaksanalys. Det konstaterades att de axiella belastningstoppar ar av stort intresse har an medelvardena for de axiella belastningar. Den metod som for att mata axiella belastningar bestar av en forsvagad lager som kunde motstaen uppskattad axiell belastning och samtidigt genomgar stammar som kan kannas av tojningsgivare. Testdriftpunkter bestod av en serie av motorhastigheter och belastningar. Rotfelsanalys resultat antydde att det fanns en minimal avdrift av 1N kraft per 10 graders temperaturokning palagren. Aven den framsta orsaken till drift i tradtojningsgivare matningar var plastisk deformation av det axiella lagret som haft en hogre axiella belastningar pagrund av transient provning. En 1D modell baserad pakontroll Volym tillvagagangssatt innefattande bade tryck och impuls bidrag forutsaga axialbelastningar modellerades. De vgas axiella kraftmtningar under strre delen av driftsfrhallandena var negativa (krafter som verkar fran turbinen mot kompressorn) med det hogsta vardet registreras som -90 N vid 1400 rpm och 80 % motorbelastning. Den hogsta vagas axialkraften (negativ) forutsags av modellen under samma arbetspunkt var -100 N. Skillnaden mellan forutsagelsen och matning befanns variera mellan 10 % 15 % bland alla driftfrhallanden vagas for korrelation. Det kan konstateras att den experimentella metod som faststallts kan anvandas for att mata axiella belastningar. Vidare skulle det framtida arbetet i termer av icke-invasiva matningar, CFD-modeller och anvandning av snabba trycksensorer pakompressorsidan for att forutsaga transienta forhallanden forbattra den nuvarande metoden.