21087

۲۱۰۸۷

بهينه سازي بدنه خودرو سواري با استفاده از روش عددي گراديان پايه با تعداد متغير زياد

كارشناسي ارشد

سازه و بدنه خودرو

۱۳۹۴

۱۳۹۷/۸/۳۰

دكتر ابوالفضل خلخالي

خودرو

مدل‌‏‌‌هاي ساده¬سازي شده ابزار مناسبي براي بهينه¬سازي سازه¬ي اصلي خودرو در مراحل اوليه طراحي مي-باشند. در اين پايان‌نامه يك مدل ساده¬سازي شده از بدنه¬ي يك خودرو سواري سدان ارائه مي¬شود؛ كه در آن، اعضاي تير مانند با المان¬هاي تير يك‌بعدي و پنل¬ها با المان¬هاي درشت جايگزين مي¬شوند در اين مطالعه، با توجه به نتايج آناليز حساسيت با نرم‌افزار MSC.Nastan sol 200، اجزاي فوقاني پوسته در طراحي مفهوم جديد انتخاب‌شده و به‌منظور بررسي دقت مدل ساده‌شده، نتايج حاصل از مدل كامل اجزاي محدود FFM خودرو نسبت به سه معيار سفتي خمشي ، پيچشي و وزن با مدل ساده‌شده مقايسه شده است. با استفاده از الگوريتم بهينه‌سازي ازدحام ذرات (MOPSO) سه تابع هدف وزن، سفتي خمشي و سفتي پيچشي به‌عنوان توابع هدف و 105 متغيير طراحي شامل ارتفاع، عرض و ضخامت‌هاي تيرهاي ساده‌سازي شده انتخاب و به بهينه‌سازي پرداخته‌شده است و درنهايت با استفاده از روش‌هاي (VIKOR) و (TOPSIS) مدل برتر انتخاب‌شده است. نتايج حاصل از 2 روش ذكرشده نشان مي‌دهد كه مدل به‌دست‌آمده از روش TOPSIS نه‌تنها مي‌تواند وزن را در حدود 40٪ كاهش دهد، همچنين مي‌تواند موجب افزايش سفتي پيچشي و خمشي در اين مدل شود.

1398/07/07

Large Scale Optimization of Automotive Body Structure using Numerical Gradient Base Methods

9/29/2019 12:00:00 AM

Automotive design engineers face the challenging problem of developing products in highly competitive markets. In this regard, using conceptual models in the first step of automotive development seems so necessary. A simplified model of body-in-prime (BIP) can solve this difficulty by reducing the number of design variables.In this study, to achieve the simplified model of BIP upper shell components have selected not only due to modified components in new concept design but also those have the high impact on Bending and Torsional stiffness according to sensitivity analysis with aim of MSC.Nastan sol 200 and combined with multi-objective optimization(MOP); In order to verify the accuracy of the simplified model, the results of full finite element (FFM) model were compared with bending and torsional stiffness calculated by the simplified model. The MOP is solved using Multi-Objective Particle Swarm Optimization (MOPSO) which weight, bending and torsional stiffness were considered as objective functions and depth, width, and thickness of the simplified beams were considered as the design variables. The solutions are combined for non-dominated sorting to obtain the non-dominated individuals of 3-objective optimization. 105 optimum points are extracted from the multi-objective optimization process. Finally, VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) method and Technique for Ordering Preferences by Similarity to Ideal Solution (TOPSIS) method are employed to find trade-off optimum design points out of all non-dominated individuals compromising all three objective functions together. The stiffness results demonstrate high accuracy of the concept models with the original model in bending and torsional stiffness prediction. Furthermore, according to Nastran sol 200, the B pillar had influenced the Bending more significantly. On the other hand, Rear compartment Had the highest effect on Torsional. In addition, the results showed that the model obtained by TOPSIS method on MOPSO could not only reduce the weight of simplified components about 40% but also significantly increase could be seen in the BIP stiffness.