Search

Nowadays, the reliability of wind turbine gearbox (WTG) is one of the biggest concerns remaining in the wind industry. WTGs have been prone to early failure rather than any mechanical part of modern wind turbines and are subjected to torsional and non-torsional loads which are not yet properly assessed. They are reported to fail within three to seven years, as opposed to the expected twenty years of operation. Their downtime and maintenance process is the most costly of any failure of subassembly of WTs. To improve the reliability of WTGs, a novel Probabilistic MultiBody Dynamic Analysis (PMBDA) framework has been developed and implemented. Gears and Gear systems, like any other mechanical system, have design parameter and loading uncertainties emanating from inherent randomness, measurement and manufacturing errors. However, state-of-the-art models of WTGs and wind turbine drive trains consider deterministic loading and design parameters. This research addresses the early life failure of gearboxes of wind turbines and inquires if a novel approach termed as PMBDA will have an impact on enhancing the reliability of gear systems of wind turbines. An implementation framework that amalgamates the two broad disciplines, probabilistic analysis and multibody dynamic analysis, has been developed, several loading, assembly and design parameters are considered as random variables and their uncertainty is quantified. Using multibody dynamics software called MSC.ADAMS, contact based rigid multibody NASA test-rig gear pairs, a rigid-body and flexible WTG’s high-speed-parallel-helical-stage gear pairs and a flexible WTG’s compound-planetary-helical-stage models, with variable loading, assembly and design parameters, were developed. An advanced fast probability integration methods and standard/advanced sampling techniques were implemented to perform a probabilistic analysis of tooth dynamic force, dynamic factor, root bending stress, tooth surface compressive stress and fatigue life of gears. Probabilistic sensitivities of the performance functions to several random variables were also determined. In addition to revealing system reliability or under-performance through probability of failure (Pf), PMBDA approach also helps designers to consider certain variables critically through probabilistic sensitivity results.

Prognostics is essential in decision making to prevent cost and downtime. Prognostics is especially useful in variable loading conditions, such as loads in wind turbine gearboxes. Literature suggests that accurate prediction of remaining useful life will improve reliability and reduce maintenance cost and that there is a need for a comprehensive framework for quantifying uncertainties in prognostics. The research question of this study is “can probabilistic analysis quantify uncertainties in prognostics?” The specific aims of the research are to construct a generic probabilistic prognostics framework, to estimate the remaining useful life of structures under variable amplitude loading, and to estimate the remaining useful life of cracked gears under variable loading. A computational framework for probabilistic prognostics was developed to perform rainflow counting, dynamic analysis of a gearbox, finite element analysis of a gear tooth, and estimation of remaining useful life. The framework analyzes uncertainties related to state, loading, and material properties. The remaining useful life of structures under variable loading was analyzed by quantifying material and loading uncertainties and employing cumulative fatigue damage. The complete remaining useful life trajectories of structures under variable loading were determined. The remaining useful life estimation of a gear to predict crack growth in the gear tooth was also analyzed. The probability of failure at any loading cycle was quantified. The proposed probabilistic prognostics framework could quantify uncertainties in remaining useful life prediction for fault mitigation. Hence, it can be used as a helpful tool to prepare maintenance schedules in advance to reduce cost and avoid downtime.

Sketches can be classified in different ways, such as personal, shared, persuasive, and handover sketches. Depending on each classification, their level of ambiguity also varies. Sketches are used in many applications, including conceptual design, eliciting user preferences, shape retrieval, and sketch-based modeling. There is a need to quantify the uncertainty in sketches in the mapping of sketches to solid models in sketch-based modeling, in eliciting user preferences, and in tuning the level of uncertainty in sketches at the conceptual design stage. This motivated the formulation of the research question: can uncertainty in communication with an engineering design sketch be quantified? To answer the research question, three specific aims were developed, namely, (1) determine the ranking of the features in a sketch, (2) determine the probability of importance of the features in a sketch, and (3) quantify the uncertainty of a sketch using Shannon’s normalized entropy. This research develops a framework for quantifying uncertainty in communication with an engineering design sketches. The framework is capable of identifying and ranking the features in the sketch, determining their probability of importance, and finally quantifying the level of uncertainty in the sketch. Ranking the features of a sketch is performed by a hierarchical approach, whereas the probability of importance is determined by assessing the probability of likeliness using a shape matching approach followed by a probability transformation. Quantification of uncertainty, on the other hand, was accomplished by normalizing the entropy of the sketch with its maximum possible value. Results on uncertainty quantification of several engineering design sketches are presented in order to illustrate the use of the framework. The framework eliminates the need for expert input in the assessment of uncertainty in sketches, and, hence, can be used by design practitioners with limited experience.