Search results

The purpose of this paper is to reduce the overall computational time of aerodynamic shape optimization that involves accurate high-fidelity simulation models.

The proposed approach is based on the surrogate-based optimization paradigm. In particular, multi-fidelity surrogate models are used in the optimization process in place of the computationally expensive high-fidelity model. The multi-fidelity surrogate is constructed using physics-based low-fidelity models and a proper correction. This work introduces a novel correction methodology – referred to as the adaptive response prediction (ARP). The ARP technique corrects the low-fidelity model response, represented by the airfoil pressure distribution, through suitable horizontal and vertical adjustments.

Numerical investigations show the feasibility of solving real-world problems involving optimization of transonic airfoil shapes and accurate computational fluid dynamics simulation models of such surfaces. The results show that the proposed approach outperforms traditional surrogate-based approaches.

The proposed aerodynamic design optimization algorithm is novel and holistic. In particular, the ARP correction technique is original. The algorithm is useful for fast design of aerodynamic surfaces using high-fidelity simulation data in moderately sized search spaces, which is challenging using conventional methods because of excessive computational costs.

Strategies for accelerated multi-objective optimization of aerodynamic surfaces are investigated, including the possibility of exploiting surrogate modeling techniques for computational fluid dynamic (CFD)-driven design speedup of such surfaces. The purpose of this paper is to reduce the overall optimization time.

An algorithmic framework is described that is composed of: a search space reduction, fast surrogate models constructed using variable-fidelity CFD models and co-Kriging, and Pareto front refinement. Numerical case studies are provided demonstrating the feasibility of solving real-world problems involving multi-objective optimization of transonic airfoil shapes and accurate CFD simulation models of such surfaces.

It is possible, through appropriate combination of surrogate modeling techniques and variable-fidelity models, to identify a set of alternative designs representing the best possible trade-offs between conflicting design objectives in a realistic time frame corresponding to a few dozen of high-fidelity CFD simulations of the respective surfaces.

The proposed aerodynamic design optimization algorithmic framework is novel and holistic. It proved useful for fast design of aerodynamic surfaces using high-fidelity simulation data in moderately sized search space, which is extremely challenging when using conventional methods due to the excessive computational cost.

Dual-rotor wind turbines (DRWTs) are a novel type of wind turbines that can capture more power than their single-rotor counterparts. Because their surrounding flow fields are complex, evaluating a DRWT design requires accurate predictive simulations, which incur high computational costs. Currently, there does not exist a design optimization framework for DRWTs. Since the design optimization of DRWTs requires numerous model evaluations, the purpose of this paper is to identify computationally efficient design approaches.

Several algorithms are compared for the design optimization of DRWTs. The algorithms vary widely in approaches and include a direct derivative-free method, as well as three surrogate-based optimization methods, two approximation-based approaches and one variable-fidelity approach with coarse discretization low-fidelity models.

The proposed variable-fidelity method required significantly lower computational cost than the derivative-free and approximation-based methods. Large computational savings come from using the time-consuming high-fidelity simulations sparingly and performing the majority of the design space search using the fast variable-fidelity models.

Due the complex simulations and the large number of designable parameters, the design of DRWTs require the use of numerical optimization algorithms. This work presents a novel and efficient design optimization framework for DRWTs using computationally intensive simulations and variable-fidelity optimization techniques.

Extreme support vector regression (ESVR) has been widely used in the design, analysis and optimization of engineering systems of its fast training speed and good computational ability. However, the ESVR model is only able to utilize one-fidelity information of engineering system. To solve this issue, this paper extends extreme support vector regression (ESVR) to a multi-fidelity surrogate (MFS) model which can make use of a few expensive but higher-fidelity (HF) samples and a lot of inaccurate but cheap low-fidelity (LF) samples, named ESVR-MFS.

In the ESVR-MFS model, a kernel matrix is designed to evaluate the relationship between the HF and LF samples. The root mean square error of HF samples is used as the training error metric, and the optimal hyper-parameters of the kernel matrix are obtained through a heuristic algorithm.

A number of numerical problems and three engineering problems are used to compare the ESVR-MFS model with the single-fidelity ESVR model and two benchmark MFS models. The results show that the ESVR-MFS model exhibits competitive performance in both numerical cases and practical cases tested in this work.

The proposed approach exhibits great capability for practical multi-fidelity engineering design problems.

A MFS model is proposed based on ESVR, which can make full use of the advantages of both HF data and LF data to achieve optimal results at same or lower cost.

The purpose of this paper is to propose an optimal predictive model for the short-term forecast of real-time non-stationary machine variables by combining time series prediction with adaptive algorithms to minimize the error and to improve the prediction accuracy.

The proposed model is applied for prediction of speed and controller set point of three-phase induction motor operating on closed loop speed control with AC drive and PI controller. At Stage 1, the trend of the machine variables has been extracted and added to auto-regressive moving average (ARMA) time series prediction. ARMA prediction has been carried out using different combinations of AR and MA methods in order to make prediction with less Mean Squared Error (MSE).

The prediction error indicates the inadequacy of the model to estimate the data characteristics, which has been resolved at the subsequent stage by cascading an adaptive least mean square finite impulse response filter to the time series model. The adaptive filter receives the predicted output including training data and iteratively adjusts its coefficients for zero error convergence.

The componentized data prediction based on time series and cascade adaptive filter algorithm decomposes the non-stationary data characteristics for predictive maintenance. Evaluation of the model with different combination of time series algorithms and parameter settings of adaptive filter has been carried out to illustrate the performance of the prediction model. This prediction accuracy is compared with existing linear adaptive filter prediction using MSE as comparison index. The wide margin in the MSE values substantiates the prediction efficiency of the proposed model for machine data.

This model predicts the dynamic machine data with component decomposition at high accuracy, which enables to interpret the system response under dynamic conditions efficiently.

The conventional design method relies on a priori knowledge, which limits the rapid and efficient development of electronic packaging structures. The purpose of this study is to propose a hybrid method of data-driven inverse design, which couples adaptive surrogate model technology with optimization algorithm to to enable an efficient and accurate inverse design of electronic packaging structures.

The multisurrogate accumulative local error-based ensemble forward prediction model is proposed to predict the performance properties of the packaging structure. As the forward prediction model is adaptive, it can identify respond to sensitive regions of design space and sample more design points in those regions, getting the trade-off between accuracy and computation resources. In addition, the forward prediction model uses the average ensemble method to mitigate the accuracy degradation caused by poor individual surrogate performance. The Particle Swarm Optimization algorithm is then coupled with the forward prediction model for the inverse design of the electronic packaging structure.

Benchmark testing demonstrated the superior approximate performance of the proposed ensemble model. Two engineering cases have shown that using the proposed method for inverse design has significant computational savings while ensuring design accuracy. In addition, the proposed method is capable of outputting multiple structure parameters according to the expected performance and can design the packaging structure based on its extreme performance.

Because of its data-driven nature, the inverse design method proposed also has potential applications in other scientific fields related to optimization and inverse design.

Aiming to address the forecasting dilemma of seasonal air quality, the authors design the novel self-adaptive seasonal adjustment factor to extract the seasonal fluctuation information about the air quality index. Based on the novel self-adaptive seasonal adjustment factor, the novel seasonal grey forecasting models are established to predict the air quality in China.

This paper constructs a novel self-adaptive seasonal adjustment factor for quantifying the seasonal difference information of air quality. The novel self-adaptive seasonal adjustment factor reflects the periodic fluctuations of air quality. Therefore, it is employed to optimize the data generation of three conventional grey models, consisting of the GM(1,1) model, the discrete grey model and the fractional-order grey model. Then three novel self-adaptive seasonal grey forecasting models, including the self-adaptive seasonal GM(1,1) model (SAGM(1,1)), the self-adaptive seasonal discrete grey model (SADGM(1,1)) and the self-adaptive seasonal fractional-order grey model (SAFGM(1,1)), are put forward for prognosticating the air quality of all provinces in China .

The experiment results confirm that the novel self-adaptive seasonal adjustment factors promote the precision of the conventional grey models remarkably. Simultaneously, compared with three non-seasonal grey forecasting models and the SARIMA model, the performance of self-adaptive seasonal grey forecasting models is outstanding, which indicates that they capture the seasonal changes of air quality more efficiently.

Since air quality is affected by various factors, subsequent research may consider including meteorological conditions, pollutant emissions and other factors to perfect the self-adaptive seasonal grey models.

Given the problematic air pollution situation in China, timely and accurate air quality forecasting technology is exceptionally crucial for mitigating their adverse effects on the environment and human health. The paper proposes three self-adaptive seasonal grey forecasting models to forecast the air quality index of all provinces in China, which improves the adaptability of conventional grey models and provides more efficient prediction tools for air quality.

The self-adaptive seasonal adjustment factors are constructed to characterize the seasonal fluctuations of air quality index. Three novel self-adaptive seasonal grey forecasting models are established for prognosticating the air quality of all provinces in China. The robustness of the proposed grey models is reinforced by integrating the seasonal irregularity. The proposed methods acquire better forecasting precisions compared with the non-seasonal grey models and the SARIMA model.

The aims of this study are to examine affective responses of university students when viewing their own predictive learning analytics (PLA) dashboards, and to analyse how those responses are perceived to affect their self-regulated learning behaviour.

A total of 42 Northern Irish students were shown their own predicted status of academic achievement on a dashboard. A list of emotions along with definitions was provided and the respondents were instructed to verbalise them during the experience. Post-hoc walk-through conversations with participants further clarified their responses. Content analysis methods were used to categorise response patterns.

There is a significant variation in ways students respond to the predictions: they were curious and motivated, comforted and sceptical, confused and fearful and not interested and doubting the accuracy of predictions. The authors show that not all PLA-triggered affective states motivate students to act in desirable and productive ways.

This small-scale exploratory study was conducted in one higher education institution with a relatively small sample of students in one discipline. In addition to the many different categories of students included in the study, specific efforts were made to include “at-risk” students. However, none responded. A larger sample from a multi-disciplinary background that includes those who are categorised as “at-risk” could further enhance the understanding.

The authors provide mixed evidence for students' openness to learn from predictive learning analytics scores. The implications of our study are not straightforward, except to proceed with caution, valuing benefits while ensuring that students' emotional well-being is protected through a mindful implementation of PLA systems.

Understanding students' affect responses contributes to the quality of student support in higher education institutions. In the current era on online learning and increasing adaptation to living and learning online, the findings allow for the development of appropriate strategies for implementing affect-aware predictive learning analytics (PLA) systems.

The current study is unique in its research context, and in its examination of immediate affective states experienced by students who viewed their predicted scores, based on their own dynamic learning data, in their home institution. It brings out the complexities involved in implementing student-facing PLA dashboards in higher education institutions.

Applies a computer model GAIA (Groups of Adaptive Inferencing Agents) to simulate the lifecycle of artificial groups directed by agendas which specify varying strategies for collective problem solving. Within GAIA, groups of artificial agents dynamically learn and interact by proposing, combining and testing inductive hypotheses in the form of genetic building blocks. Agents share and combine building block solutions to evolve decision trees to respond to environmental inputs. Effects of agendas which emphasize stages of conservative and liberal problem solving strategies over a group’s lifecycle were simulated. Conservative strategies emphasize consensus and collective memories within groups. Liberal strategies emphasize challenges to collective memory and individual agent predictions. Agendas which vary from conservative to liberal resulted in the poor group solutions. Significantly better group solutions were produced by an agenda varying from liberal to conservative and back to liberal (L‐C‐L). The L‐C‐L agenda focuses on critical evaluation and rewards for individual contribution in the beginning and ending lifecycle stages and provides a middle stage of collective exploration.

The purpose of this paper is to document an efficient and accurate approach to generate aerodynamic tables using computational fluid dynamics. This is demonstrated in the context of a concept transport aircraft model.

Two designs of experiment algorithms in combination with surrogate modelling are investigated. An adaptive algorithm is compared to an industry-standard algorithm used as a benchmark. Numerical experiments are obtained solving the Reynolds-averaged Navier–Stokes equations on a large computational grid.

This study demonstrates that a surrogate model built upon an adaptive design of experiments strategy achieves a higher prediction capability than that built upon a traditional strategy. This is quantified in terms of the sum of the squared error between the surrogate model predictions and the computational fluid dynamics results. The error metric is reduced by about one order of magnitude compared to the traditional approach.

This work lays the ground to obtain more realistic aerodynamic predictions earlier in the aircraft design process at manageable costs, improving the design solution and reducing risks. This may be equally applied in the analysis of other complex and non-linear engineering phenomena.

This work explores the potential benefits of an adaptive design of experiment algorithm within a prototype working environment, whereby the maximum number of experiments is limited and a large parameter space is investigated.