Design Of Computer Experiments Research Articles

Stochastic Computational Methods and Experiment Design

Stochastic Computational Methods and Experiment Design

Optimum design of stenter machine hot air supply chamber by coupling CFD & DOE

The efficiency of the textile hot air-drying process fundamentally requires the uniformity of the airflow that impinges on the textile fabric. Optimization of a stenter machine’s hot air supply chamber shape was conducted in this paper, based on the actual real-size model and considering both the outlet flow uniformity and the standard velocity deviation as responses. The purpose is to improve the airflow uniformity at the chamber outlet, and modifications applied to the baseline must be applicable to both old and next-generation stenter machines. Coupled Computational Fluid Dynamics and Design of Experiment (CFD-DoE) optimization on the stenter machine chamber geometry was carried out. The CFD model was validated by measuring the chamber outlet velocity using a 2D velocity scanner. After choosing seven factors based on mechanical manufacturing and fluid dynamics theories, a main effect screening design was conducted to highlight relevant factors that influence the outlet airflow distribution. By fixing the best categorical factors, a central composite design (CCD) was applied using three other continuous factors to generate a quadratic prediction model that could be used to simultaneously maximize and minimize the uniformity index (UI) and the standard deviation (SD), respectively. With the embedded JMP SAS desirability function, the optimal value was calculated based on a statistical approach. Those values were used to design an optimal stenter air supply chamber that has been simulated with Fluent CFD code to validate the statistical prediction model. Finally, optimized geometry was obtained and experimentally validated. Results show that this new chamber improved by 16.48% for the uniformity index and 38.07% for the standard deviation.

A class of space‐filling designs with low‐dimensional stratification and column orthogonality

AbstractStrong orthogonal arrays are suitable designs for computer experiments because of stratification in low‐dimensional projections. However, strong orthogonal arrays may be very expensive for a moderate number of factors. In this article, we develop a method for constructing space‐filling designs with more economical run sizes. These designs are not only column‐orthogonal but also enjoy a large proportion of low‐dimensional stratification properties that strong orthogonal arrays ought to have. Moreover, a class of proposed designs can be 3‐orthogonal. In addition, some theoretical results on regular fractional factorial designs are obtained as a by‐product.

Construction of Column-Orthogonal Designs with Two-Dimensional Stratifications

For the design of computer experiments, column orthogonality and space-filling are two desirable properties. In this paper, we develop methods for constructing a new class of column-orthogonal designs (ODs) with two-dimensional stratifications on finer grids, including orthogonal Latin hypercube designs (OLHDs) as special cases. In addition to being column-orthogonal, these designs have good space-filling properties in two dimensions. The resulting designs achieve stratifications on s2×s or s×s2 grids, and most column pairs satisfy stratifications on s2×s2 grids. Moreover, many column pairs can achieve stratifications on s4×s2 and s2×s4 grids. Furthermore, the obtained space-filling ODs can have s6 levels, s4 levels, and mixed levels, as required for different needs.

Design and Analysis of Computer Experiments with both Numeral and Distributional Inputs

Nowadays stochastic computer simulations with both numeral and distributional inputs are widely used to mimic complex systems which contain a great deal of uncertainty. This article studies the design and analysis issues of such computer experiments. First, we provide preliminary results concerning the Wasserstein distance in probability measure spaces. To handle the product space of the Euclidean space and the probability measure space, we prove that, through the mapping from a point in the Euclidean space to the mass probability measure at this point, the Euclidean space can be isomorphic to the subset of the probability measure space, which consists of all the mass measures, with respect to the Wasserstein distance. Therefore, the product space can be viewed as a product probability measure space. We derive formulas of the Wasserstein distance between two components of this product probability measure space. Second, we use the above results to construct Wasserstein distance-based space-filling criteria in the product space of the Euclidean space and the probability measure space. A class of optimal Latin hypercube-type designs in this product space are proposed. Third, we present a Wasserstein distance-based Gaussian process model to analyze data from computer experiments with both numeral and distributional inputs. Numerical examples and real applications to a metro simulation are presented to show the effectiveness of our methods.

Doubly Coupled Designs for Computer Experiments With Both Qualitative and Quantitative Factors

Doubly Coupled Designs for Computer Experiments With Both Qualitative and Quantitative Factors

AN ADAPTIVE STRATEGY FOR SEQUENTIAL DESIGNS OF MULTILEVEL COMPUTER EXPERIMENTS

Investigating uncertainties in computer simulations can be prohibitive in terms of computational costs, since the simulator needs to be run over a large number of input values. Building an emulator, i.e., a statistical surrogate model of the simulator constructed using a design of experiments made of a comparatively small number of evaluations of the forward solver, greatly alleviates the computational burden to carry out such investigations. Nevertheless, this can still be above the computational budget for many studies. Two major approaches have been used to reduce the budget needed to build the emulator: efficient design of experiments, such as sequential designs, and combining training data of different degrees of sophistication in a so-called multifidelity method, or multilevel in case these fidelities are ordered typically for increasing resolutions. We present here a novel method that combines both approaches, the multilevel adaptive sequential design of computer experiments in the framework of Gaussian process (GP) emulators. We make use of reproducing kernel Hilbert spaces as a tool for our GP approximations of the differences between two consecutive levels. This dual strategy allows us to allocate efficiently limited computational resources over simulations of different levels of fidelity and build the GP emulator. The allocation of computational resources is shown to be the solution of a simple optimization problem in a special case where we theoretically prove the validity of our approach. Our proposed method is compared to other existing models of multifidelity Gaussian process emulation. Gains in orders of magnitudes in accuracy or computing budgets are demonstrated in some numerical examples for some settings.

Multi-level emulation of tsunami simulations over Cilacap, South Java, Indonesia

Carrying out a Probabilistic Tsunami Hazard Assessment (PTHA) requires a large number of simulations done at a high resolution. Statistical emulation builds a surrogate to replace the simulator and thus reduces computational costs when propagating uncertainties from the earthquake sources to the tsunami inundations. To reduce further these costs, we propose here to build emulators that exploit multiple levels of resolution and a sequential design of computer experiments. By running a few tsunami simulations at high resolution and many more simulations at lower resolutions we are able to provide realistic assessments whereas, for the same budget, using only the high resolution tsunami simulations do not provide a satisfactory outcome. As a result, PTHA can be considered with higher precision using the highest spatial resolutions, and for impacts over larger regions. We provide an illustration to the city of Cilacap in Indonesia that demonstrates the benefit of our approach.

Effect of T- and Y-Pipes on Core Annular Flow of Newtonian/Non-Newtonian Carreau Fluid Using Computational Fluid Dynamics and Statistical Experimental Design Analysis

Effect of T- and Y-Pipes on Core Annular Flow of Newtonian/Non-Newtonian Carreau Fluid Using Computational Fluid Dynamics and Statistical Experimental Design Analysis

Been There, Done That: How Episodic and Semantic Memory Affects the Language of Authentic and Fictitious Reviews

AbstractThis article suggests a theory-driven approach to address the managerial problem of distinguishing between real and fake reviews. Building on memory research and linguistics, we predict that when recollecting an authentic experience in a product review, people rely to a greater extent on episodic memory. By contrast, when writing a fictitious review, people do not have episodic memory available to them. Therefore, they must rely to a greater extent on semantic memory. We suggest that reliance on these different memory types is reflected in the language used in authentic and fictitious reviews. We develop predictions about five linguistic features characterizing authentic versus fictitious reviews. We test our predictions via a multi-method approach, combining computational linguistics, experimental design, and machine learning. We employ a large-scale experiment to derive a dataset of reviews, as well as two datasets containing reviews from online platforms. We also test whether an algorithm relying on our theory-driven linguistic features is context independent, relative to other benchmark algorithms, and shows better cross-domain performance when tested across datasets. By developing a theory that extends memory and psycholinguistics research to the realm of word of mouth, this work contributes to our understanding of how authentic and fictitious reviews are created.

Sequential design of multi-fidelity computer experiments with effect sparsity

Sequential design of multi-fidelity computer experiments with effect sparsity

Performance enhancement of a vertical axis wind turbine using a slotted deflective flap at the trailing edge

Vertical axis wind turbines (VAWTs) exhibit dynamic stall under operational conditions which causes boundary layer separation from the surface of the blade. This hinders the turbine’s overall performance in which the turbine displays low efficiency, aerodynamic losses, and vibration and noise generation. The aim of this study is to optimize the performance of an H-type Darrieus VAWT by employing a single-slotted deflective flap at the trailing edge of the blade. To optimize the turbine’s performance, Computational Fluid Dynamics (CFD) and Design Of Experiments (DOE) methods are utilized simultaneously for time-saving purposes. The design parameters optimized for the NACA 0018 airfoil are the flap’s gap (Δy), overlap (Δx), and deflection angle (δf). The influence of each parameter on the performance of the VAWT are determined by DOE, which yields the optimum combination of design parameters for this H-rotor VAWT as Δx = 2 mm, Δy = 0.5 mm, and δf = 90 at TSR 3 where two-dimensional (2D) and three-dimensional (3D) CFD simulations yield turbine power increments of 27 % and 19 %, respectively. The flow characteristics of the 2D and 3D turbine models with slotted blades, developed using ANSYS Fluent R2020, demonstrated a delay in boundary layer separation and a reduction in vortices shedding due to the high-pressure flow through the slot, which in return reenergizes the boundary layer. The results from this study indicate an improved design of an H-rotor VAWT.

Global sensitivity analysis of CO2 storage in fractured aquifers via computer experiments

• Accelerated CO 2 storage sensitivity analysis via Computer Experiments as surrogate modes. • Rigorous dynamic models use to model non-reactive CO 2 storage mechanisms, inspired by the Teapot Dome in Wyoming dataset. • Demonstration that a network of connected fractures in an eolian sandstone shifts the importance of CO 2 storage controlling parameters. In this work, we use the “Design and Analysis of Computer Experiments” (DACE) technique to analyze the sensitivity of CO 2 storage in fractured aquifers. The aquifer static or geogical model was inspired by the Tensleep formation at Teapot Dome, a potential CO 2 storage target in Wyoming. The CO 2 storage performance in aquifers is comprehensively simulated in our dynamic model by considering all the trapping mechanisms, except for mineralization, i.e. structural, dissolution and residual, as well as local capillary trapping. Operational strategies are investigated to achieve an effective trade-off between higher CO 2 trapping efficiency and delayed CO 2 breakthrough at producers. The effect of natural fractures is explored by running single-porosity, dual-permeability and dual-porosity models. Results show that different models produce significantly different trapping efficiencies, which indicates that natural fractures and their type impact the CO 2 storage practice appreciably. To carry out the global sensitivity analysis, two 100-point computer experiments were designed using the modified space-filling method considering a bottom aquifer, matrix permeability and porosity, fracture permeability and spacing, and water Total Dissolved Solids (TDS). Results show that an active bottom aquifer reduces trapping efficiency noticeably. In addition, dissolution trapping becomes the dominant trapping mechanism. The variability of CO 2 trapping efficiency is explained primarily by fracture permeability, then depth, and finally matrix permeability. In contrast, in the absence of aquifer drive, the variability is almost completely explained by matrix permeability. Furthermore, our results provide guidance to the collection of information as to reduce the most dominant uncertainties.

Optimization of volunteer task assignments to improve volunteer retention and nonprofit organizational performance

Deploying the skills and experiences of nonprofit employees and volunteers is challenging due to resource constraints on nonprofit organizations and the salient characteristics of volunteer labor. This research introduces an integer linear program to make volunteer and employee task assignments for a planned nonprofit program that captures the impact of task assignments on volunteer retention in future periods. Assignments are made to balance the need for urgency of current tasks, organizational task preferences, volunteer trainings, and volunteer preferences (which impact retention). The deterministic optimization model is evaluated using a computational design of experiments with a food bank use case and a simulation model that incorporates uncertainty in volunteer preferences and retention. The developed model, which captures the impact of task assignment on expected volunteer retention, is compared to a simpler benchmark that ignores task assignment's impact on volunteer retention and instead makes assignments maximizing the organization's preferences. The proposed approach is found to be beneficial to nonprofits, even when an NPO has uncertainty in their estimates for volunteer preferences and volunteer retention threshold values.

Regression Modeling of surface piercing propeller performance based on trailing edge geometrical parameters using CFD method

The geometry of the surface piercing propeller has been designed experimentally, and no comprehensive studies are available to identify the influential parameters to attain systematic formulations. This study aims to evaluate the feasibility of presenting systematic relations between the geometric parameters of the blade section and its hydrodynamic characteristics through combined methods of computational fluid dynamics and design of experiments. In this regard, the propeller's parametric geometry is designed based on the cubic B-spline method. Hence, a series of SPPs with different trailing edges is generated according to the coordinates of control points based on the D-optimal response surface methodology. The propeller performance was investigated using URANS numerical simulation method in accordance with the VOF method and sliding mesh technique.ANOVA analyses of the quadratic regression models of hydrodynamic coefficients in terms of edge height (Y) and protrusion length (X), based on a 95% confidence interval, indicate these models have prediction adequacy of more than 90%. The main factor (X) is the only significant factor in both models. Increasing it will reduce the hydrodynamic coefficients. As a result, trailing edge variations can significantly affect torque and thrust coefficients by up to 40%, while efficiency changes are about 7%.

A computational design of experiments based method for evaluation of off-the-shelf total knee replacement implants

A methodology to explore the design space of off-the-shelf total knee replacement implant designs is outlined. Generic femur component and tibia plate designs were scaled to thousands of sizes and virtually fitted to 244 test subjects. Various implant designs and sizing requirements between genders and ethnicities were evaluated. 5 sizes optimised via the methodology produced a good global fit for most subjects. However, clinically significant over/underhang was present in 19% of subjects for tibia plates and 25% for femur components, reducing to 11/20% with 8 sizes. The analysis highlighted subtly better fit performance was obtained using sizes with unequal spacing.

Orthogonal designs with branching and nested factors

Computer experiments with branching and nested factors are a common class of computer experiments, but it is challenging to construct designs for this type of experiments. In this paper, we define a special type of design called branching orthogonal Latin hypercube design (BOLHD). Such a design has an appealing structure, that is, no matter at each level of a branching factor or the level‐combination of branching factors, the corresponding design points of nested factors form an orthogonal Latin hypercube design (OLHD). This structure makes it a good choice for designing computer experiments with branching and nested factors. We propose several deterministic construction methods when branching factors have the same number of levels. Based on sliced Latin hypercube designs (SLHDs), the proposed methods are easy to operate. Some construction results are tabulated for practical use.

Sequential Design of Computer Experiments: Current Status and Future Directions

A brief literature review addresses the sequential design of computer experiments. Algorithms, initial designs, surrogate modeling, and stopping criteria are presented. Sequential designs are classified in terms of deterministic/stochastic, granularity, and model-free/model-dependent. Criteria for exploitation and exploration are presented. Some hybrid algorithms that balance exploitation and exploration are introduced and compared with an illustrative example. Practitioners working on computer simulators can benefit from the review presented in this paper by implementing the sequential design of computer experiments.

A new type of robust designs for chemometrics and computer experiments

There are many difficulties that may be encountered in experiments of chemometrics as well as most computer experiments, that ought to explore an approximate model instead of the true one that is complicated. Space-filling designs including uniform designs are robust for such situation. However, when the underlying regression model is known, the D-optimal design (DOD) is the most effective on parameter estimation, but DOD is not robust against the model change. In this paper, we propose a new type of composite designs guaranteeing both robustness and effectiveness. Subsequently, we compare the prediction performance of the seven candidate composite designs, under various case studies. For the convenience of implementation, instead of chemical experiments, we adopt computer experiments as illustration involving some popular models that have been widely applied in evaluating the performance of optimization algorithms. Among all candidate composite designs, we recommend two of them based on their advantaged performance in all cases we explored. Our recommendation is also suitable for most physical experiments.

Combination of computational fluid dynamics and design of experiments to optimize modules for direct contact membrane distillation

The design of membrane distillation modules is an important issue that must be addressed for it to be practically implemented. Herein, a combination of the design of experiments (DOE) and computational fluid dynamics (CFD) was developed to evaluate the interaction between variables of module geometry and how they affect the performance of a direct contact membrane distillation (DCMD) process. A three-dimensional CFD model was developed to calculate flux, the performance ratio (PR), and the permeate-to-feed ratio (ϕp) as functions of the height, length, and width of the channel in a plate-and-frame DCMD module. The CFD model was validated using experimental results in a laboratory DCMD system. Based on a DOE with a central composite design (CCD), CFD simulations were carried out, and response values were obtained for each combination. The results show that length is the most important factor affecting flux, and height is the most important factor affecting PR and ϕp. However, the effect of module width was not significant compared with that of the other variables. Based on the results, a set of regression models were developed to estimate flux, the PR, and ϕp, and these estimations match the CFD simulation results well. Using the derived models, the dimensions of the DCMD module were optimized to simultaneously increase water productivity and energy efficiency.