Search

In this study, the procedure of surrogate-assisted optimization is constructed to solve the multi-objective design problem of a refrigeration system. In the refrigeration process, the required coefficient of performance (COP) can be varied to the required cooling loads. In this case, the optimum operating conditions for a specified COP range is required to reduce the power consumption of the system. In this situation, the problem turns into a multi-objective optimization problem to simultaneously maximize COP and minimize power consumption. A surrogate model of the COP and power consumption are generated using several kernel functions. The best model using a linear spline kernel function is selected and used in the optimization process. A comparative study of several recent and well-known multi-objective metaheuristics was performed to measure performance of the available algorithms. The Pareto fronts containing optimum operating conditions for the refrigeration plant over the entire range of COP values were obtained in this study.

Industrial refrigeration systems typically exhibit a higher energy consumption rate compared to other systems. It is a very important system, and for the brewing industry, it is at the heart of brewing as, in each process, the temperature of the beer must be controlled. Therefore, a health monitoring or faults detection system is a necessary tool for predictive and preventive maintenance to avoid faults in the system. This work presents machine learning models for faults prediction of the refrigeration system in a brewing firm. The raw data of processing parameters are collected, while the system health for each data set is identified and used for machine learning model training. Three operation cases based on the actual operation of the brewery's refrigeration system are analyzed in this work. Several machine learning models including Naïve Bayes (NB), Generalized Linear Models (GLMs), Logistic Regression (LR), Fast Large Margin (FLM), Deep Learning (DL), Decision Tree (DT), Random Forest (RF), Gradient Boosted Trees (GBT), and Support Vector Machine (SVM) are applied for refrigeration system fault detection in the three operation cases while their performances are investigated. The numerical simulation is performed based on RapidMiner Studio. According to the experiment of the three operation cases, the deep learning model was found to be the most accurate and required the least amount of time for the analysis. The accuracy percentages are 86.7%, 90.2%, and 82.5%, while the running times are 15 seconds, 27 seconds, and 15 seconds, respectively. This work can be considered as the baseline for future studies on applied machine learning models for fault detection in refrigeration systems.

This paper presents an optimal design of a sweep blade for the axial wind turbine using a hybrid surrogate-assisted optimizer. The design problem is defined to maximize the ratio of the torque coefficient to the thrust coefficient of a turbine blade at a low wind velocity of 10 m/s. Pitch angle and leading-edge blade curve are considered as the design variables. For the aerodynamic analysis of the wind turbine blade, computational fluid dynamics has been used as a high-fidelity simulation. While the surrogate models including, the Kriging model (KG), the radial basis function model (RBF), and the proposed hybrid of KG and RBF (HyKG-RBF) models are applied for function approximation or low-fidelity simulation. In this study, to obtain a set of sampling points and surrogate models development, an optimal Latin Hypercube sampling (OLHS) technique is utilized in the design of the experiment (DOE). A differential evolutionary (DE) algorithm is used to solve the proposed design problem. The performance of the proposed hybrid surrogate assisted optimization method is contrasted with two conventional surrogate assisted optimization techniques. Results demonstrate that the proposed hybrid surrogate model viz. HyKG-RBF is the most efficient surrogate-assisted optimization method for solving the sweep blade optimization problem.

Many-objective truss structure problems from small to large-scale problems with low to high design variables are investigated in this study. Mass, compliance, first natural frequency, and buckling factor are assigned as objective functions. Since there are limited optimization methods that have been developed for solving many-objective truss optimization issues, it is important to assess modern algorithms performance on these issues to develop more effective techniques in the future. Therefore, this study contributes by investigating the comparative performance of eighteen well-established algorithms, in various dimensions, using four metrics for solving challenging truss problems with many objectives. The statistical analysis is performed based on the objective function best mean and standard deviation outcomes, and Friedman's rank test. MMIPDE is the best algorithm as per the overall comparison, while SHAMODE with whale optimisation approach and SHAMODE are the runners-up. • A comparative test to measure the efficiency of eighteen state-of-the-practice methods is performed. • Small to large-scale truss design challenges are proposed for the validation. • The performance is measured using four metrics and Friedman's rank test.

A novel method for system identification of small-scale fixed-wing Unmanned Aerial Vehicles (UAVs) using a metaheuristics (MHs) approach is proposed. This investigation splits the complex aerodynamic model of UAV into longitudinal and lateral dynamics sub-systems. The system identification optimisation problem is proposed to find the UAV aerodynamic and stability derivatives by minimizing the R-squared error between the measurement data and the flight dynamic model. Thirteen popular optimisation algorithms are applied for solving the proposed UAV system identification optimisation problem while each algorithm is tested for 10 independent optimisation runs. By performing the Freidman’s rank test, statistical analysis of the experiment work was carried out while, based on the fitness value, each algorithm is ranked. The outcomes demonstrate the dominance of the L-SHADE algorithm, with mean R-square errors of 0.5465 and 0.0487 for longitudinal and lateral dynamics, respectively. It is considered superior to the other algorithms for this system identification problem.

This paper presents multi-objective topology and sizing optimization of a morphing wing structure. The purpose of this paper is to design a new aircraft wing structure with a tapered shape for ribs, spars, and skins including a torsion beam for external actuating torques, which is anticipated to modify the aeroelastic characteristic of the aircraft wing using multi-objective optimization. Two multi-objective topology optimization problems are proposed employing ground element structures with high- and low-grid resolutions. The design problem is to minimize mass, maximize difference of lift effectiveness, and maximize the buckling factor of an aircraft wing subject to aeroelastic and structural constraints including lift effectiveness, critical speed, and buckling factors. The design variables include aircraft wing structure dimensions and thickness distribution. The proposed optimization problems are solved by an efficient multi-objective metaheuristic algorithm while the results are compared and discussed. The Pareto optimal fronts obtained for all tests were compared based on a hypervolume metric. The objective function values for Case I and Case II at 10 selected optimal solutions exhibit a range of structural mass as 115.3216–411.6250 kg, 125.0137–440.5869 kg, lift effectiveness as 1.0514–1.1451, 1.0834–1.1639 and bucking factor as 38.895–1133.1864 Hz, 158.1264–1844.4355 Hz, respectively. The best results reveal unconventional aircraft wing structures that can be manufactured using additive manufacturing. This research is expected to serve as a foundation for future research into multi-objective topology optimization of morphing wing structures based on the ground element framework.

In this paper, optimum U-shaped baffles in a square channel heat exchanger using air as a working fluid were developed using surrogate-assisted optimization. The design problem is set to maximize heat transfer performance and simultaneously minimize pressure loss across the channel. Design variables determine the radii and heights of the baffles, whereas the optimization problem is treated as box-constrained optimization. The work in this paper is aimed at finding an appropriate surrogate model for designing such a heat exchanger system. Function evaluations are performed by means of computational fluid dynamics (CFD). The computations are based on the finite volume method and are carried out at a Reynolds number of 4000. It has been found that the use of U-shaped baffles as heat transfer enhancement devices improves the thermal performance of the heat exchanger. Comparative results reveal that the Kriging model is the most accurate surrogate model, however, the surrogate model giving the best result is support vector regression.

This work presents an alternative efficient means to synthesise a fixed-structure autopilot controller that is both robust and optimal using meta-heuristics (MHs). The problem is aimed at finding controllers in several sections with the objective of minimising integral square error, subject to several constraints to ensure a robust, precise, and rapid reference tracking control system. An optimum control problem was posed while several MHs were employed to solve the problem, and their performances were investigated. Based on the results, a Self-Adaptive Differential Evolution (JADE) was found to be the most efficient algorithm. The study presents a simple but effective tool for designing a robust and optimum autopilot flight controller. It also explores the performance of several MHs in the new optimisation design field of robust and optimal flight control systems. [ABSTRACT FROM AUTHOR]