Data-driven Optimization Method Research Articles

Heat transfer optimization for MH reactor using combined taguchi design and data-driven optimization method

Adding the heat transfer enhancement components into the MH reactor can improve the hydrogen absorption/desorption rate, but the vital indicators as the gravimetric and volumetric hydrogen storage density decrease at the same time, which is hardly considered in the published researches. In this study, a comprehensive hydrogen storage performance (CHSP) indicator for the MH reactor has been proposed considering the above contradiction, which can be applied to evaluate the comprehensive performance of the MH reactor and compare the reactor performances with different heat transfer configurations. Then, with the constraint of the CHSP, the heat transfer structure of the classical finned-tube MH reactor is optimized using a novel sieving optimization strategy which combines the Taguchi design and the data-driven optimization (ANN model + GA) to increase the optimization efficiency and the reliability of optimization solutions. The optimized results indicate that the optimal structural parameters of the finned-tube MH reactor are the diameter of heat exchange tube d of 5.13 mm, the fin thickness h of 0.28 mm, the fin radius r of 21.73 mm and the fin number N of 11, and the optimal CHSP is 0.547 mg/s.

Dynamic Decoupling Method Based on Motor Dynamic Compensation with Application for Precision Mechatronic Systems

Motors are widely employed in mechatronic systems, especially in precision multiple degrees of freedom motion systems. In most applications, the dynamic equation between the motor instruction and the actual driving force is simplified as a constant. Subsequently, the static decoupling method can be utilized to design the feedback controller. However, in high-precision mechatronic systems, motor dynamics cannot be neglected, and the static decoupling performance is compromised due to discrepancies between motors. In this paper, a dynamic decoupling method is developed to improve the decoupling performance of the multiple-input multiple-output systems. The effects of transmission delays, motor dynamics, and discrepancies between different motors are taken into consideration in the dynamic decoupling method. Furthermore, a data-driven optimization method is developed to estimate the parameters of the dynamic decoupling controller. The effectiveness and superiority of the proposed method are demonstrated through numerical simulations. The experimental results show that the dynamic decoupling control method can achieve a 97.75% performance improvement at least compared to the static decoupling control method.

A novel triple periodic minimal surface-like plate lattice and its data-driven optimization method for superior mechanical properties

Lattice structures can be designed to achieve unique mechanical properties and have attracted increasing attention for applications in high-end industrial equipment, along with the advances in additive manufacturing (AM) technologies. In this work, a novel design of plate lattice structures described by a parametric model is proposed to enrich the design space of plate lattice structures with high connectivity suitable for AM processes. The parametric model takes the basic unit of the triple periodic minimal surface (TPMS) lattice as a skeleton and adopts a set of generation parameters to determine the plate lattice structure with different topologies, which takes the advantages of both plate lattices for superior specific mechanical properties and TPMS lattices for high connectivity, and therefore is referred to as a TPMS-like plate lattice (TLPL). Furthermore, a data-driven shape optimization method is proposed to optimize the TLPL structure for maximum mechanical properties with or without the isotropic constraints. In this method, the genetic algorithm for the optimization is utilized for global search capability, and an artificial neural network (ANN) model for individual fitness estimation is integrated for high efficiency. A set of optimized TLPLs at different relative densities are experimentally validated by the selective laser melting (SLM) fabricated samples. It is confirmed that the optimized TLPLs could achieve elastic isotropy and have superior stiffness over other isotropic lattice structures.

Reinforcement learning as data-driven optimization technique for GMAW process

Welding optimization is a significant task that contributes to enhancing the final welding quality. However, the selection of an optimal combination of various process parameters poses different challenges. The welding geometry and quality are influenced differently by several process parameters, with some exhibiting opposite effects. Consequently, multiple experiments are typically required to obtain an optimal welding procedure specification (WPS), resulting in the waste of material and costs. To address this challenge, we developed a machine learning model that correlates the process parameters with the final bead geometry, utilizing experimental data. Additionally, we employed a reinforcement learning algorithm, namely stochastic policy optimization (SPO), with the aim to solve different optimization tasks. The first task is a setpoint‐based optimization problem that aims to find the process parameters that minimize the amount of deposited material while achieving the desired minimum level of penetration depth. The second task is an optimization problem without setpoint in which the agent aims to maximize the penetration depth and reduce the bead area. The proposed artificial intelligence-based method offers a viable means of reducing the number of experiments necessary to develop a WPS, consequently reducing costs and emissions. Notably, the proposed approach achieves better results with respect to other state-of-art metaheuristic data-driven optimization methods such as genetic algorithm. In particular, the setpoint‐based optimization problem is solved in 8 min and with a final mean percentage absolute error (MPAE) of 2.48% with respect to the 42 min and the final 3.42% of the genetic algorithm. The second optimization problem is also solved in less time, 30 s with respect to 6 min of GA, with a higher final reward of 5.8 from the proposed SPO algorithm with respect to the 3.6 obtained from GA.

A two-stage distributed optimization method for home energy management systems via multi-modal data-driven algorithm

The home energy management system (HEMS), which utilizes multi-modal data from multiple sensors to generate the knowledge about decision making, is essential to the optimization of home energy management efficiency. Load scheduling based on HEMS can improve the utilization efficiency of multi-modal data and derived knowledge, achieve power supply-demand balance, and reduce users’ electricity costs. This paper proposes a distributed load optimization scheduling method for the load scheduling problem in HEMS based on multi-modal data-driven algorithm. Additionally, a two-stage data-driven optimization method is proposed, including a first-stage optimization model based on minimizing electricity costs and a second-stage optimization model based on minimizing system load fluctuations. In the first stage, cost self-optimization is performed based on energy storage devices. In the second stage, a load optimization instruction is issued by the control center, and each user optimizes the load fluctuations based on the system load data. Compared to centralized control methods, this approach reduces the computational overhead of the control center. Finally, simulation experiments based on load scheduling in the HEMS are conducted. The results of the first optimization stage show that when the battery capacity integrated into the system increases from 3.68 kWh to 6.68 kWh, user costs can be reduced from 57.572 cents to 42.064 cents. It is not only evident that the proposed method can effectively save users on electricity costs, but the introduction of larger capacity batteries also lowers these costs. The second stage of load fluctuation optimization results show that the proposed method can effectively optimize the usage data of a group of users and decrease the absolute peak-valley difference by 8.8%.

Hierarchical power control of a large-scale wind farm by using a data-driven optimization method.

With the participation in automatic generation control (AGC), a large-scale wind farm should distribute the real-time AGC signal to numerous wind turbines (WTs). This easily leads to an expensive computation for a high-quality dispatch scheme, especially considering the wake effect among WTs. To address this problem, a hierarchical power control (HPC) is constructed based on the geographical layout and electrical connection of all the WTs. Firstly, the real-time AGC signal of the whole wind farm is distributed to multiple decoupled groups in proportion of their regulation capacities. Secondly, the AGC signal of each group is distributed to multiple WTs via the data-driven surrogate-assisted optimization, which can dramatically reduce the computation time with a small number of time-consuming objective evaluations. Besides, a high-quality dispatch scheme can be acquired by the efficient local search based on the dynamic surrogate. The effectiveness of the proposed technique is thoroughly verified with different AGC signals under different wind speeds and directions.

Effects of sensor configuration optimization on airflow estimation in urban environment: A case study with a building group model

The complicated flow field in an urban area can be properly estimated using sparse sensor measurements and several estimation methods. Because sensor measurements are the only real-time information provided for estimations, it is necessary to investigate the effects of the sensor placement on the estimation accuracy. This study compared the performance of uniform and optimized sensor configurations in estimating the streamwise velocity field for a block-arrayed building group model. The optimized configuration was produced by a QR decomposition method, which is a data-driven optimization method that uses the input POD11POD-LSE: proper orthogonal decomposition and linear stochastic estimation; POD: proper orthogonal decomposition; LSE: linear stochastic estimation; SCO: sensor configuration optimization; CFD: computational fluid dynamics; LES: large-eddy simulation; ANN: artificial neural network; RMSE: root mean squared error. (Proper Orthogonal Decomposition) modes and deploys sensors to the positions where dominant flow structures can be efficiently measured. The robustness of two estimation methods, inverse POD and POD-LSE (Linear Statistic Estimation), against different sensor configurations was analyzed. The influence of different settings, like the number of POD modes and the number of sensors, on the optimization and estimation were also explored. According to the results, the optimal configuration can reduce the root mean square error (RMSE) by about 86% in inverse POD and 10% in POD-LSE when compared to the uniform one. Applying LSE can relieve the deficiency caused by poorly placed sensors, but since it can only supplement linear information, the configuration optimization is still critical especially to nonlinear estimations.

A diagnostic analytics model for managing post-disaster symptoms of depression and anxiety among students using a novel data-driven optimization approach

Prevalent mental disorders, such as depression and anxiety, commonly manifest in students throughout the transition to early adulthood. Mental illnesses can significantly impact students’ academic and social activities. An automatic or semiautomatic health monitoring approach is very effective for diagnosing depression and anxiety. This study aims to implement and scrutinize a data-driven optimization method for identifying and providing therapy to students with symptoms of depression and anxiety. The proposed method starts with data preprocessing and operating sentiment analysis to identify mentally disordered students. An ensemble learning classifier later divides students with symptoms into three categories based on their health condition: severe, moderate, and mild. A hyperparameter optimization approach is further adopted to improve the model’s performance. Finally, a rule-based dispatching system is implemented for scheduling therapy sessions. The proposed novel data-driven method is a post-disaster intelligent and reliable method that integrates three well-adopted techniques to address students’ depression and anxiety. The findings indicate that the conventional approach to monitoring depression among students previously detected only 7 to 15% of cases. However, the performance of the offered strategy revealed a confirmed rate of 44% of depressed and anxious students.

Shear design of recycled aggregate concrete beams using a data-driven optimization method

Under the umbrella of sustainable development, recycled aggregate concrete (RAC) is a building material that has attracted a lot of interest. This study proposes a data-driven methodology for shear capacity prediction and design optimization for RAC beams. A comprehensive test dataset with 232 samples was gathered based on clear criteria. Geometric and materials factors are identified as inputs to develop four machine learning (ML) prediction models including artificial neural network, Gaussian process regression, random forest, and categorical gradient boosting (CGB). The winner of ML models is combined with the nondominated sorting genetic algorithm II (NSGA-II) to establish a dual-objective design optimization of RAC beams subjected to shear. Results show that the CGB model achieved the highest accuracy (R2 = 0.972, RMSE = 14.909, and MAPE = 0.086) among four ML models and its standard deviation was more than 80% lower than the five existing equations. The SHapley Additive exPlanations interpreted the CGB model and identified the beam width, nominal strength of stirrups, shear-span ratio, and nominal strength of longitudinal reinforcement are the four main features affecting the shear capacity of a RAC beam. Finally, the Pareto front was created by combining the CGB model with the NSGA-II algorithm. The suggested framework provides a novel route for the structural design of RAC beams subject to shear.

A fast data-driven optimization method of multi-area combined economic emission dispatch

To provide more feasible schemes for power system optimization management and operation, the interconnection of power grids in different areas is inevitable. Naturally, the multi-area combined economic/emission dispatch (MACEED) problem becomes a more important decision problem. However, as the dimensions of MACEED problems increase, existing studies may not obtain feasible scheduling decisions in a suitable time. On this background, MACEED problems are transferred into computational expensive problems. In order to solve high-dimensional, large-scale MACEED problems, a data-driven surrogate-assisted approach is proposed. First, a feature engineering-based support vector regression surrogate model is proposed to replace the traditional objective functions in high-dimensional MACEED problems. Second, knowledge distillation is applied as a freezing and fine-tuning mechanism for the improved support vector regression surrogate models, which significantly reduces the time required to build surrogate models. Third, an improved non-dominated sorting genetic algorithm-III is proposed to obtain feasible solutions to high-dimensional MACEED problem. The proposed algorithm enhances the convergence and diversity of optimal solutions. The effectiveness of the proposed data-driven approach is demonstrated through simulations of a four-area 40-unit test system under different constraints.

Data-driven design of graded composite lattice structures with multiple microstructure prototypes and materials

Lattice structures have attracted much attention from engineers due to their excellent properties, especially with the rise of additive manufacturing technology. In this paper, a homogenization-based data-driven optimization method is proposed for designing graded composite lattice structures composed of a series of composite lattice microstructures generated from different microstructure prototypes, which are represented by using corresponding basic level set functions. Using different cutting planes to cut the same basic level set function can obtain a series of basic microstructures with different relative densities and similar configurations. The composite lattice microstructures can be obtained by combining the basic microstructures generated from different basic level set functions. The homogenization approach is used to calculate the equivalent elasticity matrix of the composite lattice microstructure. The mapping relationships among its density, equivalent elasticity matrix, and the height of cutting plane are established. Numerical examples of both compliance minimization and frequency maximization problems are conducted to verify the validation and effectiveness of the proposed method. • Homogenization-based data-driven optimization method for designing graded composite lattice structures is developed. • The mapping relationships among the relative density, the equivalent elastic tensor, and the height of cutting plane are established based on the homogenization theory. • The composite lattice microstructure is composed of several basic microstructures generated from corresponding microstructure prototypes. • Both the compliance minimization and the frequency maximization problems are conducted.

A data-driven optimization method for coarse-graining gene regulatory networks

One major challenge in systems biology is to understand how various genes in a gene regulatory network (GRN) collectively perform their functions and control network dynamics. This task becomes extremely hard to tackle in the case of large networks with hundreds of genes and edges, many of which have redundant regulatory roles and functions. The existing methods for model reduction usually require the detailed mathematical description of dynamical systems and their corresponding kinetic parameters, which are often not available. Here, we present a data-driven method for coarse-graining large GRNs, named SacoGraci, using ensemble-based mathematical modeling, dimensionality reduction, and gene circuit optimization by Markov Chain Monte Carlo methods. SacoGraci requires network topology as the only input and is robust against errors in GRNs. We benchmark and demonstrate its usage with synthetic, literature-based, and bioinformatics-derived GRNs. We hope SacoGraci will enhance our ability to model the gene regulation of complex biological systems.

A Data-Driven Optimization Method Considering Data Correlations for Optimal Power Flow Under Uncertainty

This paper introduces a data-driven optimization (DDO) method based on novel strategic sampling (SS) considering data correlations for multiperiod optimal power flow (OPF) considering energy storage devices under uncertainty (OPF-ESDUU) of uncertain renewable energy and power loads (UREPL). This DDO method depends only on the uncertainty samples to yield an optimal solution that satisfies a specific confidence level, which is effective because of two resounding learning algorithms: Bayesian hierarchical modeling (BHM) and determinantal point process (DPP). Considering both the local bus information and spatial correlations over all buses, BHM learns the convex approximation of AC power flow (CAACPF) more accurately than the existing learning methods, converting the originally non-convex OPF-ESDUU to a convex optimization problem. DPP considers the correlations between samples to find a small set of significant samples by measuring the relative weight of each sample using the random matrix theory, significantly decreasing the data samples required by the existing SS. The experimental analysis in IEEE test cases shows that after considering data correlations, 1) BHM learns CAACPF better with 13–90% accuracy improvement on average, compared with the existing learning methods, and 2) the proposed DDO performs more efficiently than the existing DDO as DPP-based SS boosts the sampling efficiency by 50% at least.

A Data-Driven Optimization Method for Simulating Arbitrarily Distributed and Spatial–Temporal Correlated Radar Sea Clutter

A Data-Driven Optimization Method for Simulating Arbitrarily Distributed and Spatial–Temporal Correlated Radar Sea Clutter

Neural Inverse Design of Nanostructures (NIDN)

In the recent decade, computational tools have become central in material design, allowing rapid development cycles at reduced costs. Machine learning tools are especially on the rise in photonics. However, the inversion of the Maxwell equations needed for the design is particularly challenging from an optimization standpoint, requiring sophisticated software. We present an innovative, open-source software tool called Neural Inverse Design of Nanostructures (NIDN) that allows designing complex, stacked material nanostructures using a physics-based deep learning approach. Instead of a derivative-free or data-driven optimization or learning method, we perform a gradient-based neural network training where we directly optimize the material and its structure based on its spectral characteristics. NIDN supports two different solvers, rigorous coupled-wave analysis and a finite-difference time-domain method. The utility and validity of NIDN are demonstrated on several synthetic examples as well as the design of a 1550 nm filter and anti-reflection coating. Results match experimental baselines, other simulation tools, and the desired spectral characteristics. Given its full modularity in regard to network architectures and Maxwell solvers as well as open-source, permissive availability, NIDN will be able to support computational material design processes in a broad range of applications.

Automatized Experimental Combustor Development Using Adaptive Surrogate Model-Based Optimization

Abstract Lean premixed combustion is the state-of-the-art technology to achieve ultra low NOx emissions in stationary gas turbines. However, lean premixed flames are susceptible to thermoacoustic instabilities, lean blowout, and flashback. The design of such a combustion system is thus always related to the balancing between the levels of emissions and flame stability. Data-driven optimization methods and the adaptation of models through artificial intelligence have experienced a surge in development in the past years. The goal of this study is to show the potential of these methods for gas turbine burner development. A special pilot burner that features 61 different positions of fuel injection, manufactured by means of selective laser melting is used to modify the gas mixture close to the flame anchoring position. Each of the injector lines is equipped with an individual valve, such that the distribution of fuel-air mixture can be modified variously. Installed into an industrial MGT6000 swirl combustor, a data-driven optimization method is used to find an optimal subset of injection locations by automated experiments. The method uses a surrogate model that is based on Gaussian processes regression. It is adopted for experimental optimization, keeping measurement efforts to a minimum. The optimizer controls the fuel valves and uses live measurements to find a distribution that generates minimal NOx emissions while ensuring flame stability. The solutions found by the optimization scheme are analyzed and advantages and limitations of the approach are discussed.

Consistent Optimization of Blast Furnace Ironmaking Process Based on Controllability Assurance Soft Sensor Modeling.

The blast furnace ironmaking process is the core of steel manufacturing, and the optimization of this process can bring enormous economic and environmental benefits. However, previous data-driven optimization methods neglect the uncontrollability of part of the variables in the predictive modeling process, which brings great uncertainty to the optimization results and adversely affects the optimization effect. To address this problem, a consistency optimization framework based on controllability assurance soft sensor modeling is proposed. The method achieves the information extraction of uncontrollable variables in a process-supervised way, and improves the posterior distribution prediction accuracy. The method also proposes an integrated self-encoder regression module, which uses the regression to guide the encoding, realize the construction of latent features, and further improve the prediction accuracy of the model. Integrating the prediction module and the multi-objective gray wolf optimizer, the proposed model achieves the optimization of the blast furnace ironmaking process with only controllable variables as prediction model inputs while being capable of giving uncertainty estimates of the solutions. Empirical data validated the optimization model and demonstrated the effectiveness of the proposed algorithm.

Active learning-guided exploration of parameter space of air plasmas to enhance the energy efficiency of NO x production

Low temperature, air plasmas have shown promise for production of NO x for nitrogen fixation. However, to make nitrogen fixation via air plasmas economically viable, a major challenge arises from reducing the energy cost of NO x generation, which is a complex function of a multitude of factors including the plasma discharge type, discharge operating parameters and presence of heterogeneous catalysts. This paper presents an active learning (AL) approach for exploring the multivariable and highly nonlinear parameter space of low temperature plasmas (LTPs) in a systematic and efficient manner. The proposed AL approach relies on Bayesian optimization, which is a data-driven optimization method that is particularly suited for optimizing black-box functions constructed from noisy observations. We demonstrate the AL approach for querying the parameter space of a DC pin-to-pin glow discharge in order to enhance the energy efficiency of NO x production. It is observed that, given a fixed experimental budget, AL consistently outperforms random search of the parameter space in terms of minimizing the energy cost or maximizing the rate of NO x generation in the presence of a constraint on discharge power. AL approaches can pave the way for automated and efficient exploration of the high-dimensional parameter space of LTPs, towards establishing insights into their complex behaviors.

Comparison of state-of-the-art machine learning algorithms and data-driven optimization methods for mitigating nitrogen crossover in PEM fuel cells

Nitrogen gas crossover (NGC) and nitrogen accumulation at the anode of proton exchange membrane (PEM) fuel cells are ineluctable and it would lead to inferior performance and even irreversible damage to functional components. To mitigate this issue, multiphysics numerical models (MNMs) are established to describe NGC behaviors and further guide experimental studies. However, to obtain the optimized parameters that would suppress NGC and retain high performance, grid search conducted on MSMs would cost unaffordable computational resources and time. Therefore, we innovatively introduced a machine learning-assisted MNM (MSM-ML) as a surrogate model, in which 9 state-of-the-art machine learning algorithms were compared, to greatly boost the resolution of this engineering problem. Through the proposed MSM-ML workflow performed on an experimentally validated MSM, the cost for obtaining the best parameter combination is greatly reduced. Moreover, the impact of each parameter in this complex system is directly revealed through the application of black-box interpretation methods afterwards. As a result, a new approach was pioneered to solve engineering problems which was demonstrated to be more efficient and intelligent than traditional methods. The NGC coefficient is reduced by 49.5%, while the power density is improved by 20% through the multivariable optimization of the developed MSM-ML.

Multi-objective aerodynamic optimization of two-dimensional hypersonic forebody-inlet based on the heuristic algorithm

Nowadays, the forebody-inlet integrated design optimization becomes more and more important in improving the performance of hypersonic vehicles. And several achievements have been made to improve the performance of the two-dimensional hypersonic forebody-inlet integrated vehicle. However, all these works are faced with many defects. For example, the work adopting the gradient algorithm is easy to be trapped in the local optimum solution. The data-driven optimization method is with a low level of accuracy. Moreover, the geometry parameterization methods adopted in those optimization procedures are hardly implemented to complex configurations. In this study, a global-searching multi-objective optimization framework is proposed to overcome these limitations and to carry out multi-objective optimization of the two-dimensional hypersonic forebody-inlet shape. The framework includes the Free-Form Deformation (FFD) method, the mesh deformation method, the high-accuracy Reynolds Averaged Navier-Stokes (RANS) solver, and the global-searching heuristic algorithm. Results obtained by this framework indicate that the total pressure recovery coefficient, the pressure rising ratio, and the mass flow coefficient are increased by 3.97%, 9% and 7.35%, respectively, which suggests that the optimization framework proposed in this study is promising to be widely used in practical engineering applications.