Stochastic-based Method Research Articles

A CRITIC-BPNN APPROACH TO FRICTION STIR WELDING PARAMETRIC SELECTION AND PREDICTION USING AA6082-T6 MATERIAL

The uncontrolled friction stir welding heat generation impacts the quality of welds. However, the intuition and experience of the engineer fail to regulate the effects of excessive heat generation on the weld quality and research has not addressed this aspect yet. This paper fills the gap by introducing an integrated CRITIC-BPNN (CRiteria Importance Through Intercriteria Correlation-Back Propagation Neural Network) method to investigate the selection and optimisation characteristics of the friction stir welding process for AA6082-T6 material. In this study, two major performance characteristics i.e. ultimate tensile strength (UTS) and percentage elongation (%EL), were chosen for analysis. The input parameters for the machining were the tool rotational speed, welding speed, tool pin profile and tool shoulder diameter. For the back propagation neural network model, a four-layer network with sigmoid hidden neurons and output neurons was selected. The weight estimates of the friction stir welding parameters are determined by the CRITIC method. For further weight determination between the nodes and edges of the neural networks, the Poisson distribution model was introduced. This stochastic-based method was used to calculate the weights at the edges, between the inputs, hidden layers and outputs of the neural network. The forward pass and backward passes are then used for updating and error minimisation. The welding speed has the highest weight with a contribution of 49.72% using the CRITIC method, implying that welding speed is the best and most influential parameter of the friction stir welding process. For the 4-1-2 neural network architecture, the values of the ultimate tensile strength and percentage elongation at the optimal thresholds are 0.6457 and 0.6019, respectively, for the first forward pass and 0.6123 and 0.6356, respectively, for the second forward pass. The predicted tensile strength is 320.64 MPa and the prediction for the percentage elongation is 18.83%. The results obtained from the proposed method could be useful for planning purposes during the friction welding process.

Applying and benchmarking a stochastic programming-based bidding strategy for day-ahead hydropower scheduling

Aneo is one of the first Nordic power companies to apply stochastic programming for day-ahead bidding of hydropower. This paper describes our experiences in implementing, testing, and operating a stochastic programming-based bidding method aimed at setting up an automated process for day-ahead bidding. The implementation process has faced challenges such as generating price scenarios for the optimization model, post-processing optimization results to create feasible and understandable bids, and technically integrating these into operational systems. Additionally, comparing the bids from the new stochastic-based method to the existing operator-determined bids has been challenging, which is crucial for building trust in new procedures. Our solution is a rolling horizon comparison, benchmarking the performance of the bidding methods over consecutive two-week periods. Our benchmarking results show that the stochastic method can replicate the current operator-determined bidding strategy. However, additional work is needed before we can fully automate the stochastic bidding setup, particularly in addressing inflow uncertainty and managing special constraints on our watercourses.

Extended Kalman filter based on stochastic epidemiological model for COVID-19 modelling

Extended Kalman filter based on stochastic epidemiological model for COVID-19 modelling

Optimization of overhead transmission lines insulation and grounding costs with respect to backflashover rate

In this paper, the optimization of transmission insulation and grounding is performed considering two backflashover rate (BFR) evaluation procedures widely applied in the industry. The optimization approach is based on the Lagrange formulation introduced by Hileman in late sixties. This method used the old AIEE procedure to evaluate the BFR of transmission lines. Since then a number of BFR evaluation are available to determine BFR rates outperforming legacy procedures. Recently, the Hileman’s optimization method was updated using the EPRI two-point method to evaluate the BFR. However, despite the step-by-step EPRI method being still widely used in industry, stochastic simulation with specialized programs is becoming standard. Thus, in this paper, the insulation and grounding optimization methodology is updated using more adequate and recent BFR evaluation procedures such as the IEEE Std. 1243-97 and the Monte Carlo EMTP-based methods. Technical and economic results were compared with the existing optimization approach based on the EPRI two-point method using the EPRI 345 kV transmission line as test system. Results show that IEEE and EPRI methods yield equivalent results. The stochastic-based method (MC-EMTP) is more optimistic showing total and marginal insulation and grounding costs below the IEEE/EPRI methods.

Multi-objective optimal formation reconfiguration with scalarization and stochastic boundaries

Multi-objective optimal formation reconfiguration with scalarization and stochastic boundaries

Boosted mutation-based Harris hawks optimizer for parameters identification of single-diode solar cell models

Boosted mutation-based Harris hawks optimizer for parameters identification of single-diode solar cell models

An Implementation of Fuzzy PD Control Design for Five Flip Folders Folding Machine Using Arduino Mega2560

The arm manipulator application that has 1 degree of freedom (DOF) can be used to flip folders in folding machines. The used actuator in folding machine is a Brushed DC motor. The Fuzzy-PD control system in this application is using Arduino Mega2560 microcontroller. Application of Fuzzy-PD control system on folding machine is able to move 5 flip folders. This Fuzzy-PD control system is able to improve the system response when there is a load, where the difference is only 7.73% from the no load condition. In addition, to implement this flip folder, a stochastic based method (ARX) is also applied on this system so it can be simulated directly. The simulation results from this folding machine model have a time constant of 0.510s, rise time of 1.501s, settling time of 1.5320s and delay time of 0.353s.

Dynamic optimization of fed-batch bioprocesses using flower pollination algorithm.

There exist several optimization strategies such as sequential quadratic programming (SQP), iterative dynamic programing (IDP), stochastic-based methods such as differential evolution (DE), genetic algorithm (GA), particle swarm optimization (PSA), and ant colony optimization (ACO) for finding optimal feeding profile(s) during fed-batch fermentations. Here in the present study, flower pollination algorithm (FPA) which is inspired by the pollination process in terrestrial flowering plants has been used for the first time to find the optimal feeding profile(s) during fed-batch fermentations. Single control variable, two control variables and state variable bounded problems were chosen to test the robustness of the FPA for optimal control problems. It was observed that FPA is computationally less intensive in comparison with other stochastic strategies. Thus, obtained results were compared to other studies and it has been found that the FPA converged either to newer optima or closer to the established global optimum for the cases studied.

Ant Lion Optimization algorithm for kidney exchanges

The kidney exchange programs bring new insights in the field of organ transplantation. They make the previously not allowed surgery of incompatible patient-donor pairs easier to be performed on a large scale. Mathematically, the kidney exchange is an optimization problem for the number of possible exchanges among the incompatible pairs in a given pool. Also, the optimization modeling should consider the expected quality-adjusted life of transplant candidates and the shortage of computational and operational hospital resources. In this article, we introduce a bio-inspired stochastic-based Ant Lion Optimization, ALO, algorithm to the kidney exchange space to maximize the number of feasible cycles and chains among the pool pairs. Ant Lion Optimizer-based program achieves comparable kidney exchange results to the deterministic-based approaches like integer programming. Also, ALO outperforms other stochastic-based methods such as Genetic Algorithm in terms of the efficient usage of computational resources and the quantity of resulting exchanges. Ant Lion Optimization algorithm can be adopted easily for on-line exchanges and the integration of weights for hard-to-match patients, which will improve the future decisions of kidney exchange programs. A reference implementation for ALO algorithm for kidney exchanges is written in MATLAB and is GPL licensed. It is available as free open-source software from: https://github.com/SaraEl-Metwally/ALO_algorithm_for_Kidney_Exchanges.

Efficient computation of shortest absent words in complete genomes

Efficient computation of shortest absent words in complete genomes

A stochastic model for parameter identification of adhesive materials

In practice, there are only a very limited number of experimental data available. Therefore, the prediction of material behaviour is difficult and a statistical analysis with a stochastic-based method is nearly impossible. In order to increase the number of tests based on experimental data, we apply the method of stochastic simulation based on time series analysis. The generated artificial data have the same stochastic behaviour as the experimental data. Advantages of artificial data are the arbitrary number of data available, and as a conclusion, the process of parameter identification can be statistically analysed. Here, we especially have experiments for adhesive materials for substantial tension tests performed at two different strain rates. Artificial data provide a stochastic proved analysis of the parameter identification concerning distribution and deviations. The analysis shows the possible range of the different material parameters and, therefore, gives a detailed view of the identification process.

Watershed Modeling and its Applications: A State-of-the-Art Review

Advances in the understanding of physical, chemical, and biological processes influencing water quality, cou- pled with improvements in the collection and analysis of hydrologic data, provide opportunities for significant innovations in the manner and level with which watershed-scale processes may be explored and modeled. This paper provides a re- view of current trends in watershed modeling, including use of stochastic-based methods, distributed versus lumped pa- rameter techniques, influence of data resolution and scalar issues, and the utilization of artificial intelligence (AI) as part of a data-driven approach to assist in watershed modeling efforts. Important findings and observed trends from this work include (i) use of AI techniques artificial neural networks (ANN), fuzzy logic (FL), and genetic algorithms (GA) to im- prove upon or replace traditional physically-based techniques which tend to be computationally expensive; (ii) limitations in scale-up of hydrological processes for watershed modeling; and (iii) the impacts of data resolution on watershed model- ing capabilities. In addition, detailed discussions of individual watershed models and modeling systems with their fea- tures, limitations, and example applications are presented to demonstrate the wide variety of systems currently available for watershed management at multiple scales. A summary of these discussions is presented in tabular format for use by water resource managers and decision makers as a screening tool for selecting a watershed model for a specific purpose.

Random walk methods for scalar transport problems subject to Dirichlet, Neumann and mixed boundary conditions

Three stochastic-based methods are proposed for solving unsteady scalar transport problems in bounded, single-phase domains. The first (Method I), a local solution appropriate to problems having Dirichlet conditions, adapts a well-known local stochastic solution of a backward Fokker–Planck equation to scalar transport. Method II, a local solution applicable to Dirichlet, Neumann and/or mixed initial boundary value problems (IBVPs), and representing a time-dependent extension of a recently reported heuristic steady solution, provides a straightforward addition to the limited collection of techniques available for Neumann and mixed problems. This approach is shown to be equivalent to a long-standing, rigorous low-order solution and, in addition, allows development of a probabilistic-based analytical solution to Neumann problems, stated in terms of an exit probability. Method III, a global solution, likewise suitable for Neumann and mixed IBVPs, follows by combined application of domain boundary Taylor expansions and Method I. This approach is shown to be computationally equivalent to a global version of Method II.

Interaction Term of Tsai-Wu Theory for Laminated Veneer

A stochastic-based method of evaluating the interaction parameter (\iF\d1\d2) of the Tsai-Wu strength theory has been presented in this paper. Treating all strength parameters of the strength theory as random variables, the mean and standard deviation of \iF\d1\d2, under plane stress conditions, have been estimated for Douglas-fir laminated veneer. This estimation has been managed through a nonlinear least-squares fit of the parameters to a cumulative probability distribution of off-axis tensile data. For the purpose of comparison, a sample evaluation of \iF\d1\d2 using deterministic methods has been presented. This evaluation showed strong dependence of \iF\d1\d2 on angle to grain. A subsequent sensitivity analysis of the off-axis tests on the value of \iF\d1\d2 indicated that data from 15° off-axis tensile tests were more stable in establishing \iF\d1\d2 than that of other angles tested: 30°, 45°, and 60°.