Determination Of Optimal Parameters Research Articles

Optimization of process parameters in Nd:YAG laser micro-drilling of graphite/epoxy based polymer matrix composite using Taguchi based Grey relational analysis

This paper presents a Taguchi based grey relational analysis methodology to find out the optimum process parameters that minimize the taper, heat affected zone, and maximize the MRR in a pulsed Nd:YAG laser micro-drilling on a novel polymer matrix composite introducing Graphite as filler materials. The input process parameters such as cutting speed, air pressure, lamp current and pulse width with five levels is used to conduct the experiments based on Taguchi’s L25 orthogonal array. Grey relational grades are calculated and used for determination of optimal process parameters and their effects on the output responses. The accuracy of the preservative model is demonstrated through the confirmation test.

Search for optimal quadrature parameters by means of metaheuristic algorithms

Element stiffness matrices in the FEM are usually calculated using Gauss-Legendre quadrature. It is well known that the accuracy of the quadrature of an element depends on the shape of the element. Deep learning can be used to predict optimal quadrature parameters for each element to improve the accuracy of its numerical quadrature. To prepare training patterns for deep learning, we can use various metaheuristic algorithms, such as GA and PSO, for solving optimization problems in order to find optimal quadrature parameters for a lot of elements of various shape. Considering the use of the results of the search as training patterns for deep learning, however, additional constraints should be taken into account. In this paper, the feasibility of some metaheuristic algorithms for finding optimal quadrature parameters is investigated.

Analysis of digital filters used in time-series small heat flux measurement

• Four statistical assessment metrics are used to evaluate the filtering effects. • Filtering effects of six filters with different operating parameters are compared. • Intersection region method is firstly proposed to get optimal operating parameter. • Moving average filter shows a high repeatability and fidelity simultaneously. Noise fluctuation is the most common issue of small heat flux measurement and it disturbs the measurement precision. In the previous studies, how to select an appropriate filter and determine its optimal operating parameter was not given systematically. In this paper, the aim is to explore an effective filter selection and optimal operating parameter determination method for the small heat flux signal. Firstly, six filters with the large range operating parameters were chosen to remove the noise of a set of 10,000 numbers time-series heat flux in the range of 5–10 W/m 2 . Then, signal-to-noise ratio, standard deviation, mean absolute deviation and correlation coefficient were introduced to comprehensively assess the filtering results. Lastly, the intersection region method is firstly proposed to determine the optimal operating parameters based on the tradeoff between repeatability and fidelity. The numerical results indicated that the signal-to-noise ratio of the sampled heat flux was ranged from 13.310 db to 20.703 db. The standard deviation and correlation coefficient was decreased from 1.111 to 0.093–0.672 and from 1 to 0.640–0.883, while the mean absolute deviation was increased from 0 to 0.498–1.056. When the filter is fixed, the repeatability and fidelity levels of filtered data are contradictory. When comparing the filtering effects of filters, some filters simultaneously showed a high repeatability and fidelity levels. For the heat flux signal in this study, the moving average filter with 51 window width and Savitzky-Golay filters with 101 window width and three highest order are the best choice. The proposed intersection region method with four assessment metrics is effectively to find an appropriate filter with high repeatability and fidelity.

Моделирование нестационарной теплопередачи в многослойной биметаллической пластине

When using the method of active thermal non-destructive testing to detect latent defects in multilayer bimetallic products, the adequate choice of technological modes of testing is of great importance. One of the important parameters is the optimum observation time, at which a sufficient amplitude of the temperature signal over the defect is provided. This value depends on the controlled object geometry, its thermophysical characteristics, and also selected testing modes and procedures. Various experimental stu-dies are used to determine the optimum observation time. Preparing bimetallic plates samples for full-scale experiments is a technologically complex and expensive task. Therefore, a large amount of preparatory research is carried out in the form of a numerical experiment, during which, with certain assumptions, changing the thermal state of the control object is simulated. Aim. Determination of optimal parameters of the thermal NDT process for a three-layer steel-aluminum plate with a planar delamination defect between the metal layers. For different variants of mutual location of the temperature registration device and source of heat flux of different power it is necessary to determine the heating time at which the sufficient amplitude of temperature signal is observed. Materials and methods. The methods of mathematical and computer modeling were used. Solving the task by the finite element method with the Agros2D code is given. Results. The calculation data have been obtained, which are used as the basis for constructing the temperature signal distribution over the bimetallic plate surface above the defect under different heating modes, heat flux rate, relative position of the heat stimulation source and temperature registration device, as well as determining the temperature signal dependence over the defect on the heating time. Conclusion. On the basis of modeling the temperature dependences necessary for interpretation of the results of thermal NDT of multilayered bimetallic plates have been determined, as well as the relations between the parameters of the heating process, which contribute to obtaining a temperature signal of sufficient amplitude. The results can be used for further experimental studies of the active thermal NDT processes of multilayer bimetallic materials.

Junk-neuron-deletion strategy for hyperparameter optimization of neural networks

With the complexity of problems in reality increasing, the sizes of deep learning neural networks, including the number of layers, neurons, and connections, are increasing in an explosive way. Optimizing hyperparameters to improve the prediction performance of neural networks has become an important task. In literatures, the methods of finding optimal parameters, such as sensitivity pruning and grid search, are complicated and cost a large amount of computation time. In this paper, a hyperparameter optimization strategy called junk neuron deletion is proposed. A neuron with small mean weight in the weight matrix can be ignored in the prediction, and is defined subsequently as a junk neuron. This strategy is to obtain a simplified network structure by deleting the junk neurons, to effectively shorten the computation time and improve the prediction accuracy and model the generalization capability. The LSTM model is used to train the time series data generated by Logistic, Henon and Rossler dynamical systems, and the relatively optimal parameter combination is obtained by grid search with a certain step length. The partial weight matrix that can influence the model output is extracted under this parameter combination, and the neurons with smaller mean weights are eliminated with different thresholds. It is found that using the weighted mean value of 0.1 as the threshold, the identification and deletion of junk neurons can significantly improve the prediction efficiency. Increasing the threshold accuracy will gradually fall back to the initial level, but with the same prediction effect, more operating costs will be saved. Further reduction will result in prediction ability lower than the initial level due to lack of fitting. Using this strategy, the prediction performance of LSTM model for several typical chaotic dynamical systems is improved significantly.

Compact and Integrated High-Power Pulse Generation and Forming System

This paper presents comprehensive analytical, numerical and experimental research of the compact and integrated high-power pulse generation and forming system based on the flux compression generator and the electro-explosive forming fuse. The paper includes the analysis of the presented solution, starting from the individual components studies, i.e., the separate flux compression generator tests in field conditions and the forming fuse laboratory test, through the formulation of the extended quasi-empirical components models aimed at enabling their optimal parameters determination at the early design stage and ending with the description of the integrated system studies in field conditions. Based on detailed research, it was possible to achieve very high parameters of the generated pulses, i.e., overvoltages of up to 340 kV with the available source power reaching 25 GW. A very high convergence of the simulation and the results of experimental research has been obtained. The parameters of the presented system have been compared with other literature solutions and the selected topology of the high power pulse generation and forming system has been distinguished against other available ones, e.g., based on Marx generators and forming lines.

Reducing the flaws of 08Yu steel sheet blanks, designed for cold stamping products

The aim of the work is to develop a method of increasing the technological plasticity while maintaining the strength and preventing aging of blanks of cold-rolled sheet steel 08Yu. The task of this work is research and determination of optimal temperature-time parameters of high-speed contact recrystallization annealing and subsequent aging of 08Yu steel sheet blanks to obtain the required level of mechanical properties that would improve their deformability and ability of very deep drawing during cold stamping of products. The optimal parameters of the speed mode are set recrystallization annealing of cold-rolled 08Yu sheet steel, which provides the best structure and properties for further cold pressure treatment. The temperature-time modes of further aging and possibilities to prevent the processes of its natural aging during prolonged operation or transportation have been studied. The method and modes of heat treatment of blanks from finished 08Yu sheet steel to facilitate their deformability and improve stamping in order to reduce waste in the manufacture of products by cold deformatiion with deep and complex drawing were suggested.

Some results of voltage increase in marine electrical propulsion installations

The main voltage parameters of marine electrical propulsion installations, at various load capacities, with further stress assessment are discussed in the article. By considering voltage options, it was revealed that an increase in inductive resistance makes it possible to manufacture a machine with high power, but the quality of electrical energy in transient modes is deteriorated. During starts, reversals, resets and load surges of powerful consumers, large voltage dips will occur on the generator. To solve the problem, including determination of optimal parameters of the generator, based on the multifactorial nature of the task of creating an electric machine for a specific purpose, an analysis of all the necessary parameters was carried out. The main dimensions of the machine depend on its full power, angular speed and electromagnetic loads. The article presents technical solutions considered using the example of a modern multipurpose icebreaker-supply “Vitus Bering”. Also, as an example, an icebreaker of the Taimyr type with a limited draft is considered, for the first time in modern practice, an alternating-changing current EDMS is used. From the experience of operating the icebreakers “Taimyr” and “Vaigach”, the transition to high voltage did not lead to a complication and increase in the complexity of the maintenance of the MPM, although it required strict compliance with safety regulations during the work. During the operation of icebreakers with MPM 6.3 kV, not a single case of single-phase short circuit or electrical injury has been recorded.

Analysis of the Impact of Randomization of Search-Control Parameters in Monte-Carlo Tree Search

Monte-Carlo Tree Search (MCTS) has been applied successfully in many domains, including games. However, its performance is not uniform on all domains, and it also depends on how parameters that control the search are set. Parameter values that are optimal for a task might be sub-optimal for another. In a domain that tackles many games with different characteristics, like general game playing (GGP), selecting appropriate parameter settings is not a trivial task. Games are unknown to the player, thus, finding optimal parameters for a given game in advance is not feasible. Previous work has looked into tuning parameter values online, while the game is being played, showing some promising results. This tuning approach looks for optimal parameter values, balancing exploitation of values that performed well so far in the search and exploration of less sampled values. Continuously changing parameter values while performing the search, combined also with exploration of multiple values, introduces some randomization in the process. In addition, previous research indicates that adding randomization to certain components of MCTS might increase the diversification of the search and improve the performance. Therefore, this article investigates the effect of randomly selecting values for MCTS search-control parameters online among predefined sets of reasonable values. For the GGP domain, this article evaluates four different online parameter randomization strategies by comparing them with other methods to set parameter values: online parameter tuning, offline parameter tuning and sub-optimal parameter choices. Results on a set of 14 heterogeneous abstract games show that randomizing parameter values before each simulation has a positive effect on the search in some of the tested games, with respect to using fixed offline-tuned parameters. Moreover, results show a clear distinction between games for which online parameter tuning works best and games for which online randomization works best. In addition, the overall performance of online parameter randomization is closer to the one of online parameter turning than the one of sub-optimal parameter values, showing that online randomization is a reasonable parameter selection strategy. When analyzing the structure of the search trees generated by agents that use the different parameters selection strategies, it is clear that randomization causes MCTS to become more explorative, which is helpful for alignment games that present many winning paths in their trees. Online parameter tuning, instead, seems more suitable for games that present narrow winning paths and many losing paths.

Forecasting Monthly Rainfall using Bio-Inspired Artificial Algae Deep Learning Network

Rainfall forecasting is the most critical and challenging task because of its dependence on different climatic and weather parameters. Hence, robust and accurate rainfall forecasting models need to be created by applying various machine learning and deep learning approaches. Several automatic systems were created to predict the weather, but it depends on the type of weather pattern, season and location, which leads in maximizing the processing time. Therefore, in this work, significant artificial algae long short-term memory (LSTM) deep learning network is introduced to forecast the monthly rainfall. During this process, Homogeneous Indian Monthly Rainfall Data Set (1871–2016) is utilized to collect the rainfall information. The gathered information is computed with the help of an LSTM approach, which is able to process the time series data and predict the dependency between the data effectively. The most challenging phase of LSTM training process is finding optimal network parameters such as weight and bias. For obtaining the optimal parameters, one of the Meta heuristic bio-inspired algorithms called Artificial Algae Algorithm (AAA) is used. The forecasted rainfall for the testing dataset is compared with the existing models. The forecasted results exhibit superiority of our model over the state-of-the-art models for forecasting Indian Monsoon rainfall. The LSTM model combined with AAA predicts the monsoon from June–September accurately.

Combined Voltage and Frequency Control for Diverse Standalone Microgrid Networks Using Flexible IDC with Novel FOC: A Real-Time Validation

In this paper, a maiden application of combined voltage and frequency control ensures voltage, frequency, and power regulation in diverse standalone low voltage AC microgrid (MG) networks by combining flexible improved droop control (IDC) with a novel fractional-order controller (FOC) proposed. An effort is made to design an IDC into the control loop based on virtual power, virtual negative resistance, inductance, and FOC to ensure power decupling and compensate voltage source converter (VSC) based distributed generators feeder voltage drop. A maiden proportional fractional integral derivative filter plus double derivative controller is used to precisely monitor VSC voltage and current reference value for reducing the control errors. The controller robustness regarding system parameters uncertainties, unmodeled dynamics, and disturbance in the MG system is ensured by a new memetic salp swarm algorithm by finding optimal controller parameters. Several simulation studies in the MATLAB Simulink environment were conducted to verify the efficacy of the proposed technique. The suggested control methodology is also investigated by hardware in the loop using the OPAL-RT real-time simulator platform to demonstrate the controllers practical functional realization. The findings confirmed the efficacy of the proposed strategy.

Optimal operating parameter determination based on fuzzy logic modeling and marine predators algorithm approaches to improve the methane production via biomass gasification

The current study is related to optimizing methane production from steam gasification of palm kernel shell (PKS) utilizing coal bottom ash as a catalyst. Based on the experimental dataset, the fuzzy logic technique, represented by the Adaptive Network-based Fuzzy Inference System (ANFIS) structure, is used to create a robust model to simulate methane production via biomass gasification. Then, a Marine Predator Algorithm (MPA) is used to determine the optimal operating parameter of the gasification process. The temperature, particle size, CaO/PKS ratio, and coal bottom ash are used as the decision variables, whereas methane production is used as the objective function. The main findings confirmed that the yield of methane gasification reached 52.82 vol% when the associated temperature, particle size, CaO/PKS Ratio, and coal bottom ash are at 678 °C, 0.42 mm, 3.03, and 0.037 wt%, respectively. Accordingly, the proposed strategy provides a better result than the Analysis of Variance (ANOVA) methodology reported in the literature. Furthermore, the proposed strategy's output exceeds both the results obtained experimentally and ANOVA methods by 20.26% and 29.69%, respectively.

Differentiable molecular simulation can learn all the parameters in a coarse-grained force field for proteins.

Finding optimal parameters for force fields used in molecular simulation is a challenging and time-consuming task, partly due to the difficulty of tuning multiple parameters at once. Automatic differentiation presents a general solution: run a simulation, obtain gradients of a loss function with respect to all the parameters, and use these to improve the force field. This approach takes advantage of the deep learning revolution whilst retaining the interpretability and efficiency of existing force fields. We demonstrate that this is possible by parameterising a simple coarse-grained force field for proteins, based on training simulations of up to 2,000 steps learning to keep the native structure stable. The learned potential matches chemical knowledge and PDB data, can fold and reproduce the dynamics of small proteins, and shows ability in protein design and model scoring applications. Problems in applying differentiable molecular simulation to all-atom models of proteins are discussed along with possible solutions and the variety of available loss functions. The learned potential, simulation scripts and training code are made available at https://github.com/psipred/cgdms.

AGRICULTURAL DATA ANALYSIS

In agriculture sector where farmers and agribusinesses have to make innumerable decisions every day and intricate complexities involves the various factors influencing them. An essential issue for agricultural planning intention is the accurate yield estimation for the numerous crops involved in the planning. Data mining techniques are necessary approach for accomplishing practical and effective solutions for this problem. Agriculture has been an obvious target for big data. Environmental conditions, variability in soil, input levels, combinations and commodity prices have made it all the more relevant for farmers to use information and get help to make critical farming decisions. This paper focuses on the analysis of the agriculture data and finding optimal parameters to maximize the crop production using Machine learning techniques like random forest regressor and Linear Regression. Mining the large amount of existing crop, soil and climatic data, and analysing new, non-experimental data optimizes the production and makes agriculture more resilient to climatic change.

Performance portability through machine learning guided kernel selection in SYCL libraries

Automatically tuning parallel compute kernels allows libraries and frameworks to achieve performance on a wide range of hardware, however these techniques are typically focused on finding optimal kernel parameters for particular input sizes and parameters. General purpose compute libraries must be able to cater to all inputs and parameters provided by a user, and so these techniques are of limited use. Additionally parallel programming frameworks such as SYCL require that the kernels be deployed in a binary format embedded within the library. As such it is impractical to deploy a large number of possible kernel configurations without inflating the library size.Machine learning methods can be used to mitigate against both of these problems and provide performance for general purpose routines with a limited number of kernel configurations. We show that unsupervised clustering methods can be used to select a subset of the possible kernels that should be deployed and that simple classification methods can be trained to select from these kernels at runtime to give good performance. As these techniques are fully automated, relying only on benchmark data, the tuning process for new hardware or problems does not require any developer effort or expertise. We demonstrate that this technique gives competitive performance to vendor specific libraries when used in inference of a large neural network.

Determination of Optimal Process Parameters for Plastic Injection Molding of Polymer Materials Using Multi-Objective Optimization

Determination of Optimal Process Parameters for Plastic Injection Molding of Polymer Materials Using Multi-Objective Optimization

ИССЛЕДОВАНИЕ ПАРАМЕТРОВ ЗАДАЧИ СПЛАЙНОВОЙ АППРОКСИМАЦИИ ЗАШУМЛЕННЫХ ДАННЫХ ЧИСЛЕННЫМИ МЕТОДАМИ ОПТИМАЛЬНОГО УПРАВЛЕНИЯ

Currently, the problems of distortion of measurement data by noise and the appearance of un-certainties in quality criteria have caused increased interest in research in the field of spline approx-imation. At the same time, existing methods of minimizing empirical risk, assuming that the noise is a uniform distribution with heavier tails than Gaussian, limit the scope of application of these studies. The problem of estimating noise-distorted data is usually based on solving an optimi-zation problem with a function containing uncertainty arising from the problem of finding optimal parameters. In this regard, the estimation of distorted noise cannot be solved by classical methods. Aim. This study is aimed at solving and analyzing the problem of spline approximation of data under uncertainty conditions based on the parametrization of control and the gradient projec-tion algorithm. Methods. The study of the problem of spline approximation of noisy data is carried out by the method of approximation of the piecewise constant control function. In this case, para-metrization of the control is possible only for a finite number of break points of the first kind. In the framework of the experimental study, the gradient projection algorithm is used for the numerical solution of the spline approximation problem. The proposed methods are used to study the parameters of the problem of spline approximation of data under conditions of uncertain-ty. Results. The numerical study of the control parametrization approach and the gradient projec-tion algorithm is based on the developed software and algorithmic tool for solving the problem of the spline approximation model under uncertainty. To evaluate the noise-distorted data, numerical experiments were conducted to study the model parameters and it was found that increasing the value of the parameter α leads to an increase in accuracy, but a loss of smoothness. In addition, the analysis showed that the considered distribution laws did not change the accuracy and convergence rate of the algorithm. Conclusion. The proposed approach for solving the problem of spline approx-imation under uncertainty conditions allows us to determine the problems of distortion of measure-ment data by noise and the appearance of uncertainties in the quality criteria. The study of the model parameters showed that the constructed system is stable to the error of the initial approxima-tion, and the distribution laws do not significantly affect the accuracy and convergence of the gra-dient projection method.

Features of Heat Treatment the Ti-6Al-4V GTD Blades Manufactured by DLD Additive Technology.

Additive manufacturing of titanium alloys is one of the fastest growing areas of 3D metal printing. The use of AM methods for parts production in the aviation industry is especially promising. During the deposition of products with differently sized cross-sections, the thermal history changes, which leads to non-uniformity of the structure and properties. Such heterogeneity can lead to failure of the product during operation. The structure of deposited parts, depending on the thermal cycle, may consist of α’, α + α’ + β’, and α + β in different ratios. This problem can be solved by using heat treatment (HT). This paper presents research aimed towards the determination of optimal heat treatment parameters that allows the reception of the uniform formation of properties in the after-treatment state, regardless of the initial structure and properties, using the example of a deposited Ti-6Al-4V gas turbine blade.

Multi-objective optimization of an ORC power plant using one-dimensional design of a radial-inflow turbine with backswept rotor blades

Organic Rankine cycle (ORC) power plants are a promising technology for converting low and medium temperature energy sources into electricity. In order to fully exploit the potential of an ORC, the components of the system should be designed appropriately. One of the key elements of the ORC installation, substantially affecting the system efficiency, is a turboexpander. This study concerns a multi-objective optimization of an ORC coupled with a one-dimensional radial-inflow turbine (RIT) model. Particular attention is given to determination of optimal parameters which have an impact on the turbine geometry of the nozzle and the rotor. The nozzle pitch to chord ratio and the blade angle at the rotor inlet are investigated as one of the decision variables. The first parameter is rarely discussed in most studies and it is either a fixed value or its optimum is determined using correlations developed for axial-flow turbines. Applying the blade angle as a variable is associated with the use of backswept rotor blades, which is achievable in the ORC and results in higher turbine efficiency. By means of Non-dominated Sorting Genetic Algorithm-II (NSGA-II) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), the optimal design point has been determined. It was found that by applying the blade angle at the rotor inlet as the decision variable, the optimal blade sweep design featuring a 0.00% incidence enthalpy loss can be obtained. Based on the parametric analysis, it was reported that by making wrong design choices with respect to the RIT decision variables, the turbine efficiency may be decreased by almost 8.00%, leading to a decrease in the net power output by more than 8.00% and an increase in the electricity production cost of more than 7.00%.

A fault information-guided variational mode decomposition (FIVMD) method for rolling element bearings diagnosis

Being an effective methodology to adaptatively decompose a multi-component signal into a series of amplitude-modulated-frequency-modulated (AMFM) sub-signals with limited bandwidth, the variational mode decomposition (VMD) has received increasing attention in the diagnosis of rolling element bearings. In implementing VMD, an optimal determination of decomposition parameters, including the mode number and bandwidth control parameter, is the pivotal starting point. However, in practical engineering, heavy background noise, abnormal impulses and vibration interferences from other internal components, often bring great challenges in selecting mode number and bandwidth control parameter. These issues may lead to the performance degradation of VMD for bearing fault diagnosis. Therefore, a fault information-guided VMD (FIVMD) method is proposed in this paper for extracting the weak bearing repetitive transient. To minimize the effects of background noise and/or interferences from other components, two nested statistical models based on the fault cyclic information, incorporated with the statistical threshold at a specific significance level, are used to approximately determine the mode number. Then the ratio of fault characteristic amplitude (RFCA) is defined and utilized to identify the optimal bandwidth control parameter, through which the maximum fault information is extracted. Finally, comparisons with the original VMD, empirical mode decomposition (EMD) and local mean decomposition (LMD) are conducted using both simulation and experimental datasets. Successful fault diagnosis of rolling element bearings under complicated operating conditions, including early bearing fault signals in run-to-failure test datasets, signals with impulsive noise and planet bearing signals, demonstrates that the proposed FIVMD is a superior approach in extracting weak bearing repetitive transients.