Optimization of Hydrogen-fueled Engine Ignition Timing Based on L-M Neural Network Algorithm

Abstract

In view of the improvement measures of the optimization control algorithm for the ignition system of the hydrogen-fueled engine, the L-M neural network algorithm, Powell neural network algorithm and the Keywords: Hydrogen-fueled Engine, L-M Algorithm, Neural Network, OptimizationCopyright © 2016 Universitas Ahmad Dahlan. All rights reserved. IntroductionResearch on hydrogen as fuel of internal combustion engine began in the middle of nineteenth century. It fell behind for about 100 years compared with the research on traditional internal combustion engine. Although at this stage, the traditional gasoline engine and diesel engine technologies become more and more perfect and mature, but it's unable to directly use the traditional internal combustion engine for hydrogen combustion due to the particularity of hydrogen as a fuel. Its technology is not perfect and there is no commercial promotion. However, hydrogen is considered to be the most promising clean energy source in this century, with its advantages of low emission, renewable energy and so on. In the process of combustion, the ignition limit of hydrogen is wide, the combustion speed is high and the calorific value is high. This results pre-ignition, inlet backfire and other abnormal combustion phenomenon in the combustion process of internal combustion engine, so that the engine cannot operate normally [1]. How to avoid the abnormal combustion phenomenon without reducing the output power of the hydrogen-fueled engine is the research focus of the relevant researchers in various countries.Changwei, et al., have studied on the effect of hydrogen and methanol mixture combustion on the emission characteristics of hydrogen-fueled engine [2]. J.M. Gomes Antunes, et al., have studied on the influence of the high pressure direct injection technology on the hydrogen combustion characteristics of diesel engine, and the influence of homogeneous charge compression ignition technology and hydrogen injection timing and length on the performance of the hydrogen-fueled engine [3]. Xing-hua, Fu-shui Liu, et al., have studied the effects of hydrogen injection timing under different rotation speed and equivalence ratio on the formation of hydrogen mixture gas and its effect on the prevention of backfireby using computational fluid dynamics simulation [4]. The effect of EGR system on the combustion and performance of a compression ignition hydrogen-fueled engine was studied by Vinod Yadav Singh, et al., [5]. However, these studies are based on a large number of bench test or simulation process, and only a part of the operating conditions of the hydrogen-fueled engine characteristics are studied, which can't take into account the operation, emission and combustion characteristics of all the working conditions of the hydrogen-fueled engine.

The methods of duo output neural network ensemble for prediction of coronary heart disease

“…The DOANNE model uses a multi-layer perceptron architecture with the Levenberg-Marquard (LM) training algorithm. The LM algorithm is designed using a second derivative approach without the need to compute the Hessian matrix [22]. The LM algorithm has advantages in terms of the speed of the training process, compared to a number of other algorithms [23].…”

Research and Development of Hydrogen-Fueled Internal Combustion Engines in China

“…Yang et al 94 (2003) as the above experimental analysis and mathematical models the combustion mechanism of engines is very complex, and the high nonlinearity and imprecision of engine modeling make it difficult to establish analytical expressions in the state space; fuzzy mathematics, artificial neural networks, and computational science may be more suitable Yang et al 95 (2008) a single-cylinder hydrogen-fueled engine, which is modified from a gasoline engine; the cylinder diameter × stroke is 84 × 90 mm, and the combustion chamber is half ball, and the compression ratio is 8.5 mathematical model analysis propose a fully automatic optimization control algorithm that combines Pareto improvement with local decision-making; the multiobjective parallel optimization process in the optimization process can effectively achieve the combination of single optimization and combined performance optimization; in the fitness function, add the weighted sum of multiple objective values to the fitness function; the weighting function of each objective varies with changes in operating status; therefore, the optimization process of hydrogen fuel engines develops toward dynamic optimization of hydrogen fuel engine performance based on the operating status of the engine Yang et al 96 (2008) as the above experimental simulation propose an optimization control algorithm that combines genetic algorithm to unify the control of variables, objectives, and operating conditions into a multiobjective integrated optimization control technology Wang et al 97 (2010) as the above experimental simulation by adding genetic algorithms, the MAP calibration process can be simplified, the calibration speed and accuracy can be improved, and the global optimal solution can be obtained Wang et al 98 (2012) as the above experimental simulation a state space model for the nonlinear combustion control system of hydrogen fuel engines has been established, achieving an equivalent transformation from the state space model to the optimal control model represented by the main operating parameters and performance indicators; this lays the foundation for improving performance indicators and eliminating abnormal combustion by optimizing control operating parameters Wang et al 100 (2017) as the above control method model and genetic algorithm an optimization model and method based on genetic information fusion algorithm have been proposed; by optimizing the operating parameters of excess air coefficient and ignition advance angle, the economy and power performance of hydrogen engines will be improved; the experiment shows that the method used in this Review is reliable and practical Wang et al 99 (2016) a single-cylinder, four-stroke hydrogen-fueled engine; the engine is water-cooled horizontal type; its displacement is 589 cm 3 , bore × stroke: 94 × 85 mm control method model and neural network a hydrogen fuel engine ignition timing optimization model based on L−M neural network algorithm was constructed to achieve calibration and optimization control of hydrogen engine ignition timing under all operating conditions Yang et al 101 (2022) as the above control method model and improved genetic algorithm with the delay of hydrogen injection timing and ignition timing, the indicated power and indicated thermal efficiency first increase and then decrease; in terms of emissions, the NO emissions at medium loads are significantly higher than those at high and low loads; in addition, with the delay of hydrogen inject...…”

“…In addition, Wang et al 99 constructed the ignition timing optimization model of hydrogen-fueled engines through adopting the L–M neural network algorithm. As compared to the BP algorithm and Powell algorithm, the L–M algorithm had a better optimization effect.…”

“…The calibration and optimization control of the ignition timing of the engines at full working conditions were realized. Wang et al 100 then proposed an optimization model and optimization method based on the genetic information fusion algorithm. The experiments showed that the economic and dynamic performance of the hydrogen engine would be enhanced through optimizing two operating parameters of λ and θ ig .…”

The Analysis of Performace Model Tiered Artificial Neural Network for Assessment of Coronary Heart Disease

“…Artificial neural network with Levenberg-Marquardt training algorithm designed using the second derivative approach without having to calculate the hessian matrix [20]. The Hessian matrix can be approximated using Equation (2).…”