کاربرد شبکه عصبی مصنوعی و روش سطح پاسخ در پیش‌بینی و بهینه‌سازی پارامترهای عملکردی و آلایندگی موتور دیزل دوگانه سوز در حضور افزودنی آب

نوع مقاله : مقاله پژوهشی

نویسندگان

1 گروه مهندسی بیوسیستم- دانشگاه محقق اردبیلی

2 گروه مهندسی بیوسیستم-دانشگاه محقق اردبیلی

3 گروه مهندسی بیوسیستم، دانشگاه محقق اردبیلی، اردبیل، ایران

10.22034/jess.2022.334830.1751

چکیده

در این مطالعه، از روش‌های شبکه عصبی مصنوعی (ANN) و سطح پاسخ (RSM) برای مدلسازی و بهینه سازی آزمایش‌های تجربی در راستای بررسی تاثیر گازطبیعی (NG)، آب و بیودیزل حاصل از روغن خوراکی پسماند در فرایند احتراق موتور دیزل تک سیلندر استفاده شد. در ابتدا، آزمایش موتور با دیزل خالص انجام شد و سپس موتور برای کار در حالت دوگانه سوز آماده شد. با استفاده از یک میکسر، گاز طبیعی در منیفولد ورودی با هوا مخلوط شد و امولسیون آب، بیودیزل و دیزل به عنوان سوخت تزریق مستقیم به کارگرفته شدند. سهم انرژی گاز طبیعی در این کار تحقیقاتی از 60 تا 80 درصد متغیر بود. آب با درصد حجمی 1/0 تا 5/0 درصد با مخلوط های دیزل-بیودیزل مخلوط شد. مدل شبکه عصبی مصنوعی برای پیش‌بینی همبستگی بین پاسخ‌های خروجی موتور شامل پارامترهای عملکردی و آلایندگی و عوامل ورودی شامل بار موتور، درصد گاز طبیعی و درصد بیودیزل با استفاده از الگوریتم پس انتشار خطا توسعه داده شده است. روش سطح پاسخ به بهینه سازی پارامترهای ورودی موتور به منظور به حداقل رساندن انتشار و به حداکثر رساندن عملکرد موتور مربوط می شود. تعداد نرون 18 بالاترین دقت، کمترین RMSE (0.490، 6.522، 1.777، 1.507، 103.97 و 1.119) و بالاترین ضریب همبستگی (0.79، 0.99، 0.88، 0.92، 0.95، و 0.85) به ترتیب برای توان ترمزی، مصرف سوخت ویژه ترمزی، بازده حرارتی،BSCO، BSCO2، و BSNOx تولید کرده است. بنابراین تعداد 18 نرون در لایه مخفی به عنوان بهینه تعداد نرون در لایه مخفی انتخاب شد و با توجه به تعداد ورودی ها و خروجی ها، بهترین معماری شبکه، 6-18-3 نامیده شد.

کلیدواژه‌ها


عنوان مقاله [English]

Application of artificial neural network and response surface method in predicting and optimizing performance parameters and pollution of a dual-fuel diesel engine in the presence of water additive

نویسندگان [English]

  • Maghdid Mortaza 2
  • Bahman Najafi 3
2 Department of Biosystem engineering-university of mohaghegh ardabili
3 Faculty of Biosystem Engineering, University of Mohaghegh Ardabili, Ardabil, Iran
چکیده [English]

Abstract

Introduction
Energy use has increased in every aspect of life while the lack of fossil fuels is a major threat to us. Accordingly, the use of alternative energy sources like biofuels is increasingly observed. Vehicles with non-fossil energy sources such as electric vehicles and hybrid vehicles are the future of transportation that will eventually eliminate fossil fuel vehicles.
Water is an additive for biodiesl that has been used in recent years by various researchers in very small amounts in diesel and biodiesel blends. However, this additive has properties that can improve the emission of pollutants. The use of artificial intelligence (AI) methods to model and optimize the performance and emissions of diesel engines have been used by many researchers. In the present study, the effects of biodiesel substitution on the characteristics of a combustion engine in the presence of water and natural gas were analyzed by an artificial neural network (ANN). The present work can be considered new in terms of the strategy used because in the research literature, there is no work similar to the present study. In this study, it has been tried to supply part of the engine power from non-fossil fuels. Although many engine studies have been performed, no significant research work has been described in engine studies with water-biodiesel-diesel mixtures in the presence of natural gas. The present study uses the response level method to validate the predicted output constraints.

Methodology
This study was conducted in three general stages including fuel sample preparation, engine testing and modeling. The first part presents the method of preparing fuel samples and examining their thermal-physical properties. Span and Twin emulsifiers were used to prepare the samples. A ratio of 0.1% was used in all samples to have the same effect of these surfactants in all samples. So that in waterless samples, the mentioned surfactants were added in the same proportion. Micromaginizer technology was used to stabilize the emulsion. The samples were placed in a homogenizer at a rotation speed of 15,000 rpm. The mentioned surfactants are the most effective materials for creating emulsions with suitable stability. The results of the emulsion preparation also clearly confirmed this claim. Samples included blends of B5, B20 and pure diesel. In all three cases, different water ratios (weight percentages of zero, 0.1, 0.2, 0.3, 0.4 and 0.5%) were used. In order to test the engine, the samples were performed on a diesel engine and a dual-burner cylinder with natural gas and constant speed in the renewable energy laboratory of the Biosystem Mechanics Department of Mohaghegh Ardabili University. Were evaluated and tested. First, pure diesel fuel samples were used to determine the baseline data for comparison. Then all the samples were tested in the mentioned engine. To extract the data, the required tools and displays connected to the engine were used, including engine speedometer, torque meter, thermocouple to measure body temperature, inlet air temperature, outlet air temperature, inlet water temperature, outlet water temperature, exhaust gas temperature. It was a pollutant and a flow meter. The modeling process was performed to correlate the independent and dependent variables. The main purpose of using soft computing methods to develop a model was to eliminate the dependence of this relationship on complex mathematical models. Matlab software was used to develop the models. Modeling was done in two stages of training and network testing. The network training stage is an important step in the formation of modeling.

Conclusion
The first part of the results examines the physical-thermal properties of fuel samples. Table 2 shows the results of the study of the thermophysical properties of fuels. Based on the obtained results, it can be seen that increasing the amount of water in the fuel samples did not change the density and viscosity of the fuels. In diesel fuel, the density values decreased with increasing water content in the fuel sample from zero to 0.1%, then increased with increasing water content from 0.1 to 0.3% and then increased from 0.4 to 0.5% relative to diesel. In B5 fuel sample, the presence of water by 0.1% first increased the density compared to B5 fuel, then by increasing the amount of water from 0.1% to 0.5%, the density decreased. But in the case of the B20, no specific trend can be observed. In terms of viscosity, in diesel fuel, the presence of water generally causes an increasing trend from 0.1 to 0.3% and a decreasing trend from 0.3 to 0.5%, but in all cases, compared to diesel fuel, the viscosity of water-containing samples increases. Finds. But in the case of fuel samples B5 and B20, in general, the viscosity of the fuel sample in the presence of water is less than that of fuel samples B5 and B20, but there is no specific trend. Therefore, it can be concluded that in the presence of biodiesel in the fuel sample, the presence of water reduces the viscosity, which is not observed in pure diesel fuel. The presence of water in the fuel sample reduces the calorific value of the fuel due to the low calorific value of the water. Increasing the percentage of biodiesel and water does not have a significant effect on changes in brake specific fuel consumption. But as expected, the increase in the percentage of natural gas replacement has significantly reduced the consumption of special brake fuel. One of the main reasons is the significant calorific value of natural gas compared to biodiesel and water compared to diesel fuel. It can be seen that increasing the percentage of biodiesel slightly increases and then decreases the thermal efficiency. The highest thermal value occurs in the range of 8 to 12% of biodiesel. Increasing the water content by a very small amount initially reduces the thermal efficiency and then increases it relatively. But increasing the percentage of natural gas has significantly increased the thermal efficiency. One of the main reasons is the high calorific value of natural gas, which has a significant effect on increasing the pressure inside the cylinder chamber and, consequently, increasing the thermal efficiency.

کلیدواژه‌ها [English]

  • Biodiesel
  • Water
  • Diesel engine
  • Artificial intelligence
  • Natural gas