تخمین عملکرد موتور دوگانه سوز با بیوگاز و بیودیزل با به کارگیری روش تطبیقی عصبی-فازی

Energy consumption in developed countries increases every year by about 1% and in developing countries by about 5%, and a large share of the world's energy source is fossil fuels [1-3]. Increasing global energy demand, reducing resources. Fossil and environmental problems (such as atmospheric pollution, greenhouse effect and global warming) have led to the search for alternative renewable energy sources with systems that are more efficient and less emitting. Among renewable energy sources, biogas is one of the most suitable options for heat and electricity generation applications [3]. Depending on the amount of methane in the biogas content, it can be used in many applications such as cooking, cooling, heating and power generation. Using biogas as a renewable fuel in internal combustion engines has significant potential in meeting the increased energy demand. Today, many studies have been conducted on the thermal use of biogas as well as in internal combustion engines (4-6). In these studies, the use of biogas and biogas-fossil fuel mixture in compression combustion engines, engine performance and fuel consumption parameters have been investigated [7-13].
Chai et al. used an artificial neural network model for a gasoline engine to estimate the main engine performance parameters. A standard artificial neural network model using back propagation algorithm was developed for the engine using experimental data of engine speed, torque, fuel flow rate, average inlet manifold temperature and coolant inlet temperature. Later, specific fuel consumption, effective power and exhaust temperature were estimated by artificial neural network and the results were compared with experimental results. The coefficient of explanation for the test and training data was about 0.99. The error value was calculated to be less than 0.02 and the average error of the test data was shown to be less than 2.7%. It is concluded that using an artificial neural network model can be a good choice for predicting the performance of an engine with high accuracy [4]. To create a better blend of diesel and biodiesel to improve power, torque, hourly specific fuel consumption and brake specific fuel consumption, Ogoz et al. studied using artificial neural network. The properties of the resulting mixed fuels were determined and used as reference values ​​for training the artificial neural network. Reference values ​​obtained from experiments in artificial neural network were used to estimate power, torque, hourly fuel consumption and specific brake fuel consumption and the estimated results were compared with the experimental results. The reliability of the study was calculated to be 99.94% [15]. Parlock et al concluded that a good neural network is a fast, consistent and easy tool for solving engineering problems. Using a back-propagation learning algorithm with an artificial neural network model, they estimated the brake-specific fuel consumption values ​​and exhaust temperature of a diesel engine. In this study, the mean absolute relative error was reported to be less than 2% [16].
• Experiment setup and experimental measurements In this study, a pilot-scale biogas system for biogas production with simultaneous fermentation method was built using 40% animal manure, 35% water, 17% whey and 8% whey. Is. Washing and desulfurization processes are applied to remove polluting gases from the produced biogas. At the end of the purification process, biogas with 50, 75 and 95% methane content is obtained. The biogas is repeatedly washed in the purification unit to remove carbon dioxide until the methane concentration reaches the desired percentage. The produced biogas is tested using a biogas generator with a 10 kW spark ignition engine. Tests are performed at engine speed of 1500 rpm with generator load at 1.5-3-4.5-6 7.5-9 kW. The composition of the biogas used in the experiments is determined using a portable biogas analyzer model Geotech GA2000.To measure the cylinder pressure, an Oprand Auto PSI TC candle is used, which can measure the pressure from 0 to 200 bar and in At the same time, it also acts as a candle. Kubler coder model Sendix 5000 was used to determine the position of the piston. With the help of measured pressure and piston position data, the change in cylinder pressure relative to piston position is determined. The exhaust gas temperature is measured using a K-type thermocouple at the outlet of the exhaust manifold.The engine is measured and observed during the test. The test results are used for Anfis training. Brake-specific fuel consumption, a key parameter that determines engine characteristics, can be expressed as the amount of fuel consumed per unit of power obtained from the engine.Thermal efficiency is the ratio of the heat converted into useful work by the engine to the total heat produced by burning the fuel. Because a significant amount of heat is removed from the engine through the engine's cooling and lubrication system, only about a third of the heat generated is converted to power.
The important results of the study are given below:
A- It has been shown that the Enfis approach can be a choice for effectively predicting the performance conditions of spark ignition engines.
B- The best values ​​of R for estimating thermal efficiency, specific braking fuel consumption and volumetric efficiency are 0.9901, 0.9594, 0.9608, which are in the acceptable range.
c-c The calculated MAPE and RMSE values ​​also show that the estimated performance of the developed models is high.
D- It has been determined that Anfis can be used to estimate the specific braking fuel consumption values, thermal efficiency and volume efficiency with high accuracy without the need to perform complex and timely studies.
Further studies can be done in the following areas.
A- The results obtained with the present study can be compared with the results obtained from different educational algorithms of Anfis.
B- In addition to Anfis models, the accuracy of developed models can be improved by using other estimation methods such as integrated methods with meta-heuristic and Anfis algorithms.
C- In addition to the estimations of thermal efficiency, specific braking fuel consumption and volume efficiency, exhaust emission values ​​(CO, HC) can be estimated.
D- If a different experimental study is conducted for different methane content, engine load, intake air temperature, etc., the models can be retrained and their results checked.