چکیده مقاله
In this paper, the process of biodiesel production from the microalgae Scenedesmus in a batch reactor was modeled using a hybrid approach First, a first principles model of the process was developed using reaction rate equations and mass balance equations in the presence of an acidic catalyst and methanol Subsequently, 6,000 data points were generated using this model and selected through the Latin Hypercube Sampling method Long Short Term Memory Neural Network was then employed for data driven black box modeling of the process Finally, the process outputs of the data driven model were compared with the first principles model using error metrics such as Mean Square Error MSE and the coefficient of determination R2 The input variables for this process included time and the concentrations of triglyceride, methanol, diglyceride, monoglyceride, biodiesel, and glycerin at time t The output variables were the concentrations of the same components at time t ∆t The data driven model achieved an R2 greater than 0 99 and an MSE on the order of 10 7 for the outputs, indicating its high accuracy and potential to replace the first principles model
کلیدواژهها
نویسندگان
Mohsen Mokari
Research Assistant, Amirkabir University of Technology Tehran Polytechnic
شیوه ارجاع
Mokari, Mohsen,1403,Application of Long Short-Term Memory Neural Network in modelling batch reactor of the biodiesel production by Scenedesmus microalgae,The 24th National Conference on Civil Engineering, the 24th National Conference on Electrical, Computer and Urban Engineering Engineering,Shirvan
ارائهشده در
مجموعه مقالات بیست و چهارمین کنفرانس ملی مهندسی برق، کامپیوتر و مکانیک18 دی 1403 · شیروان