Scholarlink

<p>Improving the performance of small-scale food mills requires a predictive approach and an examination of the interrelation between milling parameters. The aim of this paper was to evaluate and optimize the performance of small scale food grinding machines by providing a multiple regression models based on various frequent used grinding parameters. To conduct this research, data on milling parameters were collected through questionnaires, observations, and measurements from food mill industries. This data was then subjected to regression analysis to establish mathematical relationships between the milling parameters. Additionally, a decision tree model was developed using scikit-learn in Python to identify the primary predictive parameters influencing mill performance. The results indicate that correlation studies between milling parameters are satisfactory. Notably, the highest correlation coefficient (R<sup>2</sup> = 0.62) was found between the power of the electrical motor and the volume of the hopper, followed by the correlation between the daily quantity of processed food and the volume of the hopper (R<sup>2</sup> = 0.52). These correlations are characterized by first-order linear models and quadratic polynomial equations. The decision tree model effectively classifies the quantity of processed food into high and low values, ranging from 3.83 tons/year to 109.50 tons/year. Furthermore, five quantitative parameters were identified as significant predictors of the daily quantity of processed food with good accuracy (F1-score = 80%): the number of working hours per day, mill energy consumption, motor power, the number of grinding machines in the mills, and grinding speed. Based on these findings, a multiple regression model was established, yielding an R<sup>2</sup> of 0.42 and a root mean square error (RMSE) of 37.45. By using the multiple regression models established on this paper, the food milling stakeholders will easily develop new strategy to improve the food milling performance and increase the data recording for a better grinding food management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Performance investigation and optimization process of small scale food mills machines

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月6日

查看全文
  • Nsah-ko Tchoumboue,
  • Robert Melono Melingui,
  • Julius Kewir Tangka

摘要

Improving the performance of small-scale food mills requires a predictive approach and an examination of the interrelation between milling parameters. The aim of this paper was to evaluate and optimize the performance of small scale food grinding machines by providing a multiple regression models based on various frequent used grinding parameters. To conduct this research, data on milling parameters were collected through questionnaires, observations, and measurements from food mill industries. This data was then subjected to regression analysis to establish mathematical relationships between the milling parameters. Additionally, a decision tree model was developed using scikit-learn in Python to identify the primary predictive parameters influencing mill performance. The results indicate that correlation studies between milling parameters are satisfactory. Notably, the highest correlation coefficient (R2 = 0.62) was found between the power of the electrical motor and the volume of the hopper, followed by the correlation between the daily quantity of processed food and the volume of the hopper (R2 = 0.52). These correlations are characterized by first-order linear models and quadratic polynomial equations. The decision tree model effectively classifies the quantity of processed food into high and low values, ranging from 3.83 tons/year to 109.50 tons/year. Furthermore, five quantitative parameters were identified as significant predictors of the daily quantity of processed food with good accuracy (F1-score = 80%): the number of working hours per day, mill energy consumption, motor power, the number of grinding machines in the mills, and grinding speed. Based on these findings, a multiple regression model was established, yielding an R2 of 0.42 and a root mean square error (RMSE) of 37.45. By using the multiple regression models established on this paper, the food milling stakeholders will easily develop new strategy to improve the food milling performance and increase the data recording for a better grinding food management.