A Factory Analysis on the Proportion of Defects Reduction in The Animal Feed Pellet Production Process using Design of Experiment Analysis

Abstract

The aim of this research is to reduce a number of defects during a feed pellet production process to improve customer satisfaction. A factory case study produces the feed pellets for several species such as food for pigs, chickens, and ducks. Production data from January to June 2017 manufacturing found that the manufacturing defects rate were about 3.32%. The data showed that the overall defects originated from different problems; 1) cracked or broken food; 2) high humidity; 3) distorted of product color; and 4) an ingredient error, respectively. Statistical methods, design analysis, and cause analysis techniques e.g. the Ishikawa diagram, Pareto chart, and FMEA (Failure Mode and Effects Analysis) were applied to help the factory to identify the main root cause of the defects and the potential failure modes of the factory case study. Due to an increasing number of complaints, this study only concentrated on the duck feed pellet production process. The study was divided into two parts: finding the root cause of the defects, which are the most critical factors for further analysis, and applying an experimental statistical design to decrease the number of defects during the duck pellet production process. The problem with cracked or broken pellets (dust) was found as the main factor affecting the production defects. Results showed that the main factors contributing to the amount of dusk in the duck feed productions came from three factors as follow: the thickness of die, distance between compression rollers and die, and time and temperature of mill machine needed during compressing the duck feed pellet production. Both the fractional factorial experimental design, 2k and 3k, were used to evaluate the influence of each factor on the duck feed production defects. The results by using the factorial 2K experimental show that the most important variable in duck pellets production were thickness of the die, distance between compression rollers and die, and temperature of mill machine needed during compressing the duck feed pellet production while time was not an interaction effect in this problem. The 3k factorial design was used to determine the interaction effects for the duck pellets production process. The experiment was ran and tested for 3 months. The final outcomes showed a significant reduction of defects from 2.51% to 1.09% (P<0.01). The results indicated that thickness -20 mm. of the die, 0.05 mm of distance between compression rollers and die, and 95 degree Celsius of temperature of mill machine needed during compressing the duck feed pellet production would be the most appropriate set of pelleting machine for the duck production process case study.