论文部分内容阅读
The profit margin in the flour milling industry is quite narrow, so high-quality raw materials and efficiency of milling operations are crucial for every company. Many flour mills, especially those which import wheat from other countries and have limited storage space for the different varieties or classes of wheat, can not afford to buy low quality wheat. Consequently, a mathematical model which can test the impact and interactions of raw materials, in technical point of view, would be a useful decision-making tool for the milling industry. A flour miller tests wheat for physical and chemical characteristics, cleanness and soundness. The miller also performs experimental milling, if available, to have some idea how the given wheat will behave during commercial milling. Based on these test results, the miller can only guess the commercial milling results such as flour yields and flour ash and protein contents. Thus, the objective of this study was to develop empirical equations to estimate commercial milling results, using the physical, chemical and experimental milling data of the given wheat blend and also, additionally, flour ash and protein specifications of the end-user. This was done by using the actual commercial milling procedures and their wheat physical, chemical, experimental milling data, and other vital data. Data were collected from a commercial mill located in East Asia that had four production lines and used wheat blend combinations from five different wheat classes, i.e. Hard Red Winter (HRW),Dark Northern Spring (DNS), Soft White (SW), Australian Soft (AS), and Australian Standard White (ASW) wheat to produce over 40 different products. The wheat physical and chemical characteristics included test weight, thousand kernel weight, ash and protein contents. The experimental milling data were straight-grade and patent flour yields, along with patent flour ash and protein contents from a Buhler experimental mill. The commercial milling results included the flour yields of patent, first clear,and second clear flours, as well as the ash and protein contents of commercial patent flours. Using multiple linear regression procedures, we have developed empirical equations to be able to estimate the commercial patent flour yields with R2 values above 0.90 for all four production lines, and commercial first clear flour yields with R2 values ranging 0.76 to 0.98, and the commercial patent flour protein contents with R2 values of 0.89 to 0.92. However, the yields of commercial second clear flours and the commercial patent flour ash contents were not able to be estimated with high coefficients of determination (R2 values). We recommend that the empirical equations for estimating commercial milling parameters should be derived using data from each individual flour milling company, for each production line of a given mill, and furthermore, tailored to specific products at a given ash and/or protein contents desired by end-users.