Statistisk analys och processförbättring vid industriell brödtillverkning
Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Today product quality improvements and fact based decision making are important to satisfy customer needs and to maintain a competitive position in the market. High quality products are normally produced near the target values of critical product specifications and with low variation. To successfully minimize variation in production knowledge of cause and effect relationships in the production process is needed. In food production ingredients as well as process settings affect the quality of the final product. These relationships are often complex and not all can be fully explained by scientific models. However, many companies in the food industry still use only these scientific models. Process understanding can also be created by empirical models that describe the existing causal relationships between independent variables of the production process and important quality characteristics (responses). Statistical methods such as regression analysis and experimental design are effective tools in creating these empirical models, but unfortunately they are not systematically used by many companies.The aim of this study was to exemplify and illustrate how statistical tools can be used to create empirical models in industrial food manufacturing, which in turn can be used for process improvement and optimization. A secondary purpose of the study was to create an understanding of a production process in a Swedish industrial bread manufacturing company. The study examined how different process settings along a production process affected various quality characteristics for a particular bread product.The study was divided into an initial pilot study followed by an experimental study. The pilot study examined historical data from the production line by the use of scatter plots and regression analysis to map preliminary cause and effect relationships. Additionally, interviews were conducted to document the hypothesized cause and effect relationships between process settings and product quality characteristics. The study was primarily focused on developing models for bread Texture day 1 to day 5 through experimentation. The experiment involved a two-level fractional factorial design with the five experimental factors: A: Amount of water, B: Amount of return dough, C: Upper oven temperature, D: Lower oven temperature and E: Baking time.Through a systematic work process empirical regression models for eight response variables related to bread quality were developed based on experimental data. The results showed which main and interaction effects that affected bread quality characteristics: Texture day 1, Texture day 2, Texture day 3, Texture Development, Height, Moisture, Water Activity and Baking level. Using the regression models in optimization indicate that, for example, a 26 percent improvement in the texture values and texture development may be possible compared to historical data; while other quality parameters were maintained at their target values.The systematic work process in this study illustrates how statistical methods can be used to create empirical models in food manufacturing. The results also show the benefits of using the obtained empirical models for quality improvement and process optimization.
Place, publisher, year, edition, pages
2016. , 99 p.
Social Behaviour Law
Samhälls-, beteendevetenskap, juridik, Processförbättring, statistiska verktyg, brödtillverkning, försöksplanering, optimering
IdentifiersURN: urn:nbn:se:ltu:diva-56308Local ID: d168e855-a748-43bb-803e-f10848b3e84bOAI: oai:DiVA.org:ltu-56308DiVA: diva2:1029695
Subject / course
Student thesis, at least 30 credits
Industrial and Management Engineering, master's level
Validerat; 20160613 (global_studentproject_submitter)2016-10-042016-10-04Bibliographically approved