3 results on '"Vuoluterä, Fredrik"'
Search Results
2. Quality inspection of multiple product variants using neural network modules
- Author
-
Vuoluterä, Fredrik and Vuoluterä, Fredrik
- Abstract
Maintaining quality outcomes is an essential task for any manufacturing organization. Visual inspections have long been an avenue to detect defects in manufactured products, and recent advances within the field of deep learning has led to a surge of research in how technologies like convolutional neural networks can be used to perform these quality inspections automatically. An alternative to these often large and deep network structures is the modular neural network, which can instead divide a classification task into several sub-tasks to decrease the overall complexity of a problem. To investigate how these two approaches to image classification compare in a quality inspection task, a case study was performed at AR Packaging, a manufacturer of food containers. The many different colors, prints and geometries present in the AR Packaging product family served as a natural occurrence of complexity for the quality classification task. A modular network was designed, being formed by one routing module to classify variant type which is subsequently used to delegate the quality classification to an expert module trained for that specific variant. An image dataset was manually generated from within the production environment portraying a range of product variants in both defective and non-defective form. An image processing algorithm was developed to minimize image background and align the products in the pictures. To evaluate the adaptability of the two approaches, the networks were initially trained on same data from five variants, and then retrained with added data from a sixth variant. The modular networks were found to be overall less accurate and slower in their classification than the conventional single networks were. However, the modular networks were more than six times smaller and required less time to train initially, though the retraining times were roughly equivalent in both approaches. The retraining of the single network did also cause some fluctuation in, Det finns övrigt digitalt material (t.ex. film-, bild- eller ljudfiler) eller modeller/artefakter tillhörande examensarbetet som ska skickas till arkivet.
- Published
- 2022
3. Analysis of material flow and simulation-based optimization of transportation system : The combination of simulation and Lean to evaluate and design a transportation system
- Author
-
Vuoluterä, Fredrik and Carlén, Oliver
- Subjects
transportation ,Production Engineering, Human Work Science and Ergonomics ,AGV ,simulering ,lean production ,event ,Produktionsteknik, arbetsvetenskap och ergonomi ,SMO ,system ,lean ,simulation ,DES ,material handling ,kaizen ,multi-objective ,time study ,SBO ,värdeflödesanalys ,transport ,discrete ,optimering ,FACTS ,discrete-event ,optimization ,simulation-based - Abstract
The thesis has been performed in cooperation with a Swedish manufacturing company. The manufacturing site of the company is currently implementing a new machine layout in one of its workshops. The new layout will increase the product flow to another workshop on the site. The goal of the thesis was to evaluate the current transportation system and suggest viable alternatives for the future product flow. By means of discrete event simulation these alternative solutions would be modelled and subsequently optimized to determine if their performance is satisfactory. An approximated investment cost of the solutions would also be estimated. By performing a literature review and creating a frame of reference, a set of relevant methodologies were selected to provide a foundation to the project. Following these methodologies, the current state of transportation was identified and mapped using Value Stream Mapping. Necessary data from the current flow was identified and collected from the company computer systems. This data was deemed partly inaccurate and further verification was needed. To this end, a combination of Genchi Genbutsu, assistance from onsite engineers and a time study was used to verify the unreliable data points. The data sets from the time study and the company data which was deemed valid were represented by statistical distributions to provide input for the simulation models. Two possible solutions were picked for evaluation, an automated guided vehicle system and a tow train system. With the help of onsite personnel, a Kaizen Event was performed in which new possible routing for the future flow was evaluated. A set of simulation models portraying the automated guided vehicle system and the tow train system were developed with the aid of simulation software. The results from these models showed a low utilization of both systems. A new set of models were developed, which included all the product flows between the workshops. The new flows were modelled as generic pallets with the arrival distribution based on historical production data. This set of models were then subject for optimization with regard to the work in process and lead time of the system. The results from the optimization indicates the possibility to reduce the overall work in process by reducing certain buffer sizes while still maintaining the required throughput. These solutions were not deemed to be ready for implementation due to the low utilization of the transportation systems. The authors instead recommend expanding the scope of the system and including other product flows to reach a high utilization.
- Published
- 2018
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.