Globalization is rising progressively, with companies facing many challenges and complexities in recent times. Also, the demand for high-quality products at a lower price is driving companies to change their production systems. Lean production is one of the proven approaches to gain an advantage. The lean approach is about eliminating activities that do not add value and delivering the best possible qualitative product to the customer. Implementation of lean is feasible in any organization irrespective of its size or industry. For initiating lean in an organization, there are many tools and techniques available. Value Stream Mapping (VSM) is one such tool that is economical and easy to implement. VSM identifies waste and discovers opportunities for improvements as it observes material and information flow from the collection of raw materials to delivery to customers (Forno et al., 2014).The current research is based on a case study performed at an aircraft heater manufacturing firm, which is willing to implement lean for an efficient production cycle. Since the company is not subjected to the lean methodology, the study attempts to understand the VSM implementation process and its favorable circumstances. VSM has been implemented and data were analyzed. The results obtained from this study concludes that, VSM is a suitable technique to initiate lean ways of work in an SME that lacks sufficient knowledge and experience on lean.In addition, successful removal of wastage and increase in productivity are the key factors, necessary to initiate lean in an organization from a technical point of view whereas management support and acceptance of change is essential from an organizational point of view. However, there are few barriers and challenges that the company may have to face. Besides, the study suggests that it is better to have a lean training or a workshop before its implementation. This study will lay the foundation for future studies related to the initiation of lean methodology through implementing VSM when the case company is not familiar with the process as there are very few studies on this topic.

Today’s knowledge creation process has become extremely complex. On the one hand, technological innovation increasingly requires the use of complementary expertise from several apparently unrelated fields. On the other, the magnitude of the challenges that research and development present often require a large amount of resources that single actors can hardly achieve by themselves. For these reasons, the popularity of collaborative research has increased greatly in recent years. Among the many forms of collaborative research adopted today, research consortia are the most used. Given the steady increase of this trend, it has become important to understand the dynamics at play in research consortia. In particular, it is crucial to understand how participants interact with each other and the implication of such interactions on the success of a consortium’s outcome. The following study utilises a proximity-based framework to the single case study of the AddLife Competence Centre, an Uppsala University based R&D consortium for the development of Additive Manufacturing technology applied to Life Science. The data was collected by conducting eleven semi-structured interviews of AddLife participants. This study begins by presenting an overview of the literature on collaborative network, collaborative research and proximity. It moves onto analysing the interactions of actors in AddLife as a triple helix arrangement. Next, linkages between actors are expressed in terms of organisational, cognitive and social proximity. The analysis is discussed and presented as follows: uncertainty and discrepancies can lower organisational proximity, while the lack of goals alignment can lower cognitive proximity. High social proximity can mitigate the effects of both, especially in case of opportunistic behaviour. Acting on these three proximity dimensions should be prioritised over other proximity trade-offs.

No description available.

No description available.

No description available.

No description available.

Bränder är väldigt destruktiva, och när de bryter ut påverkar de allt runt omkring. Det bästa sättet att minimera skada från bränder är att upptäcka de tidigtinnan de brytit ut ordetnligt, detta handlar ofta om sekunder som gör skillnad.Denna rapport utforskar AI’s potential att detektera brand och minska tidendet tar att upptäcka den i industriella miljöer genom en bildström medobjektdetektering. Objektdetektering har gjort stora framsteg under de senaste6 åren. Dessa framsteg gör detta och andra liknande realtidobjektdetekteringsprojekt möjligt. Neurala nätverket som tränades i dettaprojekt överträffade ett brandlarm med 26 sekunder i realistiska miljöer.Fem dataset skapades och testades. Alla dessa tränades på YoloV4 modellen.Alla YOLO modeller är av typen konvolutionerande neuralt nätverk (CNN).Vikterna från träningen flyttades till en Nvidia Jetson Nano enhet och kördesmed hjälp av TensorRT tkDNN. En precision på 99.3 och Phi koefficient på0.924 uppnåddes och detta kan troligtvis förbättras med bättre data och ellerfler utvärderingsbilder. / Fires have destructive power when they break out and affect theirsurroundings on a devastatingly large scale. The best way to minimizedamage is to detect the fire as quickly as possible before it has a chance togrow. This project looks into the potential of AI to detect and recognize fireand reduce detection time using object detection on an image stream. Objectdetection has made giant leaps in speed and accuracy over the last six years.This progress has made this thesis and similar real-time object detectionprojects possible. The models brought forward in this project managed tooutperform a fire detector by 26 seconds in realistic scenarios.Five data sets were created and tested. All these were trained on the YOLOv4model. All Yolo models employ convolutional neural networks (CNN). Theweights were transferred and run on an Nvidia Jetson Nano through Tensor rtwith tkDNN. A precision of 99.3 and a Phi coefficient of 0.924 were achieved.

No description available.

Open innovation is an approach to innovation that has steadily started to gain ground during the last few years, as a practice that can help companies to remain innovative.This practice allows ideas to move freely inside out and outside in of a company. This means that a company can use external ideas to enhance its internal R&D and it can also utilize internal ideas in new innovative ways. As a result developing costs are reduced, products reach the market faster and new perspectives for future growth are created. This thesis studies Ericsson and how that company uses open innovation, what processes it implements and what are the motives behind its open innovation moves. By studying a company that utilizes open innovation for many years now and is considered a pioneer in this field we can gain valuable insight on this fascinating phenomenon.

Best Estimate codes are used for licensing, but with conservative assumptions. It is claimed that the uncertainties are covered by the conservatism of the calculation. As Nuclear Power Plants are applying for power up-rates and life extension, evaluation of the uncertainties could help improve the performance, while staying below the limit of the safety margins.   Given the problem of unstable behavior of Boiling Water Reactors (BWRs), which is known to occur during operation at certain power and flow conditions, it could cause SCRAM and decrease the economic performance of the plant. Performing an uncertainty analysis for BWR stability would give better understating of the phenomenon and it would help to verify and validate (V&V) the codes used to predict the NPP behavior.   This thesis reports an uncertainty study of the impact of Thermal-Hydraulic, Neutronic, and Numerical parameters on the prediction of the stability of the BWR within the framework of OECD Ringhals-1 stability benchmark. The time domain code TRACE/PARCS was used in the analysis. This thesis is divided in two parts: Sensitivity study on Numerical Discretization Parameters (Nodalization, Time Step, etc.) and Uncertainty part.   A Sensitivity study was done for the Numerical Parameters (Nodalization and Time step). This was done by refining all possible components until obtaining Space-Time Converged Solution, i.e. further refinement doesn’t change the solution. When the space-time converged solution was compared to the initial discretization, a much better solution has been obtained for both the stability measures (Decay Ratio and Frequency) with the space-time converged model.   Further on, important Neutronic and Thermal-Hydraulic Parameters were identified and the uncertainty calculation was performed using the Propagation of Input Errors (PIE) methodology. This methodology, also known as the GRS method, has been used because it has been tested and extensively verified by the industry, and because it allows identifying the most influential parameters using the Spearman Rank Correlation. / QC 20101126