Diplomat Dörrar AB is a company that produces doors for the Swedish market. Presently, there is a quality problem within the paint shop, which also happens to be a bottleneck. The quality problems within the paint shop means that roughly 15% of all painted doors are rejected and are forced to be repainted. To repaint a door comes with a large price tag for the company. The purpose of the work is to identify why doors are being rejected, as well as to propose actions to reduce the amount of rejections, to 7.5% This thesis uses different theories connected to quality development and production control. The main theories used in this report are Lean methodology, variation reduction, maintenance and process breakdown. By exploring these theories, it has been possible to propose actions to take in order to reduce rejections to 7.5%. The information and data that has been collected largely comes from interviews with paint shop employees, as well as preexisting data. However, it has become evident that the preexisting data has been flawed when it comes to raw data, which has not allowed the processed data to provide a clear picture. To combat this, new data collecting has been done, which has given a clear picture of which doors have been affected, which colors as well as where on the door the rejection was found. Presently, the most common forms of rejections are paint flaws, plastering flaws and debris. Paint flaws can be categorized in 4 main flaws, running paint, paint drop, bad coverage and craters. The flaws are caused by faults in the paint booth, where a robot paints the door. Plastering flaws occur because of the plastering staff have missed to plaster a flawed door, and debris flaws are caused because debris has been caught underneath the top layer of paint. All these flaws give the doors a visible defect when undergoing inspection. The faults when organized in size by number of total defects are as follows: Paint flaws – 44%, Debris – 24% and plastering flaws – 18%. This yields a total of 86% of the total rejections at Diplotmat Dörrar AB. An investigation has been done to identify the root causes of the chosen rejections. When it comes to paint flaws a few different root causes have been identified that all flaws share. The root cause for paint flaws is that almost all maintenance work is reactive, which leads to satisfactory doors being rejected because of a lack of maintenance work. The root cause for debris is the lack of instructions or responsibility to remove debris, removal of splinters as well as dirty rails in the paint shop. The reason for plastering flaws is due to the fact that there is no standardized working method when it comes to plastering. The actions that are recommended to Diplomat Dörrar AB are as follows: Create a standardized working method on all different stations as well as to implement a working preventative maintenance scheme in order to reduce rejections. Keywords: LEAN methodology, Standardized working method, Preventative maintenance

Den här studien är genomförd på en tillverkningsindustri. Fabrikens verksamhet inkluderar avdelningar för montering, bearbetning, materialhantering, underhåll och teknisk support m.m. Huvudsyftet med denna studie var att bygga en simuleringsmodell över ett bearbetningsflöde med hjälp av den uppdaterade mjukvaran FACTS. Att genomföra optimering på operationsobjekt (OP) med hjälp av SCORE, jämföra och utvärdera resultat för nyckeltalen TH, LT, WIP mot befintliga modeller samt att minska totala antalet objekt med 10 % jämfört med ursprunglig FACTS modell definierades som mål för studien. Produktvariant Y användes uteslutande under denna simuleringsstudie. I projektet användes PDCA som problemlösningsmetod vilket är ett hjälpmedel för att ge struktur åt förbättringsarbeten. Bank simuleringssteg har använts som riktlinjer i genomförandefasen. Processdata för tillgänglighet, cykeltid och MTTR insamlades från befintliga modeller i FACTS och Plant Simulation. Efter lämpliga data var insamlad skapades en grundmodell byggd för att efterlikna det verkliga fysiska systemet och befintliga modellers layout. Det nya resursobjektet användes för att modellera portalsystemen och tid för ladda/lossa produkter placerades i operationsobjekt för transport. Efter färdigställande av grundmodellen verifierades och validerades denna samt lämplig simuleringshorisont för experiment fastställdes. För att beräkna antalet nödvändiga replikationer för ett trovärdigt resultat användes tillhandahållen Excel tabell från Högskolan i Skövde. Experiment för nyckeltal och optimering med SCORE utfördes och resultaten jämfördes mot befintliga modeller för att kunna utvärdera eventuella skillnader och dess orsaker. Resultaten visade på närmast likvärdiga siffror för TH men differenser på ca 10–15 % fanns för LT och WIP i ny modell för dessa nyckeltal. SCORE påvisade vissa skillnader i de primära flaskhalsarna men generellt var placeringen av flödesbegränsningar lokaliserade till produktionens avslutande del. Antal objekt i ny FACTS modell blev 327 i jämförelse med 351 i befintlig. Det fanns ett antal olika faktorer som kan ha haft påverkan på det slutgiltiga resultatet. Skillnader i inhämtade processdata och avsaknad av ladda/lossatider i befintlig FACTS modell var faktorer som gjorde systemjämförelser mellan modeller svåranalyserade. Vidare studier krävs för mer grundlig analys av orsaker till differenser i LT och WIP mellan de olika modellerna. Det satta målet för minskat antal objekt med 10 % uppnåddes inte utan objektantal minskade med enbart 6,8 %. Enligt författarna åstadkoms dock en förenklad förståelse för produkternas flöde genom tillverkningsprocessen. / This study was conducted in a manufacturing plant. The plant includes assembly-, machining-, materiel handling-, maintenance- and technical support departments. The main purpose of this study was to create a simulation model of a machining line using the updated software version of FACTS. To implement optimization of the operational objects using SCORE, to compare and evaluate KPI results for TH, LT and WIP against existing models and reduce the total number of objects with 10 % compared to original FACTS model were defined as targets for this project. Studies was conducted using only the product variant Y. PDCA was used as a method for problem solving and to give some structure for the implemented improvements. Guidelines from Banks simulation steps were utilized in the implementation phase of this study. Process data for availability, cykel time and MTTR were collected from existing FACTS and Plant Simulation models. After finishing the necessary data collecting a basic model which mimicked the real physical system was created. The new resource object was used to model the gantry systems och time for loading/unloading were added to the operational objects used for transportation.When the basic model was completed a verification and validation phase were conducted and suitable values for simulation horizon and warm-up time was decided. An Excel chart provided by the University of Skövde was used to calculate the number of replications needed for a credible experiment result. SCORE and KPI experiments were performed and the results was compared with existing models to detect any reasons for abnormalities. The results for the new FACTS model showed an almost similar value for TH but there were differences of approximately 10-15 % for LT and WIP. Analysis of SCORE results showed some differences in the primary bottlenecks but in general the most significant flow restrictions were located in the end part of the production flow. The total number of objects in the new FACTS model was 327 compared to 351 in the existing FACTS model. There were several factors that could have affected the end results of the experiments. Differences in used process data and the lack of process times for loading/unloading of products in existing FACTS model were all factors that made a correct system comparison and analysis more difficult to transact. Further studies need to be conducted to analys causes of the abnormalities between the different models regarding LT och WIP. The target of reducing the total number of objects in new FACTS model with 10 % was not achieved. Object numbers only decreased with 6,8 % but according to the authors a more simplified understanding of how the products flow in the manufacturing process was achieved.

Artificial intelligence, specifically deep learning, is a fast-growing research field today. One of its various applications is object recognition, making use of computer vision. The combination of these two technologies leads to the purpose of this thesis. In this project, a system for the identification of different crops and weeds has been developed as an alternative to the system present on the FarmBot company’s robots. This is done by accessing the images through the FarmBot API, using computer vision for image processing, and artificial intelligence for the application of transfer learning to a RCNN that performs the plants identification autonomously. The results obtained show that the system works with an accuracy of 78.10% for the main crop and 53.12% and 44.76% for the two weeds considered. Moreover, the coordinates of the weeds are also given as results. The performance of the resulting system is compared both with similar projects found during research, and with the current version of the FarmBot weed detector. Form a technological perspective, this study presents an alternative to traditional weed detectors in agriculture and open the doors to more intelligent and advanced systems.

Migu Mekaniska är ett skärande, bearbetande företag beläget i Tidan utanför Skövde. Verksamheten grundades 1980 och utgörs idag av en legotillverkning med hjälp utav 10 stycken 3- och 5-axliga CNC-maskiner. Råmaterialet vilket hanteras består av stänger i olika längder och tjocklekar.Problematiken på företaget berör processen kring hantering av råmaterial samt tre kapmaskiner vilka förser resterande produktion med material för bearbetning. Önskemålet är att utforma arbetsstationen så att kapmaskinerna inte behöver bemannas hela tiden, utan endast då kapning av material krävs samt att berörd personal kan bistå med alternativa, värdeadderande aktiviteter. Målet med arbetet är att systematiskt belysa områden för förslag på hur materialhantering och kapning av råmaterial kan utformas samt bemannas.En bred teoretisk referensram samt redogörelse för relevant litteratur och genomförda studier stärker metod och genomförande. Litteratur och studier är utvärderade samt jämförda mellan varandra och analyserade till arbetets fördel. Genom intervjuer, samtal, observationer samt genomförande av frekvensstudie och spagettidiagram, är flertalet slöseri identifierade samt aspekter för framtida förbättringsarbete.Med stöd från litteratur och teori samt resultat från insamlade data, är förbättringsförslagen validerade utifrån för- och nackdelar samt genom ett PICK-chart. Utifrån detta rekommenderades ett förslag för vidare implementering där ett dragande system baserades på ConWIP-metodiken samt genom en interaktiv och visuell lagtavla.Efter genomfört arbete diskuterades flertalet intressanta diskussionsområden där även rekommendationer för framtida arbete framfördes. En betydande aspekt vilket framkom tydligt var implementering av Lean-filosofin samt dess betydelse för långsiktiga och hållbara lösningar. Problematiken samt frågeställningarna vilka inledningsvis beskrevs och presenterades, lyckades omfamnas av arbetet enligt planering, vilket även bidrog till att samtliga delmål och dess syfte fullföljdes. / Migu Mekaniska is a cutting and processing company located in Tidan, outside of Skövde. The business was founded in 1980 and consists today of ten 3- and 5-axis CNC-machines. The company produces components based on contracts with other companies and works as a subcontractor. The raw material which is handled and used consists of rods in different lengths and thicknesses.The current problem at the company concerns the process of handling raw material and three cutting machines which provide the remaining production with materials for further processing. The intention is to design the workstation so that the cutting machines do not have to be manned all the time and that workers can staff the station only when cutting material is required. The time remaining should be used for assisting and value adding activities. The aim of the work is to systematically illuminate areas for proposals on how material handling and cutting of raw materials can be designed and staffed.A substantial theoretical frame of reference as well as an account of relevant literature and accomplished studies, contributes to strengthen the used methods and implementation. Literature and studies are evaluated and compared between each other and analyzed for the benefit of the work. Through interviews, conversations, observations and work sampling studies and spaghetti diagrams, waste is identified as well as aspects for future improvement work.With the support of literature and theory as well as results from data collected, the improvement proposals are validated based on pros and cons and through a PICK-chart diagram. Based on this methodology, a recommend proposal for further implementation is presented. The proposal is consisting of a pulling system based on the ConWIP methodology and through an interactive and visual whiteboard.After completing the work, several interesting areas were discussed, where recommendations for future work also got presented. One important aspect that became clear was the implementation of the Lean philosophy and its importance for long-term and sustainable solutions. The initially described and presented issues and sub-objectives, were clearly embraced throughout the work and fulfilled within its purpose.

Kinnarps är ett företag som tillverkar kontorsmöbler samt tillhandahåller arbetsplatslösningar för kontor och offentliga miljöer. Vid avdelningen som examensarbetet utförs på paketeras bordsben manuellt och processen bidrar till ansenliga reklamationskostnader på grund av att fel bordsben paketeras och skickas till kund. Syftet med arbetet är att ta fram förbättringsförslag som reducerar risken för att bordsbensreklamationer uppstår med hjälp av produktionstekniska verktyg och metoder. För att uppnå syftet skulle flera datainsamlingsmetoder genomföras som analys av historiska data, observationer och intervjuer. Efter utförda datainsamlingsmetoder skulle en rotorsaksanalys skapas.Teori som används för att styrka metodvalen i arbetet beskrivs i den teoretiska referensramen. Arbeten inom liknande områden har även studerats i litteraturstudien. Samtliga av de arbetena kan i något avseende relatera till projektet och dess slutsatser.Den empiriska studien utfördes genom att analysera historiska data i form av föregående års reklamationskostnader, observationer enligt gemba samt intervjuer och samtal med personal på avdelningen. Reklamationskostnaderna för det föregående räkenskapsåret var cirka 723 000 SEK. Genom metoderna kunde ett Ishikawa-diagram skapas för att hitta rotorsaken till att problemet uppstår. Med Ishikawa-diagrammet tillsammans med den teoretiska referensramen och litteraturstudien kunde tre förbättringsförslag som reducerar antalet reklamationer presenteras.Samtliga förbättringsförslag har validerats angående verkningsgrad, kostnad och hur stor insats som krävs vid en eventuell implementering. Utifrån valideringen kunde förbättringsförslaget streckkodsläsare presenteras som det mest lämpliga alternativet. Vid en eventuell implementering uppskattas reklamationerna reduceras med 92% vilket motsvarar en besparing på cirka 665 000 SEK om året i form av reducerade reklamationskostnader. / Kinnarps is a company that manufactures office furniture and provides workplace solutions for offices and public environments. This project is performed at a manual workstation that packages table legs, this station is responsible for a significant amount of customer complaint costs due to wrong table legs being packaged and sent to the customers. The purpose of this project is to present suggestions for improvement that reduce the risk of packaging errors by using production tools and methods. To achieve the purpose of this project a data collection is to be performed by analyzing historical data and performing observations and interviews. After the data collection the cause of the event is to be found.Theory used to strengthen the choice of methods performed is presented in the theoretical frame of reference. Similar projects have also been studied and are presented in the literature review. The similar projects that were studied can in some way be related to this project and its conclusions.The empirical study was performed by analyzing historical data of the last years customer complaints costs, observations according to the gemba methodology, interviews and conversations with the staff. The costs of customer complaints for the last year was around 723 000 SEK. With the results of the performed methods an Ishikawa diagram was created to identify the causes of the packaging errors. The Ishikawa diagram combined with the theoretical frame of reference and literature review was used to develop three suggestions for improvement that help reduce the amount of customer complaints caused by packaging errors.All suggestions for improvement have been validated by analyzing the effectiveness, cost and the amount of work needed to implement the suggestions. The validations process resulted in a barcode scanner being the most suitable suggestion. By implementing a barcode scanner, the customer complaints could be reduced by an estimate of 92% which corresponds to saving around 665 000 SEK per year in customer complaint costs.

No description available.

No description available.

The automation level of today’s industries is categorized as high. Some specific areas cannot be fully automated, such as manual assembly operations. Manual assembly stations often require high flexibility due to variation in products and product types, and some operations also require human finesse for conducting the operations. A collaborative robot is produced to facilitate for the worker during operations and tasks, which can be categorized as non-ergonomically and repetitive. The technical specification for the collaborative robots is not yet fully developed, and therefore it might be hard to create a safe work environment. Design and creation is the research strategy used for the project, much due to the aim of creating something physical. The project aims to establish a demonstrator, and introduce a collaborative robot, a UR5 for human-robot collaboration for a manual assembly operations with the ability to work reverse. The fictional workflows implemented in the demonstrator are established with influence from real manual assembly operations and parts included in an engine of a truck. The widgets identified and included in the workflows are created and 3D-printed. The main goal for the project is to establish a collaboration between the worker and the robot and create a baseline for a future safety evaluation conducted on the demonstrator. The project included identification of equipment and widgets necessary, the layout of the demonstrator, workflow establishment for both assembly and disassembly, together with the configuration of the equipment and programming of the collaborative robot. Safety standards concerning robots and collaborative robots, together with the technical specification not yet fully developed, worked as a base during the establishment of the workflow, configuration of the equipment, and programming of the robot. The established workflow can work both for assembly and disassembly. The workflow includes tasks that are performed separately, together, and simultaneous on the same workpiece. Experiments have been conducted on the established workflows, and observations conducted on the tasks performed. Aspects, such as time consumption for individual tasks, risk identification of quasi-static and transient contacts, and gripper position have been included during the observation.

In the fields of automated manufacturing, flexibility has been a long struggle for companies. This factor is profoundly dependable on the demands of the customer in the form of high-quality products, variations, and low-prices, which has a high impact on the manufacturing industry. With the help of an anthropomorphic tool, the flexibility factor in, some extend, be the answer for companies working with robotic automation manufacturing. This thesis aims to investigate what kinds of objects that can be picked and placed, with the help of an anthropomorphic tool, mounted on a collaborative, UR3 robot—thereby evaluating the potential of what an anthropomorphic tool can and cannot pick. Also, investigate pick and place operation with different geometrical shape objects. The project consists of identifying objects that are being picked and placed on a tailormade kitting box located on a conveyor belt. With help from the next station regarding what kinds of objects that are being delivered, a kitting box was designed and created. Stands could then be created for the kitting box and the objects. This was performed with the help of the software Creo Parametric, which was then 3D-printed. A shelf was created to successfully pick objects from, and thereby placing objects into the kitting box. In the thesis, investigation regarding what kinds of objects the tool can pick was performed. This was conducted by picking the objects used for the thesis and everyday objects, such as duct tape, knife, and coffee cup. The experiment consisted of 15 different objects, were different grasping techniques were tested and how much force the tool used. The result of the station as a whole has successfully been created in Assar innovation arena. The an-thropomorphic tool proved to be a concept to conquer the flexibility struggle in the form of success-fully grasping and placing different objects that have different geometrical shapes and weights. With the help of a conveyor, the whole station can also work in reverse mode, by picking objects from the kit-ting box and placing them into the stands, assembled on a shelf.

The main objective of the project is to identify, study, and understand potential improvements of the Sealing Applicators production line of Ecco Finishing/Carlisle Fluid Technologies Company. The ergonomic cases in the plant, established as working postures, will be analysed using manual methods and virtual simulation tools. The analyses will be based on the measurements and depth of the data that is compiled during the visits to the plant. The software used to perform the ergonomic simulations, Siemens Jack 9.0, also has its limitations for each working posture of the production line, as some data for building the postures is missing due to limitations of the measurement taking process. The final result of the project is set of data that includes a solution implementation of the working postures analysed, also developed with simulations in Siemens Jack 9.0, of the production line, that could improve daily working life of the workers in the plant as well as management of the company. The project results reveal a need to add as well as alter elements of the workstation to make sure that the posture the user performs when working adheres to ergonomic principles.