The new law about social services, that is effective since 1st January 2007, brought series of changes into the Czech social services system. These changes should contribute to modernizing and improving the quality of provided social services. Improving of the quality happens via implementing standards for quality in social services, that are sets of measurable and verifiable criteria. Verification of completing standards runs at the place where social services are provided and it is called quality inspection. Graduation theses analyses three runs of testing inspections, that proceeded in 2002, 2004 -- 2005, 2006 -- 2007 in the Czech Republic. Analysis evaluates the best and the worst completed standards, regional location of providers, where there the inspections were performed, and general fruitfulness of providers of social services at inspection. There are mentioned recommendations, which came out of the inspection analysis.

Dot peen markings are used by Scania Ferruform to maintain a traceability of their products throughout the manufacturing. Quality inspection of the markings are performed to ensure that they are added correctly and readable. This is, however, done manually by workers, which they are looking to change. Machine vision, in combination with machine learning, could prove helpful in automating this process, which is where this thesis comes in. Images of two types of dot peen markings were gathered using different experimental setups and equipment. Amazon Rekognition and MVTec Halcon were both used to predict the characters of the images, in order to determine if the two systems could be used to demonstrate that the quality inspection can be automated. To improve the result, the images were also processed with varied techniques. The pretrained version of Amazon Rekognition and MVTec Halcon, with unprocessed image, performed the best. They both predicted all the characters correctly, and showed a high confidence in their predictions, with an average confidence of 96.41% and 99.87% respectively. When processing the images before predicting the confidence of the systems decreased and predictions were also made incorrectly. Custom training a model also showed a poor result, with the best combination of average precision and overall recall being at 0.733 and 0.561 respectively.

Mobila hydrauliska ventiler används ofta i tunga industrier för att styra hydrauliska system för att hantera material. Funktionaliteten är kritisk i dessa situationer då det utgör en miljöfara och säkerhetsfara om funktionaliteten inte är uppnådd. För att säkerställa funktionaliteten kvalitetskontrollerar Nordhydraulic samtliga ventiler i en testbänk innan de skickas till kund. Läckage och manöverkraft är några av de parametrar som testas. Om en ventil inte blir godkänd omarbetas ventilen för att sedan testas om igen. Detta medför interna kvalitetsbristkostnader för Nordhydraulic. Examensarbetet syftade till att reducera ventilavvikelser vid slutprovning genom att identifiera vilka rotorsaker som bidragit till avvikelserna, utforma ett visuellt sätt att övervaka ventilavvikelserna samt att ge rekommendationer till förbättringsåtgärder för att reducera dessa. Examensarbete genomfördes med hjälp av Andersen och Fagerhaugs (2014) föreslagna metod för rotorsaksanalys. Resultatet av examensarbetet visade att slidtillverkningsprocessen, bristfällig dokumentation av data och arbetsinstruktioner var rotorsaker som har bidragit till ventilavvikelserna vid slutprovning. Enligt operatörerna var slidbyte den mest frekventa åtgärden när en ventil inte blir godkänd. Sliderna kvalitetskontrolleras under tillverkningen, men dokumenteras inte. Variationen i slidernas tillverkningsprocess kan därför inte kvantifieras eller uppskattas på ett tillfredsställande sätt, vilket innebar att det var svårt att bedöma huruvida sliderna tillverkas inom de uppsatta toleranserna. Resultatet visade även på utmaningar med dokumentationen av provdatan i enlighet med de specificerade instruktionerna. Detta betonar vikten av förstärkta dokumentationsrutiner för att säkerställa fullständighet i provdatabasen och därigenom förbättra övervakningen av ventiltesterna. Nordhydraulic rekommenderas att utöka sin statistiska processtyrning för slidtillverkningen med hjälp av ett ramverk som examensarbetet har tagit fram. Genom att definiera huruvida slidtillverkningsprocessen är stabil och duglig kan förbättringsarbeten utföras. Examensarbetet har också utformat en dashboard som Nordhydraulic kan använda för att övervaka slutprovningen visuellt och upptäcka anomalier vid slutprovningen. Sammanfattningsvis fokuserar examensarbetet på att reducera avvikelserna hos ventiler vid slutprovning för Nordhydraulic. De identifierade rotorsakerna, tillsammans med det föreslagna ramverket och den visuella övervakningsmetoden, ger värdefulla insikter för Nordhydraulic att förbättra sina kvalitetskontrollprocesser, minimera avvikelser och säkerställa högkvalitativa ventiler till sina kunder. / Mobile hydraulic valves are frequently used in heavy industries for material handling, thanks to their high power-to-weight ratio. Functionality is critical in these scenarios as it poses both environmental and safety hazards if not achieved. To ensure functionality, Nordhydraulic performs quality inspections on all valves before they are shipped to customers, utilizing a test bench. Parameters such as leakage and actuation force are among those tested. If a valve fails to meet the requirements, it undergoes rework and is retested, resulting in internal failure costs for Nordhydraulic.  The objective of this thesis was to reduce valve deviations during final testing by identifying the root causes contributing to these deviations, designing a visual monitoring approach for valve deviations, and providing recommendations for improvement measures to reduce valve deviations. This thesis utilized the root cause analysis method proposed by Andersen and Fagerhaug (2014). The results of the thesis revealed that the spool manufacturing process, insufficient data documentation, and work instructions were identified as root causes contributing to valve deviations during final testing. According to the operators, replacing the spool was the most frequent corrective action when a valve failed to meet the requirements. The variation in the spool manufacturing process could not be quantified or estimated satisfactorily, making it difficult to assess whether the components were manufactured within the specified tolerances.  The results also indicated challenges with the documentation of test data in accordance with the specified instructions. This emphasizes the importance of strengthened documentation procedures to ensure completeness in the test database and thereby enhance the monitoring of valve tests. Nordhydraulic is recommended to increase their statistical process control using the framework developed in this thesis. By defining the stability and capability of the spool manufacturing process to produce within the specified tolerances, improvement efforts can be undertaken. The thesis has also designed a dashboard that Nordhydraulic can utilize to visually monitor the final testing process and detect anomalies. In conclusion, this thesis addresses the issue of reducing valve deviations during final testing for Nordhydraulic. The identified root causes, along with the proposed framework and visual monitoring approach, provide valuable insights for Nordhydraulic to enhance their quality control processes, minimize deviations, and ensure high-quality valves for their customers.

The final thesis is focused on problematic question of social services quality inspection. First part of the thesis aids on conception of quality in context of social services. This part also describes development of social services and their practical use in law system of Czech Republic. We can also find description of inspection methods and their system. Third charter is mainly focused on dimension of inspector's ethic codex. It is aiming on systems of values, professional ethics, ethic codex and education of inspectors. Last and main chapter is trying to explain problematic areas in actual and real conditions of inspection process. Chapter contains suggestions with several solutions. Thesis uses and combines all collected data from verified sources, specialists external inspectors.

The growing customer demands for product variety have put unprecedented pressure on the manufacturing companies. To maintain their competitiveness, manufacturing companies need to frequently and efficiently adapt their processes while providing high-quality products. Different advanced manufacturing technologies, such as industrial robotics, have seen a drastic usage increase. Nevertheless, traditional quality methods, such as quality inspection, suffer from significant limitations in highly customised small batch production. For quality inspection to remain fundamental for zero-defect manufacturing and Industry 4.0, an increase in flexibility, speed, availability and decision upon conformance reliability is needed. If robots could perform in-line quality inspection, defective components might be prevented from continuing to the next production stage. Recent developments in robot cognition and sensor systems have enabled the robot to carry out perception tasks they were previously unable to do. The purpose of this thesis is to explore the usage of robotic in-line quality inspection during changeable zero-defect manufacturing. To fulfil this aim, this thesis adopts a mixed-methods research approach to qualitative and quantitative studies, as well as theoretical and empirical ones. The foundation for this thesis is an extensive literature review and two case studies that have been performed in close collaboration with manufacturing companies to investigate how in-line quality inspection is perceived and utilised to enhance industrial robots. The empirical studies also aimed at identifying and describing what opportunities arise from having robotic in-line quality inspection systems. The result of this thesis is a synthesis of literature and empirical findings. From the literature review/study, the need for enhancing quality inspection was identified and a multi-layer quality inspection framework suitable for the digital transformation was proposed. The framework is built on the assumption that data (used and collected) needs to be validated, holistic, and online, i.e. when needed, for the system to effectively decide upon conformity to surpass the challenges of reliability, flexibility and autonomy. Empirical studies show that industrial robotic applications can be improved in precision and flexibility using the in-line quality inspection system as measurement-assisted. Nevertheless, this methodological changes and robot application face the hurdle of previous and current management decisions when passing from one industrial paradigm to another (e.g. mass production to flexible production). A discussion on equipment design and manufacturing process harmony and how in-line quality inspection and management can harmonise such a system was provided.

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 the predictive accuracy, something which was not noted in the modular network. / <p>Det finns övrigt digitalt material (t.ex. film-, bild- eller ljudfiler) eller modeller/artefakter tillhörande examensarbetet som ska skickas till arkivet.</p>

In der industriellen Fertigung unterliegen die Produkteigenschaften und -parameter, unabhängig vom eingesetzten Fertigungsverfahren, gewissen Streuungen. Im Rahmen der Qualitätsprüfung wird daher ermittelt, inwieweit die festgelegten Qualitätsanforderungen an das Produkt bzw. Werkstück trotz der Fertigungsstreuungen erfüllt werden (Brunner et al. 2011) [...] Insbesondere bei einer visuellen Prüfung durch den Menschen hängt das Ergebnis jedoch sehr stark vom jeweiligen Prüfwerker ab. Die wesentlichen Faktoren für die Erkennungsleistung sind Erfahrung, Qualifizierung und Ermüdung des Prüfers, Umgebungsbedingungen, wie Beleuchtung, Schmutz oder akustische Störfaktoren, aber auch die Anzahl und Gewichtung der zu bewertenden Merkmale (Keferstein et al. 2018). Infolge dessen kann die Zuverlässigkeit und Reproduzierbarkeit der Prüfergebnisse negativ beeinflusst werden. Gleiches gilt für die vollständige und konsistente Dokumentation der Sichtprüfung [...] Vor diesem Hintergrund wird ein Mixed Reality-basiertes Assistenzsystem entwickelt, welches den Prüfwerker bei der Durchführung und Dokumentation der visuellen Sichtprüfung unterstützen soll. Die Anforderungen dieses Ansatzes sind aus einem Kooperationsprojekt in der Automobilindustrie abgeleitet. Das dargestellte Assistenzsystem ist daher Teil von übergeordneten Aktivitäten im Zusammenhang mit 3D-Master und einer zeichnungsfreien Produktdokumentation. [...aus der Einleitung]

This graduation thesis is concerned with the topic of quality standards in social care and the formation process of writen quality standards in some organizations in the civic sector, which offer social services. The theoretical section of this thesis concerns new legal arrangement of social services, their principles and international connections; as well as standards of quality of themselves and of the different tools which are related to them, for example inspection of the quality in social services, registration of providers and education of emploees in social care services. The empirical section deals with evaluation of the research focused on the process of formatting written quality standards in organizations which were under investigation and on the presentation of obtained data which were gained, especially by observation and questionnaire inquiry in organizations.