Nowadays, computer experiments are used increasingly more to solve complex engineering and technological issues. Computer experiments are analysed through suitable metamodels acting as statistical interpolators of the simulated input-output data: Kriging is the most appropriate and widely used one. We optimise the braking performance of freight trains through computer experiments and Kriging modelling by focussing on the payload distribution along the train, so as to reduce the effects of in-train forces among wagons during a train emergency braking. One contribution of this manuscript is that to improve the freight train efficiency in terms of braking performance, we consider that the train is composed of several train sections with each one characterised by its own overall payload. A suitable Latin hypercube design is planned for the computer experiment that achieves excellent space-filling properties with a relatively low number of experimental runs. Kriging models with anisotropic covariance function are subsequently applied to assess which is the best payload distribution capable of reducting the in-train forces according to the specific train-set arrangement considered. The results are very satisfactory and confirm that our approach represents a valid method to be successfully applied by interested Railway Undertakings.
In the industrial production scenario, the goal of engineering is focused on the continuous improvement of the process performance by maximizing the effectiveness of the manufacturing and the quality of the products. In order to address these aims, the advanced robust process optimization techniques have been designed, implemented, and applied to the manufacturing process of ultrasound (US) probes for medical imaging. The suggested guideline plays a key role for improving a complex multi-stage manufacturing process; it consists of statistical methods applied for improving the product quality, and for achieving a higher productivity, jointly with engineering techniques oriented to problem solving. Starting from the Six Sigma approach, the high definition of the production process was analyzed through a risk analysis, and thus providing a successful implementation of the PDCA (plan-do-check-act) methodology. Therefore, the multidisciplinary analysis is carried out by applying statistical models and by detecting the latent failures by means of NDT (non-destructive testing), i.e., scanning acoustic microscopy (SAM). The presented approach, driven by the statistical analysis, allows the engineers to distinguish the potential weak points of the complex manufacturing, in order to implement the corrective actions. Furthermore, in this paper we illustrate this approach by considering a pilot study, e.g., a process of US probes for medical imaging, by detailing all the guideline steps.
28 February 2020 in Journal of Global Fashion Marketing
This paper proposes an innovative method for estimating brand association matching and strength. It is based on a dual analytical perspective as it compares consumer-perceived with company-defined brand associations, thereby filling a research gap in the offline and online brand studies. In order to build this method, we investigate online communities as authentic interactive environments by combining text-mining with Hybrid Log-Linear models. What emerges is an interdisciplinary method that contributes to extending the use of the user-generated content (UGC) for marketing purposes. Its application is useful for verifying the conditions for customer brand equity and more systematically controlling the adopted brand communications. The analysis also takes into consideration consumers who, after experiencing the virtual setting, have become free-brand tellers and real co-creators of brand meaning. In the study, we focus on online fashion communities as digital platforms where consumer interactions are particularly intensive, and the narratives are rich in brand-related perceptions.
The number of spots to monitor to evaluate soil respiration (Rs) is often chosen on an empirical or conventional basis. To obtain an insight into the necessary number of spots to account for Rs variability in a Mediterranean pine-dominated mixed forest, we measured Rs all year long on sixteen dates with a portable gas-analyser in 50 spots per date within an area 1/3 ha wide. Linear mixed-effects models with soil temperature and litter moisture as descriptors, were fitted to the collected data and then evaluated in a Monte Carlo simulation on a progressively decreasing number of spots to identify the minimum number required to estimate Rs with a given confidence interval. We found that monitoring less than 14 spots would have resulted in a 10% probability of not fitting the model, while monitoring 20 spots would have reduced the same probability to about 5% and was the best compromise between field efforts and quality of the results. A simple rainfall index functional to select sampling dates during the summer drought is proposed.
This paper deals with a proposal for joint modeling and process optimization for split-plot designs analyzed through mixed response surface models. It addresses the following main issues: i) the building of a joint mixed response surface model for a multiple response situation, by defining only one response through which specific coefficients are included for studying the association among the responses; ii) the considering of fixed as well as random effects within a joint modeling and optimization context; iii) the achievement of an optimal solution by involving specific as well as common coefficients for the responses. We illustrate our contribution through a case-study related to a split-plot design on electronic components of printed circuit boards (PCBs); we obtain satisfactory results by confirming the validity of this contribution, where the qualitative factor PCB is also studied and optimized.
This paper studies the impact when the experimenter does not randomize the design for split-plot experiments when there is a linear trend. The paper uses both simulation and basic theory to explain the bias if we do not randomize the design and the inflation in variance if we do. The paper shows proper residual plots can detect the trend if the design is randomized, which allows the analyst to treat time as a covariate, removing the trend’s impact. The simulation study provides an excellent way to explain to practitioners the risks from not randomizing the design.