ENBIS-17 in Naples

9 – 14 September 2017; Naples (Italy) Abstract submission: 21 November 2016 – 10 May 2017

My abstracts

 

The following abstracts have been accepted for this event:

  • Optimization of Renewable Energy Sources for a Data Center Using Energy Storage

    Authors: Christophe Varnier (FEMTO-ST institute - ENSMM - UBFC), Jean-Marc Nicod (FEMTO-ST institute - ENSMM - UBFC)
    Primary area of focus / application: Other: Machine learning for energy management
    Keywords: Optimization, Renewable energy, Datacenter, Mixed integer linear programming
    Submitted at 13-Apr-2017 11:31 by Christophe VARNIER
    Accepted (view paper)
    12-Sep-2017 09:40 Optimization of Renewable Energy Sources for a Data Center Using Energy Storage
    The growing use of information technology in many sectors, such as banking, e-commerce, entertainment and health care, to name a few, led to a rapid development of data centers, due to the need to increase processing capacity. Consequently, the management of processes and IT systems and data centers is emerging as an area whose environmental impact requires increased attention to the efficient energy consumption.
    With depleting conventional energy resources, whole world is now looking for alternatives to meet energy demands. One of the solutions is renewable energy which is non-exhaustible and non-polluting, but problem is intermittent nature of these energies.
    In this communication, we propose to study the management of a hybrid energy source system that deliver power for a datacenter. The system is composed of several energy technologies and back-up as well as energy storage units. The storage devices can be a battery bank, supercapacitor bank, or a fuel cell-electrolyzer system. Other energy sources are considered such as wind turbine and/or photovoltaic panels.
    The problem addressed in this work is a hybrid renewable energy system and the point is to manage the energy production system with the objective to satisfy a load demand and to minimize the power exchanged with the grid. A mixed integer linear program is proposed to solve the addressed problem.
  • Estimating the Size of a Defective Subgroup in Industrial Mass Production

    Authors: Benjamin Sobotta (Robert Bosch GmbH)
    Primary area of focus / application: Other: Sampling
    Keywords: Defective subgroup, Industrial production, Risk analysis, Sampling
    Submitted at 13-Apr-2017 16:08 by Lennart Kann
    Accepted
    12-Sep-2017 10:50 Estimating the Size of a Defective Subgroup in Industrial Mass Production
    In industrial mass production of electronic devices, quality control measures are in place to detect even the smallest of deviations. Nevertheless, due to the sheer quantities produced, deviations cannot entirely be avoided. However, in the vast majority of cases, only a very small fraction of the production volume is affected.
    Consequently, the estimation of the affected volume is paramount during risk analysis. We propose an approach to determine the size of this defective subgroup. At the core of our so-called Test Gate Method lies production control data. Given that this data is readily available, it is leveraged to quickly and reliably obtain the size of the subpopulation of interest. Because there are otherwise few dependencies or limitations, our method may be applied to a broad spectrum of cases. This especially holds true when the subpopulation comprises only few parts per million.
    The Test Gate Method has been successfully applied on a number of occasions. We will present two representative cases to illustrate the approach and highlight its benefits.
  • Wind Turbine Performance Decline in Sweden

    Authors: Jesper Rydén (Uppsala University), Jon Olauson (Uppsala University)
    Primary area of focus / application: Reliability
    Secondary area of focus / application: Modelling
    Keywords: Wind power, Decline, Multiple regression, Bootstrap
    Submitted at 13-Apr-2017 17:15 by Jesper Rydén
    Accepted (view paper)
    12-Sep-2017 18:20 Wind Turbine Performance Decline in Sweden
    Performance loss of wind turbines on a national scale is crucial for energy planning. Often in reliability, studies are made at component levels, but focus is here on the overall reduction of performance with age. The evolution of downtime with age is examined and factors explaining differences in trends are considered. Time series are long-term using different meteorological models and to investigate the influence from various factors, regression modelling techniques are employed. To assess uncertainties, a bootstrap approach was taken. It is found that Swedish wind turbines lose around 0.10 pp/year. For units built before 2007, the median decline is 0.15 pp/year, corresponding to an 20-year energy loss of around 6%.
  • A Fresh Look at Effect Aliasing and Interactions: Some New Wine in Old Bottles

    Authors: Jeff Wu (Georgia Tech)
    Primary area of focus / application: Other: Box medal lecture
    Secondary area of focus / application: Design and analysis of experiments
    Keywords: Conditional main effects, Fractional factorial designs, Nonregular designs, Orthogonal arrays
    Submitted at 14-Apr-2017 01:33 by Jeff Wu
    Accepted
    11-Sep-2017 14:50 George Box Award: Jeff Wu. Award Talk on "A Fresh Look at Effect Aliasing and Interactions: Some New Wine in Old Bottles"
    Interactions and effect aliasing are among the most fundamental concepts in experimental design. In this paper, some new insights and approaches are provided on these time-honored subjects. Start with the two-level fractional factorial designs. In the literature, the “de-aliasing” or estimation of aliased effects is deemed to be impossible. We argue that this “impossibility” can indeed be resolved by employing a new approach which reparametrizes effects using the notion of “conditional main effects” (cme’s), then performs model selection by exploiting the properties between the cme’s and traditional factorial effects. In some sense, this is a shocking result, because the impossibility has been taken for granted since the founding work of Finney (1945). This approach can be extended beyond designed experiments to general observational data using bi-level variable selection techniques. There is a similar surprise for three-level fractional factorial designs. The standard approach is to use ANOVA to decompose the interactions into orthogonal components. Then the quandary of full aliasing between interaction components remains. Again, this can be resolved by using a non-orthogonal decomposition with the linear-quadratic parametrization. Then a model search strategy would allow the estimation of some interaction components even for designs of resolution III and IV. Moving from regular to nonregular designs, most of the interactions are not orthogonal to the main effects. The partial aliasing of effects can be exploited for the estimation of interactions. The common theme underlying the three problems is the use of reparametrization and exploitation of non-orthogonality among effects. A historical recollection is given on how these ideas were conceived and discovered over a period of nearly 30 years.
  • Statistical Intervals for Predictions of Multiple Future Observations

    Authors: Bernard G Francq (GSK), Stéphane Laurent (GSK), Dan Lin (GSK), Walter Hoyer (GSK)
    Primary area of focus / application: Metrology & measurement systems analysis
    Secondary area of focus / application: Quality
    Keywords: Prediction interval, Prediction region, Multiple predictions (multiplicity issues), Automated application, Reproducible report
    Submitted at 14-Apr-2017 15:06 by Bernard Francq
    Accepted
    13-Sep-2017 10:10 Statistical Intervals for Predictions of Multiple Future Observations
    During the life-cycle of a commercial vaccine there are many instances where process changes are introduced. Such examples include replacing or modifying a unit operation within the production chain, or a process transfer from one production site to another. Usually, a relatively large number of “historical” batches (process before change) are available. After generation of only a small number of (typically three) process-validation batches an evaluation and a decision about the success of the change are required.

    Classical statistical tests (t-tests, equivalence tests etc.) are not appropriate for use in this case due to the small number of new batches tested. As a necessary requirement for demonstrating comparability throughout the change, it is, therefore, assessed that the individual values of the new batches should lie within the prediction interval computed from the historic batches.

    In this presentation, we generalize the concept of prediction intervals to a prediction region for multiple future observations in line with the Gaussian multivariate distribution. An analogy is drawn between inference and prediction, and between the multiplicity issue in the comparison of multiple means and the prediction region for multiple future observations. Different statistical procedures are reviewed and applied for construction of a prediction region, including approximate and exact methods (i.e. Bonferroni, Dunnett, Holm, Sidak,…).

    An automated application with graphical user interface based on R will be presented. The application allows upload of data and returns a pdf report within seconds. Also other common intervals (confidence intervals for the mean, tolerance intervals, etc.) can be obtained with the tool.
  • Issues in the Revision of ISO 3951-1

    Authors: Rainer Göb (University of Würzburg)
    Primary area of focus / application: Quality
    Keywords: Statistical standard, Acceptance sampling, Proportion nonconforming, Variables sampling
    Submitted at 15-Apr-2017 00:47 by Rainer Göb
    Accepted
    13-Sep-2017 09:20 Issues in the Revision of ISO 3951-1
    The ISO (International Organization for Standardization) acceptance sampling standards and their national versions are the most widely used statistical standards. ISO 3951 is a multi-part series of standards for variables sampling where the proportion nonconforming is determined as the probability of a normally distributed measurement falling outside a specification range. ISO 3951-1, a descendent of MIL-STD-414, considers sampling plans for univariate plans, sampling plans for multivariate characteristics are provided by ISO 3952-2. The present versions of the standards exhibit deficiencies which have often been criticised by users, in particular: 1) Intransparent structure, missing separation of operational rules, tables, examples, and informative material. 2) The use of nomograms for executing sampling plans under unknown standard deviation of the quality characteristic (so-called "sigma method"). 3) Missing support of quality parameter estimation. The present contribution analyses the deficiencies of 3951-1 and outlines a plan for a suitable revision of the standard. In particular, a method for avoiding the use of nomograms in the operation of the sigma method will be presented. The revision plan is currently being discussed be the responsible technical committee TC 69 ``Application of statistical methods'' at ISO.