University of Macau / Faculty of Science and Technology / Department of Mathematics

Thesis (M.A.)--Boston University / Some rapid approximate methods of estimating the standard deviation from samples of moderate size (20 < n < 100) are presented. The emphasis is placed on solutions of problems commonly encountered in statistical quality control, especially in the electronics industry. Factors and efficiency values are given for the use of these estimators on normally distributed data. Statistical and practical engineering and administrative criteria are suggested for testing whether particular estimators are desirable in the usual industrial situation. The estimates discussed in this paper are all order statistics, i.e. statistics which are a function of only a small number of observations selected from the whole sample. These observations are selected because of the position they occupy among the other observations when all sample observations are arranged in order of magnitude. The first estimator discussed, for instance, is the range. The range of a sample is merely the numerical difference between the largest member of the sample and the smallest member of the sample. The standard deviation of the distribution from which the sample was drawn may be estimated by dividing the range by a suitable constant. The constant is a function of the sample size and of the shape of the distribution. Factors are given for sample sizes up to ten, for the normal distribution only. [TRUNCATED]

No description available.

<p>I present an application to visualization of GPS data and Linear Correlations and models. A collection of data for each vehicle is used to compute correlations. Deviating correlations can be indicative of a faulty vehicle.</p><p> The correlation values for each vehicle are computed with use linear regression algorithms using up to 4 signals in the data (with varied time window), and display the model parameters in a window next to the GPS map. Multiple measurements (multiple drive routes and multiple model parameters) are displayed at the same time, allowing tracking over time and comparison of different vehicles.</p><p> </p><p> The whole technique is demonstrated on three data which is set on first frame by user. The results are displayed with a java application and Google Map.</p>

I present an application to visualization of GPS data and Linear Correlations and models. A collection of data for each vehicle is used to compute correlations. Deviating correlations can be indicative of a faulty vehicle. The correlation values for each vehicle are computed with use linear regression algorithms using up to 4 signals in the data (with varied time window), and display the model parameters in a window next to the GPS map. Multiple measurements (multiple drive routes and multiple model parameters) are displayed at the same time, allowing tracking over time and comparison of different vehicles. The whole technique is demonstrated on three data which is set on first frame by user. The results are displayed with a java application and Google Map.

The correlation of cascade experimental data is one method for obtaining compressor stage characteristics. These correlations specify pressure loss and flow turning caused by the blades. Current open literature correlations used in streamline curvature codes are inadequate for general application to high-speed transonic axial-flow compressors. The objective of this research was to investigate and evaluate the available correlations and ultimately discover sets of correlations which best fit the empirical data to be used in streamline curvature codes. Correlations were evaluated against experimental data from NASA Rotor 1-B and NASA Stage 35. It was found that no universal set of correlations was valid for minimum-loss point predictions. The Bloch shock loss model showed promising results in the stall regime for supersonic relative inlet Mach numbers. The Hearsey and Creveling off-minimum-loss deviation angle prediction performed consistently better than all other correlations tested. / Master of Science

International Telemetering Conference Proceedings / October 29-November 02, 1990 / Riviera Hotel and Convention Center, Las Vegas, Nevada / This paper briefly reviews past techniques for measuring FM deviation and discusses the limitations of past technology. Graphs of the Bessel functions are presented in terms of decibels (dB), offering a better method of measurement when used with a modern spectrum analyzer.

<p>We present a technique to identify deviating systems among a group of systems in a</p><p>self-organized way. A compressed representation of each system is used to compute similarity measures, which are combined in an affinity matrix of all systems. Deviation detection and clustering is then used to identify deviating systems based on this affinity matrix.</p><p>The compressed representation is computed with Principal Component Analysis and</p><p>Kernel Principal Component Analysis. The similarity measure between two compressed</p><p>representations is based on the angle between the spaces spanned by the principal</p><p>components, but other methods of calculating a similarity measure are suggested as</p><p>well. The subsequent deviation detection is carried out by computing the probability of</p><p>each system to be observed given all the other systems. Clustering of the systems is</p><p>done with hierarchical clustering and spectral clustering. The whole technique is demonstrated on four data sets of mechanical systems, two of a simulated cooling system and two of human gait. The results show its applicability on these mechanical systems.</p>

Nowadays the globalization and the economy expansion of the emerging countries demand anincreasing amount of energy. Therefore, energy production as well as the efficiency of energyusage, is essential for future developments of societies. Renewable energies appear as a turnkeysolution that could support the growing demands, and at the same time not being harmful to theenvironment [1]. Within the types of renewable energies, wind energy could be considered asone with large potential.In this paper I present the study of a Swedish wind farm placed in Rockneby.Once the wind turbines were installed and working correctly, a discrepancy between the realenergy obtained and the theoretical energy indicated by the manufacturer was detected. Thestored data in the SCADA system were compared with the values provided by the manufacturerand several analyses were performed. Initially an anomaly in the power residual deviation wasdetected. It was showing an unusual behaviour at high wind speeds. The variation of the airdensity in the wind farm at hub height was considered as a possible reason of the disagreementobserved in the power parameters since the reference density used by the manufacturer was aconstant value calculated in laboratory environment. However, this idea was rejected becausethe power generated in both conditions is similar. The pitch angle was analysed after detectinga significant variations in wind speed measurements made by the anemometer in the turbinenumber three. As a result, it was found a pitch variation in the turbine which seems due to afailure in the anemometer. As a final result, the turbulences were analysed giving as aconclusion that the turbulence intensity were situated around 20%. Therefore, I mainly suggestas a possible explanation of this fact the influence of the turbulence accompanied of a badcalibration or failure in the anemometers.

We present a technique to identify deviating systems among a group of systems in a self-organized way. A compressed representation of each system is used to compute similarity measures, which are combined in an affinity matrix of all systems. Deviation detection and clustering is then used to identify deviating systems based on this affinity matrix. The compressed representation is computed with Principal Component Analysis and Kernel Principal Component Analysis. The similarity measure between two compressed representations is based on the angle between the spaces spanned by the principal components, but other methods of calculating a similarity measure are suggested as well. The subsequent deviation detection is carried out by computing the probability of each system to be observed given all the other systems. Clustering of the systems is done with hierarchical clustering and spectral clustering. The whole technique is demonstrated on four data sets of mechanical systems, two of a simulated cooling system and two of human gait. The results show its applicability on these mechanical systems.