Accelerated aging of thick glass second surface silvered reflectors under sandstorm conditions

Concentrated solar power systems are expected to be sited in desert locations where the direct normal irradiation is above 1800 kWh/m2.year. These systems include large solar collector assemblies, which account for a significant share of the investment cost. Solarreflectors are the main components of these solar collector assemblies and dust/sand storms may affect their reflectance properties, either by soiling or by surface abrasion. While soiling can be reverted by cleaning, surface abrasion is a non reversible degradation.The aim of this project was to study the accelerated aging of second surface silvered thickglass solar reflectors under simulated sandstorm conditions and develop a multi-parametric model which relates the specular reflectance loss to dust/sand storm parameters: wind velocity, dust concentration and time of exposure. This project focused on the degradation caused by surface abrasion.Sandstorm conditions were simulated in a prototype environmental test chamber. Material samples (6cm x 6cm) were exposed to Arizona coarse test dust. The dust stream impactedthese material samples at a perpendicular angle. Both wind velocity and dust concentrationwere maintained at a stable level for each accelerated aging test. The total exposure time in the test chamber was limited to 1 hour. Each accelerated aging test was interrupted every 4 minutes to measure the specular reflectance of the material sample after cleaning.The accelerated aging test campaign had to be aborted prematurely due to a contamination of the dust concentration sensor. A robust multi-parametric degradation model could thus not be derived. The experimental data showed that the specular reflectance loss decreasedeither linearly or exponentially with exposure time, so that a degradation rate could be defined as a single modeling parameter. A correlation should be derived to relate this degradation rate to control parameters such as wind velocity and dust/sand concentration.The sandstorm chamber design would have to be updated before performing further accelerated aging test campaigns. The design upgrade should improve both the reliability of the test equipment and the repeatability of accelerated aging tests. An outdoor exposure test campaign should be launched in deserts to learn more about the intensity, frequencyand duration of dust/sand storms. This campaign would also serve to correlate the results of outdoor exposure tests with accelerated exposure tests in order to develop a robust service lifetime prediction model for different types of solar reflector materials.

Likelihood prediction for generalized linear mixed models under covariate uncertainty

This paper presents the techniques of likelihood prediction for the generalized linear mixed models. Methods of likelihood prediction is explained through a series of examples; from a classical one to more complicated ones. The examples show, in simple cases, that the likelihood prediction (LP) coincides with already known best frequentist practice such as the best linear unbiased predictor. The paper outlines a way to deal with the covariate uncertainty while producing predictive inference. Using a Poisson error-in-variable generalized linear model, it has been shown that in complicated cases LP produces better results than already know methods.

Speed prediction for triggering vehicle activated signs

Accurate speed prediction is a crucial step in the development of a dynamic vehcile activated sign (VAS). A previous study showed that the optimal trigger speed of such signs will need to be pre-determined according to the nature of the site and to the traffic conditions. The objective of this paper is to find an accurate predictive model based on historical traffic speed data to derive the optimal trigger speed for such signs. Adaptive neuro fuzzy (ANFIS), classification and regression tree (CART) and random forest (RF) were developed to predict one step ahead speed during all times of the day. The developed models were evaluated and compared to the results obtained from artificial neural network (ANN), multiple linear regression (MLR) and naïve prediction using traffic speed data collected at four sites located in Sweden. The data were aggregated into two periods, a short term period (5-min) and a long term period (1-hour). The results of this study showed that using RF is a promising method for predicting mean speed in the two proposed periods.. It is concluded that in terms of performance and computational complexity, a simplistic input features to the predicitive model gave a marked increase in the response time of the model whilse still delivering a low prediction error.