China
Africa
Explore chapters and articles related to this topic
Long-term Precipitation Trends
Published in Richard J. Chorley, Introduction to Physical Hydrology, 2019
The year-to-year variability of precipitation totals may conceal long-term changes of one kind or another in a data series, and statistical techniques are necessary to suppress the short-term irregularities. The simplest method is the calculation of a running mean (or moving average), where mean values are determined for successive, overlapping periods of five, ten, or thirty years. For example, in the five-year case P1+P2+P3+P4+P55=P¯3P2+P3+P4+P5+P65=P¯4
Psychophysiological Measures of Driver State
Published in Yvonne Barnard, Ralf Risser, Josef Krems, The Safety of Intelligent Driver Support Systems, 2019
Heart rate and heart rate variability can be used instead of giving artificial secondary tasks to drivers to assess their mental workload, as these cardio-vascular measures reflect mental effort during the driving performance. The analysis of the data is mostly done by a moving average, where averages are calculated over a predefined time window. Instead of workload estimations over an entire drive, cardiac measures can give an indication of high workload periods during the drive, i.e., moments when drivers have to invest more effort. The resolution of these periods may be 10–15 seconds. Examples of studies using these measures can be found in De Waard (1996, 2002), Hoogeboom and Mulder (2004), Mulder et al. (2005), and De Waard et al. (2008).
Statistics for the Safety Professional
Published in W. David Yates, Safety Professional’s, 2015
The median is the middle value in the list of data. To find the median, your numbers have to be listed in numerical order, so you may have to rewrite your list first. If there is no “middle” number, because there is an even set of numbers, then the median is the mean (the usual average) of the middle two values. For example, given the data set {1, 2, 3, 4, 5, 6, 7, 8}, what is the median? The median would lie between 4 and 5; therefore, add 4 + 5 = 9, then divide by 2, giving a product of 4.5. Therefore, the median for the data set is 4.5, which is the central point of the data set.
A meta-analysis and synthesis of public transport customer amenity valuation research
Published in Transport Reviews, 2019
Chris De Gruyter, Graham Currie, Long T. Truong, Farhana Naznin
Type categories cover a wide range of individual amenities but serve to explore the aspects of amenities of concern to passengers. Figure 1 shows considerable variability in customer amenity values. While the 75th percentile values are all under 2 min, individual values of up to 14 min were found (W6, Cleanliness of station/stop in Appendix A). Given this would cause the calculation of ‘average’ values to be skewed towards these maximum values, the reporting of median values is deemed more appropriate. Despite wide variability in valuations by type and amenity, Figure 1 shows little difference in median values by amenity type, perhaps with the exception of facility based amenities which have a lower median value than other amenities and also a narrower range of values within the 25th to 75th percentile. Of the amenity type categories explored, environment (0.60 min), information (0.63 min) and security (0.51 min) have the highest median values. Another key observation is that median valuations are in general all below a single minute in value. The implication is that while amenities are of clear value to customers, their value is in general small compared to overall travel time (typically 30–60 min).
Can transportation network companies replace the bus? An evaluation of shared mobility operating costs
Published in Transportation Planning and Technology, 2022
There are several issues with averaged data. First, it does not provide any information on data variation. Second, it is sensitive to outliers, like NYC MTA, for example. Finally, averages do not tell us how many values are close to the average. However, averages can be useful when data is not available or when the purpose of the analysis is to look at trends, like in the case of this paper. Although using averaged data can be misleading, we can use histograms to better understand how well the averages represent the data sample and identify whether averages are influenced by outliers. Figure 2 and the explanation in Section 3.4 provide this additional information to give the averaged data a little more context.
Creep of geomaterials – some finding from the EU project CREEP
Published in European Journal of Environmental and Civil Engineering, 2022
Gustav Grimstad, Minna Karstunen, Hans Petter Jostad, Nallathamby Sivasithamparam, Magne Mehli, Cor Zwanenburg, Evert den Haan, Seyed Ali Ghoreishian Amiri, Djamalddine Boumezerane, Mehdi Kadivar, Mohammad Ali Haji Ashrafi, Jon A. Rønningen
In this paper, the Murro test embankment is revisited, this has previously been studied be Karstunen, Rezania, Sivasithamparam, and Yin (2015). Karstunen and Yin (2010) established visco-plastic parameters for the Murro clay for the EVP-SCLAY1S model. Even though the mathematical form of the model used by Karstunen and Yin (2010) is different than the proposed unified model, their parameters for the layer with depth of 3.0 to 6.7 m corresponds to a value for β of about 25. When combining this number with the compressibility parameters, this leads to a value for μ* of about 2.5e-3. Finally, using one day as reference time, the corresponding OCR implicitly used by Karstunen and Yin (2010), for this layer, is approximately 1.2. Table 1 shows the β value, the OCR, the OCRmax and μ* after converting the EVP-SCLAY1S parameters for all layers. Figure 2 shows how the two formulations compare for the layer between 3 and 6.7 m. An almost parallel shift of the curve indicates similar creep behaviour of the two formulations within this range of strain rates. Note that the β values are varying between about 10 and 30 for the different layers. This is a large variation within a similar type of clay. At the same time, the OCR used is quite low. This is an indication for sample disturbance affecting the parameter selection e.g. β = 25 gives OCR = 1.52 after 100 years, assuming linear log(OCR) – log() relationship. From the simulation results, the over-prediction of settlement in the bottom layers gives the same indication as the stress increase in this layer is moderate compared to the pre-consolidation stress. A reinterpretation with this in mind leads to parameters given by Tables 2 and 3. Note that this study uses an average parameter set for layers 2 to 5. This will have some implications on giving a perfect match to the measurements. However, using average parameters is more relevant for engineering applications where normally only limited data is available. The remaining parameters are simply estimated based on experience from other sites and/or previous studies of the Murro clay. For more details on some of the index/state and hydraulic parameters, geometry or boundary conditions of the FE model see e.g. Karstunen and Yin (2010) or Sivasithamparam et al. (2015). In the analysis, large deformations are considered using updated mesh and pore water pressures. Figures 3 and 4 give measured and calculated settlements for different locations versus time. The model and the simplified input captures the settlements reasonably well. Figure 5 shows the effect of ignoring creep on the calculated surface settlement.