This paper considers the dynamics of traffic on a ring road-based transportation network around a major city, via traffic flow time series analysis and characterization. In particular, three traffic flow time series are examined. Two of the time series are acquired from measurement stations located on highways, while one is from a station on the ring road around Ljubljana city. For the analysis and characterization of time series the novel test called 01 test for chaos is applied. Based on the outputs of the test it is concluded that the observed traffic dynamics is inherently chaotic. Additionally, a more detailed characterization of traffic dynamics is carried out on Lyapunov spectrum basis, which reveals that traffic dynamics on the highway is quantitatively quite different from the traffic dynamics on the ring road.
COBISS.SI-ID: 13074715
Until 2016 power plants within the EU will have to meet new limits on emissions as dictated by EU regulations. One of the major challenges is to reduce emissions of nitrogen oxides (NOx) due to health and ozone-formation concerns. Combustion optimisation is one of the primary measures for reducing NOx emissions from boilers burning coal, oil, or natural gas. The optimisation can be achieved by excess air control, boiler fine tuning and balancing the fuel and air flow to the various burners in order to reach minimum NOx formation. In this paper, a multi-step-ahead prediction of NOx emissions that can provide a basis for on-line control is presented. About 9 days worth of real data were acquired from an operator of a coal-based power plant for this study. It begins with a presentation of measured variables, pre-processing of the data and a definition of performance measures. Feature selection analysis follows, identifying the variables important for multi-step NOx prediction. In this respect, the impact of primary variables that are directly related to the combustion process is compared against that of other variables important in boiler operation and some transformed variables. Based on optimal features, a model comparison study including linear (ARX and ARMAX) and nonlinear (NN and SVR) modelling approaches is presented. Results of the model comparison study reveal that for the analysed boiler, nonlinear models do not improve the robust prediction performance of a linear ARX model. In the last part of the paper, an adaptive modelling approach further investigates the potential improvements in NOx prediction. A comparison of static and adaptive versions of the linear ARX model reveals that the adaptive approach does not improve prediction performance significantly. Hence the static ARX model in combination with an optimally selected set of input variables and extracted features is recommended for the multi-step NOx prediction of the coal-based boiler.
COBISS.SI-ID: 12669211
In the laser droplet generation process three different dripping regimes are experimentally observed in dependence on the detachment pulse power. Besides being nonlinear, the process is also inherently nonstationary. In order to consistently analyze all the dripping scenarios based on an experimental time series, time-frequency analysis by means of instantaneous frequency is used. For the calculation of instantaneous frequency, the most recent developments of the HilbertHuang transform are applied, i.e. ensemble empirical mode decomposition, empirical amplitude/frequency modulation decomposition, and direct quadrature. In time-frequency spectra specific patterns are associated with corresponding dripping regimes. By means of a detailed inspection of patterns, the influence of the detachment pulse power on dripping dynamics is characterized.
COBISS.SI-ID: 12970011
Organizing and optimizing production in small and medium enterprises with small batch production and many different products can be very difficult. This paper presents an approach to organize the production cells by means of clustering-manufactured products into groups with similar product properties. Several clustering methods are compared, including the hierarchical clustering, k-means and self-organizing map (SOM) clustering. Clustering methods are applied to production data describing 252 products from a Slovenian company KGL. The best clustering result, evaluated by an average silhouette width for a total data set, is obtained by SOM clustering. In order to make clustering results applicable to the industrial production cell planning, an interpretation method is proposed. The method is based on percentile margins that reflect the requirements of each production cell and is further improved by incorporating the economic values of each product and consequently the economic impact of each production cell. Obtained results can be considered as a recommendation to the production floor planning that will optimize the production resources and minimize the work and material flow transfer between the production cells.
COBISS.SI-ID: 12359195
It is known that recurrent neural networks may have difficulties remembering data over long time lags. To overcome this problem, we propose an extended architecture of recurrent neural networks, which is able to deal with long time lags between relevant input signals. A register of latches at the input layer of the network is applied to bypass irrelevant input information and to propagate relevant inputs. The latches are implemented with differentiable multiplexers, thus enabling the derivatives to be propagated through the network. The relevance of input vectors is learned concurrently with the weights of the network using a gradient-based algorithm.
COBISS.SI-ID: 10113876