Quadratic nonnegative matrix factorization

In Nonnegative Matrix Factorization (NMF), a nonnegative matrix is approximated by a product of lower-rank factorizing matrices. Most NMF methods assume that each factorizing matrix appears only once in the approximation, thus the approximation is linear in the factorizing matrices. We present a new class of approximative NMF methods, called Quadratic Nonnegative Matrix Factorization (QNMF), where some factorizing matrices occur twice in the approximation. We demonstrate QNMF solutions to four potential pattern recognition problems in graph partitioning, two-way clustering, estimating hidden Markov chains, and graph matching. We derive multiplicative algorithms that monotonically decrease the approximation error under a variety of measures. We also present extensions in which one of the factorizing matrices is constrained to be orthogonal or stochastic. Empirical studies show that for certain application scenarios, QNMF is more advantageous than other existing nonnegative matrix factorization methods.     

T-stress near the tips of a cruciform crack with unequal arms

The T-stress near the tips of a crack of cross shape embedded in an isotropic elastic solid is analyzed. The integral transform technique is employed to convert the associated boundary value problem to a system of singular integral equations. According to the stress difference method, T-stresses can be expressed as a sum of an integral involving crack opening displacement (COD) and applied loading at infinity. Obtained results indicate that, in addition to applied loading, T-stresses at the horizontal (vertical) crack tips depend on the COD of the vertical (horizontal) crack surface. COD plays a leading role in determining T-stresses, in particular for a cruciform crack of two crack-arm lengths of the same order. Moreover, T-stresses for a single-crack limiting case can be recovered from the present results as the length of one arm approaches zero. For a biaxial tension of the same magnitude, T-stresses are present for a cruciform crack, but absent for a single crack. Finally, for several cases of interest, T-stresses for a cruciform crack are evaluated and compared with those for a single crack, and the influence of the ratio of two crack-arm lengths b/a and the COD on the T-stress of a cruciform crack is examined.     

An efficient nonfullerene acceptor for all-small-molecule solar cells with versatile processability in environmentally benign solvents

A new structure of dicyanodistyrylbenzene-naphthalimide-based nonfullerene acceptor NIDCSN was synthesized and characterized to have a favorable electron accepting property and versatile processability in various organic solvents. The nonfullerene all-small-molecule solar cells comprising p-DTS(FBTTh₂)₂ as the donor and NIDCSN as the acceptor exhibited a maximum power conversion efficiency of 3.45% with a remarkable open-circuit voltage of 1.04 V, together with similar device performances when fabricated in five different solvents including environmentally benign non-halogenated ones.     

Inverse dynamic analysis type of MPPT control strategy in a thermoelectric-solar hybrid energy harvesting system

This study presents the development of a novel inverse dynamic analysis-maximum power point tracking (IDA-MPPT) scheme in a hybrid energy harvesting system between thermoelectric module (TEM) and solar array (SA). The proposed method initially changes the harvested voltage response from both sources to be the third-order exponential function. This input function selection is based on the capability of this function to stabilize the initial response system and maintain its final position despite a prolonged response time. The mV voltage value from TEM is easily boosted up to nearly 5 V using this method. With this hybridization, the total obtained voltage is doubled to become 9.7 V, which results in a total power of 0.722 W. Furthermore, the method also allows for a fast tracking system, which enables faster voltage boosting and supercapacitor charging. The supercapacitor only requires less than 5 min to complete charging and boost the voltage to almost 5 V. Thus, a satisfactory value is obtained as compared with that of the TEM system with a chosen MPPC board.     

A new model for PCBM phase segregation in P3HT:PCBM blends

The phase segregation in P3HT:PCBM blend films has been investigated from an experimental and theoretical viewpoint. Optical microscopy, atomic force microscopy, scanning electron microscopy and X-ray diffraction show that thermal annealing of P3HT:PCBM blend films leads to the formation of PCBM aggregates. These aggregates are composed of dense randomly packed ∼50 nm PCBM crystallites with an overall aggregate density of ∼0.85 g cm⁻³. By applying the critical radius of nucleation for PCBM and the Stokes-Einstein equation for mobility of PCBM in a P3HT matrix, a model is developed which explains the formation of both crystallites and aggregates.     

The effects of paints and moisture content on the indoor air emissions from pinewood (Pinus sylvestris) boards

The emissions of volatile organic compounds (VOCs) from building materials may significantly contribute to indoor air pollution, and VOCs have been associated with odor annoyance and adverse health effects. Wood materials together with coatings are commonly used indoors for furniture and large surfaces such as walls, floors, and ceilings. This leads to high surface-to-volume ratios, and therefore, these materials may participate remarkably to the VOC levels of indoor environment. We studied emissions of VOCs and carbonyl compounds from pinewood (Pinus sylvestris) boards of 10% and 16% moisture contents (MC) with three paints using small-scale test chambers (27 L). The emissions from uncoated pinewood and paints (on a glass substrate) were tested as references. The 28-day experiment showed that the VOC emissions from uncoated pinewood were lower from sample with 16% MC. Painted pinewood samples showed lower emissions compared to paints on glass substrate. Additionally, paints on 16% MC pinewood exhibited lower emissions than on drier 10% MC wood. The emissions from painted pinewood samples were dominated by paint-based compounds, but the share of wood-based compounds increased over time. However, we noticed differences between the paints, and wood-based emissions were clearly higher with the most permeable paint.     

Influence of specimen type and reinforcement on measured tension–tension fatigue life of unidirectional GFRP laminates

It is well known that standardised tension–tension fatigue test specimens of unidirectional (UD) glass-fibre-reinforced plastics (GFRP) laminates tend to fail at end tabs. The true fatigue life is then underestimated. The first objective of this study was to find for UD GFRP laminates a test specimen that fails in the gauge section. The second objective was to compare fatigue performance of two laminates, one having a newly developed UD powder-bound fabric as a reinforcement and the other having a quasi-UD stitched non-crimp fabric as a reinforcement. In the first phase, a rectangular specimen in accordance with the ISO 527-5 standard and two slightly different dog-bone shaped specimens were evaluated by means of finite element modelling. Subsequent comparative fatigue tests were performed for the laminates with the three specimen types. The results showed that the test specimen type has a significant effect on the failure mode and measured fatigue life of the laminates. A significantly higher fatigue life was measured for the laminate with the powder-bound fabric reinforcement when compared to the laminate with the stitched reinforcement.     

On the stability of conventional and nano-structured carbon-based catalysts in the oxidative dehydrogenation of ethylbenzene under industrially relevant conditions

Relevant carbon-based materials, home-made carbon–silica hybrids, commercial activated carbon, and nanostructured multi-walled carbon nanotubes (MWCNT) were tested in the oxidative dehydrogenation of ethylbenzene (EB). Special attention was given to the reaction conditions, using a relatively concentrated EB feed (10 vol.% EB), and limited excess of O₂ (O₂:EB = 0.6) in order to work at full oxygen conversion and consequently avoid O₂ in the downstream processing and recycle streams. The temperature was varied between 425 and 475 °C, that is about 150–200 °C lower than that of the commercial steam dehydrogenation process. The stability was evaluated from runs of 60 h time on stream. Under the applied reactions conditions, all the carbon-based materials are apparently stable in the first 15 h time on stream. The effect of the gasification/burning was significantly visible only after this period where most of them fully decomposes. The carbon of the hybrids decomposes completely rendering the silica matrix and the activated carbon bed is fully consumed. Nano structured MWCNT is the most stable; the structure resists the demanding reaction conditions showing an EB conversion of ∼30% (but deactivating) with a steady selectivity of ∼80%. The catalyst stability under the ODH reaction conditions is predicted from the combustion apparent activation energies.