Search

A proposed machine learning method capable of mining feature interactions reveals the mechanisms of bio-nano interactions., The development of machine learning provides solutions for predicting the complicated immune responses and pharmacokinetics of nanoparticles (NPs) in vivo. However, highly heterogeneous data in NP studies remain challenging because of the low interpretability of machine learning. Here, we propose a tree-based random forest feature importance and feature interaction network analysis framework (TBRFA) and accurately predict the pulmonary immune responses and lung burden of NPs, with the correlation coefficient of all training sets >0.9 and half of the test sets >0.75. This framework overcomes the feature importance bias brought by small datasets through a multiway importance analysis. TBRFA also builds feature interaction networks, boosts model interpretability, and reveals hidden interactional factors (e.g., various NP properties and exposure conditions). TBRFA provides guidance for the design and application of ideal NPs and discovers the feature interaction networks that contribute to complex systems with small-size data in various fields.

The multi-attribute information appears in real world, which also includes numeric and categorical attributes. However, the previous classification algorithms for both numeric and categorical data exist in some limitations on categorical data. In this paper, a supervised Kohonen network with heterogeneous value difference metric is proposed for both numeric and categorical inputs. It employs the framework of supervised Kohonen networks, adopts heterogeneous value difference metric to measure dissimilarity between numeric and categorical data, uses the frequency of each categorical item in the Voronoi set to update the reference vector of categorical attribute on the competitive layer, and updates different competitive learning rules for numeric and categorical data. The effectiveness of the proposed algorithm is verified by UCI Machine Learning Data Repository. The classification accuracy is compared with BP, k-NN, naive Bayes network, C4.5 and SVM; the dissimilarity metric is analyzed. The proposed classification algorithm is applied to the operating mode classification for wind turbines; the effectiveness is illustrated in condition monitoring for pitch system of wind turbines.

It is of great significance for load monitoring to monitor illegal use of electricity. Load monitoring can provide supervision for government and improve residents' awareness of safety. Compared with the intrusive load monitoring, non-intrusive load monitoring has the advantages of good economy, high reliability, and quickly realizing the electricity decomposition. Many scholars have carried out this research, but there still exists the problems: it is difficult to obtain key information from big data and the diagnostic results are inaccuracy. Therefore, load monitoring is conducted in this paper. To overcome the above shortcomings, firstly this paper obtains key information of sample based on harmonic analysis. Secondly, random forest based on multiple decision trees has better accuracy in recognition. So the random forest algorithm is applied to machine learning and pattern recognition. Finally, the proposed method is identified and analyzed. The results show that the accuracy of the online electrical detection based on the harmonic analysis and random forest algorithm is greater than 86%, which shows the effectiveness of the method.

High pixel per inch (PPI) and high resolution micro-LED (μLED) displays are attracting more and more attentions. The increasing pixel number requires a large amount of bonding pads and brings huge difficulties to micro-LED system design and lowers power efficiency as well. It is urgent to integrate row and column driving circuits onto the micro-LED panel. Here we report a fully integrated active matrix programmable micro-LED system on panel (SoP) with ultraviolent (UV) and blue emission wavelengths. The micro-LED SoP has a resolution of 60x60 and pixel pitch of 70μm. The micro-LED SoP was achieved by integrating micro-LED arrays with silicon-based PMOS driving panel using fine-toned flip-chip bonding technology. With fully integrated scan and data circuits, the number of bonding pads was greatly reduced from 136 to 28 and large amount of metal interconnection lines were saved. The micro-LED SoP panel was mounted on a periphery driving board and representative characters were displayed successfully.

High pixel per inch and high-resolution micro-LED displays are attracting more and more attentions. The increasing pixel number requires a large amount of bonding pads and brings huge difficulties to micro-LED system design and lowers power efficiency as well. It is urgent to integrate row and column driving circuits onto the micro-LED panel. Here, we report a fully integrated active matrix programmable micro-LED system on panel (SoP) with ultraviolet and blue emission wavelengths. The micro-LED SoP has a resolution of 60 × 60 and pixel pitch of 70 μm. The micro-LED SoP was achieved by integrating micro-LED arrays with silicon-based p-channel metal-oxide semiconductor driving panel using fine-toned flip-chip bonding technology. With fully integrated scan and data circuits, the number of bonding pads was greatly reduced from 136 to 28, and large amount of metal interconnection lines were saved. The micro-LED SoP panel was mounted on a periphery driving board, and representative characters were displayed successfully.

In this paper, we report the design and fabrication of fine-pitch and pixelated-addressed micro-light emitting diode (LED) arrays with emission wavelengths of red (R), green (G), blue (B), and infrared (IR). The arrays have a resolution of $8\times 8$ with monochromatic LED chips directly bonded to custom-designed printed circuit boards to form pixelated matrix. R, G, B, and IR micro-LED arrays with a pixel pitch of 1 mm were demonstrated. Sequential row scanning and multiple columns programming is performed by two CMOS-based application specific integrated circuit chips providing constant-sinking current and forward voltage for each pixel of the micro-LED arrays. The uniformity of operating current and forward voltage of the micro-LED pixels was measured with 1.94% coefficient of variation. The micro-LED pixel has much shorter rising time and higher response frequency than those of regular packaged blue and white LEDs. The implemented R–G–B micro-LED arrays are the excellent candidate for high-definition fine-pitch LED displays and demonstrated its versatile functionalities such as visible light communications. The IR micro-LED arrays show great potential in health medical treatment, IR detection imaging and local positioning system.

In this paper, we investigated leakage current generated by light emission of micro-light emitting diode (LED) pixels in active matrix (AM) micro-LED displays. Custom-designed structures of metal cover with different coverage lengths were designed and fabricated to completely eliminate the photon-generated leakage current, thus enhancing the performance of AM driving circuits. Working principle of two transistors and one capacitor AM pixel and requirement of its driving transistors were also systematically discussed for liquid crystal displays, organic LED displays, and LED microdisplays. Transistors with the optimized metal-cover design have an ON/OFF-current ratio of 108, which is two orders of magnitude higher than those without metal cover. This concept was successfully implemented in a high-resolution and fine-pitch AM micro-LED display with a resolution of $400\times 240$ and a pixel pitch of $30~\mu \text{m}$ . Vivid images and movies were presented to show the great potential for wearable applications.

Coal fines migration exerts negative impacts on early water drainage of undersaturated coal seam gas (CSG) reservoirs. The complicated migration process results in ineffective and inaccurate forecast of coal fines production. Hence, a robust modelling tool is required to include the mechanisms of fines migration and to predict their impacts on rock and production. In this paper, fines migration in coal is categorized into three stages: generation, migration, and deposition processes. The corresponding models for different stages are established, including (1) a fines generation model, (2) the maximum fines-carrying concentration model and deviation factor of the modified Darcy model, (3) a fines deposition model, and (4) a dynamic permeability and porosity model. The above models are coupled with a water flow model, solved numerically using the finite difference method. Then, two dewatering strategies, including fast and moderate depressurization, are compared using the proposed models to study their effects on coal properties and following production. Finally, the production history of a CSG well in the Qinshui Basin, China, is utilized for history matching in a field case study. The simulation results indicate that new fines will be generated in a fast depressurization process and the water rate decline reduces the cleat permeability significantly. The newly generated fines can enhance the permeability temporarily, but they will block the flow channels and bring serious damage to the permeability when the water rate declines. The moderate depressurization strategy can produce the coal fines in a continuous mode, and the formation damage induced by fines deposition can be reduced to the acceptable level, which is the more reliable way to maintain well productivity. In addition, multiple well shut-in can trigger the irreversible fines deposition, reduce the permeability, and decrease the production rate.

In coal seam gas (CSG) reservoirs, hydraulic connectivity usually exists between the coal and overlying permeable formation, which results in the delay of the dewatering process and a large volume of produced water. The cross-formational water flow can dramatically influence the production of CSG reservoirs, reduce the economic viability of production and increase the impact on the environment. To assess the cross-formational water flow of a CSG well, this study investigates the effects of cross-formational water flow from the overlying permeable formation on gas production. Simulation models are constructed to represent the water crossflow in four connection scenarios, including contact zone with and without penetrated fracture, and no-contact zone with and without penetrated fracture. The penetration ratio, defined as the ratio of fracture height in permeable formation to the formation thickness, is employed to characterize the extension degree of hydraulic fracture in the overlying formation. The effects of cross-formational water flow on CSG production are analyzed using dimensionless production curves and relative permeability curves calculated from the production history. The field application of this study is demonstrated by the production history of fractured CSG wells in the Qinshui Basin, China. The simulation results show that if the overlying formation is directly connected to the coal seam (contact zone scenario), the hydraulic fracture leads to the decrease of the peak gas production rate with the cumulative water production increasing dramatically. If the overlying formation is connected to the coal seam by hydraulic fractures only (no-contact zone scenario), gas production increases in two stages. Initially, the gas production rate exhibits a slow increase, followed by a fast increase before reaching the peak rate. When the hydraulic fracture penetrates the overlying formation, a 10% increase of the penetration ratio results in 0.55% and 1.82% increase of cumulative water production in the indirect and direct connection scenarios, respectively. In both scenarios, 10% penetration ratio increase leads to 7.41% reduction of gas production rate. The methods and simulation results presented in this work can be used to evaluate the impacts of cross-formational water flow on gas production and identify the cross-formational water flow in CSG wells from production history.

In this paper, we report the design and fabrication of full-color pixelated-addressable light emitting diode on transparent substrate (LEDoTS) micro-displays. The LEDoTS micro-displays were realized by integrating monochromatic LEDs on a transparent quartz substrate using chip on board (CoB) technique. The full-color pixel has a pitch of 1 mm (P1) and includes red, green, and blue monochromatic LED chips. The pixelated-addressable driving circuit was designed on a custom PCB to provide constant current and individual addressing of each LED pixel with a certain frequency of row and column scanning. With a peripheral FPGA board, the LEDoTS micro-displays were programmed to display representative images and animations.

For improving convergence rate and preventing prematurity in quantum evolutionary algorithm, an allele real-coded quantum evolutionary algorithm based on hybrid updating strategy is presented. The real variables are coded with probability superposition of allele. A hybrid updating strategy balancing the global search and local search is presented in which the superior allele is defined. On the basis of superior allele and inferior allele, a guided evolutionary process as well as updating allele with variable scale contraction is adopted. AndHεgate is introduced to prevent prematurity. Furthermore, the global convergence of proposed algorithm is proved byMarkovchain. Finally, the proposed algorithm is compared with genetic algorithm, quantum evolutionary algorithm, and double chains quantum genetic algorithm in solving continuous optimization problem, and the experimental results verify the advantages on convergence rate and search accuracy.

A micro biochemical sensor based on a SOI optical waveguide sensing platform is reported in this article. The sensing platform utilizes a Mach-Zehnder Interferometer (MZI) configuration. In order to satisfy the single-mode transmission condition of the waveguide and match the coupling size of standard single-mode fiber, a ridge waveguide structure is adopted to construct the MZI configuration. The key parameters of the waveguide are optimized by FDTD method. A process composed of contact exposure photolithography and Inductively Coupled Plasma (ICP) etching technology has been worked out for the fabrication of the optical waveguide sensing platform on SOI wafer. In the process, only one photolithographic mask is used. The mask is designed to have several patterns including an array of MZI configurations for multiple species sensing, a group of ridge waveguides with different key parameters as test references, and a group of geometric compensation schemes to test the methods of modifying the deviation which may arise during the ICP etching step. Process simulations are conducted to predict the exposure and etching results, which give strong support to the fabrication process control. Lithography problems including photoresist desquamation and line narrowing are solved to achieve the special structure.

This paper focuses on the method of setting pre-camber for steel truss girder bridge with the known pre-camber curve. The relative elevation between two adjacent joints is taken as the input variable to deduce the mathematical expressions of the length variation of each member based on the mechanical equilibrium equation of structure formed in stages. Considering the characteristics of pre-camber setting of through steel truss girder bridge, the length variation formula of each member is further developed here in three cases: (1) changing the length of upper chord member bar only,(2) changing both the length of upper chord and diagonal wed member bar, and (3) changing the length of upper chords, diagonal wed and vertical member bar simultaneously. An example study is carried out to illustrate the application of the method. The analysis results show that the unstressed pre-camber setting method is an absolutely geometrical method, which will not generate stress in members and supports during the pre-camber setting. This method, of which the concept is clear and reliable, deserves being developed to be used in the other types of bridges.

Off-axis aspheric surface is an indispensable optical device for the high precision measurement and manufacturing equipment, and the tool path generation is a very important factor for the manufacturing to obtain the surface with high precision form and nanometric surface finish. In this paper, the systemic description of tool path generation method about the off-axis parabolic surface with ultra-precision single point diamond turning is processed. Two effective methods of generating the accurate spiral tool path are proposed and compared. The proposed methods are mainly based on mathematical calculation directly from analytic surfaces, including the general process of tool path generation, cutting-tool compensation and avoiding tool interference. This work can provide the theoretical basis of choosing the cutting tool path for off-axis parabolic surface and can improve the efficiency and precision of machine. Finally, some cutting simulations are implemented to prove the proposed methods.

With the development of economy and society, urban construction rose to a new level in China. High-rise and super high-rise building increase rapidly. In the high-rise and super high-rise buildings (H≥150 m), the structural system usually adopts the conventional reinforced concrete, including frame tube structure, tube in tube structure, steel frame with reinforced concrete core tube hybrid structure and tube in tube steel - concrete mixed structure. These traditional structural systems often lead to the problems of fat beam fat column and high expenditure, which is not economic and unfriendly to the environment. This paper not only presents a new type of concrete core wall and prefabricated space steel grid cassette tube in tube structure system, but also discusses the superiority of the new type structure through the innovation and practice of a super high (266 m) five-star hotel and office building integrated engineering construction.

One-dimensional chirped silicon grating waveguide with round-corners has been designed to obtain wideband slow light. Wide bandwidth up to 9.6-nm slow light with group index of 40.7 can be obtained in an optimized waveguide.

The main gymnasium of China University of Petroleum is single-layer hyperbolic paraboloidal roof, and the center part of the dome adopts crossed web bars in order to increase stiffness of the structurs. The 1∶12 scaled model has been made according to the structure of the construction drawings and tested in the stages of vertical combined loading. The experimental results are compared with the result of numerical simulation by Abaqus. The mechanical properties of the model are studied under self-weight load and design load. The regularity of the members' axial force and joints' displacements are available, and the results show that the deflection of dome under design load conforms to the requirement of specifications, which verified the correctness of the structure design. It is found that there is no overall instability and local instability on the testing site, and the behavior of the structure is almost linear.

Using computational fluid dynamics software, the fluid flow in cooling water jacket of HPD diesel engine is studied. Results show that all parts in A Row cylinders cooling water jacket have a better cooling effect, such as in cylinder block and in fire deck. And cooling effect for each cylinder is uniform. Total pressure loss through cooling water jacket is 32.6kPa. The size of water up-holes is better for cooling cylinder head, which can also provide boundary conditions for further calculating temperature field of cylinder head.

The real-time monitoring of vehicles velocity has become essential for traffic safety. Therefore, in order to improve the method of monitoring the velocity of vehicles on road, this paper has designed a new radar velocity-measuring system which is based on ARM processor. This system is based on Doppler principle and spectral estimation algorithm, in its hardware design part, and it takes full use of the underlying resources of ARM processor, to achieve sample, digital signal processing and peripheral control in the separate ARM core. In the software design part, it analyses the main program flow, at the same time, since the ARM processor does not support hardware floating-point operations, it puts forward the usage of fixed-point numbers to store and calculate floating-point numbers, which makes the data loading and storage minimization; And it uses ARM assembly language to write the core algorithm and takes full advantage of the 5-stage pipeline of ARM processor, to achieve the right register allocation and instruction scheduling for fine-grained control, and also to avoid the pipeline interlocks. Experimental results indicate that the accuracy of this velocity measurement system can reach ±1km/h. It not only can be used in traffic enforcement agencies, but also has low cost, low power consumption advantages.

Due to multi-collinearity of the traditional aircraft's magnetic inference model, it is difficult to estimate the parameters of model accurately. And the magnetic anomaly detection (MAD) in airborne is affected too. Because of the time-domain characteristics of magnetic target, the magnetic signal of target was considered as a kind of short-time signal. Using the statistical characteristics of aircraft's background noise, the method of adaptive detection removed magnetic noise adaptively. Based on the differences between statistical characteristics of target and noise, the detection factor proposed has strong magnetic anomaly detection capability with background of strong complex noise. The method could detect a magnetic anomaly less than 1 nT in 500m away. More than one target could be detected at the same time. The average of factor in different reference signal has better resolution than single factor in detection. Besides adaptive detection simplifies the compensation process and improves the detection efficiency of aeromagnetic survey.

Traditional method of compensating aircraft magnetic inference is complicated. In order to solve coefficients of inference, aircraft must have a box wide sides study flight before survey. However due to multi-collinearity of the traditional aircraft's magnetic inference model, it is difficult to estimate the coefficient of model accurately, and the aircraft's inference couldn't be removed clearly. The adaptive filtering algorithm based on wavelet transformation uses result of wavelet transformation for measured magnetic signal to be the reference signal of adaptive filtering. The target signal isn't correlated with noise, so the filter effect works well and the aircraft's magnetic inference compensated well. The method could simplify the process of compensation and reduce the learning time in the air. By using this method to measured data, aeromagnetic survey could detect magnetic anomaly signal both aircraft in violent shaking and smooth flight without using any another equipment such as GPS.

On the base of three-dimensional model of HPD diesel engine cooling system water jacket and cylinder head, using Fire mesh module of the AVL software, the grid models were built up, the interface of the head and inner water jacket was coupled. Dependent on boundary conditions simulated on the grid model of cooling system water jacket, using the method of dispersing and resolving, the coupled grid model was calculated and the temperature of the cylinder head was obtained under the condition of rated speed and power. The simulating results are well matched with the experimental results, which can provide gist for further calculating thermal stress and thermal crack.

In this paper, we propose a new channel estimation method for IEEE802.11a and HIPERLAN/2 wireless LANs. Based on the frequency coherence of the channel for wireless LANs, the absent frequency channel responds at virtual subcarriers and DC subcarrier are reconstructed by linear interpolation firstly, then channel estimation with time domain truncation is fulfilled. Simulation results show that the proposed method achieves a slightly better performance than the conventional iterative method and LS method while keeping less complexity.