To study the representation, analysis and processing of system fault evolution process (SFEP), the extension theory is proposed, using intelligent theory and meeting the analysis requirements of system function state. Firstly, the possibility of the combination of system function state and extension theory is discussed. Secondly, the extension representation of SFEP is studied, and the basic unit of SFEP is determined, which is cause event-transfer condition-result event. It is considered that the event is a compound action element of the action element and matter element, and the transfer is a compound relation element of the relation element and the above two compound action elements, which is called the transfer element. Finally, the divergence, correlation, implication and expansibility of the transfer element are studied, and the effect of the method on the analysis of SFEP is explained. The research is the beginning of studying SFEP through extension theory, which provides intelligent analysis method and basic theory for grasping system reliability and failure, i.e. system function state.

In order to improve the efficiency of producing new product ideas and the decision-making in evaluating and selecting the best solutions scientifically, the method of generating product design ideas is refined by integrating Extenics and TRIZ. Firstly, an extension model is established by using the basic element theory to obtain the starting point of product innovation, and then through the cross analysis of technical parameters, invention principles of TRIZ theory and extension analysis and extension transformation methods of Extenics, the specific creativity of the product is obtained. Finally, the evaluation and optimization of the obtained product creativity are carried out by using the optimization evaluation method, and the creativity that meets the needs is selected, and the innovative design scheme of the product is formed. Taking the design of mobile phone charger as an example, a lot of design ideas are obtained and the effectiveness of the refined method proved.

Portfolio selection is one of the core issues in the field of quantitative finance. A diversified portfolio selection strategy is proposed by considering transaction cost and cardinality constraint. Firstly, the return rates of risky assets are regarded as fuzzy numbers and a fuzzy return rate fitting model is proposed to determine the fuzzy distribution of the assets’ return rates. Then, a new diversification measure is proposed for the portfolio and a fuzzy mean-semi-variance-diversification portfolio selection model is established. Next, a modified genetic algorithm is designed to solve the proposed models. Finally, an empirical example with real stock data is used to illustrate the proposed strategy. The results show that the proposed strategy performs better than the conventional portfolio selection strategies.

In a market consisting of both manufacturer and remanufacturer, consumers have different willingness to pay for new and remanufactured products from manufacturer and remanufactured products from remanufacturer. Against this backdrop, a study is conducted on the impacts of legislation on corporate profits, consumer surplus, environmental benefit and social welfare when formulating the constraint situation that the manufacturer should recycle the minimum proportion of new products. The results have shown that recycling rates tend to be strict, reducing manufacturer’s new product yields and increasing their prices. At the same time, the increasing recycling ratio may result in fewer remanufactured products on the market. In addition, environmental benefits are positively related to minimum recovery targets. Numerical analysis illustrates that the higher the social consumer’s willingness to pay for remanufactured products, the better the social welfare. By comparing with the recycling fund policy, the minimum recovery rate has a more significant effect on the improvement of environmental benefits.

In order to solve the problem of high cost in high-precision robot indoor positioning method, an indoor positioning method based on ground features for monocular vision robot is proposed. This method collects the ground image through the monocular fisheye camera installed on the robot chassis, and after a series of image processing, the relationship between the ground grid straight line and the pose of the camera is used as the observational measurements, then the position information of the robot is obtained by deriving the observation equation and the extended Kalman filter algorithm. Based on the current time position data of the robot and the fixed square size of the ground, a line prediction algorithm is proposed to effectively eliminate the influence of the interference line segment. The experimental results show that the positioning accuracy of the robot indoor positioning method can reach 2 cm, which has good robustness and also can provide a good technical reference for low-cost robot indoor positioning.

An improved autonomous frontier exploration method is presented to solve the problems of autonomous mobile robot autonomously exploring indoor unknown environments with low efficiency, poor versatility, and incomplete map construction due to the narrow exploration area. The curve fitting method is used to screen the safety target exploration area for the constructed initial map frontier and for the robot unreachable target exploration area, adopting a sliding window neighborhood planning method to establish a new target exploration point, guiding the robot to autonomously navigate to the target exploration point, while using simultaneous localization and mapping technology to complete the robot's autonomous exploration and mapping of unknown environments. The experimental results show that the method can better solve the problems of narrow target area and unreachable target points in the traditional algorithm. And it can complete autonomous exploration and mapping of unknown indoor complex scenes with fewer exploration times, shorter exploration paths and higher exploration efficiency.

Deep reinforcement learning (DRL) model combining reinforcement learning and deep learning is currently widely used in the field of robot control. Robot reinforcement learning needs to train the model in a 3D simulation environment. However, in the absence of prior environmental knowledge, trial and error learning in a 3D environment leads to long training cycles and high development costs. To solve this problem, a training mode from 2D to 3D is proposed. Time-consuming and computationally intensive work is completed in a 2D environment, and the results are transferred to a 3D environment for testing. Experiments show that this training mode can improve the development efficiency by about five times, so that personal computers can also do research related to robot reinforcement learning.

Text sentiment analysis is a typical task of natural language processing, but the accuracy of existing sentiment analysis is not high, and word characterization is an important reason. A combined weighting method for short text features (CWSTF) is proposed, which can effectively improve the accuracy of sentiment analysis. The CWSTF method evaluates the contribution of features to emotions based on random forests and ranks them, and then filters features based on ranks. Then, the importance of the feature in the document is calculated by TF-IDF (Term Frequency-Inverse Document Frequency), and the final weight of the feature is determined by the importance of the feature in the document and the contribution to the sentiment; Finally, four such classifiers as SVM (Support Vector Machine), NB (Naive Bayes), ME (Maximum Entropy), and KNN (K-Nearest Neighbor) are used for comparison experiments. The experimental results show that the features processed by proposed method can more effectively improve the accuracy of sentiment classification than other methods.

An improved bat algorithm combining adaptive theory, differential evolution and bat optimization algorithm is proposed to solve the typical 0-1 discrete constraints and optimization problem of logic evaluation for fault section location in power distribution network. The global optimization ability of bat algorithm is used to iteratively optimize the fault section of distribution network. At the same time, aiming at the disadvantage that single algorithm is easy to fall into local optimum, differential evolution algorithm is introduced into the process of bat algorithm, and the mutation and crossover operation of the algorithm are adaptively optimized. The simulation results indicate that the algorithm can accurately and effectively locate the fault section of distribution network, and has better fault tolerance.

Based on the fluid volume fraction (VOF) model of Fluent software, the droplet generation process in crossing microchannel was investigated by three-dimensional numerical simulation study, and the influences of continuous phase viscosity, dispersed phase viscosity, two-phase interfacial tension and wall contact angle on droplet generation were researched respectively, providing reference for practical applications. The simulation results showed that the droplet size decreased and the generation frequency increased with the increasing flow velocity of continuous phase. When the continuous phase viscosity was increased, the droplet size would decrease and the generation frequency change was opposite. If the dispersed phase viscosity exceeded the continuous phase viscosity, the jet flow phenomenon would occur and the droplets could not be generated. With the increase of the interfacial tension coefficient of two phases, the droplet size would increase, and the formation frequency would decrease. Increasing wall contact angle is beneficial to droplet generation, and when the two-phase flow rate was 0.01 m/s and 0.02 m/s, the contact angle should be set to 150° and 120° respectively to generate droplets normally.

The pipeline is the most economical water transporting tool for human daily use. However, it generates many issues, like pipe leakage and break. These damages not only cause massive expenditure in repair work but also affect water quality. It is therefore significant for engineers to investigate the water pipeline system to ensure its effectiveness. Computational Fluid Dynamic (CFD) method is applied to analyse the flow motion in a straight pipe, in terms of flow velocity profile, entrance length and head loss. The result shows that the flow motion passes two regions, entrance region and fully developed region. At the entrance area, the pressure difference over the pipe length is non-linear. Thus, to achieve a precise result, the measurement of flow properties alteration starts at the fully developed region since the fluid properties are stable. Moreover, the consequence of the 98 simulation tests indicates that the larger diameter pipe produces less head loss. It is less possible to generate pipe leakage and break since less energy is consumed. Thus, the large diameter pipes are more economical and reliable to water pipeline application.

In order to study the effect of low-pressure exhaust gas recirculation system (Exhaust Gas Recirculation, EGR) on the fuel consumption, emissions and pumping loss of a direct-injection engine in a cylinder, a set of low-pressure EGR system control strategies is proposed based on a 2.0T direct-injection engine. The test results show that under various operating conditions, as the EGR rate increases, the specific fuel consumption decreases, the pumping loss is improved, and the carbon monoxide (CO) and nitrogen oxide (NO_x) content at the exhaust end decreases. Total Hydrocarbon (THC) content increases slightly.

The selection of parameters in aluminum alloy processing is a key factor that affects the efficiency and quality of aluminum alloy parts processing, reducing manufacturing costs and improving equipment service life. Multi-objective optimization of the parameters of 6061 aluminum alloy milling process was studied. With the five process parameters of spindle speed, feed speed, axial feed, radial feed and tool diameter as experimental factors, a five-factor and five-level orthogonal experiment of 6061 aluminum alloy high-speed milling was performed. The experimental results were analyzed by GA-BP prediction model to establish the non-linear relationship between milling parameters and surface roughness. On this basis, a multi-objective optimization model was developed aiming at the maximum material removal rate and the lowest surface roughness, and the model was solved by the gamultiobj function based on the NSGA-II algorithm. The results show that the optimized process parameters of 6061 aluminum alloy high-speed milling range at the spindle speed of 12000~13000 r·min^-1, with radial feed of 0.19~0.21 mm, feed speed of 1272~1300 mm·min^-1, axial feed of 6~8 mm, and tool diameter of 4 mm.

Air negative ions reduce the concentration of fine particulate matter in the air mainly by electrostatic adsorption, and the particle size of the fine particulate matter will have an impact on the purification effect. In order to study the purification effect of negative ions on indoor particles with different particle sizes, a typical residential room in Guangzhou is taken as an example, by adding artificial pollution sources and anion purifier, to measure with corresponding instruments the concentration changes of particles with different particle sizes. And the linear fitting of the change of particle concentration with different particle size is carried out. The results show that the concentration changes of particles with different particle sizes are approximately linear, and the purification effect of anions on particles with small particle sizes is more obvious, while the survival time of anions in polluted air is very short and the concentration very small.

Based on multi-body system dynamics and lift line aerodynamic model, considering dynamic stall and aeroelastic coupling of flexible back-swept blades, a wind turbine aeroelastic coupling model is established to study the effect of extreme operating gusts and duration of the gusts on the aerodynamic performance of a 5 MW back-swept wind turbines blades. The results show that extreme operating gusts have a great influence on aerodynamic characteristics, such as the flapwise root moment, power, angle of attack, lift coefficient and thrust. The research work has an important significance for wind turbine’s structural optimization and fatigue life design.

In order to obtain the influence of different working fluids on the liquid-vapor separation in the liquid separation header, a numerical simulation was carried out with the same mass flow rate and vapor quality as the comparison benchmark. Based on the VOF(Volume of Fluid) multiphase model, the liquid-vapor separation among six working fluids was simulated. The property variables were density, viscosity and surface tension. The result demonstrated that the liquid-vapor separation state will change under the influence of liquid vapor density ration. In addition, it was found that R134a had the best performance in liquid-vapor separation, while R1234ze(E) could replace R134a well. The optimization direction of other working fluid to improve the performance of liquid-vapor separation was also given.