Abstract: Paver can finish paving, ramming and pressing in a time, with certain density, thickness, width. It is suitable for various grades of highway, airport and road construction. The auto-levelling system, as an important part of paver, makes paver simple operation, safe and reliable, greatly reduce the loss of raw material and manpower, and the control accuracy directly influences the quality of paving flatness. The problems faced of auto-levelling system are introduced firstly. Secondly, ultrasonic sensor range finder technology at home and abroad and the present situation and basic principle, characteristics of ultrasonic sensor are clarified. Third,the range affecting factors of the ranging accuracy are found and appropriate method is put forward to compensating and correcting. Eventually, the method applied in the CPLD EPM570T100C5N, based on the C8051F040 singlechip , ultrasonic ranging circuit are formed. This paper illustrates the hardware and software design of every circuit module.
Abstract: In the modeling of rolling process, the result from FEM is influenced by the accuracy of the resistance of deformation model of material which is often from the experiment using isothermal compression at elevated temperatures. But the state under high temperature is unstable in lab. And the flow stress model under the lab condition is not consistent to the hot rolling field condition. This paper will employ the RBF artificial neural networks to train the random sample from the field data and get the relation between the rolling force and the parameters of hot rolling. The array of the rolling force and temperature under special condition can be then obtained and the resistance of deformation model will be built by applying the S.EKLUND formula. Finally, the thermo-mechanical coupling FEM with the resistance of deformation model got from RBF method is used to simulate the hot rolling process. The simulated experiment shows the modeling data is fit to the field data.
Abstract: In recent years artificial neural networks (ANNs) have been applied to many geotechnical engineering problems with some degree of success. With respect to the design of Highway embankment, accurate prediction of settlement of soft clay foundation in highway is necessary to ensure appropriate structural and serviceability performance. In this paper, an ANN model is developed for predicting settlement of soft clay foundation based on the observation data of settlement. Approximately 200 data sets, obtained from the Field Tests and the published literature, are used to develop the ANN model. In addition, the paper discusses the choice of input and internal network parameters which were examined to obtain the optimum model. Finally, the paper compares the predictions obtained by the ANN with those given by a number of traditional methods. It is demonstrated that the ANN model outperforms the traditional methods and provides accurate settlement predictions for soft clay foundation in highway.
Abstract: In this paper, a new experiment procedure is proposed to study the influence of cutter parameters and clamping methods on the stability of the milling process of thin-walled blade. A dedicated fixture is designed to carry out the experiment. Simulation results show that the new clamping system can enhance the rigidity of thin-walled blade to reduce cutting deformation and chatter vibration phenomenon. Then, cutter and cutting parameters can be optimized properly to make the system obtain high rigidity and high performance stable milling process. Industrial application indicates that the new system can improve the cutting performance and ensure the cutting quality.
Abstract: During identify natural frequency of bearing rotor, due to the complex non-linear relationship among the factors which influence natural frequency, so it is hard to establish a complete and accurate theoretical model. Based on the generalization and approximation of non-linear mapping capability of support vector machine (SVM) and the powerful ability of global optimization of the genetic algorithm (GA), the paper through optimizing the SVM by GA, establishes combined Genetic Support Vector Machine (GA-SVM). The method establishes the mapping between the natural frequency of a rolling bearing rotor and the various parameters, which reduces the rotor structure for the study similar to the natural frequency of the calculation of the workload greatly. Using the model to indentify the natural frequency of bearing rotor under different parameters, then compare identification value with experimental values shows that projections in good agreement with the experimental data.
Abstract: Concrete is a mixture of the cementing material, aggregate and water in a certain proportion and is the most main materials of the civil engineering materials. It is difficult to make modeling for a highly complex material. The concrete rebound value with wide randomness is a dependent variable, while the compressive strength value with narrow randomness is an independent variable. This paper aimed to show possible applicability of artificial neural networks (ANN) to predict the compressive strength. Back propagation neural networks (BPNN) model is constructed trained and tested using the available test data of 108 different concrete specimens. The data of input parameters used in BPNN model were covered the ratio of water to cement, fine aggregate ratio, coarse aggregates, mean value of test area of rebound method measurement. The mean absolute percentage error was less then 10.19% for compressive strength. The results showed that ANNs was good at as a feasible tool for predicting compressive strength.
Abstract: Fiber grating sensor network plays an important role in evaluating the bridge health condition and gaining bridge structure characteristic.Bridge structural distortion and variation of stress is associated with temperature.So it is essential to make temperature prediction bases on the length of wave which gained from fiber grating sensor network system.Traditionally, we apply least square method to predict temperature.In this paper,we use method of relevance vector machine (RVM) to make it.
Abstract: Rear multi-link suspension is built according to hard point parameters of a car. Using multi-body dynamics and suspension kinematics theory, the suspension performance were analyzed under the simulation test of parallel wheel travel, loading brake force after the inner rubber bushing stiffness of the lateral rod is changed. The conclusion is drawn that under the test of loading brake force, changing the axial and torsion stiffness of rubber bushing has no effect on the camber angle, total track and toe angle, a little effect on ride rate and wheel rate. Under the test of wheel travel, changing the axial and torsion stiffness has also no effect on the suspension performance; but the change of the radial stiffness has effect on the suspension performance, especially the ride rate and wheel rate.
Abstract: Cylinder compression pressure reflects the air tightness of engine. A method for measuring the compression pressure of cylinder indirectly through measuring the vibration signal of cylinder head was studied and then to detect the air tightness. The air pressure signal in cylinder and vibration signals of cylinder head were measured at the same time when the diesel engine was driven by the motor. According to port timing, the vibration signal excited by cylinder pressure was separated using time domain analysis. A RBF neural network model was set up to build the relation between compression pressure and cylinder head vibration. So the air tightness of cylinder can be detected after calculating the compression pressure by use of neural network.