Volumen 17 (2017): Edición 2 (April 2017)

The article focuses on the analysis of the possibilities to model motion and dispersion of plumes of different density gas pollutants in lowspeed wind tunnels based on the application of physical similarity criteria, in this case the Froude number. The analysis of the physical nature of the modeled process by the Froude number is focused on the influence of air flow velocity, gas pollutant density and model scale. This gives an idea of limitations for this type of physical experiments in relation to the modeled real phenomena. The resulting statements and logical links are exemplified by a CFD numerical simulation of a given task calculated in ANSYS Fluent software.

This paper describes a non-invasive system for respiratory monitoring using a Micro Electro Mechanical Systems (MEMS) flow sensor and an IMU (Inertial Measurement Unit) accelerometer. The designed system is intended to be wearable and used in a hospital or at home to assist people with respiratory disorders. To ensure the accuracy of our system, we proposed a calibration method based on ANN (Artificial Neural Network) to compensate the temperature drift of the silicon flow sensor. The sigmoid activation functions used in the ANN model were computed with the CORDIC (COordinate Rotation DIgital Computer) algorithm. This algorithm was also used to estimate the tilt angle in body position. The design was implemented on reconfigurable platform FPGA.

In order to solve the dilemma between the smaller pressure loss and the larger flow measurement signal in traditional throttle flowmeters, a throttle structure with the inner-outer tube was designed and analyzed. The mathematical relationship model deduced from hydrodynamics showed there were three major parameters to determine the designed throttle structure. Furthermore, the optimal results were achieved by combining orthogonal test design and computational fluid dynamics by taking the ratio of differential pressure of inner-outer tube divided by that of anterior-posterior tube as the optimization goal. Finally, the simulation results with the best level parameters showed that the differential pressure of the anterior-posterior throttle could remain not only the smaller value among other parameters with the same structure of inner-outer tube. On the other hand, it was about one order magnitude less than differential pressure of V-cone flowmeter in the similar installation conditions with the flow velocity varying from 0.5 to 3.0 m/s. The designed inner-outer tube flowmeter can not only save manufacture costs, but also avoid the large sensitivity of pressure sensors, which may lead to a broader application in chemical and petrochemical enterprises.

In the current paper, a residual generation filter with finite memory structure is proposed for sensor fault detection. The proposed finite memory residual generation filter provides the residual by real-time filtering of fault vector using only the most recent finite measurements and inputs on the window. It is shown that the residual given by the proposed residual generation filter provides the exact fault for noisefree systems. The proposed residual generation filter is specified to the digital filter structure for the amenability to hardware implementation. Finally, to illustrate the capability of the proposed residual generation filter, extensive simulations are performed for the discretized DC motor system with two types of sensor faults, incipient soft bias-type fault and abrupt bias-type fault. In particular, according to diverse noise levels and windows lengths, meaningful simulation results are given for the abrupt bias-type fault.

The interactions between stiffness and workspace performances are studied. The stiffness in x, y and z directions as well as the workspace of a 3-UPU mechanism are studied and optimized. The stiffness of the robotic system in every single moveable direction is measured and analyzed, and it is observed that in the case where one tries to make the x and y translational stiffness larger, the z directional stiffness will be reduced, i.e. the x and y translational stiffness contradicts with the one in z direction. Subsequently, the objective functions for the summation of the x and y translational stiffness and z directional stiffness are established and they are being optimized simultaneously. However, we later found that these two objectives are not in the same scale; a normalization of the objectives is thus taken into consideration. Meanwhile, the robotic system’s workspace is studied and optimized. Through comparing the stiffness landscape and the workspace volume landscape, it is also observed that the z translational stiffness shows the same changing tendency with the workspace volume’s changing tendency while the x and y translational stiffness shows the opposite changing tendency compared to the workspace volume’s. Via employing the Pareto front theory and differential evolution, the summation of the x and y translational stiffness and the volume of the workspace are being simultaneously optimized. Finally, the mechanism is employed to synthesize an exercise-walking machine for stroke patients.

The article presents the impact of the edge detection method in the image analysis on the reading accuracy of the measured value. In order to ensure the automatic reading of the measured value by an analog meter, a standard webcam and the LabVIEW programme were applied. NI Vision Development tools were used. The Hough transform was used to detect the indicator. The programme output was compared during the application of several methods of edge detection. Those included: the Prewitt operator, the Roberts cross, the Sobel operator and the Canny edge detector. The image analysis was made for an analog meter indicator with the above-mentioned methods, and the results of that analysis were compared with each other and presented.