Xiu Yan

Experimental study on the water lubrication of non-contacting face seals for turbopumps

Purpose – A new type of hydrostatic and hydrodynamic non-contacting face seals has been designed to meet the requirements of lower leakage, longer life and more repeatedly start and stop on shaft seals raised by liquid rocket engine turbopumps. And an experimental study on the performance of the face seal in the actual liquid oxygen turbopump was completed where low-viscosity water was selected as the seal fluid for the sake of safety. The paper aims to discuss these issues. Design/methodology/approach – Different performances of face seals under preset conditions were obtained by repeatedly running tests, and the main performance parameters encompass leakage, fluid film pressure between the faces, operating power, face temperature, and so on. Findings – The results indicate that the designed face seal has a smaller amount of leakage, with a minimum value of 3 ml/s. Furthermore, the designed face seal has been proved to demand lower operating power. Since its operating power changes slightly with differen...

Analysis of two phase flow in liquid oxygen hybrid journal bearings for rocket engine turbopumps

Purpose – A hybrid bearing of advanced cryogenic rocket engine turbopump is designed. For cryogenic fluid propellants (such as liquid oxygen) as the lubrication of bearing, bearings operating close to liquid-vapor region (near the critical point or slightly sub-cooled) are likely to develop a two phase flow region. The paper aims to discuss these issues. Design/methodology/approach – In this paper, an all liquid, liquid-vapor mixture, and all vapor, i.e. a continuous vaporization bulk flow model of density and viscosity for mixture fluid, is presented, and the general Reynolds equation and energy equation with two phase flow as lubricants is solved. The static and dynamic performance of a 50-mm-radius hybrid bearing are obtained under 20,000 rpm speed and 10 MPa supply pressure. Findings – The results show that the variations of performance of bearing operating under cryogenic liquid oxygen are not bounded by the all liquid and all vapor cases in the liquid-vapor mixture range. There behaviours are attrib...

Fast human classification of 3D object benchmarks

Although a significant number of benchmark data sets for 3D object based retrieval systems have been proposed over the last decade their value is dependent on a robust classification of their content being available. Ideally researchers would want hundreds of people to have classified thousands of parts and the results recorded in a manner that explicitly shows how the similarity assessments varies with the precision used to make the judgement. This paper reports a study which investigated the proposition that Internet Crowdsourcing could be used to quickly and cheaply provide benchmark classifications of 3D shapes. The collective judgments of the anonymous workers produce a classification that has surprisingly fine granularity and precision. The paper reports the results of validating Crowdsourced judgements of 3D similarity against Purdues ESB and concludes with an estimate of the overall costs associated with large scale classification tasks involving many tens of thousands of models.

Autonomous navigation with ROS for a mobile robot in agricultural fields

Autonomous monitoring of agricultural farms and fields has recently become feasible due to continuing advances in robotics technology, but many notable challenges remain. In this paper, we describe the state of ongoing work to create a fully autonomous ground rover platform for monitoring and intervention tasks on modern farms that is built using inexpensive and off the shelf hardware and Robot Operating System (ROS) software so as to be affordable to farmers. The hardware and software architectures used in this rover are described along with challenges and solutions in odometry and localization, object recognition and mapping, and path planning algorithms under the constraints of the current hardware. Results obtained from laboratory and field testing show both the key challenges to be overcome, and the current successes in applying a low-cost rover platform to the task of autonomously navigating the outdoor farming environment.

A novel haptic model and environment for maxillofacial surgical operation planning and manipulation

This paper presents a practical method and a new haptic model to support manipulations of bones and their segments during the planning of a surgical operation in a virtual environment using a haptic interface. To perform an effective dental surgery it is important to have all the operation related information of the patient available beforehand in order to plan the operation and avoid any complications. A haptic interface with a virtual and accurate patient model to support the planning of bone cuts is therefore critical, useful and necessary for the surgeons. The system proposed uses DICOM images taken from a digital tomography scanner and creates a mesh model of the filtered skull, from which the jaw bone can be isolated for further use. A novel solution for cutting the bones has been developed and it uses the haptic tool to determine and define the bone-cutting plane in the bone, and this new approach creates three new meshes of the original model. Using this approach the computational power is optimized and a real time feedback can be achieved during all bone manipulations. During the movement of the mesh cutting, a novel friction profile is predefined in the haptical system to simulate the force feedback feel of different densities in the bone.Whether from internal sources or arising from environmental sources, thermal error in most machine tools is inexorable. Out of several thermal error control methods, electronic compensation can be an easy-to-implement and cost effective solution. However, analytically locating the optimal thermally sensitive points within the machine structure for compensation has been a challenging task. This is especially true when complex structural deformations arising from the heat generated internally as well as long term environmental temperature fluctuations can only be controlled with a limited number of temperature inputs. This paper presents some case study results confirming the sensitivity to sensor location and a new efficient offline method for determining localized thermally sensitive points within the machine structure using finite element method (FEA) and Matlab software. Compared to the empirical and complex analytical methods, this software based method allows efficient and rapid optimization for detecting the most effective location(s) including practicality of installation. These sensitive points will contribute to the development and enhancement of new and existing thermal error compensation models respectively by updating them with the location information. The method is shown to provide significant benefits in the correlation of a simple thermal control model and comments are made on the efficiency with which this method could be practically applied.Virtual Reality (VR) applications are employed in engineering situation to simulate real and artificial situations where the user can interact with 3D models in real time. Within these applications the virtual environment must emulate real world physics such that the system behaviour and interaction are as natural as possible and to support realistic manufacturing applications. As a consequence of this focus, several simulation engines have been developed for various digital applications, including VR, to compute the physical response and body dynamics of objects. However, the performance of these physics engines within haptic-enabled VR applications varies considerably. In this study two third party physics engines - Bullet and PhysXtm- are evaluated to establish their appropriateness for haptic virtual assembly applications. With this objective in mind five assembly tasks were created with increasing assembly and geometry complexity. Each of these was carried out using the two different physics engines which had been implemented in a haptic-enabled virtual assembly platform specifically developed for this purpose. Several physics-performance parameters were also defined to aid the comparison. This approach and the subsequent results successfully demonstrated the key strengths, limitations, and weaknesses of the physics engines in haptic virtual assembly environments.Communicating the knowledge and science of product engineering, analysis and manufacturing planning is an area of continued research driven by the digital economy. Virtual Reality (VR) is a generally accepted interactive digital platform which industry and academia have used to model engineering workspaces. Interactive services that generate a sense of immersion, particularly the sense of touch to communicate shape modelling and manipulation, is increasingly being used in applications that range from Design For Manufacturing and Assembly (DFMA) and Process Planning (PP) to medical applications such as surgical planning and training. In simulation, the natural way for solid modelling is the use of primitive geometries, and combinations of them where complex shapes are required, to create, modify or manipulate models. However, this natural way makes use of Booleans operands that require large computational times which make them inappropriate for real time VR applications. This work presents an insight on new methods for haptic shape modelling focused on Boolean operands on a polygon mesh. This is not meant as a contrast to point/mesh- editing methods, instead it is focused on idealising polygonal mesh modelling and manipulation for use with haptics. The resulting models retain a high level of geometric detail for visualisation, modelling, manipulation and haptic rendering.

Idealising mesh modelling for haptic enabled services and operands

This paper presents a practical method and a new haptic model to support manipulations of bones and their segments during the planning of a surgical operation in a virtual environment using a haptic interface. To perform an effective dental surgery it is important to have all the operation related information of the patient available beforehand in order to plan the operation and avoid any complications. A haptic interface with a virtual and accurate patient model to support the planning of bone cuts is therefore critical, useful and necessary for the surgeons. The system proposed uses DICOM images taken from a digital tomography scanner and creates a mesh model of the filtered skull, from which the jaw bone can be isolated for further use. A novel solution for cutting the bones has been developed and it uses the haptic tool to determine and define the bone-cutting plane in the bone, and this new approach creates three new meshes of the original model. Using this approach the computational power is optimized and a real time feedback can be achieved during all bone manipulations. During the movement of the mesh cutting, a novel friction profile is predefined in the haptical system to simulate the force feedback feel of different densities in the bone.Whether from internal sources or arising from environmental sources, thermal error in most machine tools is inexorable. Out of several thermal error control methods, electronic compensation can be an easy-to-implement and cost effective solution. However, analytically locating the optimal thermally sensitive points within the machine structure for compensation has been a challenging task. This is especially true when complex structural deformations arising from the heat generated internally as well as long term environmental temperature fluctuations can only be controlled with a limited number of temperature inputs. This paper presents some case study results confirming the sensitivity to sensor location and a new efficient offline method for determining localized thermally sensitive points within the machine structure using finite element method (FEA) and Matlab software. Compared to the empirical and complex analytical methods, this software based method allows efficient and rapid optimization for detecting the most effective location(s) including practicality of installation. These sensitive points will contribute to the development and enhancement of new and existing thermal error compensation models respectively by updating them with the location information. The method is shown to provide significant benefits in the correlation of a simple thermal control model and comments are made on the efficiency with which this method could be practically applied.Virtual Reality (VR) applications are employed in engineering situation to simulate real and artificial situations where the user can interact with 3D models in real time. Within these applications the virtual environment must emulate real world physics such that the system behaviour and interaction are as natural as possible and to support realistic manufacturing applications. As a consequence of this focus, several simulation engines have been developed for various digital applications, including VR, to compute the physical response and body dynamics of objects. However, the performance of these physics engines within haptic-enabled VR applications varies considerably. In this study two third party physics engines - Bullet and PhysXtm- are evaluated to establish their appropriateness for haptic virtual assembly applications. With this objective in mind five assembly tasks were created with increasing assembly and geometry complexity. Each of these was carried out using the two different physics engines which had been implemented in a haptic-enabled virtual assembly platform specifically developed for this purpose. Several physics-performance parameters were also defined to aid the comparison. This approach and the subsequent results successfully demonstrated the key strengths, limitations, and weaknesses of the physics engines in haptic virtual assembly environments.Communicating the knowledge and science of product engineering, analysis and manufacturing planning is an area of continued research driven by the digital economy. Virtual Reality (VR) is a generally accepted interactive digital platform which industry and academia have used to model engineering workspaces. Interactive services that generate a sense of immersion, particularly the sense of touch to communicate shape modelling and manipulation, is increasingly being used in applications that range from Design For Manufacturing and Assembly (DFMA) and Process Planning (PP) to medical applications such as surgical planning and training. In simulation, the natural way for solid modelling is the use of primitive geometries, and combinations of them where complex shapes are required, to create, modify or manipulate models. However, this natural way makes use of Booleans operands that require large computational times which make them inappropriate for real time VR applications. This work presents an insight on new methods for haptic shape modelling focused on Boolean operands on a polygon mesh. This is not meant as a contrast to point/mesh- editing methods, instead it is focused on idealising polygonal mesh modelling and manipulation for use with haptics. The resulting models retain a high level of geometric detail for visualisation, modelling, manipulation and haptic rendering.

Proceedings of the 37th international MATADOR conference

This paper presents a practical method and a new haptic model to support manipulations of bones and their segments during the planning of a surgical operation in a virtual environment using a haptic interface. To perform an effective dental surgery it is important to have all the operation related information of the patient available beforehand in order to plan the operation and avoid any complications. A haptic interface with a virtual and accurate patient model to support the planning of bone cuts is therefore critical, useful and necessary for the surgeons. The system proposed uses DICOM images taken from a digital tomography scanner and creates a mesh model of the filtered skull, from which the jaw bone can be isolated for further use. A novel solution for cutting the bones has been developed and it uses the haptic tool to determine and define the bone-cutting plane in the bone, and this new approach creates three new meshes of the original model. Using this approach the computational power is optimized and a real time feedback can be achieved during all bone manipulations. During the movement of the mesh cutting, a novel friction profile is predefined in the haptical system to simulate the force feedback feel of different densities in the bone.Whether from internal sources or arising from environmental sources, thermal error in most machine tools is inexorable. Out of several thermal error control methods, electronic compensation can be an easy-to-implement and cost effective solution. However, analytically locating the optimal thermally sensitive points within the machine structure for compensation has been a challenging task. This is especially true when complex structural deformations arising from the heat generated internally as well as long term environmental temperature fluctuations can only be controlled with a limited number of temperature inputs. This paper presents some case study results confirming the sensitivity to sensor location and a new efficient offline method for determining localized thermally sensitive points within the machine structure using finite element method (FEA) and Matlab software. Compared to the empirical and complex analytical methods, this software based method allows efficient and rapid optimization for detecting the most effective location(s) including practicality of installation. These sensitive points will contribute to the development and enhancement of new and existing thermal error compensation models respectively by updating them with the location information. The method is shown to provide significant benefits in the correlation of a simple thermal control model and comments are made on the efficiency with which this method could be practically applied.Virtual Reality (VR) applications are employed in engineering situation to simulate real and artificial situations where the user can interact with 3D models in real time. Within these applications the virtual environment must emulate real world physics such that the system behaviour and interaction are as natural as possible and to support realistic manufacturing applications. As a consequence of this focus, several simulation engines have been developed for various digital applications, including VR, to compute the physical response and body dynamics of objects. However, the performance of these physics engines within haptic-enabled VR applications varies considerably. In this study two third party physics engines - Bullet and PhysXtm- are evaluated to establish their appropriateness for haptic virtual assembly applications. With this objective in mind five assembly tasks were created with increasing assembly and geometry complexity. Each of these was carried out using the two different physics engines which had been implemented in a haptic-enabled virtual assembly platform specifically developed for this purpose. Several physics-performance parameters were also defined to aid the comparison. This approach and the subsequent results successfully demonstrated the key strengths, limitations, and weaknesses of the physics engines in haptic virtual assembly environments.Communicating the knowledge and science of product engineering, analysis and manufacturing planning is an area of continued research driven by the digital economy. Virtual Reality (VR) is a generally accepted interactive digital platform which industry and academia have used to model engineering workspaces. Interactive services that generate a sense of immersion, particularly the sense of touch to communicate shape modelling and manipulation, is increasingly being used in applications that range from Design For Manufacturing and Assembly (DFMA) and Process Planning (PP) to medical applications such as surgical planning and training. In simulation, the natural way for solid modelling is the use of primitive geometries, and combinations of them where complex shapes are required, to create, modify or manipulate models. However, this natural way makes use of Booleans operands that require large computational times which make them inappropriate for real time VR applications. This work presents an insight on new methods for haptic shape modelling focused on Boolean operands on a polygon mesh. This is not meant as a contrast to point/mesh- editing methods, instead it is focused on idealising polygonal mesh modelling and manipulation for use with haptics. The resulting models retain a high level of geometric detail for visualisation, modelling, manipulation and haptic rendering.