Search

The rotary axis is crucial for the machining accuracy of five-axis machines, and its motion accuracy is significantly influenced by the dynamic motion error. A test of circular and elliptical test piece cutting is proposed to identify and analyze the dynamic motion error of the rotary axis. The influence of the feed rate on the dynamic motion accuracy is explained. The dynamic motion error can be divided into two types, depending on its source and manifestation: dynamic error inside the servo loop (DEIS) and dynamic error outside the servo loop (DEOS). At different feed rates, elliptical test piece cutting and kinetics analyses are conducted to identify the DEIS, and circular test piece cutting together with double ball bar (DBB) tests are conducted to identify the DEOS. On-machine inspection and the coordinate measuring machine (CMM) calibration are performed. By combining the DBB test and kinetics analysis results, the DEIS and DEOS are identified and analyzed. The experimental verification shows that the DEIS and DEOS change according to the feed rate and can be determined from the machining error of the test piece. Consequently, the proposed method offers an effective reference for improving the motion accuracy of the rotary axis.

The two-phase flow patterns of ammonia in different scale channels are experimentally investigated in this study. The inner diameters of the tested tubes are 4 mm and 8 mm. The observed flow patterns in the two tubes include slug flow, stratified-wavy flow, annular flow, and annular flow with partial dryout. For both 4 mm and 8 mm tubes, the annular flow gradually expands to a wider vapor quality range with a rise in mass flux and a descent in saturation temperature at the cost of other flow patterns. The comparison of the flow patterns in the two tubes shows that the 8 mm tube has a wider range of annular flow. This is attributed to that the higher dominance of the surface tension force in the 4 mm tube is conductive to the liquid slug to holdup on the channel wall and consequently delay the transition to annular flow. The flow patterns in the two tubes have been compared to the flow pattern maps of Baker, Mandhane et al., and Wojtan et al.. The Wojtan et al. flow pattern map could give a superior prediction for the flow patterns in the 4 mm tube, whereas none of the three existing models accurately predict the flow patterns in the 8 mm tube.

A novel type of horizontal eccentric tube cooler was proposed that combines the advantages of self-adaptive capacity of thermosyphon technology and high heat-transfer efficiency of conventional double-pipe exchanger. Compared with the conventional heat-exchange tube evaporator, it has self-adjustment ability of heat transfer rate with rotation angle. Four different structures of horizontal tube cooler with eccentric axis structure were proposed and tested for verifying the self-adjustment capacity performance. Finally, one optimal structure was selected as the recommendation, which has distinct successive self-adjustment ability with rotation angle 0–135°, self-adjustment range of heat transfer rate 36.67%, and the lowest heat transfer rate loss 19.66% with about 48% filling rate. It will have revolutionary application significance in some fields requiring heat exchange regulation.

A new type of ducted hot aisle containment system for racks cooling of data center has been proposed and put into practice gradually. However, the related academic research has not been carried out, especially for the improvement of cooling performance and influencing factors for the cooling system. In this paper, one experiment platform with 8 racks in 2 rows with multiple-heat-sources was established and tested under various working scenarios. In terms of improving the cooling performance, airtight treatment on the basis of ducted hot aisle containment system is better than itself and is also better than only collar, which is better than no collar. Besdides, doors closing is better than that doors opening, "V" diffuser is better than "T" diffuser, and “T” diffuser is better than no diffuser. Properly increasing installation ratio of blanking panels and air volume have also favorable impact. The experimental data and results in this paper will be of great significance to the design and application of ducted hot aisle containment system for racks cooling.

This dissertation studies the impact of indecision on weighted voting systems. A weighted voting system is a decision-making model of voters with unequal powers. To quantify the system performance, system reliability defined as the probability of the system making the correct decision is introduced, which is based on assumptions that the system operates as 2 types of input values (0,1) and 3 types of output values (0,1,x) with 3 types of errors, and voters are unequally weighted and subject to 3 failure-modes (stuck-at-0, stuck-at-1, stuck-at-x). For any weighted voting system, a decision rule is a criterion of how to process the information to generate the final decision. Indecision effect results from one of output values x, where any voter may have no decision on given input values, and it is first considered into the decision rule. To investigate how indecision affects the system performance, this dissertation demonstrates the research in three parts including (1) a threshold weighted voting indecisive system is built, (2) dynamic process on the built system is studied, and (3) a controlled weighted voting indecisive system is generated given unknown inputs. In the first part, a new system is built by considering indecisive effects in the process of decision-making. A generalized threshold weighted voting n-unit indecisive system using a new decision rule consisting of a threshold parameter and a new indecisive parameter is created. The main results imply the high dependence of the system performance from indecisive effect via static parameter initialization, which suggests more research on the indecisive parameter. In the second part, a dynamic process on the indecisive parameter is applied to study system characterization under various situations such as weights change and input types. The results provide a full picture of system performance under the probability distribution of indecisive parameters given different values of threshold and lead the further investigation on prior information of inputs. In the third part, a threshold weighted voting indecisive system with embedded control systems on unknown inputs is designed, where three sub-systems including unknown input observers (UIO), feedforward-feedback controllers (FFC), and the weighted voting system (WVS). To informatize unknown inputs, UIO does the translation of unknown inputs which is used to produce reference inputs in the feedforward controller for disturbances elimination before generating intermediate outputs in WVS, in which, intermediate outputs are monitored in the feedback controller to identify the consistency from reference inputs based on independent final decision criterion. Two major contributions are made from this dissertation including (1) developing a threshold weighted voting system for decision-making when indecision is considered, and (2) developing a fundamental framework on the controlled threshold weighted voting indecisive system when inputs are unknown. Numerical examples are presented to illustrate the generated systems, voting methods, and criteria.