关键词: 转向架试验台, 三维测力装置, 维间耦合, 最小二乘法, BP神经网络, 耦合度

Abstract: In order to measure the force loaded on bogie test bench and improve the measurement accuracy, a three-dimensional (3-D) force measuring platform with multiple force sensors connected in parallel was designed. Its measurement principle was analyzed theoretically and a Wheatstone bridge measurement circuit was built. 3-D force measuring platform is a kind of shear stress sensor, and the analysis shows that the output voltage of its mea-surement circuit has a definite linear relationship with the force loaded. The static calibration test of 3-D force measuring platform under uniaxial load was carried out and the output characteristics under uniaxial load and the interdimensional coupling effect of load measurement were analyzed. It finds that under uniaxial load, the voltage output of the measurement circuit corresponding to the loaded axis has good linearity, but the voltage output of the non-loaded axis measuring circuit is small. In this paper, the coupling degree evaluating coupling effect between dimensions and the numerical decoupling method based on least square method and BP neural network were also proposed, and their effectiveness was verified by experiments. At the same time, the orthogonal method was used to carry out the multi-axis loading test of the three-dimensional force measuring platform. The results show that the single-axis measurement error and the overall measurement error are not more than 1.5%, thus the measurement accuracy meets the test requirements.

Key words: bogie test bench, 3-D force measuring platform, interdimensional coupling, least square method, BP neural network, coupling degree