| Accurately obtaining the output torque of the drive shaft is of great significance to the optimization of loader design,fault prediction and diagnosis.With the deepening of large data collection and application research on the operation of loader and other construction machinery,how to conveniently collect the torque signal on the drive shaft has become a difficult problem for researchers.The existing torque measurement technology has problems such as difficulty in calibration,short life,operation error,and high cost,which lead to the lack of operating torque data in the big data of loader operation.To this end,this article proposes a soft measurement method suitable for long-term torque monitoring and a torque measurement method that can eliminate the calibration test for the problem of torque measurement of the drive shaft of a loader.The main contents are as follows:Firstly,a soft-sensing model based on easy-to-measure data to characterize difficult-to-measure torque is constructed.Find out the influencing factors of the torque of the drive shaft from the energy transmission of the loader as an auxiliary variable for soft-sensing modeling.Collect the running data of a ZL50 prototype,integrate multiple filtering methods to clean the data noise,and provide data preparation for soft-sensing modeling.Mic is used as the evaluation criterion to eliminate irrelevant and redundant features.Secondly,in view of the characteristics of sampled data with multiple operating conditions,multiple disturbances and nonlinearity,this paper presents a new algorithm for large-scale sample modeling,which called distributed sparse least squares support vector machine(DSLS-SVM),and combined with CS search to train the loader torque soft sensor model.The test results of the sample set show that the scheme can meet the torque measurement with a certain accuracy,and is suitable for long-term operation status monitoring of loaders.Thirdly,improve the technology and structure to realize the calibration-free measurement of the loader drive shaft torque.The existing strain gauge torque measurement requires calibration tests due to the characteristics of the strain gauge,the pasting process,and the measurement error.The difficulty in implementation and high cost limit the promotion of torque measurement.Based on this,a solution to solidify the strain transmission loss into a known coefficient and specific design indicators are proposed to achieve the effect of calibration-free.Starting from the strain measurement mechanism,the mathematical model between the bridge voltage and the received torque is established,and the transmission efficiency is introduced for correction.With the aid of material analysis,finite element simulation and bench test,a combination of design indicators to achieve stable transmission efficiency of the sensing unit is found to verify the feasibility of calibration-free measurement,and a "voltage-torque" model to guide calibration-free measurement is given.Fourthly,realize the customized system design of soft measurement and calibration-free measurement.Design and verify the vehicle-mounted torque soft measurement system,including:multi-channel simulation,pulse,CAN signal acquisition,multi-application scenarios of USB,WIFI,RS232 communication,data interaction and command control,visual man-machine interface,soft-sensing model online prediction and other functions.Design a calibration-free torque measuring instrument and use it for bench testing,integrating strain sensing,signal scaling,wireless communication,and power supply into one,light and anti-interference. |