For the difficult problem that the shearer drum load is difficult to be detected and perceived directly, a shearer drum load indirect estimation measurement method and technique is proposed and invented. Based on the structural characteristics of the rocker arm, the basic rod system skeleton of the rocker arm is extracted. Based on the transfer route of the drum load on the rocker arm, the feasibility of using the rocker arm pin shaft load for drum load prediction is analyzed, and the correlation mechanics equation of load from the drum to the rocker arm pin shaft is constructed. To address the statically indeterminate problem in the equation, the Lagrange multiplier is introduced to construct the energy function of the rocker rod system skeleton. Based on the principle of minimum complementary energy and solving the extreme value of energy function, the statically indeterminate equation is transformed into the statically determinate equation, and a rough prediction model of drum load based on the force of the rocker arm hinged ear pin shaft is derived, and the maximum relative error of the model is 25.15 %. To address the low prediction accuracy of the model due to the dynamic characteristics of the drum rocker arm, the drum load is derived from the rocker arm hinged ear pin shaft load and the rough prediction model as input, and the experimentally measured drum load as output. The bp neural network error correction model is constructed, and the accurate estimation and perception of the drum load are realized by the prediction model and the bp neural network correction model. Finally, the accuracy of the prediction algorithm is verified by experiments. The results show that after correction by bp neural network, the average error of the drum three directional force is 0.14 × 104N ∼ 0.55 × 104N, the mean square error is 0.18 × 104N ∼ 0.65 × 104N, and the relative error is 3.99 %∼10.12 %. The average error of the drum three directional moment is 0.31 × 104N·m ∼ 0.40 × 104N·m, the mean square error is 0.16 × 104N·m ∼ 0.50 × 104N·m, and the relative error is 2.04 %∼11.44 %. The accuracy of drum load prediction is higher than 11.44 % in all cases. The research results can provide theoretical and technical support for intelligent shearer design, development, and healthy operation and maintenance.
Read full abstract