Thesis Examination Committee
Prof Shaojie SHEN, ECE/HKUST (Chairperson)
Prof Zexiang LI, ECE/HKUST (Thesis Supervisor)
Prof Fu ZHANG, CEE/HKUST
Nowadays, during precisely screwing assembly process, there is a strict requirement for the acceleration-speed and force-torque real-time precisely control. It wastes a lot of time to set these parameters during different assembly process. To improve the efficiency, it is very necessary to develop an intelligent wrench which can automatically precisely set these parameters. But now all such products are all developed by foreign companies include the Altlas Copco, DEPRAG and so on which is very expensive, costing from 120K-200K RMB. And every year, many companies like Foxconn and DJI spent a lot of money to buy such equipment for their products’ assembly process which means there is a huge potential market for such product.
For this problem, we build an intelligent torque wrench which can automatically set the best value for these parameters. Our development idea is based on the Cloud platform to automatically collect all the screwing data. And based on these data, we could optimize the best parameters which can make sure the screwing process successfully within the shortest time. According to the screwing feature, we set the whole process into 7 steps, and set the corresponding optimization priority. Based on the newton and gradient optimization algorithm, we can get the best value for every step within 10 times screwing process. And compared with the foreign companies’ product, the cost for our intelligent wrench system is cheaper, while guaranteeing the shortest time for the screwing process. Further, when we get enough screwing data for different assembly process, based on machine learning method, we can provide users with a more reliable screwing process solution, costing less money.