Synthesis of the optimal acceleration mode of the tower crane slewing mechanism

The high performance of cranes is one of the main requirements for construction of buildings and constructins. Tower cranes usually perform such operations. One of the most effective ways to achieve this requirement is to eliminate pendulum oscillations of the load on the flexible suspension for tower cranes in the shortest possible time. Therefore, this article aimed to obtain a general solution to the problem of optimal performance control during the acceleration of the crane’s slewing mechanism. A dynamic system model of the tower crane was presented, boundary conditions and optimisation criteria were described, and a torque constraint was introduced. The original problem was reduced to the problem of unconstrained optimisation of a nonlinear function. The problem was solved multiple times, in which the mass of the load m, the length of the flexible suspension l, and the boom extension range r were varied within certain limits. The VCT-PSO optimisation algorithm was used to find solutions. Based on the obtained solutions, a data set for training the artificial neural network (training set) and a data set for testing its operation (test set) were formed. There were used a feedforward neural network to solve the problem of predicting the moments of control switching. Three inputs (corresponding to the parameters m, r, l), three outputs (corresponding to the moments of control switching), and three hidden layers with five neurons in each were used. The neural network was trained using the ADAM method. The results of estimating the quality of the prediction of the control switching moments were determined, and the plots of estimation errors were presented. The high quality of the approximation of the solution surface of the problem using the artificial neural network was established. The analysis of the system motion dynamics under optimal control was carried out, which demonstrated that the system motion indicators were least affected by the mass of the load m and most affected by the length of the flexible suspension l