• Alexey Bodrov

Student thesis: Phd


In this thesis, a novel control strategy has been developed to reduce industrial electrical energy use. This solution requires little or no investment, and will not compromise production targets, product quality or environmental impact. It can be used to promote Industry 4.0 via smart information exchange on the lowest device/unit process manufacturing level. By allowing communication between motors in a multi-drive system the energy consumption can be reduced through smart scheduling of motor operation times and loading. The positioning system in metal-cutting machine tools provides an example application, where the separate axes are powered through variable-speed drives. The multi-axis positioning is a series kinematics system, where the motor drives are indirectly mechanically coupled through the tool tip/end-effector trajectory. 2D systems are considered for simplicity, where the drives’ dynamics are independent from each other. The proposition of this thesis is that overall electrical energy consumption can be reduced, by controlling the drive with the longer axis’ movement distance for minimum time, and the drive with the shorter distance for optimal energy. The latter is required to finish its move by the time the former reaches its end point, therefore an ability to compute and exchange expected finishing times between axes is required. The resulting end-effector trajectory is nonlinear, so the method is proposed for point-to-point moves. Two different approaches have been studied for optimal control of the positioning system: Variational and Model Predictive Control (MPC). Both these methods were tested to control the motor drive within its current and voltage constraints. The former is shown to be suitable for a simple plant model approximation, where the analytical open-loop solution for both control problems can be obtained. However, this approach is of limited use or even not applicable for the derivation of a close-loop control law. The variational technique is useful for reference profile generation, where the energy use is dominated by stand-by or support needs, for example, in robotics applications. This method was demonstrated on a mobile six-legged robot CORIN, where an open-loop optimal time profile was generated for each joint’s motor control, leading to the more than 20% energy saving, whilst remaining within all constraints. In the MPC approach, the control is split into two stages. First, the optimal reference state trajectory is generated off-line. It is shown that a linear MPC formulation can be used for brushed DC motor drives whereas a non-linear MPC formulation is necessary for the Surface Mounted Permanent Magnet Synchronous Motors (SMPMSM). Then the optimal state trajectory is used as reference and feed-forward to either a conventional nested-loop controller or to a MPC on-line reference tracking controller. The proposed control has been compared with conventional G00 and G01 trajectories. For the DC motor drive under test, experimental results show 13% and 8% efficiency rise compared to the G00 and G01 respectively. For the SMPMSM control, simulation results reveal 16% energy consumption reduction compared to the G00 case, and almost the same energy consumption as the G01 case, due to the field-weakening operation of the drive in the proposed control method. Since the proposed algorithm reduces the operational time, this still results in lower stand-by energy consumption, which is shown to be dominant.
Date of Award1 Aug 2019
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorAlexander Smith (Supervisor) & Judith Apsley (Supervisor)


  • industry 4.0
  • machine tools
  • MPC
  • positioning systems
  • motor drive
  • multi-drive systems
  • constraint optimization

Cite this