A. The Speed Characteristic of Sewing Machine The BLDCM with four poles and three-phase is adapted to supply the power of the ISM, and the motor is driven by the square wave with full-bridge style, so it will lead the fluctuation of the output torque [5]. Besides, four-bar mechanism and crank slide mechanism are contained in the sewing machine’s drive mechanism which appears nonlinear characteristic and leads to low control precision on any point in the movement rule. The inertia will periodically change in great range on the rotating of the sewing machine due to the drive mechanism. https://vssewingmachine.in/usha-sewing-machine-price-in-chennai/ Angle Acceleration will fluctuate periodically when the sewing machine is running driven by the motor servo system, which will lead the angle speed of the rotating when the output torque is fixed. Fig. 2 shows the velocity curve on one fixpoint of the rotating under no-load while the speed is set to 150r/min.

The velocity curve of setting up speed of 150r/min showed in Fig. 2 fluctuates periodically in a finite region. The real-time speed in the graph is gained from the calculation of the position signal detected by HALL sensor which has low precision. A simple algorithm is needed with less work and more control precision while there is not high precision position sensors such as optical-electricity encoder.

B. Increment-PI control Algorithm The improved PI control algorithm discussed in the paper is based on the normal increment-PI algorithm. Equation (1) expresses the normal increment-PI algorithm, with the parameter Kp denoting proportion coefficient and Ki denoting integral coefficient [9].

C. PID with digital filter There is lots of disturbance all around when it is working in an industry environment, so disturbance signals will be included in the collected data while ISM is collecting speed signal. Most of the disturbance signal is random disturbance signal. Hence, the digital filter algorithm is added in the speed control algorithm in order to prevent system’s natural control from disturbing signal. Arithmetical mean filter technique, one of the digital filter algorithm, has good control effect on the periodic disturbance and it is simply realized [2,7,10]. The main operation procedure is described as follows: n sampled speed data are considered as a queue, and the length of the queue n is fixed. Then the average of these sampled data Xi is considered as the effective speed dataY .

The response speed of the system can be observed by changing the valuen , and then the proper value n could be determined by the best case. The application of the digital filter, can ensure the validity of the data processing, reduce the difficulty of the system control and consequently improve the precision of the system control.

D. PI Control with Changing Dead Zone Fig. 2 shows that the speed curve fluctuates in a finite region when ISM works in the no-load condition. But it doesn’t result in great mechanical vibration, which can be seemed to be normally working. Huge workloads will be taken to improve the speed performance to linear if control algorithm is only to be improved, and the feasibility is little because of the limit of the characteristic of sewing machine and the use of HALL sensors. So, the speed is permitted to fluctuate in a limited region which will not lead a bad effect for making the algorithm much simpler. The algorithm which is called PI with dead zone means that the speed of sewing machine will not be adjusted if the speed error error(k) is no more than a fixvalue.

It is found that the speed on the same point will fluctuate in a different region while the setting up speed is different. The higher the speed is set up, the wider the speed curve fluctuates. So it requires associating the dead zone value with value of the setting up speedVS .

The application of the PI control algorithm with changing dead zone, can only improve the dynamic performance with a certain extend. However, this is a very necessary step because strong vibrations will occur while PI algorithm with constant dead zone is adopted when the setting up speed is high. These vibrations may lead to noise pollution and terrible damage of mechanical components. But it will not appear under the high setting up speed working status while the PI control algorithm with changing dead zone is adopted.

E. Multi-segment-PI control ISM is required to complete the speed adjustment within a shortest time and make the output torque invariable when the loads appeared. When the control signal is too little, the mechanism of ISM may be blocked, which will lead a great of loads. But great mechanical vibration will appear if the control signal is too large. The events mentioned above are not allowed in both aspects of quality insurance and safety production. The speed signal of the servo system is detected by HALL sensor of the BLDCM. HALL sensor can detect 12 speed signals while BLDCM works in one circuit. Therefore, to realize the high precision of ISM will be an extreme difficulty work. According to all above, improve steps below are needed in the case of long speed-detect period of control system for the stably controlling of ISM when loads appear. 1) Adjusting the time period of closed-loop Obviously, the shorter the speed-detect period of closeloop is, the faster the system’s response speed will be. However, the speed-detect period of close-loop should be longer than the time period of speed inspection because HALL sensor is adopted to detect the speed signal. 2) The parameters Kp and Ki of PI control algorithm are segmented according to the setting up speed, which is called segment-PI control algorithm.

The application of segment-PI control can improve the dynamic performance of the system and make the system have the smooth performance at every speed segment. According to the concrete status of the practical system, segment-PI can not only be divided into 3 segments, but also can be divided into 4, 5 or more. The more the number of segments are, the better the smooth performance of the system’s speed curve will be. However, the increasing segments will increase the workloads of parameter adjustment. Therefore, the changing-dead-zone multi-segment-PID algorithm with digital filter is adopted in this paper after the discussion above. Because the applied control algorithm includes integral link, the step of preventing the integral from being saturated is required to add in the control process. This step is omitted in this paper.

F. Simulation of ISM’s speed control by dSPACE According to ISM’s servo system, the steps of HILS by dSPACE are: 1) Build mathematical model of ISM servo system’s speed control and then build the simulation model by using MATLAB/Simulink and control unit of dSPACE. Fig. 4 is the simulation model of ISM servo system’s speed control which is built based on the mathematical model showed in Fig. 3. The normal control unit supplied by MATLAB/Simulink and the special control unit supplied by dSPACE are both used in the simulation model. And the control unit called Serial Receive represents the serial port of the hardware of dSPACE, and the serial port can send and receive data from or to dSPACE. 2) Translate the mathematical model into object files by using MATLAB/RTW and RTI. 3) Build visual monitor interface of ISM’s servo system by using ControlDesk. Fig. 5 shows the ISM Monitor Interface of ISM’s control system created by ControlDesk. The knobs on the left of the frame can set the setting up speed VS and the duty cycle of the BLDCM. The dialog on the bottom of the frame can change all the settings of the real-time display. The Plotter control unit in the middle of the frame shows the real-time velocity curve.

4) According to the ISM’s working principle, connect the driving motor and sewing machine to the hardware platform of dsPACE. Fig. 6 shows the simple hookup of the ISM in the HILS. The control signal is supplied by PC which MATLAB/Simulink and dSPACE’s software system are built, and the signals such as PWM signals and HALL signals are Fig. 4 Simulation Model of Algorithm Fig. 3 Schematic diagram of speed control system Arithmetical Mean Filter Segment-PID with Changing Dead Zone Y Xi VS ek Brushless DC V Motor k u Fig. 5 ISM Monitor Interface transmitted from PC to BLDCM by dSPACE’s hardware platform CP1104. The feedback signal is also transmitted by CP1104. 5) Download the object files to the hardware of dSPACE and run it, the real-time data could be monitored from the ControlDesk monitor interface, e.g. graph of velocity curve. 6) According to the real-time data that monitored, modify the mathematical model and adjust the parameter of system, then translate it and download the object files into hardware again. Repeat this step till the servo system working in a good status.