How to improve the picking accuracy of injection molding manipulator through control system?

The core of improving the picking accuracy of injection molding manipulator through control system is to optimize the three links of instruction generation, execution feedback and dynamic correction, and combine hardware adaptation and algorithm upgrade to realize "accurate instruction, execution without deviation and error compensation". Specific measures are as follows:

1. Matching the upgrade control core with the driver

Select high-precision controller and servo system.

Adopt special motion controller (instead of basic PLC): support multi-axis linkage and complex interpolation algorithm (such as linear/circular interpolation), and the operation period can be shortened to less than 0.1ms, ensuring no delay in high-speed picking.

"Closed-loop control" of servo system: the servo motor should be equipped with a high-resolution encoder (such as a 23-bit absolute encoder with an accuracy of 0.001mm) and form a "command-position feedback-correction" closed loop with the controller to eliminate errors such as step loss and slippage in real time.

For example, for thin-walled food box picking, the positioning accuracy of the servo system should reach ±0.05mm, and the controller should support 1μm pulse output to avoid deviation caused by insufficient instruction accuracy.

Adaptive optimization of driving parameters

The dynamic parameters of the servo motor are adjusted by the controller: the gain and inertia ratio are automatically adjusted according to the load weight (such as food boxes with different specifications), so as to avoid "overshoot" caused by excessive inertia when starting and stopping (such as rushing past the target position when picking up parts).

Enable "feed-forward control": predict the motion inertia of the manipulator in advance, and add compensation to the instruction to reduce the position deviation in the acceleration/deceleration stage (especially suitable for long-stroke manipulators).

Second, introduce real-time feedback and error compensation algorithm.

Multi-sensor fusion feedback

Install high-precision position sensor: install grating ruler or magnetic grating ruler on the key joints of the mechanical arm (such as waist and big arm) to directly detect the actual position of the arm (rather than relying on the indirect feedback of the motor encoder) to eliminate the "following error" caused by the transmission gap.

Visual positioning compensation: the actual position of the food box in the mold can be identified in real time by industrial camera and visual algorithm (such as template matching and edge detection) (there may be a deviation of ±0.5mm due to the injection error), and the controller can automatically correct the coordinates of the taken part according to the visual feedback, which is especially suitable for the unstable scene of product positioning (such as multi-cavity in one mold and slight mold sticking).

Force feedback: a force sensor is installed at the end of the fixture to detect the contact force (such as the suction force of the suction cup and the clamping force of the clamping jaw) when the workpiece is taken, and the controller adjusts the action according to the force signal (such as fine-tuning the position to ensure the fit when the suction force is insufficient) to avoid product deviation caused by uneven force.

Dynamic error compensation algorithm

Mechanical error modeling compensation: the inherent errors of the manipulator in different positions (such as the terminal droop caused by insufficient rigidity of the arm) are recorded through calibration experiments, and an error database is established. The controller automatically superimposes the compensation value in the instruction (for example, the height compensation of+0.2mm is preset at the maximum stroke position).

Temperature drift compensation: a temperature sensor is integrated in the controller to monitor the environmental temperature of the manipulator in real time, calculate the size change according to the coefficient of thermal expansion and contraction of metal (such as aluminum alloy α = 23× 10/℃), and dynamically correct the position instruction (such as automatically shortening the horizontal stroke by 0.1mm in high temperature environment).

Third, optimize the motion trajectory and program logic

Smooth trajectory planning

S-shaped acceleration and deceleration curve is adopted: instead of the traditional trapezoidal curve, the speed and acceleration of the mechanical arm are gradually changed when starting and stopping, and the impact vibration is reduced (especially when picking up parts at high speed, the vibration may lead to the end deviation of more than ±0.3mm).

Segmented trajectory optimization: the picking process is divided into "fast approaching-slow positioning-fine grasping" stages, and the speed is reduced within 10mm close to the product (for example, from 500mm/s to 50mm/s), and the positioning accuracy is improved through low-speed movement.

Adaptive adjustment of program parameters

Pre-set "precision mode" for different products: for example, enable "high precision mode" (reduce speed and increase feedback frequency) for high-precision food boxes (such as buckle structure) and "high efficiency mode" for ordinary products to balance precision and efficiency.

Introduce "learning function": record the optimal parameters (such as the best picking angle and the trigger time of the sucker) by picking data for many times, and automatically optimize the subsequent action instructions (similar to "iterative learning control").

Fourth, anti-interference and stability enhancement

Electrical anti-interference design

The signal wires of the controller and servo system are shielded wire and twisted pair, and are separately wired from the power wire to reduce electromagnetic interference (for example, high-frequency interference generated by the motor of injection molding machine may lead to sensor signal distortion).

Install power filter: stabilize the power supply voltage of controller and servo system (avoid fluctuation of more than 5%) to prevent abnormal motor output torque caused by voltage instability.

Fault self-diagnosis and correction

The controller has built-in real-time monitoring module: continuously detect servo alarm (such as overload, encoder failure) and sensor abnormality (such as visual signal loss), and once the deviation exceeds the threshold (such as picking position error > ＞0.2mm), immediately stop the action and trigger a correction program (such as recalibrating the origin).

Redundancy design: key signals (such as limit position switches) are detected by two channels to avoid positioning errors caused by single point failure.

V. Standardization of debugging and calibration process

Precise calibration of mechanical origin and coordinate system

The laser interferometer is used to calibrate the travel accuracy of each axis of the manipulator, and the deviation between the actual value and the theoretical value of each position is recorded. The "electronic gear ratio" parameter of the controller is corrected to ensure that the commanded position is consistent with the actual position.

Establish a "manipulator-injection molding machine" linkage coordinate system: associate the manipulator picking coordinate with the mold position of the injection molding machine, and automatically compensate the tiny displacement (such as the mold position deviation caused by the change of clamping force) when the mold is opened and closed by the controller.

Trial and error optimization before mass production

The first product was tested 100 times repeatedly, and the position deviation of each pickup was recorded by the controller, and the average value was calculated and written into the program as compensation to reduce the system error.

Simulate extreme working conditions (such as full load and maximum speed) test, and adjust parameters (such as increasing deceleration time) through the controller to ensure that the accuracy is still up to standard in the limit state.

summary

The core logic of the control system to improve the picking accuracy is "high-precision instruction output+real-time feedback correction+dynamic error compensation". By upgrading the hardware (controller, servo, sensor), optimizing the algorithm (trajectory planning, compensation model), strengthening the anti-interference ability, and combining with standardized debugging, the precision of repeated positioning can be improved from ±0.1mm to less than ±0.05mm, which can meet the production requirements of high-precision food boxes (such as products with sealed grooves and multi-cavity integrated molding).