In modern industrial production, the grinding process for gear shafts often remains labor-intensive, with manual loading and unloading operations dominating many workshops. This not only increases labor costs but also introduces variability in product quality due to human error. As a key component in transmission systems, the precision of a gear shaft directly impacts overall mechanical performance. Therefore, automating the grinding process for gear shafts is crucial for enhancing efficiency, consistency, and competitiveness. In this study, we propose and develop an automatic grinding system for gear shafts that utilizes an industrial robot to serve multiple CNC grinding machines. This system integrates programmable logic controllers (PLCs), human-machine interfaces (HMIs), and customized end-effectors to create a seamless, high-throughput production line. By focusing on the gear shaft as the central workpiece, we aim to address the challenges of automation in precision machining.
The traditional workflow for gear shaft manufacturing typically involves steps such as turning, hobbing, deburring, shaving, heat treatment, and finally grinding. The grinding stage is critical for achieving the required dimensional accuracy and surface finish. However, manual handling during grinding can lead to issues like part damage, inconsistent cycle times, and high labor dependency. With the advent of advanced robotics, it is now feasible to deploy industrial robots for repetitive tasks like loading and unloading gear shafts onto grinding machines. Our system leverages a single robot to manage two CNC grinders and a feeding mechanism, thereby optimizing space and resources. This approach not only reduces operational costs but also minimizes human intervention, leading to a more stable production environment. Throughout this article, we will detail the design, implementation, and benefits of this automated system, emphasizing the role of the gear shaft in driving our engineering decisions.
The core of our automated system lies in its holistic integration of mechanical, electrical, and software components. We began by analyzing the spatial and operational requirements for handling gear shafts. The system layout positions two CNC grinding machines (Machine A and Machine B) facing each other in a “品”-shaped arrangement, with an industrial robot at the center and a feeding machine (feeder) adjacent to them. This configuration minimizes the robot’s travel distance and maximizes accessibility to all workstations. The gear shafts are transported in specially designed pallets or boxes, each holding multiple pieces arranged in a grid pattern. The feeder is responsible for moving these pallets into and out of the workspace, ensuring a continuous supply of raw gear shafts and removal of finished ones. The robot, equipped with a multi-gripper end-effector, performs all transfer operations between the feeder and the grinding machines. This setup is encapsulated in the following system workflow diagram, which we will elaborate on in subsequent sections.
To formalize the system’s operation, we define the key states and transitions using a state machine model. Let $S$ represent the system state, which can be one of $\{ \text{Idle}, \text{Loading}, \text{Grinding}, \text{Unloading}, \text{Transferring} \}$. The transition between states depends on signals from the robot, grinders, and feeder. For instance, when a gear shaft is picked from the feeder, the system moves from $\text{Idle}$ to $\text{Loading}$. The grinding process then initiates, transitioning to $\text{Grinding}$. After completion, the robot unloads the finished gear shaft and loads a new one, cycling through these states. The overall efficiency can be modeled using the total cycle time $T_c$, given by:
$$T_c = T_p + T_g + T_r + T_f$$
where $T_p$ is the picking time, $T_g$ is the grinding time (which may vary between machines), $T_r$ is the robot movement time, and $T_f$ is the feeding mechanism cycle time. By optimizing these parameters, we aim to minimize $T_c$ and maximize throughput for gear shaft production.
The electrical control architecture is centered around a PLC that acts as the main coordinator. It communicates with the robot via Modbus TCP, with the grinding machines through digital I/O, and with the feeder via sensor inputs and actuator outputs. A touch-screen HMI provides operators with real-time monitoring and control capabilities, such as starting/stopping the system, viewing status alerts, and adjusting parameters for different gear shaft variants. The PLC program is structured into modular function blocks for each subsystem, ensuring scalability and ease of maintenance. For example, one block handles the robot’s motion commands, another manages the grinder’s start/stop signals, and a third oversees the feeder’s pallet exchange. This decoupled design allows for independent troubleshooting and upgrades. Below is a summary of the key hardware components and their specifications in the system:
| Component | Model/Specification | Role in System |
|---|---|---|
| Industrial Robot | Stäubli TX90 (6-axis) | Handles all gear shaft transfers |
| CNC Grinding Machine A | Customized for gear shaft grinding | Performs primary grinding on gear shaft |
| CNC Grinding Machine B | Customized for gear shaft grinding | Performs secondary grinding on gear shaft |
| Feeding Machine | Conveyor with pallet system | Supplies and removes gear shaft pallets |
| PLC | Siemens S7-1200 | Central controller for coordination |
| HMI | 10-inch touchscreen panel | User interface for monitoring and control |
In designing the end-effector for the robot, we focused on the unique geometry and handling requirements of the gear shaft. A gear shaft typically consists of a cylindrical shaft with integrated gear teeth, requiring careful gripping to avoid damage to the precision surfaces. Our end-effector features three independent grippers mounted on a common base: two two-finger parallel grippers (Gripper 1 and Gripper 2) and one three-finger parallel gripper (Gripper 3). This configuration allows the robot to perform multiple tasks in a single cycle. Gripper 3 is used to transfer gear shafts from the pallet on the feeder to a temporary staging area (a dedicated platform), as direct gripping from the pallet might interfere with the grinding machine’s fixtures. Grippers 1 and 2 are then employed for simultaneous loading and unloading on the grinding machines. Specifically, one gripper holds a raw gear shaft while the other is empty; after unloading a finished gear shaft, the robot rotates to load the raw one, thus minimizing idle time. The grippers are pneumatically actuated, with magnetic sensors providing feedback on open/close states. The force exerted by each gripper is calibrated based on the weight and fragility of the gear shaft, ensuring secure handling without deformation. The grip force $F_g$ can be expressed as:
$$F_g = \frac{W \cdot g \cdot k}{\mu}$$
where $W$ is the weight of the gear shaft, $g$ is gravitational acceleration, $k$ is a safety factor (typically 1.5 to 2), and $\mu$ is the coefficient of friction between the gripper jaws and the gear shaft material. For our gear shafts, which are made of hardened steel, we use $W \approx 5\,\text{kg}$, leading to $F_g \approx 100\,\text{N}$ per gripper. The end-effector is designed for quick changeover, allowing adaptation to different gear shaft sizes by swapping jaw inserts—a critical feature for flexible manufacturing.
The selection of the industrial robot was driven by payload capacity, reach, precision, and programming flexibility. After evaluating several models, we chose the Stäubli TX90 for its high repeatability ($\pm 0.03\,\text{mm}$), ample payload (up to $7\,\text{kg}$ nominal), and extensive work envelope ($1000\,\text{mm}$ reach). These characteristics make it ideal for handling gear shafts within the confined layout of our system. The robot’s six degrees of freedom enable complex motions, such as reorienting the gear shaft during transfer to align with machine fixtures. We programmed the robot using VAL3 language, which supports structured programming with loops, functions, and real-time I/O handling. The main program orchestrates a sequence of movements to service the two grinders and the feeder. For instance, to handle a pallet containing gear shafts arranged in a $5 \times 10$ grid, we use nested loops to iterate through each position. Let $p_{\text{origin}}$ denote the taught position for the first gear shaft in the pallet, with offsets $\Delta x = 33\,\text{mm}$ between columns and $\Delta y = 34\,\text{mm}$ between rows. The target position $p_{\text{target}}$ for the $(i,j)$-th gear shaft is computed as:
$$p_{\text{target}} = p_{\text{origin}} + (i \cdot \Delta x, \, j \cdot \Delta y, \, 0, \, 0, \, 0, \, 0)$$
where $i = 0,1,\dots,9$ and $j = 0,1,\dots,4$ represent column and row indices, respectively. This parametric approach allows easy adaptation to different pallet configurations for various gear shaft batches. The robot’s motion paths are optimized to avoid collisions and minimize cycle time, using linear and joint interpolations. A snippet of the VAL3 code illustrates the core logic:
// Initialize
movej(home_position, fast)
open(gripper1)
open(gripper2)
open(gripper3)
// Wait for feeder ready signal
wait(io: feeder_ready == TRUE)
// Loop through pallet positions
for row = 0 to 4
for col = 0 to 9
// Compute pick position
pick_pos = compose(base_position, {col*33, row*34, 0, 0, 0, 0})
// Call subroutines for grinding operations
call pick_gear_shaft(pick_pos)
call unload_machine_A()
call load_machine_A()
call unload_machine_B()
call load_machine_B()
call place_gear_shaft(finished_pos)
endfor
endfor
The integration of the robot with the grinding machines requires precise synchronization. Each grinder is equipped with sensors to detect the presence of a gear shaft and signal completion of the grinding cycle. The PLC polls these signals and triggers the robot accordingly. For example, when Machine A finishes grinding a gear shaft, it sends a “done” signal to the PLC, which then instructs the robot to unload the finished gear shaft and load a new one. This handshake mechanism ensures that the robot only accesses the machine when it is safe, preventing accidents. Additionally, we implemented error handling routines for scenarios like a missed pick or a jammed gear shaft. In such cases, the system pauses and alerts the operator via the HMI, maintaining safety and preventing damage to the gear shaft or equipment.
To evaluate the performance of our automated system, we conducted tests with a batch of gear shafts over multiple production runs. Key metrics included cycle time, accuracy, and defect rate. Compared to manual operation, the automated system reduced the average cycle time per gear shaft by approximately 30%, from 120 seconds to 84 seconds. This improvement stems from the robot’s consistent speed and elimination of human delays. Moreover, the precision of robot placement resulted in better alignment in the grinding machines, enhancing the geometric tolerances of the finished gear shafts. We measured critical dimensions such as diameter and tooth profile, finding that the automated process reduced variance by up to 50%. The defect rate due to handling damage (e.g., scratches or dents on the gear shaft) dropped from 5% to under 0.5%, underscoring the gentle handling of the custom end-effector. These gains translate directly into cost savings and higher product quality for gear shaft manufacturers.
Beyond immediate performance, our system offers scalability and flexibility. By modularizing the design, we can easily add more grinding machines or robots to handle higher volumes of gear shafts. The PLC program can be extended with additional function blocks, and the robot’s VAL3 code can incorporate new motion routines for different gear shaft geometries. We also explored the use of machine learning algorithms to predict maintenance needs based on robot joint temperatures and grinding machine power consumption, though this is part of future work. Another avenue for improvement is adaptive gripping force control using force-torque sensors on the end-effector, which would allow real-time adjustment based on the weight and surface condition of each gear shaft. This could further reduce the risk of damage, especially for lightweight or coated gear shafts.
In conclusion, the automation of gear shaft grinding through industrial robotics presents a significant advancement in manufacturing technology. Our system demonstrates that a well-integrated design, combining a robust robot, intelligent control, and specialized tooling, can dramatically improve productivity and quality in gear shaft production. The key innovations include the multi-gripper end-effector for simultaneous operations, the parametric robot programming for flexible pallet handling, and the PLC-based coordination that ensures seamless interaction among all components. As industries move towards Industry 4.0, such automated systems will become increasingly vital for maintaining competitiveness. We believe that our approach can be adapted to other precision machining processes involving similar components like gear shafts, paving the way for broader adoption of robotics in traditional manufacturing sectors. Future research will focus on enhancing the system’s autonomy through vision systems for random bin picking of gear shafts and integrating digital twin simulations for predictive optimization.
Throughout this study, the gear shaft has been the focal point, driving every design decision from gripping mechanisms to motion planning. By prioritizing the needs of this critical component, we have developed a solution that not only automates a tedious task but also elevates the entire production chain. We hope that our findings inspire further innovation in the automation of gear shaft manufacturing and related fields, ultimately contributing to smarter, more efficient factories.