Inverse Design Method for Automatic Overlay Welding Robot Mechanism on Membrane Waterwall

Literature Overview and Engineering Challenge

The paper by Tian Songya, Gui Pengqian, Fu Qiang, and Chang Yu from Hohai University and Suzhou Hailu Heavy Industry Co., Ltd. presents an inverse design methodology for an automatic overlay welding robot mechanism applied to membrane waterwall tubes in boiler systems. Published in Transactions of the China Welding Institution (2018, Vol. 39, No. 6), this work addresses a specific and challenging engineering problem: the automated overlay welding of corrosion- and wear-resistant materials onto membrane waterwall surfaces in power generation boilers.

Membrane waterwalls are integral components of modern boiler designs, consisting of parallel water tubes welded to flat steel plates to form a continuous wall that serves as both a heat transfer surface and a structural barrier. In service, these waterwalls are exposed to high-temperature flue gases containing corrosive species such as sulfur oxides and chlorides, as well as erosive fly ash particles. The combination of thermal corrosion and erosive wear leads to progressive thinning of the waterwall surface, ultimately resulting in tube rupture and catastrophic boiler failure.

The overlay welding of a corrosion- and wear-resistant alloy onto the waterwall surface is an established repair and protection strategy. However, the geometry of membrane waterwalls presents unique challenges for automated welding. The waterwall is typically installed in a vertical orientation, and the overlay welding must be performed from the top downward to minimize dilution and ensure proper weld bead formation. This orientation requirement imposes significant constraints on the robot mechanism design.

Mechanism Design and Kinematic Configuration

The proposed robot mechanism consists of several key components arranged in a coordinated system:

The overall mechanism provides six degrees of freedom, which is sufficient to position the welding torch at any point on the waterwall surface with the correct orientation. The well frame provides a rigid support structure that holds the waterwall in a fixed vertical position, while the portal frame and beam system provide the translational movements necessary to traverse the waterwall surface.

The three rotating motors on the welding trolley are responsible for the initial positioning of the torch. Motor 5 adjusts the torch angle in one plane, Motor 7 adjusts the torch angle in a perpendicular plane, and Motor 6 adjusts the torch position along the beam axis. Together, these motors establish the initial torch-to-workpiece geometry before the beam motors drive the welding process.

Inverse Design Methodology

The inverse design method is the core contribution of this paper. In conventional robot programming, the forward kinematics approach is used: given the joint angles, the position and orientation of the end effector are calculated. However, in the overlay welding application, the desired torch position and orientation along the weld path are known, and the required joint angles must be determined. This is the inverse kinematics problem.

The inverse design method proposed in this paper uses coordinate transformation theory to solve for the joint angles given the desired torch position. The key steps are as follows:

1. Define the desired torch position and orientation at each point along the weld path.
2. Establish the coordinate transformation chain from the base frame to the torch frame, passing through each joint.
3. Apply inverse coordinate transformations to solve for the joint angles that produce the desired torch position.
4. Determine the adjustment amounts for each motor based on the calculated joint angles.
5. Verify the solution by forward kinematics to confirm that the calculated joint angles produce the desired torch position.

The coordinate transformation approach is particularly well-suited to this application because the mechanism has a relatively simple kinematic structure with clearly defined joint axes. The transformation matrices for each joint can be expressed in standard Denavit-Hartenberg parameters, and the inverse solution can be obtained analytically without the need for numerical iteration.

Downward Vertical Welding Process Considerations

The choice of downward vertical welding (from top to bottom) is a deliberate process selection based on metallurgical considerations. In overlay welding, the dilution rate—the fraction of base metal melted into the weld pool—is a critical parameter that affects the final composition and properties of the overlay layer.

In downward vertical welding, the weld pool forms above the advancing torch, and the solidified weld metal is deposited below the weld pool. This configuration has several advantages for overlay welding:

• The weld pool is shielded by the previously deposited weld metal, reducing oxidation and contamination
• The cooling rate of the weld pool is moderated by the adjacent solidified metal, promoting a more uniform microstructure
• The dilution rate is reduced because the base metal is heated from above by the arc rather than from below by the weld pool
• The weld bead profile is more uniform and predictable, facilitating automated process control

The following table summarizes the process advantages of downward vertical welding for overlay applications:

Coordinate Transformation and Dimensional Determination

The paper demonstrates that the coordinate transformation method can be used not only to determine joint angles for a given torch position but also to determine the overall dimensions of the robot mechanism. By defining the required workspace—the area of the waterwall surface that must be covered by the welding torch—the inverse design method can be used to calculate the required stroke lengths of the beam motors, the travel distance of the trolley, and the angular ranges of the rotating motors.

This approach is particularly valuable during the design phase because it allows engineers to size the mechanism components based on the actual welding requirements rather than on conservative estimates. The method ensures that the mechanism has sufficient reach and dexterity to cover the entire weld path while minimizing unnecessary size and cost.

The dimensional determination process involves:

1. Defining the waterwall geometry, including tube spacing, plate dimensions, and total coverage area.
2. Specifying the required torch-to-workpiece distance and orientation at each weld point.
3. Applying the inverse kinematics solution to determine the joint angles at the start and end of the weld path.
4. Calculating the maximum displacement of each joint to determine the required stroke or angular range.
5. Adding appropriate safety margins to account for manufacturing tolerances and thermal deformation.

Engineering Practice Integration and Implementation Considerations

For engineers implementing this robot mechanism in a power plant or boiler manufacturing facility, several practical considerations must be addressed:

• Thermal deformation: The waterwall is at elevated temperature during welding, which causes thermal expansion and potential distortion of the weld path. The robot control system must compensate for these thermal effects, either through real-time sensing or through pre-programmed compensation based on temperature measurements.
• Surface irregularities: Membrane waterwalls may have surface irregularities due to manufacturing tolerances, previous repairs, or scale buildup. The robot mechanism must have sufficient compliance or the control system must have adaptive capabilities to maintain consistent torch-to-workpiece distance.
• Shielding gas management: In vertical welding, shielding gas flow must be carefully managed to ensure adequate protection of the weld pool. The gas flow rate and direction must be optimized for the downward vertical position to prevent gas blowout and oxidation.
• Safety considerations: The robot mechanism operates in close proximity to high-temperature surfaces and may be exposed to flue gases and combustion products. Adequate safety measures including thermal shielding, gas detection, and emergency stop systems are essential.

Key Questions and Reflective Insights

The inverse design method presented in this paper is a rigorous and systematic approach to robot mechanism design that is directly applicable to other automated welding applications. The key advantage of the inverse design approach is that it starts from the known requirements—the desired weld path and torch geometry—and works backward to determine the mechanism configuration and dimensions. This is in contrast to the conventional approach of designing the mechanism first and then programming the robot to follow the weld path, which may result in a mechanism that is inadequately sized or configured for the actual welding requirements.

Several questions arise from this work that merit further investigation. First, the accuracy of the inverse design method depends on the accuracy of the kinematic model. In practice, manufacturing tolerances, joint play, and thermal deformation introduce errors that may reduce the accuracy of the torch positioning. A study of the sensitivity of the torch position to errors in each joint parameter would provide valuable insight into the tolerance requirements for the mechanism components.

Second, the paper does not address the dynamic performance of the robot mechanism. During welding, the torch must maintain a consistent speed and position, and the mechanism must be able to accelerate and decelerate smoothly at the start and end of each weld pass. The dynamic analysis of the mechanism, including inertia, friction, and damping, is essential for designing a control system that maintains the required welding quality.

Third, the method is demonstrated for a specific waterwall geometry. The generalization of the inverse design method to other complex geometries, such as curved waterwalls or waterwalls with irregular tube patterns, would extend the applicability of the approach. The coordinate transformation method is inherently general and should be applicable to any geometry that can be described mathematically, but practical implementation may require additional considerations.

The study by Tian and colleagues demonstrates that the inverse design method based on coordinate transformation theory is an effective and accurate approach for designing automated overlay welding robot mechanisms for membrane waterwalls. The method provides a systematic framework for determining mechanism dimensions, joint angles, and motor adjustment amounts based on the welding requirements, and it offers a rigorous alternative to the trial-and-error approach commonly used in robot programming.

In conclusion, the inverse design methodology presented in this paper provides engineers with a powerful tool for designing automated overlay welding systems that are specifically tailored to the geometry and requirements of the application, and the coordinate transformation approach is a fundamental technique that should be incorporated into the standard toolkit of welding engineers working on automated welding system design.