Abstract: This paper presents a method for developing an offline programming system for robot arc welding on the advanced modeling system SolidWorks. A fully functional and practical offline programming system for robot arc welding has been successfully developed. This paper presents the object model of the system and introduces the functions of each basic functional module. Introduction With the continuous improvement of microcomputer performance, many large-scale CAD design software have launched their microcomputer versions, and software such as SolidWorks that runs directly on microcomputers has emerged. In the field of robotics, offline programming software that used to run on high-end graphics workstations, such as ROBCAD, IGRIP, and WORKSPACE, have launched Windows-based offline programming systems. These software are all general-purpose offline programming software and lack support for special applications of arc welding. For example, WORKSPACE focuses more on the development of calibration functions [1]. When programming, users often have to establish robot motion paths through graphical teaching, which is quite a lot of work for programming large workpieces or complex paths. From the perspective of development direction, offline programming systems are developing towards intelligence, making user programming simpler and more convenient, and automatic programming technology has become the goal pursued by people. The offline programming system platform AutoCAD2000 developed by the Welding Laboratory of Harbin Institute of Technology has been found to have some shortcomings due to inherent limitations in its development. These include insufficient solid modeling capabilities in 3D design, inability to support modeling of complex work units, and limitations in practical application. Currently, offline programming systems with a certain degree of autonomous planning capability would significantly reduce user workload. The acquisition of workpiece CAD information becomes the data input for subsequent planning, requiring the offline programming system platform to have strong modeling capabilities, capable of modeling welding workpieces in various welding tasks. Because AutoCAD itself has weak 3D modeling capabilities, models imported from other modeling tools cannot obtain sufficient workpiece modeling information, hindering subsequent automatic planning functions. Developing a cost-effective and professional arc welding offline programming system suitable for Chinese enterprises is of great significance for promoting the practical application of offline programming. 1. System Development Environment SolidWorks, due to its low price and functionality required by general users, has become the software product with the most users, highest user satisfaction, and highest sales volume in the global mainstream 3D design market, boasting over 325,000 users worldwide. Furthermore, its secondary development function is very powerful, and it can develop an arc welding offline programming system. The developed offline programming system will have strong modeling capabilities, thus having a strong ability to acquire workpiece and weld information, and to complete the automatic programming task to the greatest extent. Such an arc welding offline programming system is easily accepted by small and medium-sized enterprises. This paper applies the programming method of COM technology, uses the SolidWorks API secondary development interface and VC++ 6.0 and ATL as development tools, and develops an arc welding offline programming system on the SolidWorks platform. The SolidWorks API provides a large number of COM objects for secondary development [2]. These COM objects cover the entire SolidWorks data model. By calling the SolidWorks COM objects, users can implement the same functions as SolidWorks in their own developed system. Figure 1 shows the main SolidWorks API object model. [align=center] Figure 1 SolidWorks API object model [2][/align] 2 Object Modeling of Robot Arc Welding Offline Programming System The object model represents the static, structured system "data" properties. It is a mapping of objects simulating objective world entities and the relationships between them, describing the static model of the system. The relationships between classes and objects can be summarized as inductive, compositional, and associative relationships. This paper presents the main data structures in a three-layer robot work unit: work unit class, describing the combination of objective objects such as robots and the various connections and interactions between these objects; equipment class and joint class. To visually represent the motion process of the welding torch in a graphical environment, path objects are defined, containing multiple label point objects. A path object has a restricted association with a component object. Based on the above analysis, an object model for the offline programming and simulation system is established, as shown in Figure 2. [align=center] Figure 2 System Object Model[/align] 3 System Structure Functionally, the robot arc welding offline programming system adds arc welding functionality compared to a general offline programming system. Therefore, it includes the functions of a general offline programming system while also emphasizing support for arc welding. This paper improves the functionality of each module. The overall structure of the robot execution-level arc welding offline programming system is shown in Figure 3. [align=center]Figure 3 Functional Module Division[/align] The following is a general introduction to the functions of each module. 1) Equipment Modeling This module is used to define peripheral equipment such as robots, welding workpieces, and positioners: it defines assemblies created in SolidWorks as various equipment, defines the joint motion types of the equipment, defines the tool parameters of the equipment, and sets the kinematic parameters of the equipment. It creates a library of robot equipment and a library of peripheral equipment such as positioners. 2) Work Cell Layout The system should provide the ability to create work cells composed of multiple devices, i.e., the work cell layout function. This module provides the function of calling existing equipment, defining the pose relationship between equipment, and assembling coordinated motion equipment. It creates a work cell library. Figure 4 shows an example of a built work cell. [align=center]Figure 4 Work Cell Simulation Example[/align] 3) Motion Simulation This module mainly provides the following functions: joint motion of various devices, return to zero position motion; target point motion, circular motion, and path movement of single robot devices and coordinated motion equipment; path motion motion mode and simulation step size setting, etc. 4) Path Functionality: Path functionality is a key feature of offline programming, enabling visual recording of the robot's motion process for easy editing by the user. For welding tasks, this paper implements the function of recording on-site information using paths, including the joint angle values ​​of the robot's auxiliary axes at each path point, the TCP pose in the robot's base coordinates, welding parameters, etc. 5) Programming: Based on existing programming languages ​​and combined with welding applications, this paper provides an offline arc welding programming language (OLPWL), which users can use for offline programming. This language is similar to the programming language of actual arc welding robots, making it convenient for programmers familiar with robot operation and also suitable for training programmers. This module is an offline programmer integrating editing, debugging, and simulation functions. 6) Data Conversion: General offline programming systems provide a module for converting data to the actual robot program, i.e., a program conversion module. However, currently, due to different robot manufacturers using different programming languages, the biggest difficulty lies in the conversion of robot motion data, i.e., converting motion path point data into path point (or joint angle) data used in the actual robot program. Because different robot manufacturers use different description methods and storage formats, this information may be unreadable. The IGRIP software only provides direct program conversion with IGM and MotoMan robots. This paper proposes a method to convert motion path point data into a data format easily understood by all robot programmers, and the actual robot program is generated separately based on this data. This paper specifies a general path data description method. 7) Data Upload The offline programming system should also be able to convert the robot program obtained from teaching into a path in the simulation working cell. For the same reasons mentioned above, this paper proposes a data upload function to upload the general path data specified in this paper to the offline programming system to generate the path. 8) State Monitoring This module includes joint over-limit checks, joint motion speed and acceleration over-limit checks, and collision detection. Previous offline programming systems detected whether joint angles exceeded limits through pre-checks. This method does not allow users to form an intuitive sense of the joint angle changes throughout the robot's path motion. It does not help with the pose adjustment of path points at joint over-limit locations. This paper adopts a method of monitoring robot state changes during dynamic motion. It provides prompts and does not terminate the robot's motion process. 9) Calibration To ensure the consistency between the graphical working cell model and the actual working cell model, the actual working cell needs to be calibrated. 10) For specific applications of arc welding, this paper developed an arc welding parameter database management function. 4. Conclusion This paper developed a low-cost and professional offline programming system for arc welding running on the SolidWorks platform. This system has strong modeling capabilities, solving the problem that previous systems, due to insufficient modeling capabilities, could only extract workpiece geometric information from some simple workpieces. Using object-oriented technology, an object model for a robot arc welding offline programming and simulation system was established, and a relatively complete robot arc welding offline programming and simulation system was developed.