Fredrik Danielsson - Academia.edu (original) (raw)
Papers by Fredrik Danielsson
The International Journal of Advanced Manufacturing Technology
Industries of low-batches or one-off manufacturing aim for automation that is competitive enough ... more Industries of low-batches or one-off manufacturing aim for automation that is competitive enough to adapt to new or modified products daily through in-house knowledge that focuses on manufacturing processes and not on machine function programming. To solve this, a complete set of actions that utilize seamless data transfer from product design in CAD to a Plug & Produce automation concept is proposed together with a generic structure for the integration of standard industrial robots into agents. This structure enables agents to handle their local reference coordinate systems and locations relative to a global perspective. Seamless utilization of data from product designs to Plug & Produce will simplify and shorten the time of digital development through concurrently usable text-based and graphical configuration tools of a configurable multi-agent system. Needed data extracts directly from the product design as requirements of operational goals. Extraction of data from the product des...
Industries of low batches or one-off manufacturing aim for automation that is competitive enough ... more Industries of low batches or one-off manufacturing aim for automation that is competitive enough to adapt to new or modified products daily through in-house knowledge. A knowledge that focuses on manufacturing processes, not on machine function programming. To solve this a complete set of actions that utilize seamless data transfer from product design to a Plug & Produce automation concept is proposed together with a generic structure for the integration of standard industrial robots into agents. This structure enables agents to handle their local reference coordinate systems and locations relative to a global perspective. Seamless utilization of data from product designs to Plug & Produce will simplify and shorten the time of digital development through concurrently usable text-based and graphical configuration tools of a configurable multi-agent system. Needed data extracts directly from the product design as requirements of operational goals. Extraction of data from the product d...
Job shop scheduling problems are computationally complex combinatorial optimization problems. Gen... more Job shop scheduling problems are computationally complex combinatorial optimization problems. Genetic algorithms have been used in various forms and in combination with other algorithms to solve job shop scheduling problems. A partially flexible job shop with precedence constraints increases this complex behaviour. There are two main parts to optimizing ajob shop, the routing and the scheduling. The objective here is to get consistent optimal makespan using a genetic algorithm. This paper firstly, presents a simulation approach for the considered partially flexible job shop scheduling problem. Which take into account the precedence constraints and reduce situations of deadlock. To solve the partially flexible job shop scheduling problem a genetic algorithm was used and improved. It utilise a genetic crossovers for routing and a new random shuffle feature is introduced for the scheduling. The computational results have shown that the algorithm performs well in terms of finding a cons...
Journal of Advanced Manufacturing Systems, 2002
... and evaluated. The main tool is a virtual press station, developed by the first author, consi... more ... and evaluated. The main tool is a virtual press station, developed by the first author, consisting of an emulated control system for a feeder/extractor robot which communicates with 3D-simulated production equipment. Moreover ...
Proceedings of 2012 IEEE 17th International Conference on Emerging Technologies & Factory Automation (ETFA 2012), 2012
In order to perform efficient geometrical simulation and virtual commissioning in stamping, three... more In order to perform efficient geometrical simulation and virtual commissioning in stamping, three fields are investigated namely: simulation building time, collision detection time and optimization time. Hence, reducing time is the main theme of this paper. To reduce simulation building time and optimization time, an efficient stamping simulation model is built and tested. Collision detection time is examined by a relative motion method based on 3D to 2D geometrical collision detection. The presented results mean that simulation and virtual commissioning can be performed at least ten times faster compared to standard approaches.
Proceedings of the 24th International Conference on Flexible Automation & Intelligent Manufacturing, 2014
Simulation-based optimisation often considers computationally expensive problems. Successfully op... more Simulation-based optimisation often considers computationally expensive problems. Successfully optimising such large scale and complex problems within a practical timeframe is a challenging task. Optimisation techniques to fulfil this need to be developed. A technique to address this involves decomposing the considered problem into smaller subproblems. These subproblems are then optimised separately. In this paper, an efficient algorithm for simulation-based optimisation is proposed. The proposed algorithm extends the cooperative co-evolutionary algorithm, which optimises subproblems separately. To optimise the subproblems, the proposed algorithm enables using a deterministic algorithm, next to stochastic genetic algorithms, getting the flexibility of using either type. It also includes a constructive heuristic that creates good initial feasible solutions to reduce the number of fitness calculations. The extension enables solving complex, computationally expensive problems efficiently. The proposed algorithm has been applied on automated sheet metal press lines from the automotive industry. This is a highly complex optimisation problem due to its non-linearity and high dimensionality. The optimisation problem is to find control parameters that maximises the line's production rate. These control parameters determine velocities, time constants, and cam values for critical interactions between components. A simulation model is used for the fitness calculation during the optimisation. The results show that the proposed algorithm manages to solve the press line optimisation problem efficiently. This is a step forward in press line optimisation since this is to the authors' knowledge the first time a press line has been optimised efficiently in this way.
Advances in Intelligent and Soft Computing, 2010
Abstract. Computer Aided Process Planning has received more attention re-cently due to considerab... more Abstract. Computer Aided Process Planning has received more attention re-cently due to considerable progress in the aspects of both technology and the-ory. Beside the traditional trends and efforts to integrate the product design and process planning activities usually referred to as ...
ABSTRACT Determining the control parameters for sheet metal press lines is a large scale and comp... more ABSTRACT Determining the control parameters for sheet metal press lines is a large scale and complex optimisation problem. These control parameters determine velocities, time constants, and cam values of critical interactions between the equipment. The complexity of this problem is due to the nonlinearities and high dimensionality. Classical optimisation techniques often underperform in solving this kind of problems within a practical timeframe. Therefore, specialised techniques need to be developed for these problems. An existing approach is simulation-based optimisation, which is to use a simulation model to evaluate the trial solutions during the optimisation. In this paper, an efficient simulation-based optimisation algorithm for large scale and complex problems is proposed. The proposed algorithm extends the cooperative coevolutionary algorithm, which optimises subproblems separately. Hence, the optimisation problem must be decomposed into subproblems that can be evaluated separately. To optimise the subproblems, the proposed algorithm allows using embedded deterministic algorithms, next to stochastic genetic algorithms, getting the flexibility of using either type. It also includes a constructive heuristic that creates good initial feasible solutions to expedite the optimisation. The extension enables solving complex, computationally expensive problems efficiently. The proposed algorithm has been applied on an automated sheet metal press line from the automotive industry. The objective is to find control parameters that maximise the line’s production rate. The results show that the proposed algorithm manages to find optimal control parameters efficiently within the practical timeframe. This is a step forward in press line optimisation since to the authors’ knowledge this is the first time a press line has been optimised efficiently in this way.
ABSTRACT This paper presents a successful off-line optimisation case study of a sheet-metal press... more ABSTRACT This paper presents a successful off-line optimisation case study of a sheet-metal press line for the automotive industry. The proposed off-line optimisation method results not only in increased production rate but also in obtaining smoother material handling motions. The off-line optimisation method contains two parts. The first part is a parameter study, combining factorial design with production line simulation including emulated control systems and resulting in a simplified performance model based on multiple linear regression. The second part is an optimisation on the simplified performance model using the Nelder Mead simplex method aiming at production optimisation. Due to the complexity in a modern production line, it is not possible to create a useful model representation of the entire control system code/function. Therefore, the proposed optimisation method is based on using production line simulation including emulated control systems with the real IEC 61131-3 control code. Our approach in this work is a parameter variation strategy for optimisation, i.e. adjusting the control system parameters in the same way as the line operators do, and not changing any control strategy.
Robotics and Computer-Integrated Manufacturing, 2015
ABSTRACT In a truly flexible manufacturing system the description of the control strategy must be... more ABSTRACT In a truly flexible manufacturing system the description of the control strategy must be updated every day. Hence, a new way to handle changes in the environment down to control system deployment and production is required. This paper presents a novel approach, based on P-SOP, to handle multi-agent based control and verification. The P-SOP approach addresses flexibility, robustness and deployment in the best possible manner with the least waste of time and effort. P-SOP includes a description language where the control strategy based on actual circumstances easily can be defined. Based on the description multi-agents, to control the manufacturing, are automatically generated. An industrial advantage is that the multi-agent generator creates IEC 61131-3 PLC code that can be executed on standard PLC's. This feature eliminates the need for experts in PLC programming and reduce deployment time to become more efficient. Hence, this flexibility enables small series down to one off production in a competitive way. With multi-agent control it is also possible to handle rebalancing due to market changes, scheduling of available humans, introduction of new part types, and rerouting due to a machine break down or planned service. The generated agents are not optimised for a final solution with specific timings. All decisions are made on-line and the generated solution adapts to the circumstances that arise. With the P-SOP multi-agents it is easy to manually remove or introduce parts to the manufacturing cell without disturbing the system, e.g. for manually random inspections, removal of parts due to restart. The formulated description language and the multi-agent generator has been successfully tested and evaluated in an industrial environment.
2009 IEEE Conference on Emerging Technologies & Factory Automation, 2009
ABSTRACT In this paper a concept for Virtual Machine Vision is proposed using a commercial Comput... more ABSTRACT In this paper a concept for Virtual Machine Vision is proposed using a commercial Computer Aided Robotics software called RobCad. The system utilizes ideal virtual cameras and lights for the simulation of a real vision system. Sensory data is sent to a vision software for data analysis. The Virtual Machine Vision together with the simulation model can be used to offline programming of a vision system. Experiments have been performed by capturing images of a test piece both in the virtual environment and in a physical experimental rig. To evaluate the concept, image analysis has been performed on these images using the same vision software. The results from the vision analysis of both the virtual and the real images are compared and show good agreement. The proposed system seems to be very promising and further development is ongoing.
2014 IEEE International Conference on Automation Science and Engineering (CASE), 2014
ABSTRACT Simulation-based optimisation carries the burden of computationally expensive fitness ca... more ABSTRACT Simulation-based optimisation carries the burden of computationally expensive fitness calculations. It is very often used to tackle large-scale optimisation problems with a relatively high level of complexity. Therefore, it is of interest to have optimisation techniques dedicated to simulation-based optimisation. This paper proposes a simulation-based optimisation approach, called Constructive Cooperative Coevolutionary (C3) search procedure, to optimise the control of interacting production stations. An optimisation algorithm is embedded in the C3 search procedure to optimise subproblems separately. It includes a novel constructive heuristic that creates a feasible solution for the considered problem efficiently. It also incorporates an extended version of the existing cooperative coevolutionary method that can handle large-scale optimisation problems. Furthermore, this paper presents a case study considering a sheet metal press line as an example of interacting production stations. In this case study, the performance of the proposed C3 search procedure is evaluated and compared with other optimisation algorithms. This shows that the C3 search procedure is able to successfully optimise the press line within a given number of fitness calculations, outperforming existing algorithms. Also, it is shown that C3 can be embedded with either stochastic or deterministic optimisation algorithms, without sacrificing performance.
2009 IEEE Conference on Emerging Technologies & Factory Automation, 2009
ABSTRACT An off-line optimization of a sheet-metal press line is performed with improved producti... more ABSTRACT An off-line optimization of a sheet-metal press line is performed with improved production performances, both in terms of increased production rate and smoother robot motions. Smooth motions prevent the sheet-metal to slide out of position or even fall out, thereby causing lengthy down times. The simulation based method use a process optimizer connected to a time synchronized virtual manufacturing model including real industrial control systems, e.g. PLC. The key benefit herein is that all tuned control system parameters can be directly downloaded to the real press line without any post-processing or transformations. The challenge to find suitable optimization algorithms for a press line that handle highly nonlinear, discontinuous functions; considerable number of parameters; and long evaluation times is reached.
The International Journal of Advanced Manufacturing Technology
Industries of low-batches or one-off manufacturing aim for automation that is competitive enough ... more Industries of low-batches or one-off manufacturing aim for automation that is competitive enough to adapt to new or modified products daily through in-house knowledge that focuses on manufacturing processes and not on machine function programming. To solve this, a complete set of actions that utilize seamless data transfer from product design in CAD to a Plug & Produce automation concept is proposed together with a generic structure for the integration of standard industrial robots into agents. This structure enables agents to handle their local reference coordinate systems and locations relative to a global perspective. Seamless utilization of data from product designs to Plug & Produce will simplify and shorten the time of digital development through concurrently usable text-based and graphical configuration tools of a configurable multi-agent system. Needed data extracts directly from the product design as requirements of operational goals. Extraction of data from the product des...
Industries of low batches or one-off manufacturing aim for automation that is competitive enough ... more Industries of low batches or one-off manufacturing aim for automation that is competitive enough to adapt to new or modified products daily through in-house knowledge. A knowledge that focuses on manufacturing processes, not on machine function programming. To solve this a complete set of actions that utilize seamless data transfer from product design to a Plug & Produce automation concept is proposed together with a generic structure for the integration of standard industrial robots into agents. This structure enables agents to handle their local reference coordinate systems and locations relative to a global perspective. Seamless utilization of data from product designs to Plug & Produce will simplify and shorten the time of digital development through concurrently usable text-based and graphical configuration tools of a configurable multi-agent system. Needed data extracts directly from the product design as requirements of operational goals. Extraction of data from the product d...
Job shop scheduling problems are computationally complex combinatorial optimization problems. Gen... more Job shop scheduling problems are computationally complex combinatorial optimization problems. Genetic algorithms have been used in various forms and in combination with other algorithms to solve job shop scheduling problems. A partially flexible job shop with precedence constraints increases this complex behaviour. There are two main parts to optimizing ajob shop, the routing and the scheduling. The objective here is to get consistent optimal makespan using a genetic algorithm. This paper firstly, presents a simulation approach for the considered partially flexible job shop scheduling problem. Which take into account the precedence constraints and reduce situations of deadlock. To solve the partially flexible job shop scheduling problem a genetic algorithm was used and improved. It utilise a genetic crossovers for routing and a new random shuffle feature is introduced for the scheduling. The computational results have shown that the algorithm performs well in terms of finding a cons...
Journal of Advanced Manufacturing Systems, 2002
... and evaluated. The main tool is a virtual press station, developed by the first author, consi... more ... and evaluated. The main tool is a virtual press station, developed by the first author, consisting of an emulated control system for a feeder/extractor robot which communicates with 3D-simulated production equipment. Moreover ...
Proceedings of 2012 IEEE 17th International Conference on Emerging Technologies & Factory Automation (ETFA 2012), 2012
In order to perform efficient geometrical simulation and virtual commissioning in stamping, three... more In order to perform efficient geometrical simulation and virtual commissioning in stamping, three fields are investigated namely: simulation building time, collision detection time and optimization time. Hence, reducing time is the main theme of this paper. To reduce simulation building time and optimization time, an efficient stamping simulation model is built and tested. Collision detection time is examined by a relative motion method based on 3D to 2D geometrical collision detection. The presented results mean that simulation and virtual commissioning can be performed at least ten times faster compared to standard approaches.
Proceedings of the 24th International Conference on Flexible Automation & Intelligent Manufacturing, 2014
Simulation-based optimisation often considers computationally expensive problems. Successfully op... more Simulation-based optimisation often considers computationally expensive problems. Successfully optimising such large scale and complex problems within a practical timeframe is a challenging task. Optimisation techniques to fulfil this need to be developed. A technique to address this involves decomposing the considered problem into smaller subproblems. These subproblems are then optimised separately. In this paper, an efficient algorithm for simulation-based optimisation is proposed. The proposed algorithm extends the cooperative co-evolutionary algorithm, which optimises subproblems separately. To optimise the subproblems, the proposed algorithm enables using a deterministic algorithm, next to stochastic genetic algorithms, getting the flexibility of using either type. It also includes a constructive heuristic that creates good initial feasible solutions to reduce the number of fitness calculations. The extension enables solving complex, computationally expensive problems efficiently. The proposed algorithm has been applied on automated sheet metal press lines from the automotive industry. This is a highly complex optimisation problem due to its non-linearity and high dimensionality. The optimisation problem is to find control parameters that maximises the line's production rate. These control parameters determine velocities, time constants, and cam values for critical interactions between components. A simulation model is used for the fitness calculation during the optimisation. The results show that the proposed algorithm manages to solve the press line optimisation problem efficiently. This is a step forward in press line optimisation since this is to the authors' knowledge the first time a press line has been optimised efficiently in this way.
Advances in Intelligent and Soft Computing, 2010
Abstract. Computer Aided Process Planning has received more attention re-cently due to considerab... more Abstract. Computer Aided Process Planning has received more attention re-cently due to considerable progress in the aspects of both technology and the-ory. Beside the traditional trends and efforts to integrate the product design and process planning activities usually referred to as ...
ABSTRACT Determining the control parameters for sheet metal press lines is a large scale and comp... more ABSTRACT Determining the control parameters for sheet metal press lines is a large scale and complex optimisation problem. These control parameters determine velocities, time constants, and cam values of critical interactions between the equipment. The complexity of this problem is due to the nonlinearities and high dimensionality. Classical optimisation techniques often underperform in solving this kind of problems within a practical timeframe. Therefore, specialised techniques need to be developed for these problems. An existing approach is simulation-based optimisation, which is to use a simulation model to evaluate the trial solutions during the optimisation. In this paper, an efficient simulation-based optimisation algorithm for large scale and complex problems is proposed. The proposed algorithm extends the cooperative coevolutionary algorithm, which optimises subproblems separately. Hence, the optimisation problem must be decomposed into subproblems that can be evaluated separately. To optimise the subproblems, the proposed algorithm allows using embedded deterministic algorithms, next to stochastic genetic algorithms, getting the flexibility of using either type. It also includes a constructive heuristic that creates good initial feasible solutions to expedite the optimisation. The extension enables solving complex, computationally expensive problems efficiently. The proposed algorithm has been applied on an automated sheet metal press line from the automotive industry. The objective is to find control parameters that maximise the line’s production rate. The results show that the proposed algorithm manages to find optimal control parameters efficiently within the practical timeframe. This is a step forward in press line optimisation since to the authors’ knowledge this is the first time a press line has been optimised efficiently in this way.
ABSTRACT This paper presents a successful off-line optimisation case study of a sheet-metal press... more ABSTRACT This paper presents a successful off-line optimisation case study of a sheet-metal press line for the automotive industry. The proposed off-line optimisation method results not only in increased production rate but also in obtaining smoother material handling motions. The off-line optimisation method contains two parts. The first part is a parameter study, combining factorial design with production line simulation including emulated control systems and resulting in a simplified performance model based on multiple linear regression. The second part is an optimisation on the simplified performance model using the Nelder Mead simplex method aiming at production optimisation. Due to the complexity in a modern production line, it is not possible to create a useful model representation of the entire control system code/function. Therefore, the proposed optimisation method is based on using production line simulation including emulated control systems with the real IEC 61131-3 control code. Our approach in this work is a parameter variation strategy for optimisation, i.e. adjusting the control system parameters in the same way as the line operators do, and not changing any control strategy.
Robotics and Computer-Integrated Manufacturing, 2015
ABSTRACT In a truly flexible manufacturing system the description of the control strategy must be... more ABSTRACT In a truly flexible manufacturing system the description of the control strategy must be updated every day. Hence, a new way to handle changes in the environment down to control system deployment and production is required. This paper presents a novel approach, based on P-SOP, to handle multi-agent based control and verification. The P-SOP approach addresses flexibility, robustness and deployment in the best possible manner with the least waste of time and effort. P-SOP includes a description language where the control strategy based on actual circumstances easily can be defined. Based on the description multi-agents, to control the manufacturing, are automatically generated. An industrial advantage is that the multi-agent generator creates IEC 61131-3 PLC code that can be executed on standard PLC's. This feature eliminates the need for experts in PLC programming and reduce deployment time to become more efficient. Hence, this flexibility enables small series down to one off production in a competitive way. With multi-agent control it is also possible to handle rebalancing due to market changes, scheduling of available humans, introduction of new part types, and rerouting due to a machine break down or planned service. The generated agents are not optimised for a final solution with specific timings. All decisions are made on-line and the generated solution adapts to the circumstances that arise. With the P-SOP multi-agents it is easy to manually remove or introduce parts to the manufacturing cell without disturbing the system, e.g. for manually random inspections, removal of parts due to restart. The formulated description language and the multi-agent generator has been successfully tested and evaluated in an industrial environment.
2009 IEEE Conference on Emerging Technologies & Factory Automation, 2009
ABSTRACT In this paper a concept for Virtual Machine Vision is proposed using a commercial Comput... more ABSTRACT In this paper a concept for Virtual Machine Vision is proposed using a commercial Computer Aided Robotics software called RobCad. The system utilizes ideal virtual cameras and lights for the simulation of a real vision system. Sensory data is sent to a vision software for data analysis. The Virtual Machine Vision together with the simulation model can be used to offline programming of a vision system. Experiments have been performed by capturing images of a test piece both in the virtual environment and in a physical experimental rig. To evaluate the concept, image analysis has been performed on these images using the same vision software. The results from the vision analysis of both the virtual and the real images are compared and show good agreement. The proposed system seems to be very promising and further development is ongoing.
2014 IEEE International Conference on Automation Science and Engineering (CASE), 2014
ABSTRACT Simulation-based optimisation carries the burden of computationally expensive fitness ca... more ABSTRACT Simulation-based optimisation carries the burden of computationally expensive fitness calculations. It is very often used to tackle large-scale optimisation problems with a relatively high level of complexity. Therefore, it is of interest to have optimisation techniques dedicated to simulation-based optimisation. This paper proposes a simulation-based optimisation approach, called Constructive Cooperative Coevolutionary (C3) search procedure, to optimise the control of interacting production stations. An optimisation algorithm is embedded in the C3 search procedure to optimise subproblems separately. It includes a novel constructive heuristic that creates a feasible solution for the considered problem efficiently. It also incorporates an extended version of the existing cooperative coevolutionary method that can handle large-scale optimisation problems. Furthermore, this paper presents a case study considering a sheet metal press line as an example of interacting production stations. In this case study, the performance of the proposed C3 search procedure is evaluated and compared with other optimisation algorithms. This shows that the C3 search procedure is able to successfully optimise the press line within a given number of fitness calculations, outperforming existing algorithms. Also, it is shown that C3 can be embedded with either stochastic or deterministic optimisation algorithms, without sacrificing performance.
2009 IEEE Conference on Emerging Technologies & Factory Automation, 2009
ABSTRACT An off-line optimization of a sheet-metal press line is performed with improved producti... more ABSTRACT An off-line optimization of a sheet-metal press line is performed with improved production performances, both in terms of increased production rate and smoother robot motions. Smooth motions prevent the sheet-metal to slide out of position or even fall out, thereby causing lengthy down times. The simulation based method use a process optimizer connected to a time synchronized virtual manufacturing model including real industrial control systems, e.g. PLC. The key benefit herein is that all tuned control system parameters can be directly downloaded to the real press line without any post-processing or transformations. The challenge to find suitable optimization algorithms for a press line that handle highly nonlinear, discontinuous functions; considerable number of parameters; and long evaluation times is reached.