Diagnosis and maintenance of Relay Ladder Logic programs and PLC Ladder Logic Diagrams using artificial neural networks
Introduction
In July 2006 and June 2010, the American and Egyptian PI’s started the activities of the research grant number INF9-001-007 respectively. The goal of this research project The main objective of the project is to develop an intelligent and fault debugger scheme for discrete manufacturing system.
Project objectives
The vision of this research is introducing new solutions to the supervisory control problems of event and time manufacturing systems using artificial intelligence
The main objective of the project is to develop an intelligent and fault debugger scheme for discrete manufacturing system that has the following characteristics:
All these features keep the discrete event control system to be healthy, reducing the regular maintenance cost, and preserving the environment from the effect of its faults. In addition, this will enhance the life of discrete event control equipment. This step represents one of the best methodologies for technology transfer to the Egyptian industry especially for manufacturing sector. This will enhance the status of production lines that utilize hydraulic or pneumatic systems as their actuators. Moreover, we look forward to preparing a complete package including software and hardware components (like sensors, some circuits, microcontroller, etc.). Whether it is a production line or a heavy equipment machine, this package can be easily used for improving the control performance of manufacturing systems of pneumatic and hydraulic equipment which operates in a sequence manner.
The PIs focused their activities in the third year (2007-2008) on introducing a new design approach to control the time and event manufacturing systems. The outcome of these activities is proposing a new design algorithm based on the temporal neural network to supervisory control the time and event based manufacturing systems.
Design and methodology
Recurrent neural network has one or more feedback loops that can originate from the hidden or output neurons not only during the current time step, but also during some number of previous time steps. This makes an RNN specifically suitable for temporal and sequential processing. The RNN has been applied as a sequential controller for flexible systems.
In our project, a controller of manufacturing systems using RRN has been established. The proposed controller considers the fault detection and the controller reconfiguration simultaneously. In addition, the proposed reconfigurable controller is able to work in real time and improves the control system performance. While drastic failures such as sensor failure are considered, partial failures such as drift in sensors or small changes in dynamics of actuators are not considered here as their behavior is not reflected by manufacturing system models.
Prototypes or Products
A new course coded MCT 620 and entitled “Supervisory Control Design of Discrete Event System” has been developed and added to the latest upgraded curriculum of ASU Mechatronics postgraduate program. The content of this course is related mainly to our project and is prepared by the Egyptian PI. The basic courses of Mechatronics Engineering Department (13) at ASU are listed in table below:
Basic Courses - 12 credit hours:
In July 2006 and June 2010, the American and Egyptian PI’s started the activities of the research grant number INF9-001-007 respectively. The goal of this research project The main objective of the project is to develop an intelligent and fault debugger scheme for discrete manufacturing system.
Project objectives
The vision of this research is introducing new solutions to the supervisory control problems of event and time manufacturing systems using artificial intelligence
The main objective of the project is to develop an intelligent and fault debugger scheme for discrete manufacturing system that has the following characteristics:
- Preventing and isolating faults of the control scheme,
- Enhancing the performance of the rehabilitated discrete even system by introducing the LLD based control languages with its new diagnostic package.
- Constantly improving their operation.
All these features keep the discrete event control system to be healthy, reducing the regular maintenance cost, and preserving the environment from the effect of its faults. In addition, this will enhance the life of discrete event control equipment. This step represents one of the best methodologies for technology transfer to the Egyptian industry especially for manufacturing sector. This will enhance the status of production lines that utilize hydraulic or pneumatic systems as their actuators. Moreover, we look forward to preparing a complete package including software and hardware components (like sensors, some circuits, microcontroller, etc.). Whether it is a production line or a heavy equipment machine, this package can be easily used for improving the control performance of manufacturing systems of pneumatic and hydraulic equipment which operates in a sequence manner.
The PIs focused their activities in the third year (2007-2008) on introducing a new design approach to control the time and event manufacturing systems. The outcome of these activities is proposing a new design algorithm based on the temporal neural network to supervisory control the time and event based manufacturing systems.
Design and methodology
Recurrent neural network has one or more feedback loops that can originate from the hidden or output neurons not only during the current time step, but also during some number of previous time steps. This makes an RNN specifically suitable for temporal and sequential processing. The RNN has been applied as a sequential controller for flexible systems.
In our project, a controller of manufacturing systems using RRN has been established. The proposed controller considers the fault detection and the controller reconfiguration simultaneously. In addition, the proposed reconfigurable controller is able to work in real time and improves the control system performance. While drastic failures such as sensor failure are considered, partial failures such as drift in sensors or small changes in dynamics of actuators are not considered here as their behavior is not reflected by manufacturing system models.
Prototypes or Products
- An implemented laboratory discrete event manufacturing system with its supervisory controller.
- A software package to design and train the ANN based supervisory controller.
- A new course coded MCT 620 and entitled “Supervisory Control Design of Discrete Event System” has been developed and added to the latest upgraded curriculum of ASU Mechatronics postgraduate program. Details are given inside this report.
- A new hardware and software package of dSPACE has been added to the mechatronics lab at Ain Shams University which is supervised and directed by the Egyptian PI. Details are explained in the section of Purchased Equipment.
- The American PI has been invited to present a lecture about the discrete event system control
- Abdelhameed, M.M.; Darabi, H., “Diagnosis and debugging of programmable logic controller control programs by neural networks” IEEE International Conference on Automation Science and Engineering, 2005, 1-2 Aug. 2005 Page(s):313 – 318.
- Abdelhameed, M.M.; Darabi, H., “Neural network based design of fault-tolerant controllers for automated sequential manufacturing systems” Submitted to publication in Mechatronics at October, 2007- Some modifications are required.
- Abdelhameed, M.M.; Darabi, H., “Temporal neural network supervisory control for time and event driven manufacturing systems” Under preparation.
A new course coded MCT 620 and entitled “Supervisory Control Design of Discrete Event System” has been developed and added to the latest upgraded curriculum of ASU Mechatronics postgraduate program. The content of this course is related mainly to our project and is prepared by the Egyptian PI. The basic courses of Mechatronics Engineering Department (13) at ASU are listed in table below:
Basic Courses - 12 credit hours:
|
Marks |
Credit Hours |
Course |
No. |
Code |
No. |
|
100 |
3 |
Advanced Engineering Mathematics (2) |
675 |
PHM |
1 |
|
100 |
3 |
Engineering Numerical Analysis |
676 |
PHM |
2 |
|
100 |
3 |
Design of Mechatronics Systems |
610 |
MCT |
3 |
|
100 |
3 |
|
|
MCT* |
4 |
|
|
12 |
Sum |
|||
* Student should choose one of the following courses in addition to the first three courses to study 12 credit hours:
MCT 620: Supervisory Control Design of Discrete Event System
MCT 621: Motion Control and Servo Systems
MCT 622: Mechatronics Systems Modelling and Identification
Conclusions
1- The PI’s succeeded to achieve the goals of the project.
2- The Egyptian PI has gained many managerial skills in managing the activities of this project.
3- Some adjustments in the project budget have been done to full fill all the activities of the project.
4- Flexibility of the US-Egypt Science and Technology Joint Fund committee helped the PI’s to achieve the goals of the project activities.
5- Discreet even systems- Technology Transfer: One of the high technologies has been transferred to the Egyptian universities as a research area and to industry through the private manufacturing sector to enhance the machining of mirrors (metal and scanner mirrors), lenses, various optical components.
MCT 620: Supervisory Control Design of Discrete Event System
MCT 621: Motion Control and Servo Systems
MCT 622: Mechatronics Systems Modelling and Identification
Conclusions
1- The PI’s succeeded to achieve the goals of the project.
2- The Egyptian PI has gained many managerial skills in managing the activities of this project.
3- Some adjustments in the project budget have been done to full fill all the activities of the project.
4- Flexibility of the US-Egypt Science and Technology Joint Fund committee helped the PI’s to achieve the goals of the project activities.
5- Discreet even systems- Technology Transfer: One of the high technologies has been transferred to the Egyptian universities as a research area and to industry through the private manufacturing sector to enhance the machining of mirrors (metal and scanner mirrors), lenses, various optical components.