K. Manabe

Data fusion of distributed AE sensors for the detection of friction sources during press forming

Abstract The authors proposed a friction source detection system using multi-acoustic emission (AE) sensors and data fusion system in this study. In this work, three AE sensors are positioned at the metal forming tool for identification of the location of friction sources. The AE signals were processed by FFT and discrete wavelet transform (DWT), in order to reduce the noise and extract features of signals due to friction. The Fujimori method was employed to measure the arrival time from AE source to the distributed sensors. Furthermore, a database was constructed and employed to estimate the source location. The detection system was applied for measuring the friction during deep drawing of sheet metal. Several experiments were carried out by using the friction detection system and the results show that the location of friction source can be successfully detected by using the multi-AE sensors system even if location of friction resource on interface of the die and the workpiece is invisible.

Artificial intelligence identification of process parameters and adaptive control system for deep-drawing process

Abstract A new in-process identification method of material properties and lubrication condition in the deep-drawing process of anisotropic sheet metals is proposed and applied to the adaptive process control of the blank holding force (BHF). The method is based on a combination model of artificial neural network (ANN) and elastoplastic theory. Three delegated plastic deformation properties, i.e. n value, F value and plastic anisotropic coefficient r, were identified using the measured process information at the beginning of the process by means of ANN. The friction coefficient μ and the optimal BHF control path were then calculated from the theoretical model. Furthermore, the friction coefficient was monitored during the entire process, and a closed-loop control was applied to modify the BHF path corresponding to the frictional variation. Experimental results show that the artificial intelligence (AI) control system can cover a wide range of both materials and influential parameters, such as friction and ambient temperature automatically. It is confirmed that the newly developed system is a valid alternative for the quick responsible control system with high flexibility.

Development of real-time process control system for precision and flexible V-bending with an on-line database

Abstract An advanced process control system based on expertise is developed for the V-bending process. An experimental database and a fuzzy inference model are employed as the expertise for real-time process control. The punch force-stroke curves and other process parameters stored in the database are retrieved in-process and evaluated by the fuzzy model for precise prediction control. Also, a learning process, which is similar to the way craftsmen acquire knowledge, is applied for database and fuzzy model improvements in order to cope with new materials. Several experiments were carried out to confirm the validity of the developed control system. The results show that the process control is capable of coping with differences in the material properties and other forming parameters automatically even if the operator has no information on the material properties and the process.

Intelligent design architecture for process control of deep-drawing

A concept of design architecture with a database for an intelligent sheet metal forming system was proposed to enable designing of a process control system without experts who are skilled and experienced in the forming process. In this study, the proposed architecture was applied to the variable blank holding force control technique for circular-cup deep-drawing. The system is available for three objective functions which are typical process requirements, cup wall uniformity, cup height improvement and energy saving. The availability of this design architecture is confirmed by experiments on aluminum alloy sheets.

Intelligent control scheme of divided blank holder for square-cup deep-drawing process

To extend the potentially applicable process of the database- and Finite-Element-Method (FEM)-assisted system for the intelligent control (DFMASIC) of press systems for circular-cup deep-drawing, previously developed by the authors, the control system concept was applied to the divided and variable Blank-Holding-Pressure (BHP) control in the process. The basic idea proposed is that each divided controllable part, the straight part and corner part of the blank holder, may be considered and controlled individually as a circular-cup-drawing portion. For the divided and variable BHP control in the square-cup drawing, the results of the DFMASIC system show an enhancement of more than 40% for the maximum drawn cup height, compared with experimental results. Consequently, it is verified that the DFMASIC system is applicable to the square-cup drawing process as well as for circular-cup drawing.

Adaptability to frictional change of fuzzy adaptive blank holder control for deep drawing

Validity of the fuzzy adaptive variable blank holder force (BHF) control deep drawing system to frictional change in the process has been studied for steel sheet. The circular-cup deep-drawing experiment has been carried out using steel sheet (SPCD) with high anisotropy (r=1.57). To change the lubrication condition, the partial lubrication method was adopted using fluorine lubricant. The friction coefficient was evaluated by the plastic deformation model. It is confirmed for steel sheet that the BHF was properly controlled corresponding to friction change at the flange part.

Development of Fuzzy Inference System With Learning Algorithm for Determining Press-Forming Conditions

A fuzzy inference system with an associated database was developed in order to determine the forming conditions required to obtain the objective height of a drawn cup using magnesium alloy sheet. The results from both experimental and finite element analyses (FEA) have been used for constructing the database which contains information on the forming conditions used and the resultant height of the drawn cup. A case study utilizing circular-cup deep-drawing was used to test the inference system, whereby experimental results for the height of the drawn cup using specific forming conditions (temperature, blank size and blank holding force) agreed with those predicted using the inference system. This confirms the validity of the inference system. Furthermore, a learning method was developed in order to improve the system using further experimental results, the FEA results and a fuzzy algorithm. The experimental analysis showed good agreement between the required objective height and the experimentally measured height of the drawn cup.© 2010 ASME