-
公开(公告)号:JP2001070846A
公开(公告)日:2001-03-21
申请号:JP2000219097
申请日:2000-07-19
Applicant: ABB RESEARCH LTD
Inventor: EICKMEYER DIETMAR , BOERNER GUNTER
Abstract: PROBLEM TO BE SOLVED: To cope with the precise distribution of coating material thickness by fixing a spray parameter to be set from a desired spray pattern to be attained by utilizing an artificial neural net work learning by using practically measured data. SOLUTION: A coating film is formed by the movement of a spray device in the horizontal or vertical direction and the distribution of the coating film thickness is measured. The profile curve of the distribution of the coating film thickness is shown is the form of 200 individually measured values and a variable to be inputted to the artificial neural net work is expressed by using the measured value. As the input parameter to the artificial neural net work, amplitude is used together with the Fourier analysis of the profile of the coating material thickness distribution, and the spray parameter is fixed from the desired spray pattern to be attained as the output parameter such as the quantity of a coating material, the directing-air quantity and high voltage.
-
公开(公告)号:JP2001062358A
公开(公告)日:2001-03-13
申请号:JP2000229258
申请日:2000-07-28
Applicant: ABB RESEARCH LTD
Inventor: EICKMEYER DIETMAR , BOERNER GUNTER
Abstract: PROBLEM TO BE SOLVED: To determine the thickness distribution of the coating layer formed by an electrostatic coating material spraying apparatus. SOLUTION: A phenomenological and mathematical model of a semi-static three-dimensional spraying pattern is set and used, and specified parameters as fixed input parameters are directly inputted in the phenomenological model. Practical physical input parameters whose effects on the spraying results are not known precisely are supplied to an artificial neural network in which simulation is previously carried out using practical input data. The artificial neural network carries out conversion of the added input parameters into model input parameters to be supplied to the phenomenological model and the spraying patterns formed by the phenomenological model are accumulated in a further functional unit as functions of the movement data of a spraying apparatus 2 included in the input parameters for forming the entire coating layer and the output of the thickness distribution of the coating layer is obtained.
-
公开(公告)号:DE19936148A1
公开(公告)日:2001-02-01
申请号:DE19936148
申请日:1999-07-31
Applicant: ABB RESEARCH LTD
Inventor: BOERNER GUNTER , EICKMEYER DIETMAR
IPC: B05D1/04 , B05B5/025 , B05B12/00 , G05B13/02 , G06F9/44 , G06N3/00 , G06N7/02 , B05B5/04 , G05B13/00 , G05B17/00
Abstract: Electrostatically sprayed paint parameters suitable as input values for a paint spraying device are determined via at least one artificial neuronal network where each parameter is available as out output of the network. Initially, a learning phase involves inputting a number of realistic measured values into the network or into several networks to provide a paint thickness distribution and the associated realistic spray parameters. User phase input values are applied to the network or networks and result from an analysis of the paint thickness distribution of a specified spray result.
-
公开(公告)号:DE19920654A1
公开(公告)日:2000-11-09
申请号:DE19920654
申请日:1999-05-05
Applicant: ABB RESEARCH LTD
Inventor: EICKMEYER DIETMAR , FANTANA NICOLAIE L , BOEHME HELMUT
Abstract: The method involves automatically deriving the conductor impedance threshold value from a predefined impedance threshold value depending on a determined degree of dependability of the current determined conductor impedance value.
-
公开(公告)号:DE19936146A1
公开(公告)日:2001-02-01
申请号:DE19936146
申请日:1999-07-31
Applicant: ABB RESEARCH LTD
Inventor: EICKMEYER DIETMAR , BOERNER GUNTER
IPC: B05D1/04 , B05B5/00 , B05B5/025 , B05B12/00 , G05B17/02 , G05B23/02 , B05B5/04 , G05B13/00 , G05B17/00
Abstract: Determining the distribution of sprayed paint thickness of paint layers formed according to a set spraying parameter using an electrostatic spraying device requires use of a phenomenological mathematical model of a three-dimensional spray image with input of specific parameters such as the rotational angle of the spray device electrodes together with the latter's rate of movement. In addition, actual physical input parameters such as color volumes, pilot air data and high voltage value are supplied to an artificial neuronal network, although their influence on the result of paint spraying is not precisely known. The spray images from the phenomenological model are integrated to a total paint layer in a further functional unit, independent of the movement data.
-
公开(公告)号:DE19933811A1
公开(公告)日:2001-02-01
申请号:DE19933811
申请日:1999-07-20
Applicant: ABB RESEARCH LTD
Inventor: EICKMEYER DIETMAR , FANTANA NICOLAIE LAURENTIU , BOEHME HELMUT , DIWES ANDREAS
Abstract: During medium- and high-voltage power supply the fluctuating demand of consumer loads leads to differing power flows in the system which need to be influenced by the tap-changer operation in order to provide optimal operation, and with the tap-changer no longer being sufficiently capable of controlling the power flows under the prevailing conditions of the deregulated market, the secondary voltage is controlled by means of a controller (RE), a controlled converter (UM) and a three-phase transformer (UT), leading to less mechanical wear .
-
公开(公告)号:DE19954950A1
公开(公告)日:2001-05-17
申请号:DE19954950
申请日:1999-11-16
Applicant: ABB RESEARCH LTD
Inventor: EICKMEYER DIETMAR , BOEHME HELMUT
Abstract: The method involves determining a threshold for at least one parameter, monitoring for threshold violation and using this to detect short circuits. An uncertainty range is determined for the parameter values, and values within this range are assessed for the existence of a short circuit using artificial neural networks and a conclusion regarding the existence of a short drawn by combined evaluation of the threshold and uncertainty monitoring.
-
-
-
-
-
-