Abstract: In an optical fiber drawing device of the present invention, as an operational parameters corresponding to the real drawing speed in the duration from the moment of starting the operation of the optical fiber drawing device till the moment of reaching the target drawing speed are extracted from operational parameters stored in a operational parameter storage unit and applied to the optical fiber drawing device. The operational parameters include a base preform feeding speed, a target variation of real drawing speed, a gain multiplied by the speed element related with the variation of the real drawing speed, an operation cycle, a drawing furnace temperature, a coating resin pressure to coat the optical fiber, and flow of cooling gaseous helium, etc.

Abstract: In an optical fiber drawing device of the present invention, as an operational parameters corresponding to the real drawing speed in the duration from the moment of starting the operation of the optical fiber drawing device till the moment of reaching the target drawing speed are extracted from operational parameters stored in a operational parameter storage unit and applied to the optical fiber drawing device. The operational parameters include a base preform feeding speed, a target variation of real drawing speed, a gain multiplied by the speed element related with the variation of the real drawing speed, an operation cycle, a drawing furnace temperature, a coating resin pressure to coat the optical fiber, and flow of cooling gaseous helium, etc.

Abstract: The value of a variable reflecting the operating condition of mechanical equipment is continuously detected as operation condition data. A long period average value of the operating condition data detected in long reference period, and a short period average value of data detected in a short analysis period ending in the present time are calculated. Just how much the short period average value differs from the long period average value is adopted as an index of abnormality and using this index, it is possible to ascertain abnormal operating conditions regardless of the absolute value of the variable detected. The above short period average value, long period average value and standard deviation thereof are applied to a first fuzzy membership function for calculating the degree of stability of the mechanical equipment therefrom.

Abstract: The present invention provides a roller mill control method capable of automatically controlling a roller mill which is comparatively difficult to operate for operation in an optimum operating condition, and a roller mill controller for carrying out the roller mill control method. The roller mill controller stores beforehand a roller mill characteristic model representing the relation between changes in the values of the operating condition variables detected by a detecting unit (2) and operating parameters, restrictions on the values of the operating condition variables and the operating parameters, an index of performance for evaluating the values of the operating condition variables and the operating parameters.