会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 2. 发明专利
    • METHOD AND APPARATUS FOR AUTOMATIC CONSTRUCTION OF ABNORMALITY DIAGNOSING ALGORITHM
    • JPH05203544A
    • 1993-08-10
    • JP1214492
    • 1992-01-27
    • HITACHI LTDHITACHI ENG SERVICE
    • SATO YOSHIYUKIYONEYAMA TAKAOSATO KAZUYATAGUCHI YUJI
    • G01M99/00G01N29/14G05B23/00
    • PURPOSE:To construct an abnormality diagnosing algorithm for performing the highly accurate diagnosis by operating the specified value for the normal average value again as the virtual measured value of each measured value of a normal product when the average value of the amounts of deviations of the normal product is lower than a reference value, and forcibly making the computed value of the amount of deviation zero when the individual measured value is smaller with respect to the average value of the normal product. CONSTITUTION:Analysis objects 10 are classfied into the objects having the normal property and the objects having the abnormal property. Values with regard to various kinds of diagnostic parameters are detected with a detector 12. The output signal of the detector 12 is sent into a computor 18 through an amplifier 14 and an A/D converter 16. In the computor 18, the values with regard to a plurality of the diagnostic parameters undergo statistic processing based on the detected signal of the detector 12, and the diagnostic level is determined based on the result of the processing. When an algorithm for discriminating the normal and abnormal products is constructed, the specified value with respect to the average value of the normal product is operated again as the virtual measured value of each measured value of the normal product, when the average value of the amounts of deviations of the products of the normal group is less than a reference value. The computed value of the amount of deviation is forcibly made to be zero when the individual value is smaller with respect to the average value of the normal product.
    • 7. 发明专利
    • METHOD AND DEVICE FOR AUTOMATIC CONSTRUCTION OF LEARNING TYPE ABNORMALITY DIAGNOSIS ALGORITHM
    • JPH0743352A
    • 1995-02-14
    • JP23898892
    • 1992-08-14
    • HITACHI LTDHITACHI ENG SERVICE
    • YONEYAMA TAKAOSATO KAZUYATAGUCHI YUJI
    • G01N29/14
    • PURPOSE:To conduct an accurate diagnosis for each type of diagnosis target by updating the combination of parameters for diagnosis and a judgment level until the judgment correct answer rate of conforming and non-conforming articles reach a target correct answer rate. CONSTITUTION:Even if an algorithm is constructed by an algorithm construction method 50, the judgment correct answer rata of conforming and non-conforming articles may not always indicate the maximum value according to selected four effective parameters and a diagnosis level. Therefore, a correct answer rate is calculated 200 first. Then, when the correct answer rate does not reach a target correct answer rate, it is instructed so that a combination of effective parameters which are different from those selected by the method 50 can be selected and also the judgment level is automatically changed to increase the correct answer rate. It is repeated until the target correct answer rate can be obtained. Further, when the target correct answer rate cannot be obtained by the processing, the combination of parameters and the judgment level are stored constantly by a maximum correct answer rate calculation means 203, thus determining 202 the combination of each parameter indicating the maximum correct answer rate to be the abnormality diagnosis algorithm.
    • 10. 发明专利
    • JPH05312682A
    • 1993-11-22
    • JP12170192
    • 1992-05-14
    • HITACHI LTDHITACHI ENG SERVICE
    • YONEYAMA TAKAOSATO KAZUYASATO MINORUNAKAZAWA TOSHIOSATO YOSHIYUKI
    • G01M13/04G01N29/14
    • PURPOSE:To non-destructively and correctly diagnose damage of a slide part by placing an AE sensor and determining a defective part in the slide part when an output level of an AE signal exceeds a predetermined level. CONSTITUTION:An AE sensor 2 is attached to a scroll compressor 1, and after an AH signal is amplified by an amplifier 3, the AE signal is detected by-a wave detecting circuit 4. Then an AE signal characteristic extracting part 5 is used to analyze the AE detected signal. Further a determining part 6 compares the analyzed and processed data value with a reference value to diagnose whether it is good or defective and diagnosis results are displayed on a display part 7. In this case the AE signal characteristic extracting part 5 calculates an average value of the detected AE signal, rotation primary frequency component and an overall value. The determining part 6 compares the average AE value with a first reference value K1, and one lower than the reference value is determined to be good while the rotation primary frequency component is compared with a second reference value K2 for one exceeding the reference value. If the second reference value K2 is exceeded, the part is determined to be defective, while the overall value is compared with a third reference value K3 for one not exceeding the second reference value. If the third reference value is not exceeded, it is determined to be good, otherwise it is determined to be defective.