Noise And Vibration Big Data Monitoring And Analysis Platform On 600MW Unit Auxiliary Equipment

With the continuous progress and development of power industry,600 MW unit has become the main unit of the power grid.Auxiliary equipment is an important part of the unit,but it is the weakest link in the state of the smart grid monitoring.therefore,to strengthen the auxiliary equipment condition monitoring is the most urgent problem needed to be solved.It is found that the monitoring data of auxiliary equipment presents large data features.However,there are still no large data applications research or application examples in the field.To solve this problem,based on some existing theoretical research of large data technology in the field of smart grid,combined with state monitoring technology,early warning and diagnosis technology,digital signal processing technology,filter technology and embedded technology,we design a noise and vibration big data monitoring and analysis platform on 600 MW unit auxiliary equipment.The platform is designed to acquire,calculate and process the massive noise and vibration monitoring data rapidly,and establish an early-warning model for the auxiliary equipment in specific operating conditions and environmental conditions to determine the potential malfunction of each equipment accurately.If potential malfunction exists,we can deal with it early to avoid accidents.For the equipment with potential malfunction,we can use large data analysis techniques to mining data,thus find rich and valuable information hidden in massive data.Firstly,the paper introduces the whole structure and design principle of the platform in detail.Secondly,the paper introduces the digital signal processing algorithms of calculating the noise frequency A,C,Z and vibration acceleration,velocity and displacement.The algorithms include IIR digital filter algorithm and FFT domain calculating algorithm.What's more,the MATLAB simulation proves that the deviation is in line with the national I-tolerance standards;Then,the paper introduces the principle and realization of intelligent early-warning model based on fuzzy comprehensive evaluation and Bayesian discriminant algorithm.The noise vibration data collected in the field are input into the model for analysis,and the result verifies the model's accuracy.Finally,the paper introduces the comprehensive and detailed spectrum analysis of noise signal vibration signal.For the noise spectrum analysis,an improved 1/3 octave method is proposed,in which the spectrum value of a band is expressed by the spectral effective value in a band,and the method of adding Blackman window and attenuation compensation method is adopted to solve the traditional 1/3 octave energy leakage and amplitude attenuation problems;For the vibration spectrum analysis,FFT is used to get panoramic spectrum;then,for the local frequency range you want to refine,using adaptive spectral refinement algorithm(ZOOM-FFT algorithm)based on the complex analytic band-pass filter to analyze.The algorithm sovles the problems:refinement times are limited,calculation precision is bad,the calculation amount is large.At present,the system has been applied to Huaneng Group Yueyang power plant.It achieves all of expected targets,and acquires good application effect.