Water quality monitoring was a basic work for water quality resource management, there were some problems in the current detection means for oily waste water, such as low measurement accuracy, off line detection, an so on. Based on the principle of surface plasmon resonance (SPR), a optical fiber sensor was designed, which could monitor the interaction of biological molecules online and real-time, with label-free, without separation and purification. It was a non-contact detection technology and it was easy to implement remote monitoring. The designed optical fiber SPR sensor system was tested through oily waste water with different concentrations. The experimental results showed that the optical fiber SPR sensing system could measure the oil in waste water with high accuracy, fast speed, and good stability, the method achieved on-line monitoring and improved the real-time of detection and the ease of operation.

The fiber sensor based on SPR, surface plasmon resonance, could effectively overcome the shortcomings of the existing measurement methods of oil in waste water, such as outdated detection ways, low measurement accuracy, poor adaptive ability environment, and so on, and it had extensive research and application value. On the basis of analyzing the principle of SPR, the fiber sensor based on SPR was designed and the system structure model was given. The solutions of oil and waste water with different concentrations were tested and the relation between resonant wavelength of SPR spectra and refractive index of the tested solutions were obtained. The experiment results show that it is feasible that the fiber testing technology is used to measure the oil in waste water, and the study has certain significance in sewage disposal for the petrochemical industry and water reuse.

In order to provide theoretical basis for studying the corona discharge problem deeply, this paper analyzed advantages and disadvantages of the wavelet method and the spatial and temporal method based on empirical mode decomposition (EMD). It combined the threshold technique of wavelet method and the EMD technique of spatial and temporal processing method, and then proposed the threshold processing method based on EMD. The paper used the three methods to analyze a simulated signal of corona discharge polluted by white noise, the results show that the signal-noise ratio of the paper’s method is biggest and the root mean square error of the paper’s method is smallest. It proves de-noising effects of the paper’s method are better than the other two methods. In the end, the paper reconstructed corona discharge signal submerged in strong background noise from actual corona discharge by the paper’s method, it proves the method can extract small signal from complex background noise.

High Voltage Direct Current (HVDC) system was often the link between major power grids. In order to ensure the safe operation of HVDC, fault diagnosis method was studied. The system structure of HVDC Light system was analyzed, and the normal and fault conditions of the transmission system were simulated in the Matlab platform. This method was very convenient and simple, and it could restore the real situation. The use of simulated waveform could effectively identify the fault types of actual situation, and then it was a flexible, efficient and practical method of fault diagnosis in HVDC Light system.

In order to overcome the limitation of common methods in modeling test system of broadband electromagnetic pulse (EMP), the paper provided a stepping identification arithmetic to solve the problem of building model in whole frequency range. It obtained low-frequency model and high-frequency model respectively in low-frequency domain and high-frequency domain by identification using import signal and output signal measured by test, it obtained the identification model of the test system by connecting the low-frequency model and high-frequency model, the model was simplified by offsetting adjacent poles and zeros. The results of simulation and test show that the papers method can reflect the dynamic characteristic of the system in whole frequency range, and it is feasible in practical application.