Search

Estuarine ecosystem is the ecotone between the river and ocean,which is under tremendous influences of concentrated human activities.Now,most estuarine ecological risk studies are based on the terrestrial ecosystem model,ignoring spatial heterogeneity distributing laws.This paper takes Daliao River as the study area and establishes a targeted estuary ecological risk evaluation model.Considering the scale request,the assessing technology mainly includes four steps: confirmation and analysis of study area,identification and analysis of risk sources,analysis of exposure and ecotoxic effects,and integrated estimation of ecological risks.Then the evaluation system is built up,which includes indices of biodiversity,afflux material,comprehensive pollution,and pressure index of socio-economic development.Combining the method of AHP,Principal Component Analysis and Polarized Standardization,the integrated index of ecological risk can afterwards be figured out.This model takes salinity and pollution sources as the major factors in the divisions of different sections and different water periods.And then this article effectively assesses the spatial distribution of ecological risk in Daliao River estuary,and explores the spatial and temporal distribution laws of estuarine ecological risk.The results show that,on the whole,the ecological risk of Daliao River estuary area is relatively high(the average risk value of three water periods is 0.56),which is over the middle level in 0 to 1 of the ecological risk index space;on time scale,the runoff is the main factor resulting in the ecological risk difference;on the spatial scale,the ecological risk index is affected by the pollutants carried by runoff of upstream,downstream pollution emissions and dilution water at the mouth the river together.In order to improve the environmental condition of the estuary area,more attention should be paid to the adverse effects of human activities on the estuarine ecosystem and improve the ecological conditions of the whole valley,thus the environmental problems of estuarine ecosystem can be solved fundamentally.

在现有拉曼光纤放大器优化设计中，对拉曼散射耦合模型的求解是制约各种优化算法求解效率的瓶颈。提出运用最小二乘支持向量回归机，建立多抽运拉曼光纤放大器的输入输出回归模型，并利用该模型直接替代对拉曼散射耦合模型的求解，实现了优化速率数量级的提升。采用该方法对C 波段的多波长反向抽运拉曼光纤放大器进行了优化实验，仅使用2 个抽运源就可以迅速得到增益波动小于±0.5 dB 的优化设计，并且其平均开关增益可以直接使用该模型进行快速调节。实验结果表明，相比传统的打靶法和平均功率法，通过该拉曼光纤放大器回归模型可以直接得到输出增益，极大地提高了设计的时效性，具有一定的参考价值。 Solving the coupled Raman amplification is the main limitation to the efficiency of the existing optimization method for Raman fiber amplifier design. The least squares support vector regression model is applied to address this issue. Instead of directly calculating the Raman coupled equation, a multi-input multi-output model for Raman fiber amplifier is established. A C-band Raman fiber amplifier with a gain ripple level of ± 0.5 dB is designed using only two pumps. Additionally, the on-off gain can be quickly and directly adjusted by the proposed model. Compared with the shooting algorithm and average power analysis technique, the results indicate that the proposed regression model greatly improves the solving efficiency.

鉴于横向应力和温度对光纤布拉格光栅的交感耦合呈现非线性关系，将横向局部应力与温度的交感问题转化为非线性系统参数辨识问题: 首先根据光纤布拉格光栅横向应力与温度的传感原理构造一个理论意义上的反射光谱，然后利用理论光谱和采样光谱之间的差异度建立系统参数辨识模型，通过差分进化算法对该优化辨识模型进行求解，最终实现横向局部应力与温度的同时测量。实验和仿真结果表明，本文提出的方法具有较高的辨识精度，得到的横向应力误差绝对值一般小于4．0 ×10 － 3N，温度误差绝对值一般小于3． 0 × 10 － 3℃，传感器对温度的灵敏度为11． 7 pm/℃。该方法能够解决光纤布拉格光栅横向局部应力和温度的交叉敏感问题，提升光纤布拉格光栅传感网络在实际工程中的传感精度。Considering that fiber Bragg grating ( FBG) sensors are sensitive to both transverse stress and temperature simultaneously, which leads to the difficulty in the independent measurement of these two measurands, we transform this problem into nonlinear system parameter identification with identification parameters of temperature and stress. A theoretical reflection spectrum is first constructed based on the sensing principle of FBG transverse stress and temperature. Then a novel identification model is established based on the discrepancy between actual measured spectrum and the constructed spectrum. The differential evolution (DE) algorithm, a simple yet powerful evolutionary algorithm, is applied to solve the proposed model and obtain the appropriate values of temperature and stress simultaneously. Experiment and simulation results show that the measurement error of transverse stress is typically smaller than 4.0 × 10 － 3N and the measurement error of temperature is smaller than 3. 0 × 10 － 3℃, the temperature sensitivity of the sensor is 11.7 pm/℃. The result indicates that the proposed approach is effective in discriminating the cross-sensitivity of FBGs, while maintaining high measurement accuracy. Published version