基于分段改进S变换和随机森林的复合电能质量扰动识别方法

各类分布式设备和智能设备接入电力系统,使得电力系统对电能的波动越来越敏感,这导致对电能质量扰动(PQD)的识别和处理变得越来越重要。通过将分段改进S变换(SMST)和随机森林(RF)算法相结合,提出了一种用于复杂噪声环境下PQD识别的新方法。首先,基于检测误差和峰度对SMST的不同频段进行分别调参,并使用SMST提取待检测信号的75种时频特征,构成原始特征集。然后,改进分类回归树(CART)的节点分裂过程,加入了离散值处理策略并使用Gini指数的下降作为新的节点分裂规则。同时,在下次节点分裂前,将基尼指数下降值为零的特征从特征集中删除。最后,使用改进的CART算法构建了RF分类器并对复合PQD信号进行分类。实验证明,在不同的信噪比条件下,新方法均能有效识别多数单一PQD信号和常见的双重复合PQD信号。虽然新方法在运行效率方面仍有一定的改进空间,但其在不同层面上的改进均能有效提升PQD识别精度,且平均分类精度明显高于各类传统PQD识别方法。

引入调节因子改进S变换电压暂降源识别

配电网电压暂降的正确分类识别对抑制和缓解电压暂降具有重大意义。分析了分别由三种不同相的接地短路故障、感应电动机启动和变压器投运引起的五种电压暂降现象及其典型特征,在标准S变换的高斯窗函数上添加幅度调节系数和指数调节系数两个调节因子得到改进的S变换。采用改进S变换的方法得到不同类型电压暂降信号的基频幅值曲线和频率幅值包络线,提取暂降深度、突变点个数、基频幅值上升和下降斜率、二次谐波含有率和暂降时间比六个特征指标。将这些特征指标数据进行归一化处理后输入支持向量机,实现对配电网不同类型电压暂降的识别。最后,仿真结果表明,将基于引入调节因子得到的改进S变换用于特征指标提取,对电压暂降进行分类识别的正确率相比标准S变换更高。

Thymoquinone Suppresses Cellular Proliferation, Inhibits VEGF Production and Obstructs Tumor Progression and Invasion in the Rat Model of DMH-Induced Colon Carcinogenesis

A myriad of medicinal effects has been attributed to Thymoquinone (TQ), the major biological-active component of Nigella sativa. TQ has been shown to exhibit potent anti-tumor activities. The present work was undertaken to further explore TQ’s chemopreventive efficacy against 1, 2-dimethylhydrazine (DMH)-induced colon carcinogenesis in the rat model through a two-phase study (initiation and post-initiation) and to evaluate its potential impact on tumor progression and invasion in vivo. TQ treatment in the initiation phase significantly reduced tumor incidence, multiplicity and mean tumor volume. However, although mean tumor volume and multiplicity were decreased upon TQ treatment in the post-initiation phase, TQ did not reduce incidence significantly. Cellular proliferation, as assessed by expression of colonic PCNA, was shown to be inhibited in consequence to TQ treatment in both phases, with a more pronounced reduction in the initiation phase. In addition, our results demonstrated an appreciable negative impact of TQ on vascular endothelial growth factor (VEGF) production in tumor-bearing rats. Furthermore, we provided evidence that TQ-treatment, in both phases, tended to considerably suppress tumor progression and invasion. Taken together, the present study demonstrated that TQ, at an orally daily dose of 10 mg/kg, has a chemopreventive effect in the initiation phase, and has the potential to attenuate tumor burden, suppress progression of pre-neoplastic lesions and to inhibit tumor growth in the post-initiation phase of DMH-induced colon carcinogenesis, We surmise that such effects of TQ may be due to suppression of cellular proliferation and inhibition of VEGF production. The results could provide an effective chemopreventive approach in the primary prevention of colon cancer in humans in the next future, and illuminate a promising horizon to combat progression of benign colonic pre-neoplastic lesions into malignant metastatic tumors and to manage colon cancer.

基于改进S变换和GA-SVM的电能质量扰动识别与分类

电能质量扰动的准确识别与有效分类是改善与治理电能质量问题的前提,针对当前电能质量扰动识别与分类存在的不足,提出一种基于改进S变换和遗传算法(genetic algorithm,GA)优化支持向量机(support vector machine,SVM)的电能质量扰动识别与分类方法。首先,在S变换高斯窗函数中引入调节因子,依据信号特征分频段选择调节因子,以提高电能质量扰动分析的时频分辨率。其次,结合S变换模时频矩阵提取时频特征曲线,构建电能质量扰动初始特征集。最后,基于遗传算法进行扰动特征优选和支持向量机参数优化,并利用GA-SVM完成电能质量扰动分类。对14种电能质量扰动的分类进行测试,验证所提方法的有效性与鲁棒性。

Ethanolic Extract of <i>Nigella sativa</i>Seeds Lacks the Chemopreventive Efficacy in the Post Initiation Phase of DMH-Induced Colon Cancer in a Rat Model

Nigella sativa, belonging to the Ranunculacea family, is a versatile phytochemical mine bestowed with multi-dimensional medicinal effects. The current study was performed to investigate the chemopreventive efficacy, if any, of the ethanolic extract of indigenous Nigella sativa seeds (ENS) and to evaluate its potentials on tumor progression during the initiation and post initiation phases of Dimethylhydrazine-induced colon carcinogenesis in a rat model. ENS treatment during the initiation phase unveiled chemopreventive effect manifested by significant reductions in tumor incidence, multiplicity and tumor volumes. Histopathological findings and modified Duke’s classification of tumors provided evidence that ENS, administered in the initiation phase, is capable of delaying progression, restricting invasion and attenuating aggressiveness of colon tumors. These results imply that ENS may be a promising candidate in the primary prevention of colon cancer. On the contrary, we demonstrated that ENS lacked chemopreventive and tumorigenesis inhibitory effects in the post initiation phase. We speculate that the chemopreventive effect of ENS might be due to the synergistic actions of various constituents present in the extract. However, extensive studies are warranted and more efforts need to be dedicated to fractionate, analyze and to further appraise the anticancer effect of ENS before any definitive conclusions can be drawn.

一种基于时频域多特征量的电能质量混合扰动分类新方法

针对电能质量混合扰动分类问题,提出一种基于时频域多特征量的分类新方法。首先利用聚类经验模型分解方法和改进不完全S变换对扰动信号进行处理,并提取9个时频域特征值;然后,将特征量输入到分块化的自动分类系统中进行扰动识别。该方法充分考虑单一扰动之间的相互干扰,并通过互补的时频域特征量进行有效的抑制。仿真结果表明,在一定的噪声条件下,所提方法可有效分类电压暂降、电压暂升、电压短时中断、脉冲暂态、振荡暂态、谐波和闪变等电能质量扰动及其组合而成的混合扰动。