基于多模型融合的CNN-LSTM-XGBoost短期电力负荷预测方法

短期电力负荷的精准预测可以有效指导机组组合调度、经济调度与电力市场运营。针对输入数据特征量受限时负荷预测的低精度问题,提出一种基于多模型融合的CNN-LSTM-XGBoost短期电力负荷预测方法。通过建立融合局部特征预提取模块的LSTM（long short term memory）网络结构,并将其与XGBoost（eXtreme boosting system）预测模型并行结合,之后结合MAPE-RW（mean absolute percentage error-reciprocal weight）算法进行模型融合初始权重设置,对最佳权重进行搜索,构建最佳融合模型。通过运用电力负荷数据对所提方法进行预测实验,结果表明CNN-LSTM- XGBoost模型的MAPE（mean absolute percentage error）与RMSE（root mean square error）分别为0.377%与148.419 MW,相比于单一网络模型与融合模型结构实现了误差指标的显著降低,验证了基于多模型融合的CNN-LSTM-XGBoost短期电力负荷预测方法具有较快的模型训练速度、较高的预测准确度与较低的预测误差。     

轮胎炼胶生产能源管理系统

研究烷基二硫代磷酸锌TPZ在丁苯橡胶（SBR）/白炭黑体系中的应用。结果表明：在SBR/白炭黑体系中以TPZ等量替代促进剂DPG,TPZ具有类似促进剂DPG的活化作用,可降低胶料的滚动阻力,且更加环保;少量TPZ作为偶联活化剂用于SBR/白炭黑体系的母炼中,可以减小硅烷偶联剂用量,降低成本,提高填料分散性和偶联效率,从而使胶料获得更佳的物理性能和动态力学性能。     

基于WNN与FCM的电动汽车动态充电负荷预测方法

随着电动汽车动态无线充电(EV-DWC)技术的发展,针对目前EV-DWC负荷建模理论工作不全面的现状,以交通流量作为影响充电负荷的主要因素,以天气、典型日期、季节等因素为次要影响因素,根据路况建立负荷模型,通过电动汽车型号和状态的聚类不同对汽车分配不同的功率,完成动态充电负荷的建立。采用小波神经网络(WNN)对时序信息进行处理预测,再同误差反向传播神经网络(BPNN)相结合预测充电道路上的车流,短期车流预测精度为85%,用模糊C聚类(FCM)算法对电动汽车的充电类型以及该类型所对应的充电功率进行划分,将进入充电道路的电动汽车分为7种类型。根据各种充电类型分配相应的充电功率,完成日负荷建模。     

特殊工业粉尘地区复合绝缘子腐蚀失效过程研究

硅橡胶复合绝缘子因其优异的憎水性及憎水迁移性而具备较好的防污闪能力,往往作为应对污闪事故的首选方案。但在部分特殊工业粉尘地区,复合绝缘子在短时间运行后会出现爬电及蚀损现象,并发展为绝缘失效。为研究特殊工业粉尘地区复合绝缘子的腐蚀失效过程,测试了某特殊工业园区内运行的复合绝缘子的表面污秽度及污秽成分、污闪电压梯度、憎水性能及微观性能,并分析了其腐蚀失效过程。主要结论如下:绝缘子表面等值盐密为0.1~0.2 mg/cm²,其污秽成分与化工污源类似;存在自然污秽时绝缘子表面憎水性良好,但其憎水性的减弱、恢复及迁移特性均不能满足运行要求;绝缘失效的起因是电晕放电导致憎水性降低,进而导致伞裙电蚀损,最后发展为绝缘失效。     

光热电站容量效益评估及影响因素研究

光热发电作为一种清洁能源,配置储热后具有良好的调节性能,近年得到大力发展。依据自身经济性,光热电站往往配备较大的储热装置,能够平移所吸收的光热能,且具有跨日调节能力,使得光热机组有一定替代常规机组带负荷的能力,即容量效益。本文提出一种基于等可靠性的光热电站容量效益评估方法,考虑了调峰方式、储热时长、新能源规模、光热电站规模对光热电站容量效益发挥的影响,采用基于数学优化的生产模拟仿真程序,以周为尺度,计算全年8760 h的系统运行状态,计及了机组启停、水电跨日调节、抽蓄跨日调节、光热电站的跨日调节等因素。最后通过对西北地区实际电网的仿真,验证了所提方法的有效性。     

基于改进K-means聚类的配电网电压暂降频次估计法

传统考虑保护动作特性的电压暂降频次估计法需要获取详尽的保护配置信息,然而配电网保护配置多样,在不同因素如过渡电阻、运行方式、故障类型等的影响下,阶段式保护各级保护区可能产生较大变化,采用传统方法对电压暂降持续时间进行评估可能会产生较大误差。文中提出一种基于改进K-means聚类的配电网电压暂降频次估计方法,在未知线路保护配置基础上,基于电压暂降历史监测数据与保护动作信息,采用改进K-means聚类算法,对电压暂降幅值-持续时间进行聚类分析,推断线路保护配置情况,计算保护动作时间与保护动作电压。根据计算结果,在考虑不同故障类型、不同运行方式及不同过渡阻抗的情况下进行配电网电压暂降频次估计。在IEEE RBTS-6母线测试系统的母线5配电网中进行仿真,验证了文中方法的有效性和优越性。     

多类型源储协调互动的配电网分布鲁棒优化调度

分布式光伏、风电等可再生能源出力具有较强的随机性和不确定性,其大量并网给配电网的运行带来了极大的挑战。本文提出了一种多类型源储协调互动的配电网分布鲁棒优化调度方法,基于分布鲁棒优化理论,对配电网中传统离散设备、可再生能源以及储能决策进行协调优化,提高配电网运行的经济性和安全性,实现了配电网运行决策保守性与鲁棒性的有效平衡。首先,考虑多类型源储资源的协调互动,建立配电网优化调度模型,并将其转化为混合整数二阶锥规划的形式;其次,利用可再生能源出力场景集进行不确定性刻画,建立分布鲁棒优化模型,并通过列和约束生成算法进行求解;最后,在PG&E69节点系统上进行算例分析,验证了所提方法的可行性与准确性。     

Importance of Low-Temperature Melt-Mixing on the Construction of Stereocomplex Crystallites with Superior Nucleation Efficiency in Asymmetric Poly(l-lactide)/Poly(d-lactide) Blends

The nucleation efficiency (NE) of stereocomplex crystallites (SCs) formed in asymmetric poly(L-lactide)/poly(D-lactide) (PLLA/PDLA) blends is generally unsatisfactory because the competition between stereocomplexation and chain mixing involved in the melt-mixing process can cause low formation efficiency and even severe aggregation of SCs. Herein, it is attempted to achieve high-efficient formation of finely dispersed SCs particles by designing a unique melt-mixing procedure, where the mixing of PLLA with 0.75 wt% PDLA is first performed at elevated temperatures (far above the melting temperature of SCs) to allow the homogeneous mixing of PLLA/PDLA chains and then at a low temperature (slightly above that of homocrystallites) to permit the full stereocomplexation of the premixed chains. It is found that the SCs formed in the blends exhibit unexpectedly low NEs (e.g., 54.5%), much inferior to that (73.6%) in the counterpart without undergoing premixing. This is because the introduction of premixing leads to a remarkable deterioration in the amount of SCs particles formed, despite decreased particle size, highlighting that the direct mixing at low temperatures of 170–180 °C (about 20–30 °C lower than that used in common melt-processing of PLA) is more effective for the construction of SCs with superior NE. The mechanisms for these striking findings are discussed.     

Stereocomplex Crystallization Induced Significant Improvement in Transparency and Stiffness–Toughness Performance of Core-Shell Rubber Nanoparticles Toughened Poly(l-lactide) Blends

Toughening modification of poly(l -lactide) (PLLA) with rubber particles is often realized at the cost of transparency, mechanical strength, and modulus because high rubber loadings are generally required for toughening. In this work, a promising strategy to simultaneously improve the transparency and stiffness–toughness performance of poly(butyl acrylate)-poly(methyl methacrylate) (BAMMA) core-shell rubber nanoparticles toughened PLLA blends by utilizing the stereocomplex (SC) crystallization between PLLA and poly(d -lactide) (PDLA) is devised. The results reveal that the construction of SC crystallites in PLLA matrix via melt-mixing PLLA/BAMMA blends with PDLA can prevent BAMMA nanoparticles from aggregation and promote them to form network-like structure at lower contents. As a result, not only higher toughening efficiency with less rubber contents but also superior transparency is achieved in the PLLA/PDLA/BAMMA blends as compared with the PLLA/BAMMA ones where large aggregated BAMMA clusters are formed. Moreover, the outstanding reinforcement of SC crystallites network for PLLA can impart an enhanced tensile strength and modulus to PLLA/PDLA/BAMMA blends, thus improving the stiffness–toughness performance of PLLA/PDLA/BAMMA blends to a higher degree. This work demonstrates that SC crystallization is a promising solution to solve the contradiction between transparency and mechanical properties and then obtain superior comprehensive performances in rubber toughened PLLA blends.     

Boron Nitride Nanosheets Strengthened PVA/Borax Hydrogels with Highly Efficient Self-Healing and Rapid pH-Driven Shape Memory Effect

Self-healing hydrogels often possess poor mechanical properties which largely limits their applications in many fields. In this work, boron nitride nanosheets are introduced into a network of the poly(vinyl alcohol)/borax (PVA/borax) hydrogels to enhance the mechanical properties of the hydrogel without compromising the self-healing abilities. The obtained hydrogels exhibit excellent mechanical properties with a tensile strength of 0.410 ± 0.007 MPa, an elongation at break of 1712%, a Young's Modulus of 0.860 ± 0.023 MPa, and a toughness of 3.860 ± 0.075 MJ m⁻³. In addition, the self-healing efficiency of the hydrogels is higher than 90% within 10 min at room temperature. Benefiting from the excellent self-healing properties, the shapeability of the hydrogel fragments is observed using different molds. In addition, the hydrogels display rapid pH-driven shape memory effects and can recover to their original shape within 260 s. Overall, this work provides a new approach to hydrogels with integrated excellent mechanical properties, self-healing abilities, and rapid pH-driven shape memory effects.