作为配电网中能量传输的主要载体，地下电缆的安全、可靠运行已经成为建设未来高可靠性配电网的关键。复杂的运行环境使得电缆绝缘常出现不可逆的老化，最终导致系统故障。目前，大多数技术致力于识别故障电缆并将其从系统中切除，仍然难以避免部分地区的停电，属于一种“被动”的保护方式。如何实现高精度的电缆绝缘在线状态感知，从而对老化电缆进行及时的检修，从根本上避免故障发生已经成为未来配电网运维模式由“被动”转向“主动”的关键。为了实现配电网地下电缆的高精度状态感知，本文从传感技术、监测技术和老化定位技术三个方面开展研究，主要提出以下三个核心技术：（1）基于磁屏蔽原理的共模漏电流高精度测量技术；（2）基于共模监测信号注入的定量在线监测技术；（3）基于端部稳态电信号测量的局部老化定位技术。所提技术可以实现对电缆绝缘老化的针对性检修，提高配电网可靠性。在传感技术的研究中，首先提出能够在较大的负载电流干扰下实现微小漏电流测量的磁屏蔽原理并建立了相应的解析分析理论。在提出的磁屏蔽原理基础上，进一步得到共模漏电流测量模型，并基于多目标优化算法实现了对漏电流传感器参数的最优设计。本论文是首次建立了磁屏蔽原理的解析理论并提出基于其的传感器参数优化设计方法，从根本上改变了传统传感器设计中需要有限元仿真和迭代试错的模式，提高了测量精度和设计效率。在监测技术方面，提出通过系统中性点注入监测信号，测量系统中的共模漏电流从而实现在线监测的方法。设计了不同频率的监测信号注入装置，使得提出的监测技术具有非侵入性且易于实现。分析了系统中共模漏电流和绝缘参数之间的关系，实现了对绝缘状态的在线、定量评估。在局部老化定位技术方面，针对不同的老化类型提出了多种定位方法。对于老化长度较短的情况，通过分析不同频率监测信号在电缆上的电压分布得到漏电流频谱分布特征，从而实现定位。对于老化长度较长的情况，推导了电缆局部老化模型，并提出基于模型及数据驱动的定位方法。相比于传统方法，所提方法不依赖于局部老化段的参数突变，因而对于多种局部老化类型均能实现有效的定位。
As the main power carrier in distribution grids, the safe and reliable operation of underground power cables has become a critical issue for future distribution grids of high reliability. Operating in complicated environment, cable insulation undergoes irreversible ageing and will eventually lead to system faults. To date, most researches focus on detection and removal of faulty cables, which still endures regional power outages. These methods can be regarded as “passive” protection strategies. By contrast, online condition monitoring of power cables can lead to timely maintenance of aged cables, which can avoid system faults in an early stage. How to realize high-accuracy online condition sensing has become the key to realize “proactive” maintenance for future distribution grids, instead of normal “passive” protection.To realize high-accuracy online condition sensing for underground power cables in distribution grids, this dissertation is arranged from three aspects including sensing, online monitoring and segmental ageing location techniques. Three core technologies are developed, which are: (1) High-accuracy measurement of common-mode (CM) leakage current based on magnetic shielding theory, (2) Quantitative and online monitoring of underground power cables based on CM monitoring signal injection, and (3) Segmental ageing location method based on steady-state measurement at cable terminals. The proposed methods can lead to timely and targeted maintenance of insulation ageing in power cables, improving the reliability of distribution grids.In the studies of sensing techniques, the magnetic shielding theory is firstly proposed, which can realize minor leakage current measurement under large load current disturbances. The analytical theory is also derived for the magnetic shielding effect. The CM leakage current measurement model is then proposed based on magnetic shielding theory and the parameters of the leakage current sensor is optimized using multi-objective optimization algorithm. For the first time in relevant researches, this dissertation establishes the analytical theory for magnetic shielding effect and proposes a sensor parameter optimization method. The proposed technique will change the sensor design procedure ultimately with a higher sensitivity and design efficiency, while finite element analysis with trial and error is commonly adopted in conventional sensor development.In the studies of monitoring methods, the dissertation has proposed novel online monitoring methods with monitoring signal injection from system neutral point, the CM leakage current is then measured to realize insulation state evaluation. Monitoring signal injection device is designed, ensuring the non-intrusiveness and easy implementation of the proposed method. The relationship between induced CM leakage current and insulation parameters is also derived to realize online and quantitative monitoring.Finally, in the studies of segmental ageing location, several methods are proposed for different types of segmental ageing. For insulation ageing with shorter length, the voltage distribution on the cable is used for the location and ageing location feature is extracted from the induced leakage current distribution pattern. For longer segmental ageing, relevant model is established. Model-based ageing location method and date-driven method are both proposed. Compared with traditional methods, the proposed techniques do not rely on abrupt changes of electrical parameter on the aged segment, which can realize effective location for various ageing patterns.
