Quantifying the Reliability of a PMU Network

Spread knowledge

he wide area measurement system (WAMS) is gradually becoming an important guarantee of security and stability in smart transmission grids. In the utility industry, WAMS is sometimes referred to as a PMU (phasor measurement unit) network, as synchronized PMUs are the most crucial elements in collecting real-time monitoring information. A PMU network provides much better observability and controllability in smart grid operations. However, like any other physical system, a PMU network itself can fail. The consequences of PMU network failure are serious and can include a large blackout. Therefore, the reliability of PMU networks should be quantitatively evaluated and assured.

A PMU network is composed of PMUs at substations and generation stations, phasor data concentrators (PDCs), local communication networks, backbone communication networks and a control center. The reliability of a PMU network can be quantified using reliability evaluation methods. A PMU network is divided into three types of substructures: the phasor measurement devices, the regional communication network and the backbone communication network. The basic procedure consists of two main steps. The failure modes of modules in each substructure are analyzed first to build an equivalent two-state reliability model of substructure using different evaluation techniques. Then the equivalent reliability models of the substructures are combined to assess the reliability of the whole PMU network using a fault tree analysis method.

In the hierarchical structure, a PDC and multiple PMUs constitute a PMUs-PDC working group, in which communication is carried through a regional network. Multiple PMUs-PDC groups are connected to the control center through a backbone communication network that is composed of fabric links and ring interface units.

A PMU device can be divided into seven modules in light of their operational functions for reliability evaluation. Each module can be further broken down into sub-components. Markov models for the sub-components and modules, which are based on state space diagrams and transitions between states, are developed first and then these models are converted into an equivalent two-state model, which can be used to quantify the reliability indices of the PMU device and easily incorporated into the reliability evaluation of PMUs-PDC working groups.

A regional communication network transmits information between PMUs and PDC. Regional communication networks can be classified into three categories. In the first category, PMU measurement information is transmitted through the utility’s own existing facilities—on a carrier wave or microwave communication channels, for example. In the second category, a commercial optic fiber communication network is used. In the third category, a communication network is built by utility specifically for PMU information.

The reliability of a regional communication network is associated with connectivity identification between multiple inputs (many PMUs) and a single output (one PDC) under contingency conditions. A network survival mechanism refers to the way of recovering normal data transfer in the network after a contingency event such as a link failure. Network survival mechanisms can be classified into static protection and dynamic restoration. In the static protection scenario, a backup path is pre-established together with the primary path for each PMU. In dynamic restoration, no backup path is pre-specified. When a contingency happens, a search process starts to dynamically find a possible backup path. A set of reliability evaluation techniques can be used to quantify the reliability of a regional communication network. These include graph theory, set theory and minimum cutsets (minimum combinations of component failures that can cause system failure).

The backbone communication network transmits information between PDCs and the control center. It is often designed as a synchronous optical network with a synchronous digital hierarchy ring configuration and dual-passages. In general, there are two optic fiber rings in the network. One is the primary optic fiber ring to transmit working digital signals in the normal operation state, and the other is a standby optic fiber ring that can transmit the same digital signals only in a contingency situation through successful switching operation. This is called the 1+1 backup mode.

From a reliability evaluation viewpoint, the backbone communication network can be modeled using communication interfaces in series with an optic fiber system module. A combined method of Markov models and the state enumeration technique has been developed to quantify the reliability of the optical fiber system and whole PMU network.

Data uncertainty is a challenge for PMU network reliability assessment. PMUs have been installed in power systems only in recent years. The statistical failure data of PMUs are still sparse, which introduces imprecision in estimation of reliability parameters of PMU’s components. A solution to take account of the uncertainty is the application of combined statistical and fuzzy Markov methods.

The methods and models described above have been applied to an IEEE test system and an actual project at BC Hydro, Canada. The applications demonstrated that PMU network reliability can be quantified using the presented methods and models. An equivalent reliability two-state model for the whole PMU network can be obtained from quantified reliability assessments. Such an equivalent model provides flexibility for the reliability evaluation of an integrated smart transmission grid that is composed of a traditional electric power system and PMU network. This is a new topic in smart grid reliability.