Article written by Phil Kreveld Power Parameters Australia
The question mark in the title is relevant—because it brings up choices that need to be made for there to be a future for the NEM.
The future as discussed here relates to the system of baseload thermal, hydro and pumped hydro, gas-fired, wind and solar farm generation, AND the fast-growing private sector of virtually exclusive solar photovoltaic (PV) generation.
What are the problems facing the NEM?
The major issue affecting the future of the NEM, is the range of engineering problems arising from the marriage of (in)convenience between legacy generation, transmission, and distribution, with non-dispatchable generation.
The likely progeny of this marriage does not get much publicity. Future physical, information, and control systems requirements of the NEM grid and associated distribution networks are invariably not part of plenary sessions at electrical power industry conventions and engineering expositions.
The ‘midwives’ to whom the task will fall, to nurse the progeny resulting from old and new technologies, congregate in specialist engineering sessions — in truth far removed from the considerations of those in the main meetings where the market and commercial issues get the main, if not the only exposure.
Many of the fraternity and sorority in these specialist sessions hail from schools of electrical engineering where they were trained in power flow calculations, slack bus-referred, iterative calculations of voltage angles of generator and load buses as well as small and transient disturbances in synchronous generators.
They seem out of place in this future NEM. However, we ignore the physics underlying their engineering training at the peril of inheriting a national network whose stability is balanced on a knife-edge. And hence this examination of requirements for a stable future.
What are the easy fixes that the NEM can implement?
Let’s begin by examining one of the ‘easy’ fixes – batteries.
In the popular press, batteries are mentioned as alternatives to retiring baseload. As to the size of batteries, some easy maths will do for the purposes of this article.
A 25-gigawatt NEM demand, subject to a 60% penetration of non-dispatchable generation, would require of the order of 10 gigawatts in battery power.
That’s not all that is needed—even if the aggregate battery capacity was achievable, a battery bank, situated some 300 kilometres distant from synchronous generators, is a different proposition from one connected to the generator bus.
The NEM’s long transmission line network and widely dispersed baseload synchronous generators are in conflict with large incursions of non-dispatchable generation. There are big differences in voltage angles along its 5000-kilometer stretch.
Short term load following for battery banks is easier than for synchronous generators, but of limited use, unless there is simultaneity in the observation of voltage angles—they are the indicators of instantaneous power generated (and at the battery bank inverter).
If batteries are a challenging option for replacing baseload, what are the other considerations?
Power flow calculations are made easier by programs such as DigiSilent, but are based on static conditions—and complicated enough as it is; involving uncoupled admittance matrices, Jacobeans, and converging iterations such as Newton-Raphson.
Such a program is suitable for network design or contingency planning when demand variations are slow. But are not suitable for operational control—particularly when latencies are shortened due to wind and solar. In this case multi-nodal measurements and appropriate controls are essential to ward off voltage collapse and to keep the entire system in synchrony.
What about ‘Control’ as a part of the solution?
Admittedly, the word ‘control’ rolls of the tongue—but what precisely needs to be controlled? In reality and, broadly speaking, they are the same controls that were necessary when Charles Proteus Steinmetz published his treatise on AC circuits in 1894.
Getting down to business, a multi-nodal measurement system needs to control and balance total reactive power. Balancing generated with absorbed reactive power, and maintain the stability of voltage angles (or real power balance).
We now have the tools to achieve this level of control.
There are geostationary satellites to provide precise synchronous timing signals, highly capable synchrophasor and power quality monitors, instrument-class current and voltage transformers with very low ratio and phase errors, intelligent electronic devices to operate all manner of protective gear, and high-speed communication networks to send control signals to a multitude of inverters.
The distribution of side electrical power is a critical item in maintaining NEM stability.
Reverse power flow during periods of high insolation, turn substations into highly variable loads and even into ‘negative loads’, the latter having to be avoided at all costs as it is an almost guaranteed factor for instability.
Monitoring within networks and control of distributed generation is therefore unavoidable.
The ‘plug and play’ nature of solar inverters makes the task of controlling power at the substation level, very difficult, if not impossible unless there are methods in place to control their outputs.
Politically this is a fraught question and alternatives such as static var compensators can help in voltage support, but not in real power control at the edge of the network. Only limiting power infeed will reduce voltage.
Presently distribution networks are concerned with high voltages in branch circuits because these limit inverter output — but that is not solving the basic issue of substation stability, which will be increasingly compromised.
Where to from here for the NEM?
For reason difficult to fathom, the idea of grid-wide monitoring has not gained a foothold as yet.
However, the argument that synchronized, granular, multi-nodal information has to be the basis for any number of control paradigms is easy to support—even if, depending on initial expenditure considerations, limitations are imposed on control protocols and there is a superfluity of information.
As is well known, what you don’t or can’t measure, cannot be managed or controlled. Grid-wide, multi-nodal information has to be the starting point if informed decisions are to be taken.
Article written by Phil Kreveld Power Parameters Australia
