By Phil Kreveld
How not to design a mainly solar and wind powered grid is to go about it as we do in Australia.
No chief of network engineering as in the UK, a bunch of regulatory authorities good at writing rules but little else, an energy market operator with plenty of engineering talent but limited authority over networks, state governments with their own renewable agendas and a Commonwealth Government with CO2 reduction dreams.
Early on, we had an ideal investment climate for wind, solar, and pumped hydro but then renewables started to grow and grow and loss factors on remote energy zone links began to interfere with their economics. These days, loss factors are the least of profit reduction problems. The investment climate is souring because the legacy networks to which most this renewable stuff is being attached cannot handle the power flow resulting in oscillations of voltage and power. We can, therefore, confidently expect more instances like Victoria’s Rhombus of Regret, where solar and wind farms are severely cut back in power output in order to attenuate these oscillations.
Related article: Gas exports could be diverted to fend off winter shortage
We have mainly ignored the impacts on the grids by grid following and grid forming inverters that are essential for wind and solar. The former are in the majority, the latter are coming on stream—but in microgrids that can connect and disconnect to and from medium- and high-voltage lines. The inverter technologies are good; what remains very much in doubt is their influence on grid stability when they become pretty much the ‘only game in town’.
Dig below the publicity of renewable enthusiasts—and one finds that engineering concerns don’t rate. However, AEMO is seriously concerned because there is no definitive solution to prospective grid stability problems. And it’s easy to understand their worries. No one, anywhere in the world, has physically demonstrated how to operate an inverter-based grid of our size—basically a radial grid of 5000km stretching from Queensland to South Australia.
The challenge is to break the nexus between synchronous generation and fossil fuels because their popularly applauded departure is very much adding to future stability problems. AEMO was warned by the now toothless Energy Security Board not to be so enthusiastic in its renewable energy contribution planning, building up solar and wind energy based on the rapid departure of synchronous generation. It might have been more of a tug of the forelock to the State and Commonwealth Governments in the latest version of AEMO’s integrated systems plan. Wind, solar, batteries—good! Coal and gas—bad!
In a talk to Melbourne University late last year, AEMO’s Daniel Westerman said we should be planning on some 40 synchronous condensers, precisely because coal power stations are shutting down. He was plugging synchronous machines to help prop up grid strength and, hopefully, get those oscillations as well voltage control steadied. They are needed for voltage stability and inertia, Westerman said. Syncons are half of a generator-turbine set but with less than half the inertia of the combination so that flywheels often have to be added—but there’s more, namely the excitation of their rotors for which more secure infrastructure is required—and there’s more yet, because the conventional control systems for syncons need to be redesigned in order to work with inverter generation. Are we nationally going to fork out a couple of billion dollars without any real confidence that the syncons will do the job of stabilising 80 to 100% renewable grids?
Academic work in grid stability constraints is of great value but it is research work that unfortunately has had little or no bearing on physical operations. Research is pointing out the inherent problems of grid following inverters in low grid strength networks—in short, instability in their phase-locked loops to repress current, power and voltage oscillation. Research work is being done to design grid-following and grid-firming inverters without the need for phase-locked loop voltage following. However, it is not part of commercialisation as we rely on overseas suppliers of technology. Similarly, the design of load sharing schemes between grid (voltage) forming inverters and their stability constraints are a completely new ball game.
Synchronous generators keep in synchrony by interchanging synchronous torque and individual droop controls allow them to deliver whatever power (up to their maximum rating) they are asked to commercially deliver. They do this while maintaining constant frequency. Inverter sources can perform a similar task but without synchronising torque, obviously. They have only been proven in microgrids. The experience does not translate into large, multi-generator and dynamic load centre grids. Of course, AEMO is perfectly well aware of this and for that reason has put ‘hold stages’ on the engineering-issues plan accompanying its ISP.
The hold stages are sensible because they would allow for testing and verification of operability. The national problem we face is that we have not been working on an integrated design for the NEM grid. Furthermore, we have not actively trialled parallel operation of BESS-inverter combos to mimic, for example, large rating generator-turbine sets equivalent to thermal power stations. We have not evaluated alternatives to syncons, including static compensators and their ability to control interarea and local area oscillations. It is a given that hold stages will therefore be prolonged. Given the absolute need for grid security, the long hold periods will result in disincentives for renewable investment plans.
What might an integrated plan look like? The first step is to analyse the dynamic performance of distribution networks. Second by second measurement of power, reactive power and positive sequence voltage angle would be a good start. Presently all the attention is on interconnectors while the distribution networks are often and for longer periods close to energy independent. Yet this is ignored in transmission planning! AEMO, when questioned on this, provides a fuzzy answer along the lines of sort of having considered it but then admitting that it is not part of engineering considerations.
An integrated plan would first determine which distribution networks should be given a time line for a planned maximum transmission power flow, and on the basis of that time line plan, would be required to design and plan auxiliary generation to support distributed energy resources. Electrical Engineering 101 tells us that generators close to the load are the most energy efficient as well stable solution. Given the focus on DER, particularly rooftop solar and community batteries this is a ‘no brainer’. It is not a particularly easy job involving resynchronisation and protection engineering technology but a highly desirable, sensible economic step. However, as matters stand, neither AEMO nor anyone else can direct this first step to take place!
Related article: Call for nominations from generators for Reliability Panel
The next step would be the re-examination of commitments to interconnectors in the light of revised maximum power flow requirements. South Australia is an interesting case for examining the need to upgrade interconnectors or constructing new ones when the obvious first step is to upgrade local synchronous generation and/or large-scale inverter based voltage and frequency stabilisation. Next would be to take AEMO’s investigation on system strength and on that basis devise future allocations for allowable generator connections, i.e. based on detailed grid impedance databases. Requirements for new generators ought to be on the basis of adding to system strength by virtue of required overload capacity—and obviously more than the 120 to 150% of full load typical of BESS-inverter systems, therefore requiring their paralleling.
AEMO should be given the authority to be the nation’s Head of Networks and to commence directing its bevy of stakeholders to follow its strict guidelines in grid design and generator development and allocation. Although unprovable in that we would have to run the experiment also on the present ad-hoc basis, there can be no doubt that an integrated, centrally directed design approach will serve us better and be more economical than the path we are on at present.
