Power Generators Presented by Samson Yu School of Electrical and - - PowerPoint PPT Presentation

Dynamic State Estimation of Power Generators

Presented by

Samson Yu

School of Electrical and Electronic Engineering (EECE), University of Western Australia (UWA)

Future grids:

“A seamless, cost-effective electricity system, from generation to end- use, capable of meeting all clean energy demands and capacity requirements”

• Societal, economic and environmental.

0. . Motivations

Future grids:

1. Scale-up of clean energy (renewables, natural gas, nuclear, fuel cells).

Worldwide: invested AUD $350 billion in clean energy 2015, higher than in fossil fuel power plants. DO GOOD and DO WELL. AU: invested AUD $4.3 billion in clean energy 2016. AU RET: 23.5% of AU’s electricity generated by clean energy by year 2020. Year 2015, 14.7%.

0. . Motivations

Future grids:

2. Consumer participation and choice (distributed generation, electric vehicle, etc.). 3. Holistic design solutions (AC-DC transmission and distribution solutions, microgrids, and centralized-decentralized control). 4. Two-way flows of energy and information (Internet Of Thing) 5. Reliability and security (cyber)

0. . Motivations

Questions:

1. What are “dynamic states” of power generators? 2. Why do we want to estimate the dynamic states? 3. How can we perform the dynamic state estimation? 4. How can the estimated dynamic states be used?

0. . Motivations

1. . Dynamic states of f power generators

Slower Dynamics (All the generators) Faster Dynamics (Transmission and distribution network with all electrical loads) 𝑦 = 𝑔(𝑦, 𝑣) 0 = 𝑕(𝑦, 𝑣)

• We have a Differential Algebraic Equation (DAE) formulation of a power system.

𝑦 = 𝑔 𝑦, 𝑣 , 0 = 𝑕 𝑦, 𝑣 .

Examples:

• 1. Rotating power generators

I. Synchronous generators II. Asynchronous generators

2. Stationary power generators

I. Fuel cells II. Solar PVs III. Nuclear power

1. . Dynamic states of f power generators

Synchronous generator

1. . Dynamic states of f power generators

Asynchronous generator

1. . Dynamic states of f power generators

Fuel cells (e.g., SOFC)

1. . Dynamic states of f power generators

Reason 1—Hard to directly measure. Reason 2—Some states we want to control, and we need to know them before we can control them. (e.g., gas partial pressure of fuel cells). Reason 3—More information may help us design a more effective

• controller. (e.g., in DFIG knowledge of flux helps with rotor current

regulation).

2. . Estimating the dynamic states—Reason

• S. Yu, T. Fernando, Tat Kei Chau and H. H.-C. Iu, “Voltage control strategies for

solid oxide fuel cell energy system connected to complex power grids using dynamic state estimation and STATCOM,” IEEE Transactions on Power Systems, DOI: 10.1109/TPWRS.2016.2615075. 04-Oct-2016

• S. Yu, T. Fernando, and H. H.-C. Iu, “Dynamic State Estimation Based Control

strategy for DFIG wind turbine connected to complex power systems,” IEEE Transactions on Power Systems, vol.32, no. 2, p.p. 1272-1281.

• Internal dynamic states are hard to acquire directly—

use the knowns to estimate the unknowns. where 𝜗 is measurable input vector and 𝜂 is measureable output vector.

• In a power system, voltages, currents and their phase angles are

measureable signals, and can be measured by PMUs.

3. . Estimating the dynamic states—How

𝑦 = Υ 𝑦, 𝜗 , 𝜂 = Ψ 𝑦, 𝜗 , 𝑦 = 𝑔 𝑦, 𝑣 , 0 = 𝑕 𝑦, 𝑣 ,

• Use different mathematical methods to perform dynamic state
• estimation. So far, we have developed several algorithms:
• State and Functional Observers to acquire internal dynamic states of

synchronous generators.

• Extended Kalman Filter for synchronous, asynchronous and stationary

power generators.

• Unscented Kalman Filter for synchronous, asynchronous and stationary

power generators.

• Particle Filter for synchronous, asynchronous and stationary power

generators. ISSUES: Model validation, model discretization, probabilistic model, PMU measurement noises, transmission time delays, etc.

3. . Estimating the dynamic states—How

The estimated dynamic states of power generators provide more information of the operating conditions, with which we could develop better control strategies to achieve a more favorable control performance.

4. . Estimating the dynamic states—Use

IEEE Transactions on Power Systems

• 1. FO-based LFC approach

System Control Signal

4. . Estimating the dynamic states—Use

• 1. FO-based LFC scheme

Schematic of the IEEE standard (New England) 39-bus test system studied Disturbances: Case 1: An increase in load demands on busbar 1, 12, 24. Case 2: An decrease in load demands on busbar 7, 21,24.

4. . Estimating the dynamic states—Use

• 1. FO-based LFC scheme

Important simulation results

4. . Estimating the dynamic states—Use

IEEE Transactions on Power Systems

• 2. DSE-based DFIG control strategy

IEEE Transactions on Power Systems

4. . Estimating the dynamic states—Use

• 2. DSE-based DFIG control scheme

Schematic of the IEEE standard (New England) 39- bus test system integrated with wind turbine power generation system Disturbance (fault): Transmission line between bus 21 and 22 is disconnected.

4. . Estimating the dynamic states—Use

Simulation results of DSE

• 2. DSE-based DFIG control scheme

Simulation results of DSE-based control

4. . Estimating the dynamic states—Use

• 3. DSE-based control strategies for solid oxide fuel cell power plant connected to a

complex power grid.

• 4. DSE-based sliding mode control strategies for wind turbine system.
• 5. DSE-based sliding mode control strategies for traditional power generators,

regulating busbar frequencies and voltage magnitudes.

• 6. DSE-based frequency restoration strategy for solar PV farm connected to power

grids. … See our official website for detailed information on our publications. pace.ee.uwa.edu.au

We are pacing towards a clean energy future !