State Estimation of Power System

Reference:
John J. Grainger, William D.
Stevenson, Jr. "Power System
Analysis
What is a State?
 The complete solution of the power system is known if all voltages and angles are
identified at each bus. These quantities are State Variables of the system.

Define State Estimation.

 State Estimation is a process to estimate the electrical state of a network by
eliminating inaccuracies and errors from measurement data.
Why Estimate?
 Meters are not perfect.
 Meters are not everywhere.
 Very few phase measurements.
 State estimation suppresses bad measurements and uses the measurement to the
fullest extend.
 Why state estimation is necessary?
•Power system control needs operational data which are
measured and sent to the server on‐line through SCADA.

•These are random noise and/or gross error corrupted

•Cleaning up is essential

•State estimator does it applying a statistical criterion to a

redundant set of measurements (i.e. M > N; N is the number
of state variables basically V,θ).

•It first estimates the best set of system states and using
those recalculates a set of consistent measurements close to
the exact values with a certain level of confidence.
 Load flow also computes unknown V, θ. Then what is the difference between LF and SE?

n number of
[Δθ]i+1 = [JθP]‐1 [P‐f1(V,θ)]i+1
Pinj and Qinj n No. of [θ, V]
measurements which were
or database [ΔV]i+1 = [JθP]‐1 [Q‐f2(V,θ)]i+1 previously
values unknown
Load flow
(no filtering action on measurement noise)

M number (M>N)of
measurements N number of
Z ={injections, line flows, bus [ΔX]i+1 = [G]‐1 [H]T [R]‐1[Z‐h(x)]i+1
states X ={θ, V
voltage magnitudes, at all buses}
phase angles from PMU if any} State estimator using WLS criterion
 Voltmeter
PT
Ammeter Analog to Digital
Converter
CT Watt meter
Transmission
Line/ Transformer/ Bus VAR meter
Communication
Link
Accept or Reject data if
certain measures of
accuracy are exceeded. Energy Control
System Operator State Estimation Center

Contains data with

random error and noise
Method of Least Square: The best estimates
are chosen as those which minimize the
weighted sum of the squares of the
measurement errors.
 Method of Least Square:

Explanation with DC Circuit Example

 Measurements

Unknown Unknown
(To be estimated) (To be estimated)

x1 x2

Measurement Measurement

z → Quantities being measured

x → Quantities to be estimated
 From Eq. (1) and (2),
1 2
𝑥1 = 3 3𝑧4 − 𝑥2 − 𝑧4 = 8𝑧4 − 3𝑥2
I1 I2 I3
1 3
𝑧4 = 𝑥1 + 𝑥2 … … … … … … … … … (5)
I4 8 8

x1 x2 From Eq. (1) and (5),

I5
1 3
𝑥1 = 3𝑧3 − 𝑥1 − 𝑥2
8 8
Applying KCL at node- 1: 3 1
𝑧3 = 𝑥1 + 𝑥2 … … … … … … … … … (6)
8 8
𝐼1 + 𝐼2 = 𝐼4
(𝑥1 −𝑧3 ) + (𝑧4 − 𝑧3 ) = 𝑧3 From Eq. (4) and (5),
𝑥1 = 3𝑧3 −𝑧4 … … … … … … … … … (1) 1 3
𝑧2 = 𝑥2 − 𝑥1 + 𝑥2
Applying KCL at node- 2: 8 8
5 1
𝐼2 + 𝐼5 = 𝐼3 𝑧2 = 𝑥2 − 𝑥1 … … … … … … … … … (7)
𝑧4 − 𝑧3 + 𝑧4 = 𝑥2 − 𝑧4 8 8
𝑥2 = 3𝑧4 −𝑧3 … … … … … … … … … (2) From Eq. (3) and (6),
Again, 3 1
𝑧1 = 𝑥1 − 𝑥1 − 𝑥2
𝑥1 − 𝑧3 = 𝑧1 … … … … … … … … … (3) 8 8
5 1
𝑥2 − 𝑧4 = 𝑧2 … … … … … … … … … (4) 𝑧1 = 𝑥1 − 𝑥2 … … … … … … … … … (8)
8 8
 Labelling the coefficients of measurement Eqns.:

x1 x2

Meter readings contains errors that generally follow a

statistical pattern. So the measurement equations: Where, z j,true denote the true value of the measured
quantity z j.

Now, rearranging the above equations in vector-

matrix form:

In more compact form.:

𝐞 = 𝐳 − 𝐳𝐭𝐫𝐮𝐞 = 𝐳 − 𝐇𝐱 … … … (𝟗)
 𝐞 = 𝐳 − 𝐳𝐭𝐫𝐮𝐞 = 𝐳 − 𝐇𝐱 … … … (𝟗)
The Eq. (9) above represents the error between actual measurement z and the true (but unknown) values
ztrue ≜ 𝐻𝑥 of the measured quantities.
The true value of x1 and x2 cannot be determined. But we can calculate the estimates of x1 and x2 which
are- 𝑥1 and 𝑥2

Hence, substituting the actual values of x with their estimates in Eq. (9),

𝒆 = 𝐳 − 𝒛 = 𝐳 − 𝐇𝒙 = 𝒆 + 𝑯𝒙 − 𝐇𝒙 = 𝐞 − 𝐇(𝒙 − 𝒙) … … … (𝟏𝟎)
Expanding the above Eq. in vector-matrix form:

The quantities with circumflex are the estimated values.

 Criterion for calculating the estimates- 𝒙𝟏 and 𝒙𝟐

Minimize the following weighted square error objective function:

Why weighted square error objective function?

 The positive and negative errors may cancel each other, which may result wrong estimates. That’s
why error terms squared before summation instead of direct algebraic summation.
 To ensure the measurements from meters of greater accuracy to be treated more favorably than
less accurate measurements, weights are introduced with each error term.
 Meter with higher accuracy is assigned higher weight than the meter with lower accuracy.
 Minimization of Objective function
Weighted square error objective function:

According to the usual necessary conditions for minimizing f, the estimates 𝒙𝟏 and 𝒙𝟐 are those values of
x1 and x2 that satisfy the following criterions:

The notation 𝒙 indicates that the equations have to be evaluated from the state estimates 𝒙 = [𝒙𝟏 𝒙𝟐 ]
since the true values of the states are not known. The unknown actual errors ej are then replaced by
estimated errors 𝒆𝒋 , which can be calculated once the state estimates 𝒙𝒋 are known.
 Let’s recall the system of Eqns.
Necessary conditions for minimizing the objective function, f:

Differentiating the above Eqns.

w.r.t. x1:
The above equations in vector-matrix form: 𝜕𝑒1 𝜕𝑒2 𝜕𝑒3
= ℎ11 = ℎ21 = ℎ31
𝜕𝑥1 𝜕𝑥1 𝜕𝑥1
𝜕𝑒4
= ℎ41
𝜕𝑥1
Differentiating w.r.t. x2:
𝜕𝑒1 𝜕𝑒2 𝜕𝑒3
= ℎ12 = ℎ22 = ℎ32
𝜕𝑥2 𝜕𝑥2 𝜕𝑥2
𝜕𝑒4
= ℎ42
𝜕𝑥2
 Calculation of Least Square Estimate 𝜕𝑒1
= ℎ11
𝜕𝑒2
= ℎ21
𝜕𝑒3
= ℎ31
𝜕𝑥1 𝜕𝑥1 𝜕𝑥1
Necessary conditions for minimizing the objective function, f in
vector-matrix form: 𝜕𝑒4
= ℎ41
𝜕𝑥1

𝜕𝑒1 𝜕𝑒2 𝜕𝑒3

= ℎ12 = ℎ22 = ℎ32
𝜕𝑥2 𝜕𝑥2 𝜕𝑥2
𝜕𝑒4
= ℎ42
𝜕𝑥2

Replacing the derivative terms:

𝐞 = 𝐳 − 𝐳𝐭𝐫𝐮𝐞 = 𝐳 − 𝐇𝐱 … … … (𝟗)
𝒆 = 𝐳 − 𝒛 = 𝐳 − 𝐇𝒙 … … … … (𝟏𝟎)

⇒ 𝐻 𝑇 𝑊𝑧 = 𝐻 𝑇 𝑊𝐻𝑥 ⇒ 𝑥 = 𝐻 𝑇 𝑊𝐻 −1 𝐻𝑇 𝑊𝑧

Using the compact notation yields: G

⇒ 𝒙 = 𝑮−𝟏 𝑯𝑻 𝑾𝒛 … … … … (𝟏𝟏)
𝐻𝑇 𝑊𝑒 = 𝐻𝑇 𝑊 𝑧 − 𝐻𝑥 = 0
 Calculation of Difference Between Estimated State and Actual State
𝒙 = 𝑮−𝟏 𝑯𝑻 𝑾𝒛 is the least square estimate of the State variables.

𝐺 = 𝐻 𝑇 𝑊𝐻 −1 is called the Gain Matrix.

𝒙 is the close estimate of the state variable x. The difference between actual and estimated value of the
state variable can be found by the following procedure:

⇒ 𝐱 = 𝐱 + 𝐆−𝟏 𝐇 𝐓 𝐖𝐞

Reminder :

𝐞 = 𝐳 − 𝐳𝐭𝐫𝐮𝐞 … … … … (12)
= Measurement – True value

Hence, any one or more of the errors can influence the difference between each state estimate and it’s
true value.
 Calculation of Difference Between Estimated Measurement and Actual Measurement

Let ′ s recall Eqn. (10)

𝒆 = 𝐳 − 𝒛 = 𝐳 − 𝐇𝒙 = 𝒆 + 𝑯𝒙 − 𝐇𝒙 = 𝐞 − 𝐇(𝒙 − 𝒙) … … … (𝟏𝟎)
Now, put 𝒙 − 𝒙 = 𝑮−𝟏 𝑯𝑻 𝑾𝒆

𝒆 = 𝐞 − 𝐇𝑮−𝟏 𝑯𝑻 𝑾𝒆 = 𝐈 − 𝐇𝑮−𝟏 𝑯𝑻 𝑾 𝒆 … … … (𝟏𝟑)

Here, I is an identity matrix

The use of Eqn. 12 and 13 is not possible without knowing the actual error e. These
equations are used for analytical purpose
Conceptual example on WLS state estimation using a DC circuit

Solution:
Let x1 = V1, x2 = V2
By applying basic circuit theorem e.g.
superposition principle the measurements
Z can be related to state variables and network
parameters (resistances) as shown.

[Z] = [H] [X]

[R]‐1 = [W]