Electrical power system and a multi-timescale coordinated optimization scheduling method therefor
A multi-timescale coordinated optimization scheduling method for an electrical power system that includes a number of distributed energy resources (DERs) is disclosed. The method includes performing long-timescale optimization scheduling for the electrical power system based at least on renewable energy generation forecast data of the DERs to obtain long-timescale operation planning data. The method further includes performing mid-timescale optimization scheduling for the electrical power system based on the long-timescale operation planning data and measured data of the DERs to obtain mid-timescale operation planning data. The method further includes performing at least close to real-time optimization scheduling for the electrical power system based on the mid-timescale operation planning data, the measured data of the DERs and grid signals of the electrical power system to obtain short-timescale power setpoints for the DERs. An electrical power system in which the method is used is also disclosed.
1. A multi-timescale coordinated optimization scheduling method for an electrical power system including a plurality of distributed energy resources (DERs), the method comprising:
performing, via cloud computing using a cloud optimization engine, a long-timescale optimization scheduling for the electrical power system based at least on forecast data to obtain a long-timescale operation planning data;
determining if the long-timescale operation planning data is available;
performing, via edge computing using a model predictive control engine, a mid-timescale optimization scheduling for the electrical power system by solving a first optimization problem based on the long-timescale operation planning data and measured data of the plurality of DERs to obtain a mid-timescale operation planning data when it is determined that the long-timescale operation planning data is available;
performing, via edge computing using the model predictive control engine, the mid-timescale optimization scheduling for the electrical power system by solving a second optimization problem based on forecast data and the measured data of the plurality of DERs to obtain the mid-timescale operation planning data when it is determined that the long-timescale operation planning data is not available, the second optimization problem is different from the first optimization problem;
performing, via edge computing using a real-time predictive control engine, at least close to real-time optimization scheduling for the electrical power system based on the mid-timescale operation planning data, the measured data of the plurality of DERs and grid signals of the electrical power system to obtain a short-timescale power setpoints for the plurality of DERs; and
sending the short-timescale power setpoints to the plurality of DERs and controlling the short-timescale power setpoints;
wherein the long-timescale is longer than the mid-timescale, and the mid-timescale is longer than the short-timescale.
2. The multi-timescale coordinated optimization scheduling method according to claim 1 , wherein
the first optimization problem is defined by:
Q
UG
,
E
+
Q
UG
,
unc
+
∑
i
∈
ES
(
Q
ES
,
cyc
i
+
Q
ES
,
dev
i
)
+
∑
i
∈
CPV
Q
CPV
,
pen
i
+
∑
i
∈
CL
Q
CL
,
cur
i
+
∑
i
∈
GT
(
Q
GT
,
G
i
+
Q
GT
,
ref
i
)
,
and the second optimization problem is defined by:
Q
UG
,
E
+
Q
UG
,
unc
+
∑
i
∈
ES
(
Q
ES
,
cyc
i
-
Q
ES
,
FR
i
)
+
∑
i
∈
CPV
Q
CPV
,
pen
i
+
∑
i
∈
CL
Q
CL
,
cur
i
+
∑
i
∈
GT
(
Q
GT
,
G
i
+
Q
GT
,
SU
i
)
wherein
Q UG,E is an energy procurement cost;
Q UG,unc is an uncontracted capacity cost;
Q i CPV,pen is a curtailable photovoltaic (CPV) penalty cost;
Q i CL,cur is a controllable load (CL) curtailment cost;
Q i ES,cyc is a battery energy storage system (BESS) cycling aging cost;
Q i ES.FR is a BESS regulation capacity revenue;
Q i Es,dev is a BESS cost of deviation from engine charging/discharging power references for long-timescale optimization scheduling;
Q i GT,G is a gas turbine (GT) operation cost;
Q i GT,SU is a GT startup cost; and
Q i GT,ref is a GT cost of deviation from engine power references for long-timescale optimization scheduling.
3. The multi-timescale coordinated optimization scheduling method according to claim 1 , further comprising:
obtaining interpolated long-timescale operation planning data based on the long-timescale operation planning data; and
wherein performing, via edge computing, mid-timescale optimization scheduling for the electrical power system by solving a first optimization problem based on the long-timescale operation planning data and measured data of the plurality of DERs to obtain mid-timescale operation planning data comprises
performing, via edge computing, the mid-timescale optimization scheduling for the electrical power system by solving the first optimization problem based on the interpolated long-timescale operation planning data, and the measured data of the plurality of DERs to obtain the mid-timescale operation planning data.
4. The multi-timescale coordinated optimization scheduling method according to claim 3 , further comprising:
obtaining interpolated forecast data based on the forecast data; and
obtaining blended forecast data based on the interpolated forecast data and the measured data of the plurality of DERs; and
wherein performing, via edge computing, the mid-timescale optimization scheduling for the electrical power system by solving the first optimization problem based on the interpolated long-timescale operation planning data and the measured data of the plurality of DERs to obtain the mid-timescale operation planning data comprises
performing, via edge computing, the mid-timescale optimization scheduling for the electrical power system by solving the first optimization problem based on the interpolated long-timescale operation planning data, the blended forecast data and the measured data of the plurality of DERs to obtain the mid-timescale operation planning data.
5. The multi-timescale coordinated optimization scheduling method according to claim 4 , wherein obtaining at least one of the interpolated forecast data and the interpolated long-timescale operation planning data is performed asynchronously from the mid-timescale optimization scheduling.
6. The multi-timescale coordinated optimization scheduling method according to claim 1 , wherein performing the mid-timescale optimization scheduling comprises performing Model Predictive Control (MPC)-based mid-timescale optimization scheduling.
7. The multi-timescale coordinated optimization scheduling method according to claim 1 , wherein the long-timescale operation planning data comprises at least an hour-ahead planning data.
8. The multi-timescale coordinated optimization scheduling method according to claim 7 , wherein the mid-timescale operation planning data comprises at least a minute-ahead operation planning data corresponding to the at least an hour-ahead planning data.
9. The multi-timescale coordinated optimization scheduling method according to claim 1 , wherein the forecast data comprises at least one of price, load and renewable energy generation forecast data.
10. The multi-timescale coordinated optimization scheduling method according to claim 1 , wherein the long-timescale is at least half an hour, the mid-timescale is at least one minute and the short-timescale is at least four seconds.
11. An electrical power system comprising:
a plurality of distributed energy resources (DERs); and
an energy management system comprising:
a cloud computing-based long-timescale scheduler using a cloud optimization engine for a long-timescale optimization scheduling of the electrical power system based on forecast data to obtain a long-timescale operation planning data;
an edge computing-based mid-timescale scheduler using a model predictive control engine for a mid-timescale optimization scheduling of the electrical power system to obtain a mid-timescale operation planning data, wherein
the mid-timescale optimization scheduling is performed by solving a first optimization problem based on the long-timescale operation planning data and measured data of the plurality of DERs if it is determined that the long-timescale operation planning data is available;
the mid-timescale optimization scheduling is performed by solving a second optimization problem based on forecast data and the measured data of the plurality of DERs if it is determined that the long-timescale operation planning data is not available, the second optimization problem is different from the first optimization problem; and
an edge computing-based real-time scheduler using a real-time predictive control engine for at least close to real-time optimization scheduling of the electrical power system based on the mid-timescale operation planning data, the measured data of the plurality of DERs and grid signals of the electrical power system to obtain a short-timescale power setpoints for the plurality of DERs and send the short-timescale power setpoints to the plurality of DERs and controlling the short-timescale power setpoints;
wherein the long-timescale is longer than the mid-timescale, and the mid-timescale is longer than the short-timescale.
12. The electrical power system according to claim 11 , wherein
the first optimization problem is defined by:
Q
UG
,
E
+
Q
UG
,
unc
+
∑
i
∈
ES
(
Q
ES
,
cyc
i
+
Q
ES
,
dev
i
)
+
∑
i
∈
CPV
Q
CPV
,
pen
i
+
∑
i
∈
CL
Q
CL
,
cur
i
+
∑
i
∈
GT
(
Q
GT
,
G
i
+
Q
GT
,
ref
i
)
,
and the second optimization problem is defined by:
Q
UG
,
E
+
Q
UG
,
unc
+
∑
i
∈
ES
(
Q
ES
,
cyc
i
-
Q
ES
,
FR
i
)
+
∑
i
∈
CPV
Q
CPV
,
pen
i
+
∑
i
∈
CL
Q
CL
,
cur
i
+
∑
i
∈
GT
(
Q
GT
,
G
i
+
Q
GT
,
SU
i
)
wherein
Q UG,E is an energy procurement cost;
Q UG,unc is an uncontracted capacity cost;
Q i CPV,pen is a curtailable photovoltaic (CPV) penalty cost;
Q i CL,cur is a controllable load (CL) curtailment cost;
Q i ES,cyc is a battery energy storage system (BESS) cycling aging cost;
Q i ES,FR is a BESS regulation capacity revenue;
Q i ES,dev is a BESS cost of deviation from engine charging/discharging power references for long-timescale optimization scheduling;
Q i GT,G is a gas turbine (GT) operation cost;
Q i GT,SU is a GT startup cost; and
Q i GT,ref is a GT cost of deviation from engine power references for long-timescale optimization scheduling.
13. The electrical power system according to claim 11 , further comprising:
a data interpolator for obtaining interpolated long-timescale operation planning data based on the long-timescale operation planning data; and
wherein performing the mid-timescale optimization scheduling by solving the first optimization problem based on the long-timescale operation planning data and the measured data of the plurality of DERs if it is determined that the long-timescale operation planning data is available comprises
performing the mid-timescale optimization scheduling by solving the first optimization problem based on the interpolated long-timescale operation planning data, and the measured data of the plurality of DERs if it is determined that the long-timescale operation planning data is available.
14. The electrical power system according to claim 13 , wherein:
the data interpolator further obtains interpolated forecast data based on the forecast data; and
wherein performing the mid-timescale optimization scheduling by solving the first optimization problem based on the interpolated long-timescale operation planning data and the measured data of the plurality of DERs if it is determined that the long-timescale operation planning data is available comprises
performing the mid-timescale optimization scheduling by solving the first optimization problem based on the interpolated long-timescale operation planning data, the interpolated forecast data and the measured data of the plurality of DERs if it is determined that the long-timescale operation planning data is available.
15. The electrical power system according to claim 14 , wherein the data interpolator obtains at least one of the interpolated forecast data and the interpolated long-timescale operation planning data asynchronously from the mid-timescale optimization scheduling.
16. The electrical power system according to claim 11 , wherein the mid-timescale optimization scheduling comprises Model Predictive Control (MPC)-based mid-timescale optimization scheduling.
17. The electrical power system according to claim 11 , wherein the long-timescale operation planning data comprises at least one hour-ahead planning data.
18. The electrical power system according to claim 17 , wherein the mid-timescale operation planning data comprises at least a minute-ahead operation planning data corresponding to the at least one hour-ahead planning data.
19. The electrical power system according to claim 11 , wherein the forecast data comprises at least one of price, load and renewable energy generation forecast data.
20. The electrical power system according to claim 11 , wherein the long-timescale is at least half an hour, the mid-timescale is at least one minute and the short-timescale is at least four seconds.