Parallel computation of dynamic state estimation for power system
View Patent ↗Techniques are disclosed for performing the estimation and/or prediction of the dynamic system state of large power system networks, using a multi-processor approach. More particularly, a large power system network is divided into smaller sub-systems, each sub-system having associated dynamic state variables. Each sub-system is assigned to one or more of a plurality of processing elements, and dynamic state variables for each sub-system are estimated or predicted independently, using the processing elements and sensor measurements. In several embodiments, the dynamic state of each sub-system is computed through the construction of a set of dedicated observers, such as linear parameter-varying (LPV) observers, which are designed to reduce the effects of other sub-systems on the state estimation problem.
1. A method for monitoring the state of a power system network that includes one or more power generators, a plurality of power buses, and a plurality of monitoring sensors, the method comprising:
dividing the power system network into a plurality of sub-systems;
assigning each of one or more of the sub-systems to one or more of a plurality of processing elements, each sub-system having a plurality of associated dynamic state variables; and
estimating or predicting the associated dynamic state variables for each sub-system at a least real-time speeds using the plurality of processing elements and measurements from the plurality of monitoring sensors, such that the estimating or predicting for each sub-system is independent of the estimating or predicting for the remaining sub-systems.
2. The method of claim 1 , wherein said dividing of the power system network into the first plurality of sub-systems is based on one or more of the following:
the number of state variables in the power system network;
the number and locations of monitoring sensors in each potential sub-system;
a processing capability for one or more of the processing units; and
a number of processing units available.
3. The method of claim 1 , wherein said estimating or predicting the associated dynamic state variables for each sub-system comprises modeling the sub-system using a dynamic observer, the dynamic observer including a designed-observer gain that attenuates effects on the estimated dynamic state variables from unknown state estimates for other sub-systems.
4. The method of claim 3 , wherein the dynamic observer comprises a proportional integral (PI) dynamic observer.
5. The method of claim 3 , wherein the dynamic observer comprises a linear parameter varying (LPV) dynamic observer.
6. The method of claim 5 , further comprising designing the LPV dynamic observer for each of the sub-systems using a graphical observer design approach.
7. The method of claim 1 , further comprising selecting the number of sub-systems to match a number of available processing elements.
8. The method of claim 1 , wherein two or more of the processing elements are processor cores in a multi-processor apparatus.
9. The method of claim 1 , wherein two or more of the processing elements are distributed among two or more processing platforms.
10. The method of claim 1 , further comprising dynamically controlling one or more parameters in the power system network based on the estimated or predicted dynamic state variables.
11. The method of claim 1 , further comprising alerting a system operator to a system disturbance based on the estimated or predicted dynamic state variables.
12. The method of claim 1 , further comprising:
re-dividing the power system network to obtain re-divided sub-systems, wherein said re-dividing reflects changes in sensor locations or the availability of measurements from sensors, or both;
assigning each of the re-divided sub-systems to the processing elements; and
estimating or predicting the associated dynamic state variables for each re-divided sub-system at a least real-time speeds, using the plurality of processing elements and measurements from the plurality of monitoring sensors, such that the estimating or predicting for each re-divided sub-system is independent of the estimating or predicting for the remaining re-divided sub-systems.
13. A computerized system for monitoring the state of a power system network that includes one or more power generators and a plurality of power buses, the computerized system comprising:
a plurality of monitoring sensors distributed through the power system network; and
a plurality of processing elements, wherein each processing element is configured to have assigned to it a sub-system of the power system network, each sub-system having a plurality of associated dynamic state variables, wherein each processing element is configured to estimate the associated dynamic state variables for the sub-system at a least real-time speeds using the measurements from the plurality of monitoring sensors, and such that the estimating for each sub-system is independent of the estimating for the remaining sub-systems.
14. The computerized system of claim 13 , further including a sub-system identification processor configured to divide the power system network into a plurality of sub-systems, based on a number of available processing elements.
15. The computerized system of claim 14 , wherein the sub-system identification processor is configured to select the number of sub-systems to match a number of available processing elements.
16. The computerized system of claim 13 , further comprising a scheduling processor configured to assign sub-systems to processing elements, wherein said assigning is based on one or more of the following:
the number of sub-systems;
the number of monitoring sensors for each of one or more of the sub-systems; and
the processing capabilities of each of one or more of the processing elements.
17. The computerized system of claim 13 , wherein the processing elements are configured to estimate the associated dynamic state variables for each sub-system by modeling the sub-system using a dynamic observer, the dynamic observer including a designed-observer gain that attenuates effects on the estimated dynamic state variables from unknown state estimates for other sub-systems.
18. The computerized system of claim 17 , wherein the dynamic observer comprises a proportional integral (PI) dynamic observer.
19. The computerized system of claim 17 , wherein the dynamic observer comprises a linear parameter varying (LPV) dynamic observer.
20. The computerized system of claim 19 , wherein the processing elements are configured to design the LPV dynamic observer for each of the sub-systems using a graphical observer design approach.
21. The computerized system of claim 13 , wherein two or more of the processing elements are processor cores in a multi-processor apparatus.
22. The computerized system of claim 13 , wherein two or more of the processing elements are distributed among two or more processing platforms.