Root cause and system enhancement analysis

• US 8,239,070 B1
• Filed: 09/19/2011
• Issued: 08/07/2012
• Est. Priority Date: 06/27/2003
• Status: Expired due to Fees

• Current Assignee: Intellicon, Inc.
• Original Assignee: Intellicon, Inc.
• Inventors: Minshall, Benjamin A., Hunt, Ryan J., Schlueter, Robert A.
• Primary Examiner(s): Cottingham, John R.
• Assistant Examiner(s): LEE, DOUGLAS S
• Application Number:
  US13/236,441
• Time in Patent Office: 323 Days
• Field of Search: 700/286, 700/292, 700/293, 700/295, 702/182, 703/18
• US Class Current: 700/286
• CPC Class Codes: H02J 2203/20 Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
• CPC Class Codes: H02J 3/00 Circuit arrangements for ac mains or ac distribution networks
• CPC Class Codes: H02J 3/0012 Contingency detection
• CPC Class Codes: Y02E 60/00 Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
• CPC Class Codes: Y04S 40/20 Information technology specific aspects, e.g. CAD, simulation, modelling, system security

Abstract

A method of assessing an electrical power system is provided that includes performing a plurality of contingencies using a mathematical model of the system and determining a representative solution for each contingency. For each bus and contingency, symptoms of system stress exhibited in the representative solution are tabulated. Each contingency is categorized into a contingency cluster based on its representative solution, and those buses having a highest number of the symptoms of system stress for each contingency categorized within a respective contingency cluster are sorted into a group associated therewith. The contingency clusters are arranged into a ranking based on its effect on the network, and a location for a network enhancement is selected from a group of buses associated with a contingency cluster of the ranking that has a largest effect on the network and also includes buses as a result of the sorting.

Claims

1. A method of assessing a condition of an electrical power system including a network of buses, at least certain of the buses including a respective reactive source, the method comprising:

(I) performing a plurality of contingencies using a mathematical model of the network manipulated by a processor;

(II) determining a representative solution for each contingency of the plurality of contingencies;

(III) for each bus in the network, tabulating symptoms of system stress exhibited by the bus in the representative solution for each contingency;

(IV) categorizing each contingency into one of at least two contingency clusters based on its representative solution;

(V) sorting those of the buses having a highest number of the symptoms of system stress summed over those contingencies categorized within a respective contingency cluster into a respective bus group associated with the respective contingency cluster;

(VI) arranging each of the at least two contingency clusters into a ranking based on its effect on the network; and

(VII) selecting a location for a network enhancement from a group of buses associated with a contingency cluster of the ranking that has a largest effect on the network and also includes buses as a result of the sorting.

2. The method of claim 1 wherein

performing a plurality of contingencies comprises performing a plurality of single contingencies and performing a plurality of multiple contingencies, each multiple contingency obtained by combining at least two of the single contingencies of the plurality of single contingencies;

and wherein

categorizing each contingency comprises categorizing each single contingency into one of at least one single contingency cluster based on its representative solution and categorizing each multiple contingency into one of at least one multiple contingency cluster based on its representative solution.

3. The method of claim 2 wherein

the at least one single contingency cluster comprises an unsolved single contingency cluster and a solved single contingency cluster and the at least one multiple contingency cluster comprises an unsolved multiple contingency cluster and a solved multiple contingency cluster;

wherein

the ranking comprises an order of the unsolved single contingency cluster, the unsolved multiple contingency cluster, the solved single contingency cluster and the solved multiple contingency cluster;

wherein

the solved single contingency cluster includes single component contingencies that do not solve using a first solution method but solve using a second solution method;

wherein

the solved multiple contingency cluster includes multiple component contingencies that do not solve using the first solution method but solve using the second solution method;

wherein

the unsolved single contingency cluster includes single component contingencies that do not solve using either the first solution method or the second solution method; and

wherein

the unsolved multiple contingency cluster includes multiple component contingencies that do not solve using either the first solution method or the second solution method.

4. The method of claim 3, further comprising:

grouping a single contingency that is a common component of a plurality of multiple component contingencies that do not solve using either the first solution method or the second solution method in the unsolved single contingency cluster.

5. The method of claim 1 wherein determining the representative solution for each contingency comprises:

representing the contingency in the mathematical model as an equivalence model of equipment associated with the contingency, the equivalence model having parameters proportional to a percentage of the contingency removed;

attempting a solution to the mathematical model with the percentage of the contingency removed in a single step being 100%;

using the solution to the mathematical model with the percentage of the contingency removed in the single step being 100% as the representative solution for the contingency when the mathematical model converges;

when the mathematical model does not converge with the percentage of the contingency removed in the single step being 100%,

A) selecting the percentage of the contingency removed as a new value of greater than 0% and less than 100%;

B) attempting a solution to the mathematical model with the percentage of the contingency removed being the new value;

C) if the mathematical model converges with the percentage of the contingency removed being the new value, increasing the new value if the new value is below 100%;

D) if the mathematical model does not converge with the percentage of the contingency removed being the new value, decreasing the new value such that the new value remains above 0%;

E) repeating B), C) and D) at a step until the mathematical model does not converge for two steps where the percentage of the contingency removed for the two steps is equal to a predefined minimum step value or until the mathematical model converges with the new value being 100%;

F) if the mathematical model converges with the new value being 100%, using the solution to the mathematical model with the new value being 100% as the representative solution; and

G) if the mathematical model does not converge in E), selecting a solution where the mathematical model converges for the new value closest to 100% as a boundary case solution, and predicting the representative solution based on the boundary case solution.

6. The method of claim 5 wherein

performing a plurality of contingencies comprises performing a plurality of single contingencies and performing a plurality of multiple contingencies, each multiple contingency obtained by combining at least two of the single contingencies of the plurality of single contingencies;

wherein

categorizing each contingency comprises categorizing each single contingency into a solved single contingency cluster when its representative solution is the solution to the mathematical model with the percentage of the single contingency removed in a single step being 100% and otherwise categorizing it into an unsolved single contingency cluster and comprises categorizing each multiple contingency into a solved multiple contingency cluster when its representative solution is the solution to the mathematical model with the percentage of the multiple contingency removed in a single step being 100% and otherwise categorizing it into an unsolved multiple contingency cluster; and

wherein

the ranking comprises an order of the unsolved single contingency cluster, the unsolved multiple contingency cluster, the solved single contingency cluster and the solved multiple contingency cluster.

7. The method of claim 5 wherein predicting the representative solution comprises scaling at least certain values obtained in the boundary case solution based on the percentage of the contingency removed in the boundary case solution.

8. The method of claim 5 wherein predicting the representative solution comprises predicting a loadflow variable by calculating a factor equal to:

a per unit change in a value of the loadflow variable in a base case solution to

a value of the loadflow variable in the boundary case solution/the percentage

of the contingency removed in the boundary case solution as a fraction; and

adding the value of the loadflow variable in the boundary case solution to the factor multiplied by the value of the loadflow variable in the boundary case solution.

9. The method of claim 8 wherein the at least one loadflow variable comprises at least one of a bus voltage, a current value, an active power flow on a branch, a reactive power flow on a branch or reactive power out of a generator.

10. The method of claim 5 wherein B) comprises attempting the solution to the mathematical model by enabling at least one of the following voltage control options:

(i) tap changers;

or

(ii) switchable capacitors; and

by electing one of an active power mismatch simulation including;

(a) governor power flow;

(b) AGC power flow;

(c) optimal power flow;

or

(d) dynamic simulation approximation snapshots.

11. The method of claim 1 wherein the symptoms of system stress exhibited by the bus in the representative solution comprise at least one of a voltage violation of the bus, a thermal overload of a branch connected to the bus, a surge impedance loading of a branch connected to the bus or an outage of the branch connected to the bus.

12. The method of claim 1 wherein tabulating symptoms of system stress exhibited by the bus in the representative solution for each contingency comprises incrementing a count for the bus:

A) when the bus is exhibiting a voltage violation in the representative solution;

B) for each branch connected to the bus exhibiting a thermal overload in the representative solution; and

C) for each branch connected to the bus exhibiting surge impedance loading in the representative solution; and

D) for each branch connected to the bus that is associated with the contingency; and

wherein sorting is performed using the count.

13. The method of claim 1, further comprising:

arranging each group of buses associated with the respective contingency cluster into one or more subgroups of buses, wherein each bus of a subgroup is related to another bus of the subgroup.

14. The method of claim 13 wherein each bus of the subgroup is physically coupled to another bus of the subgroup, the method further comprising:

arranging the subgroups of buses in order within each contingency cluster based on their effect on the network; and

wherein selecting a location for a network enhancement from a group of buses associated with a contingency cluster of the ranking that has a largest effect on the network and also includes buses as a result of the sorting comprises selecting the location from a subgroup of the contingency cluster that has the largest effect on the network.

15. The method of claim 1, further comprising:

using the representative solution to determine a set of reactive sources of the network exhausted by each contingency;

defining a plurality of agents, each agent comprising a plurality of buses such that when a reactive load is applied to any bus within the plurality of buses, a same at least one reactive source generally completely exhausts its reactive reserves;

arranging the plurality of agents into a hierarchy such that agents at a higher level in the hierarchy include more sources of reactive reserves than agents at a lower level in the hierarchy and wherein at least some of the agents at lower levels in the hierarchy have reactive reserves that are subsets of the reactive reserves of agents at higher levels in the hierarchy; and

wherein family lines of agents are defined based on agents at lower levels in the hierarchy having the subsets of the sources of reactive reserves contained in agents at higher levels in the hierarchy and a patriarch agent has the largest set of sources of reactive reserves in one or more family lines of agents; and

defining subareas of the network as at least one agent where a same set of reactive sources of the network is exhausted by a contingency.

16. The method of claim 15 wherein selecting a location for a network enhancement comprises selecting the location based on an arrangement of the subareas of the network, the subareas forming at least one of:

a subregional subarea that comprises a family line of agents that experiences propagation of voltage instability in successively larger patriarch agents in families of agents;

a regional subarea that comprises families of subareas with multiple family lines of subareas, each with a separate root cause subarea and each with a patriarch subarea; and

an inter-regional subarea that comprises a starfish shape having arms comprised of different family lines or different families of subareas, each with a patriarch subarea, and having a body comprised of a network of patriarch subareas.

17. The method claim 15 wherein selecting a location for a network enhancement comprises selecting the location from a lowest agent of a subarea for the contingency cluster.

18. The method of claim 1, further comprising:

selecting the network enhancement;

testing the network including the network enhancement by repeating (I) through (VI) using the mathematical model of the network updated with the network enhancement.

19. The method of claim 18, further comprising:

if the testing shows that the at least two contingency clusters include buses, performing at least one of selecting an additional network enhancement or replacing the network enhancement with a different network enhancement and testing the network using the mathematical model of the network updated with at least one of the additional network enhancement or with the different network enhancement in place of the network enhancement.

20. The method of claim 1, further comprising:

selecting a type and a size for the network enhancement at the location.

21. The method of claim 1 wherein the network enhancement comprises at least one of an additional line, a re-conductored line, a raised voltage rating of a line, series compensation of a line, static VAR compensation (SVC), voltage rescheduling, active rescheduling or load shedding.