Back to results
US 12,607,662 B2Granted
Phase identification of power distribution systems
AssigneeMitsubishi Electric Research Laboratories, Inc.
InventorsHongbo Sun, Zong-Jhen Ye, Jianlin Guo, Ye Wang
Life of this patent
Estimating the term…
Abstract
Method and system for phase detecting of hybrid wye-connected and delta-connected loads in an unbalanced power distribution system. A data-driven event-triggered phase identification algorithm is presented in which the exceeding current fluctuation events are used to build the relationship between the feeder head and the undetermined phase load and mutual information index is used to assess the dependence level of the feeder head on the undetermined phase load. The connected and non-connected phase of wye-connected and delta-connected single-phase load are determined as the phase having highest and lowest dependences of the phases of feeder head on the load, respectively.
Claims21
A phase detecting system for detecting phases of loads in a power distribution system, comprising:
- a signal interface configured to receive measurements from sensors connected to feeder heads and loads on the power distribution system, wherein the measurements include a real power, a reactive power, and a voltage magnitude for each phase of the feeder heads and each phase of the loads, wherein the loads include a set of loads with undetermined phases;
- at least one processor;
- and
- a non-transitory computer-readable medium having stored thereon a set of instructions for detecting phases of the loads, which when performed by the at least one processor, causes the at least one processor to performs steps of:
- storing, by selecting from the measurements, for a set of time intervals, current magnitude timeseries for each phase of the feeder heads and each phase of the loads with undetermined phases in a memory, wherein the current magnitude timeseries include current magnitudes at measured time intervals, wherein the current magnitudes are computed from the real powers, the reactive powers, and the voltage magnitudes, wherein each of the real powers, each of the reactive powers, and each of the voltage magnitudes corresponds to an interval among the set of measuring intervals;
- selecting a load indicating a greatest current magnitude from the set of loads with undetermined phases as a study load by a statistic detecting algorithm;
- determining exceeding current fluctuation events from the current magnitude timeseries at each phase of the feeder heads and the study load based on a threshold current range;
- computing a first marginal probability distribution for the exceeding current fluctuation events on each phase of the feeder heads, a second marginal probability distribution for the exceeding current fluctuation events at the study load, and a joint probability distribution for joint exceeding current fluctuation events for each combination of phases of the feeder heads and the study load;
- determining a phase for the study load based on dependence of each phase of the feeder heads on the study load according to mutual information index determined by the first and second marginal probability distributions of the exceeding current fluctuation events and the joint probability distribution of the joint exceeding current fluctuation events;
- updating the current magnitude timeseries at the feeder heads by a current disaggregation algorithm for equivalent current of the study load based on measured real powers, reactive powers and voltage magnitudes at the study load, and removing the study load from the set of loads with undetermined phases stored in the memory;
- and
- determining if there is a load indicating an undetermined phase in the set of loads with undetermined phases in the memory, then continuing the steps of the selecting through the updating until the set of loads with undetermined phase becomes empty, otherwise generating and outputting a dataset of the determined phases for loads.
9 more elements
A phase detecting method for detecting phases of loads in a power distribution system, comprising steps of:
- receiving measurements from sensors via a signal interface, wherein the sensors are connected to feeder heads and loads on the power distribution system, wherein the measurements include a real power, a reactive power, and a voltage magnitude for each phase of the feeder heads and each phase of the loads, wherein the loads include a set of loads with undetermined phases;
- storing, by use of the measurements, for a set of time intervals, current magnitude timeseries for each phase of the feeder heads and each phase of the loads with undetermined phases in a memory, wherein the current magnitude timeseries include current magnitudes at measured time intervals, wherein the current magnitudes are computed from the real powers, the reactive powers, and the voltage magnitudes, wherein each of the real powers, each of the reactive powers, and each of the voltage magnitudes corresponds to an interval among the set of measuring intervals;
- selecting a load indicating a greatest current magnitude from the set of loads with undetermined phases as a study load by a statistic detecting algorithm;
- determining exceeding current fluctuation events from the current magnitude timeseries at each phase of the feeder heads and the study load based on a threshold current range;
- computing a first marginal probability distribution for the exceeding current fluctuation events on each phase of the feeder heads, a second marginal probability distribution for the exceeding current fluctuation events at the study load, and a joint probability distribution for joint exceeding current fluctuation events for each combination of phases of the feeder heads and the study load;
- determining a phase for the study load based on dependence of each phase of the feeder heads on the study load according to mutual information index determined by the first and second marginal probability distributions of the exceeding current fluctuation events and the joint probability distribution of the joint exceeding current fluctuation events;
- updating the current magnitude timeseries at the feeder heads by a current disaggregation algorithm for equivalent current of the study load based on measured real powers, reactive powers and voltage magnitudes at the study load, and removing the study load from the set of loads with undetermined phases stored in the memory;
- and
- determining if there is a load indicating an undetermined phase in the set of loads with undetermined phases in the memory, then continuing the steps of the selecting through the updating until the set of loads with undetermined phase becomes empty, otherwise generating and outputting a dataset of the determined phases for loads.
6 more elements
Dependencies are inferred from the claim text.