Mitsubishi Electric Research Laboratories, Inc.

Mitsubishi Electric Corporation

Hongbo Sun

Kyeong Jin Kim

Jianlin Guo

Kyeong-Jin Kim

Shahil Shah

Ariana Minot

Anamika Dubey

Zhao Wang

Shiva Poudel

Tomihiro Takano

Zhenhua Wang

Yishen Wang

Chanaka Keerthisinghe

Sriramya Bhamidipati

Zhenyu Tan

Tong Huang

Gang Wang

Shengyi Wang

Fangyuan Chang

Ashwin Shirsat

Imtiaj Khan

Vanja Svenda

Guangyu Feng

Bowen Zhang

A diagnosis system is provided for diagnosing the partial discharge faults in a power transformer. The system includes a feature extractor, a fault classifier and a domain discriminator and whose parameters are optimized by minimizing a combination of feature loss, classifier loss and domain discrepancy discriminator loss. The feature loss is a weighted sum of cosine similarity distance loss, batch-based instance separation loss, and batch-based feature decorrelation loss. The measured applied voltages and partial discharge voltages are pre-processing using sliding window method to represent the each partial discharge event as a series of statistics based moments for a set of overlapped time blocks within the duration of fault event, wherein statistics based moments include an average value of applied voltage for representing the applied voltage magnitude and fault event occurring moment for each time block, and the mean, standard deviation, Kurtosis, and Skewness of partial discharge voltages for representing the variations of partial discharging within each time block.

Transfer learning-based power transformer partial discharge diagnosis

A control system collects a graph-based representation of the power distribution system as a radial tree of nodes connected by edges. Each node represents a section of the power distribution system isolated by switches and each edge represents a switch that connects a section represented by the node with a neighboring section represented by the neighboring node. Each node has a property of power demand of a corresponding section governed by a difference between demanded and available energy for the section, and each edge has a property constraining throughput of energy through the switch in a direction of a power flow. The control system solves an optimization problem to determine the states of the edges and controls the switches based on the optimized states of the corresponding edges.

Systems and methods for blackout rotation enabled control of power distribution systems

Disclosed is a method and system for identifying an existence, location and type of a weak-signal fault in an islanded inverter-based microgrid. The weak-signal fault includes a high impedance fault, an inverter DC-side short-circuit fault, and an inverter tripping fault, and usually fails to be detected by conventional relay methods due to small magnitude of fault current. Upon received voltage and current measurements from intelligent electronic devices installed in the microgrid, the variation mode decomposition algorithm is firstly applied to detect the existence of fault based on denoised time series of measurements using discrete wavelet transform algorithm. After detecting the presence of fault, the correlation-based matrix is applied to locate the suspicious fault locations, and then K-nearest neighbors model is utilized to identify the faulty branch among those locations using dynamic time warping algorithm to measure the distance between neighbors. Following fault localization, fault classification is done by observing sequence components and phasor measurements and feeding the observational inputs to a fault classification logic circuit model.

Weak-signal fault identification of inverter-based microgrids

A computer-implemented method is provided for performing energy disaggregation of a distribution system-level net-load measurements using continuous-point-on-wave (CPOW) measurement units. The method uses a processor coupled with a memory stored instructions implementing the method using neural networks including an encoder network, a feature extractor, a separator network, a decoder network stored in the memory, wherein the instructions, when executed by the processor carry out at steps of the method include generating net-load time series data from voltage and current measurements via the CPOW measurement units, generating a compressed latent space representation from the net-load time series, converting the net-load time series into time-frequency domain, passing time domain cotextual information with the converted time-frequency domain representation of net-load time series to the feature extractor, estimating two weight matrices to be multiplied with an output from the encoder network and learning temporal features of a native load and a photovoltaic (PV) generation, transforming weighted latent representation corresponding the native load and the PV generation into time-domain representations, and predicting the native load and the PV generation at distribution system-level from the transformed time domain representations corresponding to the native load and PV generations.

Extraction of instantaneous renewable generation from net load measurements of power distribution systems

Embodiments for distributing energy for an energy system having an energy generation source, an energy storage system and a load. The method including identifying a risk level for the energy system, the risk level having objectives prioritized relative to one another. Calculating an objective function based on values including energy market values, electricity rates, and power producing, storing and consumption. Identifying an Optimized solution for charging or discharging the energy storage system based on the objective function. Controlling the distribution of energy to the energy storage system for charging according to the Optimized solution based on the objective function and discharging according to the Optimized solution. The objective function is Optimized using feasibility constraints, again Optimized using technical constraints and additional constraints. The Optimized solution maximizes an expected total pool market revenue and an expected total future market revenue, while minimizing an expected total energy cost for the energy system.

Risk-constrained optimization of virtual power plants in pool and future markets

Disclosed a method and system for fault location and protection of Microgrids which are operated under islanded mode by being disconnected from distribution systems under large disturbance or disaster scenarios. The disclosed method is invented to meet the protection needs for disconnected ungrounded microgrids without appropriate protection systems that are usually installed only at distribution substations. Moreover, the disclosed protection method or scheme is targeted at inverter-dominated microgrids in which fault current limiters are installed to protect inverters introduced by the increasing penetration of inverter-based distributed generators (IBDGs). Based on transient analysis and dynamic simulation of islanded ungrounded microgrids during different types of faults, it is realized that of the first terminal and second terminal sensors of the branch different types of faults can be effectively detected based on sequent components of currents flowed through terminals at faulted branches or sections, even with multiple IBGDs deployed in the Microgrids under different control strategies. The phase angle differences of zero sequence currents, and magnitude differences of negative sequence currents, and sign changes of phase currents are used to locate and protect against unsymmetrical and symmetrical faults within the islanded microgrids with lower fault currents, respectively.

Method and system for fault location and protection of inverter-dominated islanded ungrounded microgrids

Systems and methods for identifying a fault condition in an Ungrounded Electrical Distribution (UED) system, the system receives measurements with instantaneous values and effective values associated when a fault event is identified, measured transient waveforms and a fault type. A processor applies an empirical mode decomposition to the measured transient waveforms to extract a dominant vibration mode and an associated derived waveform corresponding to the dominant vibration mode. A Hilbert transform is applied to the associated derived waveform to obtain a set of feature attributes. Subsets are computed from the set, at a pre-fault time, at a fault inception time, and at a post-fault time, and inputted into the fault type trained neural network model. An output of the model are locational parameters used to determine a fault section, a fault line segment and a fault location point with a topology connectivity analysis of the UED system.

Transient based fault location method for ungrounded power distribution systems

Systems and methods performed by a fault detection apparatus for fault detection and localization in distribution feeders having branches and nodes. The method including receive feeder raw data in a feeder of a power system. Process the feeder raw data with given operational electrical characteristics of the feeder to generate a branch attribute dataset for each branch separated by a pair of nodes for all branches. Generate a node attribute dataset for each node for all the nodes in the feeder. Input the branch and node attribute datasets into a trained neural network to determine whether a branch has a fault and a fault location within the branch, to output a classification of the fault and the fault location. Generate an alert signal based upon determining the classified fault and fault location in response to the alert signal to an outage response system.

Distribution fault location using graph neural network with both node and link attributes

A system for a region including a processor to receive data including weather and component data. Generate a dataset for the region to identify weather events and a plurality of model forcings converted into tabular form. A machine learning (ML) model for each model forcing for each component uses dataset. Iteratively, for each weather event: identify for each component a corresponding model forcing with weather variables. Generate an output value for ML model corresponding to the identified components to the set of weather variables and update ML model. Receive observed data over time periods of an impending weather (IW) event. Iteratively, for each time period: identify, for each component a corresponding model forcing with weather variables and update ML model. Generate for the updated ML model, an output value predicting a component status as a failed or not failed for the time period.

Weather-related overhead distribution line failures online forecasting

A time synchronization system an input interface configured to receive pseudorange measurements from a set of antennas of at least one Global Navigation Satellite System (GNSS) receiving system that operates the antennas at known positions according to a clock with an unknown time bias to receive signals from satellites in sight of the antennas and an output interface configured to output the determined time bias. The time synchronization system also includes a processor configured to compare the pseudorange measurements of different antennas to produce an antenna-specific clock offset for each of the antennas, correct the pseudorange measurements of the antennas according to the antenna-specific clock offsets of corresponding antennas to produce corrected pseudorange measurements, and determine the time bias of the clock using the corrected pseudorange measurements.

GNSS-based timing synchronization

Systems and methods for coordinating network and control parameters of a power distribution system (PDS) with interconnected microgrids in response to a subset of interconnected microgrids entering island-mode due to a predicted future disaster, generating samples of network and control parameter combinations, determining optimal adjustments of network and control parameters with respect to the disaster condition, determining optimal set of network and control parameters to be reinforced or adjusted, activating the parameter adjustments and reinforcements on the determined tie lines and PCCs of the microgrids.

Methods and systems for a holistic framework for parameter coordination of interconnected microgrid systems against disasters

Systems and methods to control activation of a generation unit. Compute an offer amount, representative of a value at which power is available to be supplied by a generation unit for a current or each subsequent upcoming time period in a time window. Construct energy supply curves using a set of points for each time period in the time window, to obtain associated amounts of power to supply with corresponding prices per each time period. Use a gradual decreasing price staircase approach to iteratively determine multiple offer amount scenarios. Submit the multiple offer amount scenarios along with corresponding values representative of available quantities of power to an operator computer that controls a market-based resource allocation system. Receive back a submitted supply value for the current or upcoming time period. Compare the submitted supply value to the submitted offer amount, and activate or deactivate the generation unit based on the comparison.

Day-ahead joint generation scheduling and bidding optimization for power plants

Systems and methods are configured for detecting and locating simultaneous faults in a distribution network having primary and secondary buses connected with two-terminal sections. The measured three-phase voltages and currents from the primary buses are received via a communication network, and the series impedances and shunt admittances and the pre-fault connectivity topology are retrieved from the storage. Simultaneous faults are determined to occur by verifying at least one phase of at least three buses having phase current mismatches determined as the differences between the measured values and calculated values using normal topology and phase voltages that are greater than a threshold. A location for each fault is determined individually if the faults are occurring at non-adjacent sections in the distribution network, or jointly if the faults are occurring at adjacent sections in the distribution network. Faulted line sections are isolated by activating switching operations for connected switches to the faulted line sections that communicatively linked to distribution system grid via communication network.

Simultaneous fault detection and location of power distribution systems

A control system for controlling a power distribution network including a set of devices for providing power commands the devices to reach a consensus on an average of uniquely weighted states of the devices based on iterative exchange of a uniquely weighted state of each device with its neighboring devices. The system determines the values of states of each of the devices corresponding to the received average of uniquely weighted states, determines a new value of the state of at least on device allowing to reroute distribution of the power, and commands to the device to change the state to the new value.

Post-disaster topology detection and energy flow recovery in power distribution network

Systems and methods for configuring micro-grids to restore some power in a power distribution grid (PDG) in response to a power disruption over the PDG. A computing system configured to receive current condition information from devices in the PDG. Form a minimum spanning forest (MSF) to identify a set of micro-grids, each spanning tree in a forest is a self-sustained islanded micro-grid network. Assign a ranking to each inter-bus link within each micro-grid according to constraints, to identify some inter-bus links above a high-ranking threshold to be switched off during a restoration period. Identify switches that restore power to some critical loads of a subset of critical loads with different forest configurations, based on buses that are switched on, to determine a subset of micro-grids less susceptible for link failures during the restoration period. Upon receiving a power disruption, activate the switches to restore some power to the PDG.

Growing community of inventors

Lexington, MA, United States of America

Hongbo Sun