Characterization of electrical grids and prediction of fault conditions using inverters

Abstract

The goal is to quickly diagnose the malfunction. [Solution] An inverter coupled to the power grid characterizes the power grid. The inverter outputs multiple electrical signals of different frequencies to the power grid, measures the power grid's response to the multiple electrical signals to acquire measurement data, processes the measurement data to generate predictive data characterizing one or more fault conditions in the power grid. The inverter adjusts its operating settings based on the predictive data. The operating settings affect the power grid's response to the fault conditions.

Description

Technical Field

[0001] (Cross - Reference to Related Applications) This application claims priority to U.S. Patent Application No. 18 / 339,757, filed on June 22, 2023, the entire disclosure of which is incorporated herein by reference.

[0002] (Field of the Invention) This specification generally relates to characterizing a power grid using an electric inverter.

Background Art

[0003] An electric inverter is a device used to convert direct - current (DC) electricity into alternating - current (AC) electricity. Inverters are commonly used in a variety of applications, including solar power generation systems, backup power systems, and electric vehicles. In a solar power generation system, for example, solar panels generate DC electricity, which is converted into AC electricity by a conversion device for use or storage. Similarly, vehicles with backup power systems and on - vehicle electrical systems often use inverters to convert DC power from a battery into AC power for use by electrical appliances and other devices.

[0004] A power grid is a network of power plants, substations, transformers, transmission lines, and other electrical assets used to supply power from generation facilities to users. A fault in a power grid refers to a problem or disruption in the flow of electricity caused by a malfunction or failure in any of the grid's components.

Summary of the Invention

[0005] This specification describes techniques for characterizing a power grid. This characterization can be used, for example, to predict conditions that may potentially cause a fault in the power grid. This characterization is achieved by using measurement data obtained from one or more electric inverters integrated into the power grid.

[0006] Power grid failures can cause significant damage and disruption to the system, making rapid detection, location, diagnosis, and repair crucial. Furthermore, it is desirable to be able to predict the likelihood of failures occurring within the power grid within a timeframe (for example, based on voltage, current, and frequency variations at different locations within the power grid, as well as environmental factors) so that protection and prevention strategies can be applied in a timely manner.

[0007] However, diagnosing and predicting power grid faults can be a complex and challenging task. Modern power grids are highly complex systems with numerous interconnected components, including power lines, transformers, generators, and switchgear. Faults can occur at any point within the system, making it difficult to pinpoint their exact location. Diagnosing power grid faults may require collecting and analyzing large amounts of data, including diagnostic test and measurement data from various locations within the power grid, including hard-to-access areas. This makes rapid fault diagnosis difficult.

[0008] An inverter is a high-speed response solid-state device that can quickly adapt its output current and voltage based on downstream demand. Modern power grids can include a large number of electric inverters. For example, each home solar system may have an inverter to convert the DC power generated by the solar panels into AC power. Other examples of electric inverters already integrated into power grids include batteries and electric vehicles connected to the power grid. Inverters can also include communication units for communicating with computer networks.

[0009] The technology described herein can utilize electric inverters, which are already widely incorporated into modern power grids, to characterize and predict power grid failure conditions. Furthermore, the use of machine learning models can leverage available benchmark measurement data and simulation data to predict grid failure conditions. As a result, the system can make accurate and timely predictions of failure conditions, which makes it possible to reconfigure or adjust the operating parameters of the power grid to avoid or mitigate the predicted failure conditions.

[0010] In one innovative embodiment, this specification describes a method for characterizing a power grid. This method can be implemented by an electric inverter electrically coupled to the power grid. The inverter outputs multiple electrical signals of different frequencies to the power grid, measures the power grid's response to the multiple electrical signals to obtain measurement data, and processes the measurement data to generate predictive data characterizing one or more fault conditions of the power grid. Each of the characterized fault conditions is associated with one or more grid conditions that potentially cause the fault condition. The inverter adjusts its operating settings based on the predictive data. The operating settings affect the power grid's response to at least one grid condition associated with at least one of the characterized fault conditions.

[0011] In some implementations of this method, the measurement data includes one or more of the following: each voltage measurement or each current measurement.

[0012] In some implementations of this method, in order to adjust the operating settings, the inverter determines that the predicted data includes a potential overcurrent condition and, in response, adjusts the inverter's operating settings to cause an adjustment to the phase angle of the current in the power grid adjacent to the inverter.

[0013] In some implementations of this method, in order to adjust the operating settings, the inverter determines that the predicted data contains inverter malfunctions, and in response, adjusts the inverter's operating settings to electrically disconnect the inverter from the power grid.

[0014] In some implementations of this method, to output multiple electrical signals, the inverter sequentially generates a voltage waveform with a specific frequency for each of the different frequencies. For example, the frequency range of the different frequencies can include the range of 1 Hz to 10 kHz.

[0015] In some implementations of this method, to process the measurement data, the inverter generates impedance spectra of the power grid at different frequencies based on the measurement data, and processes the impedance spectra using a predictive model to generate predictive data. For example, the impedance spectrum may include amplitude spectra and phase angle spectra. To process the impedance spectrum, the inverter processes one or more of the amplitude spectrum or phase angle spectra to generate predictive data.

[0016] In some implementations of this method, the inverter processes the input, which includes the measured data, using a predictive model to generate predictive data. For example, the input to the predictive model may include additional data characterizing one or more of the following: current weather conditions, future weather conditions, timestamps, power grid maintenance records, or the occurrence of fires in the power grid area.

[0017] In some implementations of this method, the prediction data includes, for each of the characterized failure states, one or more output values ​​that characterize the occurrence of each failure state. For example, one of the output values ​​may characterize the likelihood of each failure state occurring currently, or the likelihood of each failure state occurring within a given period.

[0018] In some implementations of this method, the inverter uses the measurement data to identify a fault state from a predetermined set of fault states in order to process the measurement data and generate predictive data. Each predetermined fault state is stored in the inverter as being associated with a respective predetermined grid state. A grid state includes one or more of the following: inverter malfunction, failure of a component of the power grid, or an anomaly at the load location. Failures of components of the power grid may include transformer failures, circuit breaker failures, capacitor failures, transmission line failures, or distribution line failures.

[0019] In some implementations of this method, one of the output values ​​further characterizes the type of fault in the fault state. The fault type may include one or more of the following: short circuit fault, open circuit fault, ground fault, overload fault, phase-to-phase fault, or phase-to-ground fault.

[0020] In another innovative aspect, this specification describes an electric inverter configured to perform the power grid characterization method described above when connected to a power grid.

[0021] In another innovative aspect, this specification describes one or more computer-readable storage media that, when executed by one or more processors of an electrical converter, store instructions causing one or more processors to control one or more circuits of an inverter to perform the power grid characterization method described above.

[0022] In another innovative aspect, this specification describes another power grid characterization method that uses measurement data collected from multiple electrical inventors. This method can be implemented by a computing system including one or more computers. The system receives measurement data from each inverter of a plurality of inverters electrically coupled to the power grid at a plurality of locations via a network. The system processes the measurement data to generate predictive data that characterizes one or more fault conditions of the power grid. In processing the measurement data, the system processes the input generated from the measurement data using a machine learning model trained on training data, which includes a plurality of training examples. Each of the plurality of training examples includes a training input that characterizes its respective voltage and / or current measurement and a training output that characterizes each of at least one of the fault conditions. Based on the predictive data, the system identifies adjustments to be made in at least one grid component in the power grid to reduce the probability of occurrence of at least one of the characterized fault conditions. The system sends instructions to the power grid to adjust the operating settings of the power grid based on the identified adjustments.

[0023] In some implementations of this method, each measurement data received from each inverter includes one or more of the voltage measurements or current measurements that the inverter has measured over multiple frequencies.

[0024] In some implementations of this method, the adjustment includes at least one of the following: replacing some of the faulty equipment in the grid, disconnecting a specific inverter from the grid, or adjusting the setpoint of a controller in the grid.

[0025] In some implementations of this method, multiple training examples include one or more training examples generated from simulation data.

[0026] In some implementations of the method, the plurality of training examples includes one or more training examples generated from measurement data.

[0027] In some implementations of the method, the input further includes data characterizing the grid topology of the power grid.

[0028] In some implementations of the method, the input further includes data characterizing one or more of the current weather state, future weather state, time stamp, maintenance records of the power grid, or occurrence of a fire in the area of the power grid.

[0029] In some implementations of the method, the prediction data includes, for each respective fault state in the characterized fault state, one or more respective output values characterizing the occurrence of each respective fault state.

[0030] In some implementations of the method, one of the one or more respective output values characterizes the likelihood of the respective fault state currently occurring, or the likelihood of the respective fault state occurring within a predetermined period.

[0031] In some implementations of the method, the prediction data identifies a fault state from a set of predetermined fault states. Each predetermined fault state is stored in the inverter as being associated with a respective predetermined grid state. The grid state includes one or more of malfunctions of the inverter, failures of components of the power grid, or abnormalities at the load location. For example, failures of components of the power grid can include transformer failures, circuit breaker failures, capacitor failures, transmission line faults, or distribution line faults.

[0032] In some implementations of the method, each respective output value includes a value identifying or locating the faulty component.

[0033] In some implementations of this method, the instructions for adjusting the operating settings of the power grid include, in response to detecting an inverter malfunction, one or more of the following: an instruction to disconnect the inverter from the power grid; an instruction to disconnect a portion of the power grid affected by the fault condition from other portions of the power grid; an instruction to bypass a portion of the power grid using an alternative route; or an instruction to restart one or more control nodes of the power grid.

[0034] In another innovative aspect, this specification describes a system comprising one or more computers and one or more storage devices that, when executed by one or more computers, store instructions causing one or more computers to perform the power grid characterization method described above. In another innovative aspect, this specification describes one or more computer storage media that, when executed by one or more computers, store instructions causing one or more computers to perform the power grid characterization method described above.

[0035] Details of one or more embodiments of the subject matter described herein are given in the accompanying drawings and the following description. Other features, aspects, and advantages of the subject matter will become apparent from the description, drawings, and claims. [Brief explanation of the drawing]

[0036] [Figure 1] An example of a power grid including an electrical inverter configured to characterize the power grid is shown. [Figure 2] An example of a power grid characterization system is shown. [Figure 3] This flowchart illustrates an exemplary process performed by electric inverters to characterize the power grid. [Figure 4] This flowchart illustrates an exemplary process performed by computing systems to characterize the power grid. [Figure 5] An exemplary computer system is shown. Similar reference numbers and names in various drawings refer to similar elements. [Modes for carrying out the invention]

[0037] Figure 1 shows an example of a power grid 100 including an electric inverter 110 configured to characterize the power grid 100. The electric inverter 110 is electrically coupled to one or more other components 105 of the power grid 100.

[0038] Generally, the inverter 110 is configured to convert DC electricity from a DC power supply 115 into an AC output waveform. The conversion is performed by a set of electronic circuit units, which include, for example, a voltage regulator 120 configured to adjust the input DC voltage to a stable level, an oscillator 130 configured to generate an AC signal (e.g., a square wave or pulse train waveform), a filter circuit 140 configured to remove unwanted harmonics and noise distortion and shape the output waveform, and a control circuit 170 configured to control the frequency of the output waveform and other operating parameters of the inverter 110. The inverter 110 further includes one or more processors 160, for example, a microprocessor, which is programmed to perform operations to characterize the power grid 100, such as detecting and predicting fault conditions in the power grid 100.

[0039] To characterize the power grid 100, the processor 160 is configured to control the circuit unit of the inverter 115 to output an AC electrical signal 145 to the power grid 100. The response of the power grid 100 to the AC signal 145 can be measured by a measuring device, for example, a current (and / or voltage) meter 180 of the electric inverter 110. The processor 160 processes the measurement data to generate predictive data characterizing one or more fault conditions of the power grid.

[0040] The AC electrical signal 145 may be in the form of an AC voltage signal or an AC current signal. Generally, the AC signal 145 is a low-amplitude signal (e.g., having a voltage or current magnitude below a threshold) that does not interrupt the normal operation of the power grid 100. In some implementations, the inverter 100 is controlled by the processor 160 to perform a frequency sweep when generating the AC electrical signal 145. That is, the inverter 100 changes the frequency of the AC signal 145 at different points in time, for example, by changing the operating parameters of the oscillator 130 and / or filter 140. In exemplary examples, the frequency range of the sweep may include a range of 1 Hz to 10 kHz.

[0041] The inverter 110 can sweep a frequency range using a specific number of sweep frequencies with any suitable frequency sweep scheme, such as linear sweep, logarithmic sweep, or step sweep. The frequency range, the number of frequencies sampled, and / or the frequency sweep scheme can be predefined or determined by the processor 160, depending on factors such as the current state of the grid 100, the current time, historical measurement data of the grid, and / or historical prediction data of the grid.

[0042] For each sweep frequency, the response of the power grid 100 to the AC signal 145 is measured by one or more measuring devices (including, for example, an ammeter 180). The measured response may include the amplitude and / or phase angle of the voltage and / or current measured for each sweep frequency. The measured response may further include the voltage and / or current response at one or more harmonic frequencies for one or more of the sweep frequencies.

[0043] The processor 160 receives measurement data from a measurement device and processes the measurement data to generate predictive data 190 that characterize one or more fault conditions in the power grid. For example, the predictive data 190 may include output values ​​that characterize the occurrence of a fault condition, such as a score that characterizes the likelihood of a fault condition (e.g., within a given set of fault conditions) that is currently occurring or will occur within a given period.

[0044] Failures can occur for a variety of reasons, including equipment malfunctions, environmental events or factors, and human error. Common types of failures in a power grid include short-circuit failures, open-circuit failures, ground faults, overload failures, phase-to-phase failures, and phase-to-ground faults.

[0045] One or more grid conditions may lead to one or more specific faults. For example, a grid condition such as equipment failure or undesirable power line connection (e.g., resulting from lightning or fallen tree branches) may lead to a short circuit fault; a grid condition such as power line damage or equipment failure may lead to an open circuit fault; a grid condition such as damaged insulation, faulty equipment, or undesirable earthing connection may lead to a ground fault; a grid condition such as increased demand or faulty equipment may lead to an overload fault; a grid condition such as faulty equipment or improper installation may lead to a phase-to-ground fault; and a grid condition such as damaged insulation or faulty equipment may lead to a phase-to-ground fault.

[0046] The prediction data may include one or more fault conditions and one or more grid conditions associated with each fault condition. For example, the prediction data may include a malfunction of the inverter 110 itself, a failure of an asset component of the power grid 100 (e.g., a transformer failure, circuit breaker failure, capacitor failure, transmission line failure, or distribution line failure), or an anomaly at a load location in the power grid 100. The prediction data 190 may include data that identifies the fault component as part of the grid condition (e.g., component index, location coordinates, etc.). The prediction data 190 may specify the type of fault condition, such as a short-circuit fault, open-circuit fault, ground fault, overload fault, phase-to-phase fault, or phase-to-ground fault.

[0047] In some implementations, the processor 160 can calculate the characteristic impedance spectrum of the grid 100 at the grid connection port with the inverter 110 based on the voltage and current measured at the sweep frequency. The impedance spectrum may be a complex spectrum, i.e., a spectrum that includes both the amplitude and phase angle of the impedance at frequency. The processor 160 can process the amplitude spectrum and / or phase angle spectrum to generate prediction data 190.

[0048] In some implementations, the processor 160 uses a predictive model to create predictive data 190 based on at least the measured data. That is, the processor 160 generates predictive data 190 by processing the model input, which includes the measured data, using the predictive model. In some implementations, the predictive model can be a machine learning model, such as a linear regression model, a Bayesian classifier, a support vector machine, a K-means clustering model, and / or a neural network. The parameters of the machine learning model can be determined by a training process based on training data. The training data can include a set of training examples, each of which includes its respective training input and corresponding training label.

[0049] In some implementations, in addition to the measured response data from the power grid 100, the processor 160 predicts the failure state based on other information potentially relevant to the state of the power grid 100, such as weather conditions (past, present, and / or forecast), timestamps of the measured data, maintenance records of the power grid 100, and the occurrence of events such as fires, floods, or earthquakes within the power grid area. The additional information may be part of the input to the prediction model (in addition to the measured data).

[0050] After generating the prediction data 190, the processor 160 can take action according to the prediction data 190. For example, in some implementations, the processor 160 can output the prediction data 190 to another system via the input / output interface 150, or present it to a user, such as an operator of the power grid 100. For example, if the prediction data 190 predicts that a particular fault condition is likely to occur (for example, if the likelihood score corresponding to a particular fault condition exceeds a threshold), the inverter 110 can send a report or warning message containing information about the fault condition to a device accessible by the operator (for example, via a network), thereby enabling timely action to prevent the fault condition from occurring or to mitigate its impact.

[0051] In some implementations, the processor 160 can use the prediction data to identify adjustments to the operating settings of the power grid 100. These adjustments involve reducing the likelihood of at least one fault condition occurring, or mitigating the adverse effects of at least one fault condition predicted in the prediction data. The inverter 110 may take action based on the prediction data to influence the power grid's response to at least one grid state associated with at least one of the characterized fault conditions.

[0052] For example, if the predicted data 190 indicates a malfunction of the inverter 110 itself, the inverter 110 can output a control signal (e.g., to a switch) to either turn itself off or electrically disconnect the inverter 110 from the power grid 100.

[0053] In another example, if the predicted data 190 indicates a potential overcurrent condition due to an excessively large reactive load, the inverter 110 may change its operation to adjust the phase angle of the current in the power grid 100 adjacent to the inverter 110. For example, the inverter 110 may change its operating configuration so that the phase angle between the current and voltage at the inverter's output terminals is opposed to the phase angle of the power grid 100.

[0054] In another example, if the forecast data 190 indicates a failure of an asset component in the power grid, the processor 160 can generate a message and send it to the control unit of the power grid 100 to notify the control unit of the asset component failure. The control unit can then employ strategies to mitigate the component failure. For example, the control unit can determine whether the failure can be avoided using an alternative route in the power grid 100. If it is determined that the failure can be avoided using an alternative route, the control unit can output a control signal to electronically control one or more switches to switch the power grid 100 to use the alternative route. In some examples, the control unit is part of or can communicate with the power grid characterization system 230, which is described below with reference to Figure 2.

[0055] In addition to characterizing the power grid, the inverter 110 can perform various functions within the power grid 100. For example, the inverter 110 may be used in a photovoltaic power generation system to convert DC power generated by solar panels into AC power that can be used by household appliances and other components of the power grid 100. In another example, the inverter 110 may be used in a wind turbine to convert variable-frequency AC power generated by the turbine into constant-frequency AC power that can be used by the power grid 100. In yet another example, the inverter 110 may be used in an uninterruptible power supply (UPS) to convert DC battery power into AC power. In yet another example, the inverter 110 may be used in an electric vehicle to convert DC power stored in a battery into AC power. In yet another example, the inverter 110 may be used in an air conditioning system to control the speed of a compressor motor. As described in this disclosure, the inverter 110 has additional functions for characterizing the power grid 100. In some implementations, the inverter 110 can perform its primary functions, such as DC-AC power conversion and characterizing the power grid 100, in parallel.

[0056] The inverter 110 may be any suitable type of inverter that fits the intended application and the functions described below. In some implementations, the inverter 110 may include electronic equipment using silicon carbide (SiC) and / or gallium nitride (GaN) as semiconductor materials, which offers advantages such as fast switching speed and low switching loss. These characteristics allow the inverter 110 to have a high switching frequency and produce a high-quality output waveform with a wide frequency range.

[0057] As described above, the inverter 110 processes the measurement data locally within the inverter 110 using the processor 160. Alternatively or additionally, the inverter 110 may be implemented to transmit the measurement data to an external computing system, such as a central server, for processing by an external system. Data transmission can be performed via wired or wireless connection over a network, such as the Internet. In fact, the external computing system can process measurement data from multiple inverters installed at multiple locations in the power grid to predict grid failure conditions.

[0058] Figure 2 shows a power grid characterization system 230 for characterizing the power grid 200. System 230 is an example of a system implemented as a computer program on one or more computers in one or more locations, which can implement the systems, components, and techniques described later.

[0059] The power grid 200 includes grid components 205 for generating, distributing, and / or transmitting electricity to multiple load locations. The power grid 200 includes multiple electric inverters 210 (e.g., 210a, 210b, 210c, etc.) electrically coupled to the power grid 200 at multiple locations. The inverters 210 can perform different functions within the power grid, such as DC-AC power conversion for solar panels, UPS batteries, or electric vehicles, AC-AC power conversion for wind turbines, or speed control for electric motors.

[0060] Each inverter is further configured to collect measurement data at its respective location within the power grid 200. For example, each inverter can output a low-amplitude signal to the power grid 200 (e.g., by performing a frequency sweep) that does not interrupt the normal operation of the power grid 100, and collect measurement data (e.g., voltage and / or current measurements) that characterize the power grid's response to the low-amplitude signal at each location of the inverter. Additional details and examples of signal generation and data collection by each specific inverter are described above with reference to Figure 1. The inverters transmit the collected measurement data to the system 230 via the network 220.

[0061] System 230 receives measurement data from each of the inverters 210 via the network 220. System 230 includes a machine learning model 232 configured to process the measurement data and generate predictive data characterizing the fault conditions of the power grid 200. Specifically, System 230 takes measurement data as input, processes the input using the machine learning model 232, and generates an output that includes predictive data characterizing the fault conditions of the power grid 200.

[0062] In some implementations, in addition to the measurement data received from the inverter 210, the input to the machine learning model 232 may further include data characterizing the grid topology of the power grid. The grid topology data can have any suitable data format and generally includes data characterizing the physical layout and / or connectivity of the components constituting the power grid 200.

[0063] In some implementations, the input to the machine learning model 232 may further include data characterizing other information that is potentially relevant to the state of the power grid 200 and could lead to failure conditions on the grid. Examples of such information may include weather conditions (e.g., past, present, and / or expected weather conditions), timestamps of measurement data, maintenance records of the power grid 200, or the occurrence of events such as fires, floods, or earthquakes in the area of ​​the power grid.

[0064] The machine learning model 232 can be any suitable type of machine learning model, including, for example, linear regression models, Bayesian classifiers, support vector machines, K-means clustering models, and / or neural networks.

[0065] The predictive data in the output of the machine learning model 232 may include values ​​that characterize the occurrence of failure conditions. For example, the values ​​may be in the form of a score that characterizes the likelihood of a particular failure condition currently occurring (e.g., of a given set of failure conditions) or a particular failure condition occurring within a given future period.

[0066] The grid conditions that cause the corresponding fault condition may include malfunction of one or more inverters 210, a failure of an asset component of the power grid 200 (e.g., transformers, circuit breakers, capacitors, transmission lines, or distribution lines), or an anomaly at the load location of the power grid 200. The prediction data may include data identifying the component that will experience the fault (e.g., component index, location coordinates, etc.). The prediction data may further specify the type of fault condition, such as a short-circuit fault, open-circuit fault, ground fault, overload fault, phase-to-phase fault, or phase-to-ground fault.

[0067] System 230, or another system communicating with System 230, may include a training engine 234 for training a machine learning model 232. The training engine determines the parameters of the machine learning model 232 based on the training data.

[0068] Generally, training data can include a set of training examples. Each training example includes a training input that characterizes voltage and / or current measurements, and a training label that indicates whether the training input leads to a fault condition, and if so, characterizes at least one fault condition of the training input.

[0069] The set of training examples may include training examples generated from measurement data and / or simulation data. For example, training examples generated from measurement data may include historical data of power grid 200 or another power grid, voltages and / or measurements recorded by one or more inverters in the power grid can be used as training inputs, and one or more fault conditions (or normal conditions) observed for the power grid can be used as corresponding training labels.

[0070] In some implementations, the training engine 234 can repeatedly update the training of the machine learning model 232 based on new data collected about the power grid 200, including new measurement data received from the inverter 210 and new observed fault conditions. That is, the training engine 234 can dynamically train the machine learning model 232 based on more recently available data. The updated training can be performed periodically at predetermined intervals or based on other criteria. For example, the training engine 234 can repeatedly evaluate the prediction error of the machine learning model 232 by comparing predicted fault conditions with observed fault conditions, and update the training of the machine learning model 232 when the prediction error exceeds a threshold.

[0071] Based on the predictive data, system 230 or another system may decide to take actions such as sending a report or warning message containing information about the fault condition to a device accessible by the operator, thereby enabling timely measures to prevent the fault condition from occurring or to mitigate its impact.

[0072] In some implementations, if predictive data indicates a high probability of one or more failure conditions occurring, system 230 or another system communicating with system 230 may identify adjustments to be performed in at least one grid component within the power grid 200 to reduce the probability of at least one of the failure conditions occurring, or to mitigate the adverse effects of at least one failure condition. Adjustments may include replacing a faulty piece of equipment in the grid, disconnecting a specific component, such as a specific inverter, from the grid, or adjusting the setpoints of controllers in the grid. The system may then send commands to the power grid to adjust the operating settings of the power grid based on the identified adjustments.

[0073] Figure 3 is a flowchart illustrating an exemplary process 300 for characterizing a power grid. For convenience, the process 300 is described as being carried out by an electric inverter. For example, the inverter 110, as described with reference to Figure 1 and appropriately programmed according to this specification, can perform the process 300.

[0074] In step 310, the inverter outputs an electrical signal to the power grid. In step 320, the inverter measures the power grid's response to the electrical signal and obtains measurement data. The response includes voltage and / or current measurements. The inverter can repeat steps 310 and 320 to output multiple electrical signals of different frequencies to the power grid and measure the power grid's response to each output signal. For example, the inverter can perform a frequency sweep by sequentially generating a voltage waveform having each of the different frequencies and measure the response at each frequency (e.g., the current in response to the voltage waveform).

[0075] In 330, the inverter processes the measurement data to generate predictive data that characterizes one or more fault conditions in the power grid. For example, the inverter can generate predictive data by (i) using the measurement data to generate impedance spectra of the power grid at different frequencies, and (ii) processing the impedance spectra using a predictive model.

[0076] In some implementations, the predictive data may include, for each failure condition in a given set of failure conditions, one or more output values ​​that characterize the occurrence of each failure condition. The characterization of failure conditions may be for currently occurring failure conditions and / or failure conditions that will occur within a given future period.

[0077] The predictive data may include one or more grid states associated with each of the one or more characterized failure conditions. For example, grid states may include inverter malfunctions, failures of power grid components, anomalies at load locations, or other conditions affecting the operation of the power grid.

[0078] In 340, the inverter adjusts its operating settings based on predictive data. The operating settings affect the power grid's response to at least one grid state associated with at least one of the characterized fault states.

[0079] For example, if the predicted data indicates a malfunction of an electrical inverter, the inverter's processor may adjust the inverter's operating settings to electrically disconnect the inverter from the power grid. In another example, if the predicted data indicates a potential overcurrent condition, the inverter's processor may adjust the inverter's operating settings to cause an adjustment to the phase angle of the current in the power grid adjacent to the inverter.

[0080] Figure 4 is a flowchart illustrating an exemplary process 400 for characterizing a power grid. For convenience, the process 400 is described as being carried out by a system of one or more computers located at one or more locations. For example, the system 230 described with reference to Figure 2, appropriately programmed according to this specification, can perform the process 400.

[0081] In 410, the system receives measurement data from each inverter of a plurality of electric inverters electrically coupled to the power grid at multiple locations via a network. Each measurement data received from each inverter may include each voltage measurement and / or each current measurement measured by the inverter over multiple frequencies.

[0082] In 420, the system processes measurement data to generate predictive data characterizing one or more failure conditions in the power grid. Specifically, the system uses a machine learning model to process input generated from measurement data to generate predictive data. The input may further include other data related to grid performance, such as grid topology data, weather condition data, timestamp data, grid maintenance data, or data characterizing events that may affect the operation of the power grid.

[0083] The parameters of a machine learning model can be determined using a training process based on training data, which includes a set of training examples. Each training example includes a training input that characterizes each voltage and / or current measurement, and a training output that characterizes at least one of each fault conditions.

[0084] In 430, the system identifies, based on predictive data, adjustments to be performed in at least one grid component within the power grid to reduce the likelihood of at least one fault condition occurring in a characterized fault condition, or to mitigate the adverse effects of at least one fault condition. Adjustments may include, for example, replacing a faulty piece of equipment in the power grid, disconnecting a particular inverter from the power grid, or adjusting the setpoint of a controller in the power grid. In 440, the system sends an instruction to the power grid to adjust the operating settings of the power grid based on the identified adjustments.

[0085] Figure 5 shows an exemplary computer system 500 that can be used to perform the specific operations described above, for example, the operation of the inverter 110 in Figure 1 or the operation of the system 230 in Figure 2. The system 500 includes a processor 510, memory 520, storage device 530, and input / output device 540. Each of the components 510, 520, 530, and 540 can be interconnected, for example, using a system bus 550. The processor 510 can process instructions to be executed within the system 500. In one implementation, the processor 510 is a single-threaded processor. In another implementation, the processor 510 is a multi-threaded processor. The processor 510 can process instructions stored in memory 520 or on storage device 530.

[0086] Memory 520 stores information within the system 500. In one implementation, memory 520 is a computer-readable medium. In another implementation, memory 520 is a volatile memory unit. In yet another implementation, memory 520 is a non-volatile memory unit.

[0087] The storage device 530 can provide high-capacity storage for the system 500. In one implementation, the storage device 530 is a computer-readable medium. In various different implementations, the storage device 530 may include, for example, a hard disk device, an optical disk device, a storage device shared over a network by multiple computing devices (e.g., a cloud storage device), or some other high-capacity storage device.

[0088] The input / output device 540 provides input / output operation for the system 500. In one implementation, the input / output device 540 may include one or more network interface devices, such as an Ethernet card, a serial communication device, such as an RS-232 port, and / or a wireless interface device. In another implementation, the input / output device may include a driver device configured to receive input data and transmit output data to other input / output devices, such as a keyboard, printer, and display device 560. However, other implementations, such as mobile computing devices, mobile communication devices, and set-top box television client devices, may also be used.

[0089] An exemplary system is shown in Figure 5, but the subject matter and functional operations described herein can be implemented in other types of digital electronic circuits, including the structures disclosed herein and their structural equivalents, or in computer software, firmware, or hardware, or a combination of one or more of these.

[0090] In this specification, the term “configured” is used in relation to systems and computer program components. One or more computer systems are configured to perform a particular operation or action to mean that the system has installed software, firmware, hardware, or a combination thereof that causes the system to perform the operation or action while in operation. One or more computer programs are configured to perform a particular operation or action to mean that one or more programs contain instructions that, when executed by a data processing device, cause the device to perform the operation or action.

[0091] The subject matter and functional operating embodiments described herein can be implemented in digital electronic circuits, tangibly embodied computer software or firmware, computer hardware, or one or more combinations thereof, including the structures disclosed herein and their structural equivalents. Embodiments of the subject matter described herein can be implemented as one or more modules of computer programs, i.e., computer program instructions encoded on a tangible non-temporary storage medium for execution by a data processing device or for controlling the operation of a data processing device. The computer storage medium may be a machine-readable storage device, a machine-readable storage board, a random or serial access memory device, or one or more combinations thereof. Alternatively or additionally, the program instructions may be encoded on artificially generated propagating signals, such as machine-generated electrical, optical, or electromagnetic signals generated to encode information to be transmitted to a receiving device suitable for execution by a data processing device.

[0092] The term "data processing device" refers to data processing hardware and encompasses all kinds of devices, machines, and equipment for processing data, including, for example, programmable processors, computers, or multiple processors or computers. A device may also be, or further include, dedicated logic circuits such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits). Optionally, in addition to hardware, a device may include code that creates an execution environment for computer programs, such as processor firmware, protocol stacks, database management systems, operating systems, or code comprising one or more of these.

[0093] A computer program, which may also be called or described as a program, software, software application, app, module, software module, script, or code, may be written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and may be deployed in any form, including standalone programs or modules, components, subroutines, or other units suitable for use in a computing environment. A program may, but is not required, correspond to a file in a file system. A program may be stored in a part of a file that holds other programs or data, for example, in a markup language document, in a single file dedicated to the program in question, or in a set of collaborative files, for example, in one or more scripts stored in a file that stores one or more modules, subprograms, or parts of code. A computer program may be deployed to run on one computer, or on multiple computers located in one site, or distributed across multiple sites and interconnected by a data communication network.

[0094] In this specification, the term “engine” is used broadly to refer to a software-based system, subsystem, or process programmed to perform one or more specific functions. Generally, an engine is implemented as one or more software modules or components and installed on one or more computers in one or more locations. In some cases, one or more computers are dedicated to a particular engine. In other cases, multiple engines may be installed and run on the same computer (one or more).

[0095] The processes and logic flows described herein can be executed by one or more programmable computers running one or more computer programs to perform their functions by acting on input data and producing outputs. The processes and logic flows can also be executed, for example, by dedicated logic circuits such as FPGAs or ASICs, or by a combination of application-specific logic circuits and one or more programmed computers.

[0096] A computer suitable for running computer programs may be based on a general-purpose or dedicated microprocessor, or both, or any other type of central processing unit. Generally, the central processing unit receives instructions and data from read-only memory, random-access memory, or both. Essential elements of a computer are a central processing unit for executing or running instructions, and one or more memory devices for storing instructions and data. The central processing unit and memory may be complemented by or incorporated into special-purpose logic circuits. Generally, a computer may also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or be operablely coupled to receive data from them, transfer data to them, or both. However, a computer is not required to have such devices. Furthermore, computers can be integrated into other devices, such as mobile phones, personal digital assistants (PDAs), mobile audio or video players, game consoles, Global Positioning System (GPS) receivers, or portable storage devices, such as Universal Serial Bus (USB) flash drives.

[0097] Computer-readable media suitable for storing computer program instructions and data include, for example, semiconductor memory devices such as EPROM, EEPROM, and flash memory devices, magnetic disks such as internal hard disks or removable disks, magneto-optical disks, and all forms of non-volatile memory, including CD-ROM and DVD-ROM disks.

[0098] To provide user interaction, embodiments of the subject matter described herein may be implemented on a computer, which may have a display device for displaying information to the user, such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, and a keyboard and pointing device, such as a mouse or trackball, to which the user can provide input to the computer. Other types of devices may be used similarly to provide user interaction, for example, the feedback provided to the user may be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback, and input from the user may be received in any form, including acoustic, voice, or tactile input. In addition, the computer may interact with the user by sending documents to and receiving documents from devices used by the user, for example, by sending a web page to a web browser on the user's device in response to a request received from a web browser. The computer may also interact with the user by sending text messages or other forms of messages to a personal device, such as a smartphone running a messaging application, and receiving response messages from the user as replies.

[0099] Data processing equipment for implementing machine learning models may also include, for example, dedicated hardware accelerator units for handling the general, computationally intensive parts of machine learning training or production, i.e., inference workloads.

[0100] Machine learning models can be implemented and deployed using machine learning frameworks such as the TensorFlow framework, the Microsoft Cognitive Toolkit framework, the Apache Singa framework, or the Apache MXNet framework.

[0101] Embodiments of the subject matter described herein can be implemented in a computing system that includes, for example, a backend component as a data server, or a middleware component, such as an application server, or a frontend component, such as a graphical user interface, a web browser, or a client computer having an application that allows a user to interact with the implementation of the subject matter described herein, or any combination of one or more such backend, middleware, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication, such as a communication network. Examples of communication networks include local area networks ("LANs") and wide area networks ("WANs"), such as the Internet.

[0102] A computing system can include clients and servers. Clients and servers are generally geographically separated and typically interact via a communication network. The client-server relationship arises from computer programs running on each computer that have a client-server relationship with each other. In some embodiments, the server sends data, such as an HTML page, to a user device for the purpose of displaying data to a user interacting with a device acting as a client and receiving user input from the user. Data generated on the user device, such as the results of user interactions, can be received from the device to the server.

[0103] This specification contains details of many individual implementations, which should be interpreted as descriptions of features specific to a particular embodiment, rather than limiting the scope of any feature or the scope of the claims. Certain features described herein in the context of a separate embodiment may also be realized in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may be realized individually in multiple embodiments or in any preferred partial combination. Furthermore, features may be described above as acting in a particular combination, and may even be initially claimed as such, but one or more features from a claimed combination may, in some cases, be removed from the combination, and the claimed combination may be directed towards a partial combination or a variation of a partial combination.

[0104] Similarly, although the operations are depicted in a specific order in the diagrams, this should not be understood as requiring that such operations be performed in a specific or sequential order shown, or that all exemplified operations be performed, in order to achieve the desired result. In certain circumstances, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged in multiple software products.

[0105] Accordingly, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. In some examples, the operations described in the claims can be performed in a different order and still achieve the desired results. Furthermore, the processes shown in the accompanying drawings do not necessarily require the specific order or sequence shown to achieve the desired results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

1. A power grid characterization method performed by an inverter electrically coupled to the power grid, wherein the method is The inverter outputs multiple electrical signals of different frequencies to the power grid, The inverter measures the response of the power grid to the plurality of electrical signals and acquires measurement data. The inverter processes the measurement data to generate predictive data characterizing one or more fault conditions of the power grid, wherein each of the characterized fault conditions in the one or more fault conditions is associated with one or more grid conditions that potentially cause the fault condition. The inverter adjusts the operating settings of the inverter based on the predicted data, such that the operating settings affect the response of the power grid to at least one grid state associated with at least one of the characterized fault states. A method for characterizing power grids, including the following.

2. The power grid characterization method according to claim 1, wherein the measurement data includes one or more of the respective voltage measurements or the respective current measurements.

3. Adjusting the aforementioned operating settings The prediction data is determined to include a potential overcurrent condition, and in response to this, A power grid characterization method according to claim 1, comprising adjusting the operating settings of the inverter to cause an adjustment to the phase angle of the current in the power grid adjacent to the inverter.

4. Adjusting the aforementioned operating settings The prediction data is determined to include a malfunction of the inverter, and in response to this, A method for characterizing a power grid according to claim 1, comprising adjusting the operating settings of the inverter to electrically disconnect the inverter from the power grid.

5. Outputting the aforementioned multiple electrical signals The power grid characterization method according to claim 1, comprising sequentially generating voltage waveforms having each of the aforementioned different frequencies.

6. The power grid characterization method according to claim 4, wherein the frequency range of the different frequencies includes the range of 1 Hz to 10 kHz.

7. Processing the aforementioned measurement data Based on the measurement data, generate the impedance spectrum of the power grid at the different frequencies, A power grid characterization method according to claim 1, comprising processing the impedance spectrum to generate the predicted data by using a predictive model.

8. The impedance spectrum includes an amplitude spectrum and a phase angle spectrum. The power grid characterization method according to claim 7, wherein processing the impedance spectrum includes processing one or more of the amplitude spectrum or the phase angle spectrum in order to generate the prediction data.

9. The process of processing the aforementioned measurement data to generate the aforementioned prediction data is to A power grid characterization method according to claim 1, comprising processing an input including the measurement data by using a predictive model to generate the predictive data.

10. The input to the prediction model is Current weather conditions, Future weather conditions, Timestamp, Maintenance records of the aforementioned power grid, or The power grid characterization method according to claim 8, further comprising additional data characterizing one or more of the following: the occurrence of a fire in the area of ​​the power grid.

11. The power grid characterization method according to claim 1, wherein the prediction data includes, for each of the characterized failure states, one or more output values ​​that characterize the occurrence of each of the characterized failure states.

12. The power grid characterization method according to claim 11, wherein one of the one or more output values ​​characterizes the likelihood of each of the failure conditions currently occurring, or the likelihood of each of the failure conditions occurring within a predetermined period.

13. The process of processing the aforementioned measurement data to generate the aforementioned prediction data is to Using the measurement data, one or more fault states are identified from a predetermined set of fault states, wherein each predetermined fault state is stored in the inverter as being associated with a predetermined grid state, and the grid state is Malfunction of the aforementioned inverter, Failure of a component of the power grid, or A method for characterizing a power grid according to claim 11, comprising one or more of the following: an anomaly at the load location.

14. The failure of the component of the power grid is Transformer malfunction, Circuit breaker failure, Capacitor failure, A power line failure, or A method for characterizing a power grid according to claim 13, including faults in power distribution lines.

15. The power grid characterization method according to claim 11, wherein one of the one or more output values ​​further characterizes the type of failure of the failure state.

16. The power grid characterization method according to claim 15, wherein the type of fault includes one or more of the following: short-circuit fault, open-circuit fault, ground fault, overload fault, phase-to-phase fault, or phase-to-ground fault.

17. An inverter comprising one or more processors, wherein the one or more processors Controlling one or more circuits of the inverter to output multiple electrical signals of different frequencies to the power grid, wherein the inverter controls one or more circuits of the inverter that are electrically coupled to the power grid. Receiving measurement data characterizing the response of the power grid to the plurality of electrical signals, The process involves processing the measurement data to generate predictive data characterizing one or more failure states of the power grid, wherein each of the characterized failure states in the one or more failure states is associated with one or more grid states that potentially cause the failure state. Adjusting the operating settings of the power grid based on the prediction data, such that the operating settings affect the power grid's response to at least one grid state associated with at least one of the characterized fault states. An inverter configured to perform operations including those mentioned above.

18. One or more computer-readable storage media for storing instructions, wherein when an instruction is executed by one or more processors of an electrical converter, the one or more processors of the inverter, Controlling one or more circuits of the inverter to output multiple electrical signals of different frequencies to the power grid, Receiving measurement data characterizing the response of the power grid to the plurality of electrical signals, The process involves processing the measurement data to generate predictive data characterizing one or more failure states of the power grid, wherein each of the characterized failure states in the one or more failure states is associated with one or more grid states that potentially cause the failure state. One or more computer-readable storage media that cause the power grid to adjust its operating settings based on the prediction data, such that the operating settings affect the power grid's response to at least one grid state associated with at least one of the characterized failure states.

19. A method for characterizing a power grid performed by a computing system, wherein the method is Receiving measurement data from each inverter of multiple inverters electrically coupled to the power grid at multiple locations via a network, The process involves processing the measurement data to generate predictive data characterizing one or more fault conditions of the power grid, wherein the processing includes processing inputs generated from the measurement data using a machine learning model trained on training data comprising a plurality of training examples, each of which generates the predictive data comprising a training input characterizing its respective voltage and / or current measurement and a training output characterizing at least one of the one or more fault conditions. Based on the prediction data, identify adjustments to be performed in at least one grid component within the power grid to reduce the probability of occurrence of at least one of the characterized failure conditions. Sending a command to the power grid to adjust the operating settings of the power grid based on the identified adjustment, A method for characterizing power grids, including the following.

20. The power grid characterization method according to claim 19, wherein each measurement data received from each inverter includes one or more of the voltage measurements or current measurements measured by the inverter over multiple frequencies.

21. The power grid characterization method according to claim 19, wherein the adjustment includes at least one of replacing some of the faulty equipment in the grid, disconnecting a specific inverter from the grid, or adjusting the setpoint of a controller in the grid.

22. The power grid characterization method according to claim 19, wherein the plurality of training examples include one or more training examples generated from simulation data.

23. The power grid characterization method according to claim 19, wherein the plurality of training examples include one or more training examples generated from measurement data.

24. The power grid characterization method according to claim 19, wherein the input further includes data that characterizes the grid topology of the power grid.

25. The aforementioned input is Current weather conditions, Future weather conditions, Timestamp, Maintenance records of the aforementioned power grid, or The power grid characterization method according to claim 19, further comprising data characterizing one or more of the following: the occurrence of a fire in the area of ​​the power grid.

26. The power grid characterization method according to claim 19, wherein the prediction data includes one or more output values ​​that characterize the occurrence of each of the characterized failure states.

27. The power grid characterization method according to claim 26, wherein one of the one or more output values ​​characterizes the likelihood of each of the failure conditions currently occurring, or the likelihood of each of the failure conditions occurring within a predetermined period.

28. The prediction data identifies one or more fault states from a predetermined set of fault states, and each predetermined fault state is stored in the inverter as being associated with a predetermined grid state, and the grid state is Malfunction of the aforementioned inverter, Failure of a component of the power grid, or A method for characterizing a power grid according to claim 27, comprising one or more of the following: an anomaly at the load location.

29. The failure of the component of the power grid is Transformer malfunction, Circuit breaker failure, Capacitor failure, A power line failure, or A method for characterizing a power grid according to claim 28, including faults in power distribution lines.

30. The power grid characterization method according to claim 26, wherein each of the one or more output values ​​includes a value that identifies or locates a component of a fault.

31. The command for adjusting the operating settings of the power grid is, In response to detecting an inverter malfunction, a command is issued to disconnect the inverter from the power grid. An instruction to disconnect a portion of the power grid affected by the aforementioned fault condition from the rest of the power grid, An instruction to use an alternative route to bypass a portion of the power grid, or The power grid characterization method according to claim 19, comprising one or more instructions for restarting one or more control nodes of the power grid.

32. A system comprising one or more computers and one or more storage devices for storing instructions, wherein when an instruction is executed by one or more computers, the system provides the following to the one or more computers: Receiving measurement data from each inverter of multiple inverters electrically coupled to the power grid at multiple locations via a network, The process involves processing the measurement data to generate predictive data characterizing one or more fault conditions of the power grid, wherein the processing includes processing inputs generated from the measurement data using a machine learning model trained on training data comprising a plurality of training examples, each of which generates the predictive data comprising a training input characterizing its respective voltage and / or current measurement and a training output characterizing at least one of the one or more fault conditions. Based on the prediction data, identify adjustments to be performed in at least one grid component within the power grid to reduce the probability of occurrence of at least one of the characterized failure conditions. A system that causes the power grid to perform an operation including sending a command to the power grid to adjust the operating settings of the power grid based on the identified adjustment.

33. One or more computer storage media for storing instructions, and when executed by one or more computers, the instructions are stored in the one or more computers. Receiving measurement data from each inverter of multiple inverters electrically coupled to the power grid at multiple locations via a network, The process involves processing the measurement data to generate predictive data characterizing one or more fault conditions of the power grid, wherein the processing includes processing inputs generated from the measurement data using a machine learning model trained on training data comprising a plurality of training examples, each of which generates the predictive data comprising a training input characterizing its respective voltage and / or current measurement and a training output characterizing at least one of the one or more fault conditions. Based on the prediction data, identify adjustments to be performed in at least one grid component within the power grid to reduce the probability of occurrence of at least one of the characterized failure conditions. One or more computer storage media that perform an operation including transmitting a command to the power grid to adjust the operating settings of the power grid based on the identified adjustment.

Images