Power Daemon NILM Disaggregation for Energy Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monitoring the power consumption of aggregated electrical devices connected to a shared power distribution point is impractical with existing methods, as they fail to accurately determine which devices are ON or OFF, leading to inefficiencies in energy management and potential security breaches.
Innovation Solution
A Non-invasive Load Monitoring (NILM) module, referred to as the Power Daemon, processes measurements of electrical node power to determine the optimum probability mass function (PMF) for device states, optimizing the disaggregation of power features such as current, apparent power, and reactive power to infer the ON/OFF states of individual devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a power monitor is coupled to each individual piece of equipment to directly monitor power consumption, then measurement precision is improved, but device complexity and cost increase significantly
Solution Approach 1:
The invention segments the power consumption signal into individual device components by analyzing the aggregated power signal at the power node. Instead of using multiple physical monitors, the system mathematically segments the total power consumption into contributions from each individual device, achieving precise measurement without additional hardware per device.
Solution Approach 2:
The system uses an intermediary computational approach rather than direct physical measurement at each device. By introducing signal processing algorithms and mathematical models as intermediaries, the system infers individual device power consumption from the aggregated signal, avoiding the need for multiple physical monitors.
2Device complexity
If NILM methods are used to infer device states from aggregated power signals, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The system performs preliminary actions by collecting and storing power consumption data over time before making inference decisions. By accumulating historical power signals and device state information in advance, the system builds a comprehensive dataset that improves the accuracy of subsequent device state predictions without requiring complex real-time processing.
Solution Approach 2:
The invention implements feedback mechanisms where the system continuously compares inferred device states with actual power consumption patterns and adjusts its inference algorithms accordingly. This feedback loop refines the measurement precision over time by learning from historical data and correcting inference errors, maintaining accuracy while keeping the system relatively simple.
3Ease of operation
If aggregation of devices to share a power node is implemented, then ease of operation is improved, but loss of information occurs regarding individual device states
Solution Approach 1:
The system extracts individual device state information from the aggregated power consumption signal by identifying unique signal characteristics and patterns associated with each device. Through mathematical decomposition and pattern recognition, the system pulls out hidden information about individual device states that would otherwise be lost in the aggregated signal.
Solution Approach 2:
The invention transitions from analyzing only the magnitude of power consumption to examining multiple dimensions of the power signal including temporal patterns, frequency characteristics, and phase relationships. By adding these dimensional analyses, the system recovers individual device information that exists in different signal dimensions within the aggregated measurement.
Data Source
AI summary
Apparatus for disaggregating a plurality of aggregated electrical devices that receive power via a same power node, the apparatus comprising: a memory having values of at least one power feature for each of the aggregated electrical devices that characterizes power consumption of the device; and a processor configured to: receive a measure of a value for the at least one power feature of node power distributed to the aggregated devices via the power node; determine an optimum probability mass function (PMF) based on the values for the at least one power feature for each of the electrical devices that is optimized to provide an optimum disaggregation state vector for the aggregated electrical devices that satisfies a predetermined criterion to provide a value for the at least one power feature that agrees with the received measure of the at least one power feature.