Monitoring Device Reduces Training Feature Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As the number of monitoring target electrical devices increases, the number of training feature amounts required for inferring power consumption also increases significantly, leading to a substantial processing load due to the need for numerous combinations of training feature amounts, which is inefficient and costly.
Innovation Solution
A monitoring device and system that acquires unit-specific waveform data to infer operation states using two distinct feature amount groups: one for determining power-on or power-off states and another for estimating power consumption of electrical devices, reducing the number of training feature amounts needed by focusing on specific power states and device combinations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of monitoring target electrical devices increases, then the monitoring coverage is improved, but the number of training feature amounts required increases significantly leading to increased processing load
Solution Approach 1:
The patent segments the inference process into two distinct stages: first inferring power states (on/off) of electrical devices, then inferring power consumption only for devices in the on state. This segmentation reduces the overall processing load by dividing a complex single-stage inference into manageable sequential steps, directly addressing the contradiction between monitoring coverage and processing complexity
Solution Approach 2:
The patent applies partial action by not inferring power consumption for all electrical devices simultaneously, but only for those currently in the on state. This selective approach reduces the number of training feature amounts needed while maintaining accurate monitoring coverage, as it processes only the necessary subset of devices at any given time
2Adaptability or versatility
If the number of monitoring target electrical devices increases, then the monitoring coverage is improved, but the cost burden increases due to the increase in measurement instruments
Solution Approach 1:
The patent employs a single measuring instrument at the power supply facility that serves multiple functions: measuring total current consumption and enabling inference of both power states and power consumption for multiple electrical devices. This universal approach eliminates the need for separate measurement instruments for each device, reducing cost burden while maintaining comprehensive monitoring coverage
Solution Approach 2:
The patent introduces an inference mechanism that acts as an intermediary between the single measuring instrument and multiple electrical devices. This intermediary process allows one instrument to effectively monitor many devices by inferring their individual states from aggregate measurements, thereby reducing the number of physical instruments needed and lowering overall cost
3Measurement precision
If all electrical devices are monitored simultaneously for power consumption, then the monitoring precision is improved, but the processing load increases due to numerous combinations of training feature amounts
Solution Approach 1:
The patent segments the monitoring process into two precise stages: first determining power states with high accuracy, then inferring power consumption only for active devices. This segmentation maintains measurement precision for power consumption while reducing processing load by avoiding simultaneous inference for all devices regardless of state
Solution Approach 2:
The patent applies partial action by focusing power consumption inference only on electrical devices currently in the on state, rather than processing all devices equally. This approach maintains high precision for actual power consumption measurements while significantly reducing the computational combinations required, as inactive devices are excluded from consumption inference
Data Source
AI summary
A monitoring device (10) includes: a unit-specific waveform data acquisition unit (12) that acquires unit-specific monitoring waveform data which is waveform data of at least one among total current consumption, a total input voltage, and total power consumption in a unit in which monitoring target electrical devices are installed; a first inference unit (13) that infers operation states of at least some of the monitoring target electrical devices based on a 1st feature amount group including at least one kind of feature amount extracted from the unit-specific monitoring waveform data, and a training feature amount which is a feature amount of each of the monitoring target electrical devices in a predetermined operation state; and a second inference unit (14) that infers the operation states of some of the monitoring target electrical devices based on a 2nd feature amount group including at least one kind of feature amount extracted from the unit-specific monitoring waveform data, and different from the 1st feature amount group and the training feature amount.