CN-122020192-A - Multi-device energy consumption monitoring and management method based on Internet of things

CN122020192ACN 122020192 ACN122020192 ACN 122020192ACN-122020192-A

Abstract

The invention relates to the technical field of energy management of the Internet of things, and discloses a multi-device energy consumption monitoring and management method based on the Internet of things. The method comprises the steps of collecting multi-equipment energy consumption data through intelligent sensing nodes, dynamically dividing low, medium and high energy consumption state intervals after processing, generating a multi-dimensional characteristic matrix, monitoring the switching frequency of the state intervals in real time, comparing a moving average value of the state intervals with a self-adaptive threshold value to trigger an abnormal mark, extracting abnormal characteristic components after matching the characteristic matrix with a history mode, calibrating electrical parameters, finally constructing an equipment energy consumption correlation network, and executing reverse path searching to locate abnormal source nodes and generating a regulation command based on the calibrated characteristics. The invention can effectively capture the transient state abnormality of the equipment to realize early warning, accurately position the root cause equipment of the system energy consumption abnormality, and improve the monitoring accuracy and management efficiency.

Inventors

WU BINGXIN
ZHONG YIHAO
Weng chen
HUANG JINLONG

Assignees

深圳市盘古数据有限公司

Dates

Publication Date: 20260512
Application Date: 20260108

Claims (10)

1. The multi-device energy consumption monitoring and management method based on the Internet of things is characterized by comprising the following steps of: continuously acquiring power data, current data and voltage data of multiple devices through intelligent sensing nodes arranged at the device end, performing time synchronization and abnormal value cleaning on the acquired data, and outputting a standardized energy consumption data stream; based on the standardized energy consumption data flow, dynamically dividing energy consumption state intervals comprising a low energy consumption interval, a medium energy consumption interval and a high energy consumption interval according to a preset energy consumption threshold, extracting time domain energy consumption characteristics, frequency domain energy consumption characteristics and statistical energy consumption characteristics for each energy consumption state interval, and generating a multi-dimensional energy consumption characteristic matrix in a fusion mode; monitoring switching frequency among energy consumption state intervals in real time, triggering an abnormal state marking process by calculating the comparison between a moving average value of the switching frequency and a self-adaptive threshold value, and marking the current energy consumption state interval when the switching frequency is abnormal; Performing similarity matching on the multi-dimensional energy consumption characteristic matrix and a historical energy consumption mode library of the equipment, performing characteristic decomposition on the multi-dimensional energy consumption characteristic matrix in a normal energy consumption state interval according to a matching result, and extracting energy consumption abnormal characteristic components; According to impedance characteristics and energy efficiency characteristics in the equipment electrical parameter library, carrying out amplitude calibration and phase compensation on the energy consumption abnormal characteristic component to obtain a calibrated energy consumption abnormal characteristic component; and constructing an equipment energy consumption correlation network, performing reverse path search in the equipment energy consumption correlation network based on the calibrated energy consumption abnormal characteristic components, positioning an equipment node of an energy consumption abnormal source, calculating an optimization priority according to the influence degree of the node, and generating an equipment regulation and control instruction.
2. The intelligent diagnosis and regulation method for multi-equipment energy consumption based on the internet of things according to claim 1, wherein the time synchronization and outlier cleaning comprises: Respectively stamping time stamps on the power data, the current data and the voltage data, and aligning the data with different sampling rates to a unified time axis by adopting a linear interpolation method; calculating the variance of a sliding window of the aligned data, and eliminating data points with variances exceeding three times of standard deviation; And smoothing the removed data by using a minimum mean square error filter in the adaptive filtering algorithm to generate a standardized energy consumption data stream.
3. The intelligent diagnosis and regulation method for multi-device energy consumption based on the internet of things according to claim 1, wherein the extracting time domain energy consumption characteristics, frequency domain energy consumption characteristics and statistical energy consumption characteristics comprises: For the power data in each energy consumption state interval, calculating the mean value, variance and peak factor in the time domain as the time domain energy consumption characteristics; Performing fast Fourier transform on the power data, and extracting the amplitude of the main frequency component as a frequency domain energy consumption characteristic; Calculating the skewness and kurtosis of the current data as the characteristic of the energy consumption; And combining the time domain energy consumption characteristics, the frequency domain energy consumption characteristics and the statistical energy consumption characteristics according to rows to form a multi-dimensional energy consumption characteristic matrix.
4. The intelligent diagnosis and regulation method for multi-equipment energy consumption based on the Internet of things according to claim 3, wherein the time domain energy consumption characteristic further comprises a zero crossing rate of power data, the frequency domain energy consumption characteristic further comprises a frequency spectrum center of gravity of current data, and the statistical energy consumption characteristic further comprises a variation coefficient of voltage data.
5. The intelligent diagnosis and regulation method for multi-device energy consumption based on the internet of things according to claim 1, wherein the triggering abnormal state marking process comprises: setting a dynamic time window, counting the switching times of the energy state interval in the window, and calculating the switching frequency; Updating the self-adaptive threshold value by an exponential weighted moving average method, and starting an abnormal state mark when the switching frequency exceeds the self-adaptive threshold value; And comparing cosine similarity of the multidimensional energy consumption characteristic matrixes of the adjacent energy consumption state intervals, and marking the current energy consumption state interval as abnormal if the similarity is lower than a preset threshold.
6. The intelligent diagnosis and regulation method for multi-device energy consumption based on internet of things according to claim 5, wherein the comparing cosine similarity of the multi-dimensional energy consumption feature matrix of adjacent energy consumption state intervals comprises: extracting a power characteristic vector and a current characteristic vector from the multidimensional energy consumption characteristic matrix; Respectively calculating the cosine value of an included angle between the power characteristic vector and the current characteristic vector in the adjacent interval; and carrying out weighted average on cosine values of the included angles of the power and the current to obtain the comprehensive similarity index.
7. The intelligent diagnosis and regulation method for multi-equipment energy consumption based on the internet of things according to claim 1, wherein the feature decomposition comprises: Calculating Euclidean distance between the multidimensional energy consumption feature matrix and a template in the equipment historical energy consumption pattern library, and taking the Euclidean distance as a similarity matching result; If the Euclidean distance is smaller than the matching threshold, reducing the dimension of the multidimensional energy consumption feature matrix by adopting a principal component analysis method, and extracting principal components as energy consumption abnormal feature components; if the Euclidean distance is larger than or equal to the matching threshold value, the abnormal characteristics are separated by using independent component analysis, and the energy consumption abnormal characteristic components are generated.
8. The intelligent diagnosis and regulation method for multi-equipment energy consumption based on the internet of things according to claim 1, wherein the amplitude calibration and phase compensation comprises: According to the electrical impedance frequency response in the equipment electrical parameter library, adjusting the amplitude attenuation of the power characteristic in the energy consumption abnormal characteristic component; based on the power factor in the energy efficiency characteristic, carrying out phase alignment on the current characteristic; And re-integrating the calibrated power characteristics, the current characteristics and the voltage characteristics, and outputting the calibrated abnormal energy consumption characteristic components.
9. The intelligent diagnosis and regulation method for multi-device energy consumption based on the internet of things according to claim 1, wherein the reverse path search comprises: analyzing the edge weight and node connection relation in the equipment energy consumption association network; Starting from the equipment node corresponding to the calibrated energy consumption abnormal characteristic component, traversing the upstream node along the edge weight reverse direction; Calculating the accumulated abnormal score of each upstream node, and locating the abnormal source equipment node of energy consumption according to the score; and calculating influence degree by combining the node depth and the abnormal score, and determining the optimization priority.
10. The intelligent diagnosis and regulation method for multi-device energy consumption based on internet of things according to claim 5, wherein the size of the dynamic time window is adaptively adjusted according to the type of device and the running time, wherein for high power devices, the time window is shortened to improve sensitivity, and for low power devices, the time window is lengthened to smooth fluctuation.

Description

Multi-device energy consumption monitoring and management method based on Internet of things Technical Field The invention relates to the technical field of energy management of the Internet of things, in particular to a multi-device energy consumption monitoring and management method based on the Internet of things. Background Current equipment energy consumption monitoring systems generally rely on the installation of smart meters or sensors at the equipment end to collect basic power, current, voltage, etc. parameters. These systems typically use a fixed threshold setting to alert to overrun or use a simple statistical analysis method to identify points of energy consumption that deviate significantly from normal levels. For group monitoring of multiple devices, it is common practice to analyze the data of each device independently, or to perform a simple summation calculation, lacking in-depth analysis of inter-device correlations and effects. The prior art solutions have drawbacks. The fixed threshold alarm mechanism is insensitive to slow energy consumption drift or instantaneous and intermittent abnormal states, and is easy to generate a missing report. The analysis mode based on the independent equipment cannot effectively cope with systematic faults which cause the linkage reaction of the related equipment due to the abnormality of one equipment. Conventional static analysis methods have difficulty capturing such dynamic unstable behavior when anomalies manifest as frequent switching of the device between different energy consumption levels, rather than steady exceeding a certain threshold. In the aspect of locating an abnormality source, the prior art often stays at the site of finding an abnormality phenomenon, and cannot quickly and accurately locate original source equipment causing the abnormality of the whole energy consumption network from a system level, so that the maintenance efficiency is low, and the regulation and control measures lack pertinence. In complex operating scenarios of multiple devices, early signs of energy consumption anomalies often manifest as rapid oscillations in operating conditions rather than absolute out-of-limits in values. The existing method lacks special monitoring and analyzing capability for the dynamic behavior characteristics, and can not send out early warning before the fault is enlarged. An anomaly of one node may propagate along a particular path due to electrical connections or process logic coupling between devices. The conventional technology does not build the correlation network model, so that operation and maintenance personnel still need to check one by experience after the abnormality occurs, automatic and rapid positioning of an abnormality source cannot be realized, and the requirements of modern intelligent energy consumption management on accuracy and efficiency are difficult to meet. Disclosure of Invention The invention aims to provide a multi-equipment energy consumption monitoring and management method based on the Internet of things, which aims to solve the problems in the background technology. In order to achieve the above purpose, the present invention provides a multi-device energy consumption monitoring and management method based on the internet of things, the method comprising: continuously acquiring power data, current data and voltage data of multiple devices through intelligent sensing nodes arranged at the device end, performing time synchronization and abnormal value cleaning on the acquired data, and outputting a standardized energy consumption data stream; based on the standardized energy consumption data flow, dynamically dividing energy consumption state intervals comprising a low energy consumption interval, a medium energy consumption interval and a high energy consumption interval according to a preset energy consumption threshold, extracting time domain energy consumption characteristics, frequency domain energy consumption characteristics and statistical energy consumption characteristics for each energy consumption state interval, and generating a multi-dimensional energy consumption characteristic matrix in a fusion mode; monitoring switching frequency among energy consumption state intervals in real time, triggering an abnormal state marking process by calculating the comparison between a moving average value of the switching frequency and a self-adaptive threshold value, and marking the current energy consumption state interval when the switching frequency is abnormal; Performing similarity matching on the multi-dimensional energy consumption characteristic matrix and a historical energy consumption mode library of the equipment, performing characteristic decomposition on the multi-dimensional energy consumption characteristic matrix in a normal energy consumption state interval according to a matching result, and extracting energy consumption abnormal characteristic components; According to impedance character