Patent Grant
12253548
The present invention relates to the field of “Smart Grids”, i.e. electricity distribution networks that integrate and efficiently manage the behaviour and actions of all users connected to the network, with the aim of guaranteeing an economically efficient operation of the electrical system, with low losses and with a high level of safety, continuity and quality of supply.
In particular, the invention relates to a system for monitoring and analyzing electrical operating parameters of a load in an electrical network.
Various devices are currently known, called in jargon “smart sockets”, capable of measuring electrical quantities, in particular mains voltage and current absorbed by the load. Although they perform their function very well, they are, however, very specific devices that are able to process the measured values locally to provide general indications about some electrical parameters, such as the power absorbed by the load, which can be used by the operators of the electric network to offer customers differentiated contracts based on specific consumption characteristics.
To date, in order to conduct advanced analyzes on the characteristics of the electrical system, network analyzers and/or oscilloscopes are used, positioned at strategic points of the electric network. These devices make it possible to meet the shortcomings of smart sockets but, at the same time, introduce problems related to:
EP3549227 describes a smart socket that can monitor and control consumption within an electrical network, periodically measuring the voltage values of the network and the current absorbed by a load connected to that network.
However, this system is limited to reading these values to remotely control the switching on and off of the load in order to optimize consumption and/or other performance indices of the load itself. On the other hand, there is no intelligent analysis of the parameters detected in order to carry out a network and load diagnostics.
In documents US2017336444, US2006119368, WO2015160779, US2016164288, US2013138651 some diagnostic systems of an electrical network are described, in which a network monitoring and pattern detection of the voltage and/or current trends is performed to compare them with predetermined patterns, generally simplified by extrapolating only some threshold values.
However, none of these systems is able to provide recognition of the specific anomaly detected on the network, rather limiting itself to the recognition of a generic anomaly and thus slowing down the process of diagnosing the origin of the disorder. Furthermore, none of these systems provides the possibility of increasing the diagnostic capabilities through continuous training of the microcontroller.
It is therefore a feature of the present invention to provide a system for monitoring and analyzing electrical operating parameters of a load in an electric network that allows to provide data processing that can be used in the Smart Grid environment for the active management of energy consumption of an electrical network.
It is also a feature of the present invention to provide such a system which allows to describe the operation of the electrical network in real time.
It is still a feature of the present invention to provide such a system that allows to conduct different types of network analysis and diagnostics according to the needs of a user.
These and other objects are achieved by a system for monitoring and analyzing electrical operating parameters of a load in an electric network according to claims from 1 to 10.
According to another aspect of the invention, a method for monitoring and analyzing electrical operating parameters of a load in an electric network that makes use of the system according to claims from 1 to 10.
Further characteristic and/or advantages of the present invention are more bright with the following description of an exemplary embodiment thereof, exemplifying but not limitative, with reference to the attached drawings wherein:
With reference to
In particular, the smart socket 110 comprises a voltage detection module arranged to measure a voltage value in the network, as an electric potential difference between the ends of the load 10 and a current detection module in the electric network 20 arranged to measure a current value adsorbed by the load 10, when the load 10 is connected to the electric network 20.
The smart socket 110 also comprises a microcontroller connected to the voltage detection module and to the current detection module.
In particular, the microcontroller comprises an artificial intelligence such as, for example, an artificial neural network.
With reference even at
In particular, the training provides a first step where the neural network acquires a plurality of patterns pij of predetermined current and/or voltage trends that are associated, according to the instructions provided, to respective events Ei. More in detail, for each event Ei the neural network acquires a number mi of patterns pij of current and/or voltage trend that highlight the disturbance or anomaly represented by this event Ei.
On the basis of the set of mi patterns pij associated with the event Ei, the neural network is therefore capable of classifying this event E; and of extrapolating a plurality of characteristic parameters Cik distinguishing the patterns Pij of this set. Such characteristic parameters Cik, chosen independently by the neural network, allow to evaluate whether a new current and/or voltage trend pattern acquired by the neural network is attributable or not to the same set and therefore is, possibly, attributable to the event Ei associated with that set.
After placing the smart socket 110 in series between the load 10 and the electric network 20, the microcontroller makes it possible to set the kind of diagnostic to carry out in the network. The microcontroller proceeds then to carry out a periodic acquisition, for example with frequency f=10 kHz, of the voltage and current values in the electric network, obtaining the current and voltage trends over time. Such current and voltage trends over time, as well as the current and voltage trends used in the training, can be shown graphically by the waveforms, some examples of which are shown in
The microcontroller then proceeds to carry out a comparison between the measured voltage and current trends and some predetermined voltage and current trends corresponding to the correct functioning of the network, to identify the presence of possible anomalous patterns.
In case that at least one anomalous pattern is detected, the microcontroller proceeds, through the neural network, to the search of characteristic parameters Cik that allow to verify whether the anomalous pattern is attributable to one of the events E; classified during the step of training. In particular, in case that the identified anomalous pattern comprises at least one predetermined number of characteristic parameters Cik of a classified event Ei, the microcontroller confirms the presence of this event Ei in the acquired voltage and/or current trend and proceeds to communicate this disturbance or anomaly, for example by means of a sound and/or acoustic alarm emitted by to smart socket 110 and/or by the graphic interface 120.
If, on the other hand, the anomalous pattern does not include at least the aforementioned predetermined number of characteristic parameters Cik of a classified event Ei, the microcontroller confirms the presence of an unclassified event in the acquired voltage and/or current trend.
In this second case, the microcontroller can then proceed in different ways, depending on the instructions provided during the step of training.
In particular, in a first embodiment, the microcontroller can simply issue an alarm that notifies the detection of an unclassified event, possibly requesting an update of the training step by inserting the unclassified event among the events being defined.
Alternatively, the microcontroller can proceed with the analysis of the electrical network 20 until a predetermined number of unclassified events associated with patterns having a certain number of common characteristic parameters cik is reached. The commonality of these characteristic parameters Cik indicates in fact that the anomalous patterns detected do not show disturbances different from each other but all highlight the same disturbance and must therefore be included in a set associated with a new event Ei to be classified. This classification operation of the new event Ei can be carried out by an external operator, upon signaling from the microcontroller, or be carried out independently by the microcontroller itself.
Therefore, the system for monitoring and analyzing electrical parameters, according to the present invention, allows, thanks to the training step and the continuous updating of this step, an extremely higher capacity to recognize the disturbance or operating anomaly compared to prior art systems.
In particular, the comparison between the detected voltage and/or current trends and the predetermined voltage and/or current trends can take place both numerically, processing the voltage and/or current values detected by the smart socket, and at graphic level, processing the spectrogram obtained from the detected voltage and/or current waveforms and verifying their correspondence with a predetermined spectrogram.
In the first case, the microcontroller can make use, for example, of recurrent neural networks with the Long Short Term Memory algorithm which allows to process sequences of values and is divided into three main cells:
At each step the input signal is processed and an output is defined starting from the information deriving from the previous inputs and from the current one.
In the second case, the microcontroller can for example make use of convolutional neural networks that recognize graphic patterns of the spectrograms of sampled waveforms.
Furthermore, the microcontroller periodically sends the network diagnostics to the graphic interface 120, which includes both the operating status of the network and some quantities of interest derived from the voltage and current trends, such as the active power, the reactive power and phase shift (cos φ) of the network. In addition, the microcontroller can send an alarm signal in case the network operating status is anomalous.
In this way, an operator can constantly monitor the status of the network and the processed data can be used for the optimization of the network itself or for the identification and location of faults on the network or in the load itself.
By way of example,
The foregoing description some exemplary specific embodiments will so fully reveal the invention according to the conceptual point of view, so that others, by applying current knowledge, will be able to modify and/or adapt in various applications the specific exemplary embodiments without further research and without parting from the invention, and, accordingly, it is meant that such adaptations and modifications will have to be considered as equivalent to the specific embodiments. The means and the materials to realise the different functions described herein could have a different nature without, for this reason, departing from the field of the invention. it is to be understood that the phraseology or terminology that is employed herein is for the purpose of description and not of limitation.
| Number | Date | Country | Kind |
|---|---|---|---|
| 102019000025855 | Dec 2019 | IT | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/IB2020/062594 | 12/31/2020 | WO |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2021/137193 | 7/8/2021 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 20050194936 | Cho | Sep 2005 | A1 |
| 20060119368 | Sela et al. | Jun 2006 | A1 |
| 20100301784 | Tagome | Dec 2010 | A1 |
| 20130138651 | Lu et al. | May 2013 | A1 |
| 20160164288 | Yang et al. | Jun 2016 | A1 |
| 20170111000 | Saito | Apr 2017 | A1 |
| 20170271915 | Quinn | Sep 2017 | A1 |
| 20170336444 | Sela | Nov 2017 | A1 |
| Number | Date | Country |
|---|---|---|
| 3549227 | Oct 2019 | EP |
| 2 560 032 | Aug 2018 | GB |
| 2015160779 | Oct 2015 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20230054387 A1 | Feb 2023 | US |
