PL248440B1 - Device and method for detecting changes in operating mode and identifying electrical receivers - Google Patents

Abstract

The subject of the application is a device for detecting a change in operating mode and identifying electrical receivers and a method for detecting a change in operating mode and identifying receivers, in particular with learning systems and detecting changes in these states. The device for detecting a change in operating mode and identifying OEE electrical energy receivers in an electrical power network comprises a voltage path that includes a voltage measurement module (22). The input of the voltage measurement module (22) is connected in parallel to the power network, and the output via a phase detection module (10) to a sample selection module (9). The device comprises a current path that includes a current measurement module (32), the input of which is connected in series to the power network and the output to the sample selection module (9). The output of the sample selection module (9) is connected to a matrix forming module (11), wherein the sample matrix from the matrix forming module (11) contains a predetermined number of the last K periods of the voltage signal and m samples of the instantaneous current value for each of the last K periods. The output of the matrix forming module (11) is connected to the change vector determination module (15) via the filtration module (12) and the potential event detection module (13). The potential event detection module (13) is adapted to transmit to the input of the change vector determination module (15) the sample matrix from the matrix forming module (11) and additional information about the period number in which a potential change in the OEE operating mode occurred. The change vector determination module (15) is adapted to determine change vectors, which change vectors are transmitted via the output of the change vector determination module (15) to the input of the decision module (18). Furthermore, the output of the OEE pattern memory module (17) is connected to the input of the decision module (18). The output of the decision module (18) provides information on the detection of a change in the operating mode or the identification of OEE electrical energy consumers.

Description

Description of the invention

The subject of the invention is a device for detecting a change in the operating mode and identifying electrical receivers and a method for detecting a change in the operating mode and identifying electrical receivers, in particular with systems for learning and detecting changes in these states.

NILM (nonintrusive load monitoring) devices, or NIALM (nonintrusive appliance load monitoring), provide information on the energy consumption of individual electrical appliances (OEE) in a given circuit of a low-voltage (LV) power network. A significant feature of NILM systems is that measurements are taken at only one point in such a circuit, usually near the energy meter or fuse box. The purpose of NILM devices and systems is to provide end users with detailed information on the energy consumption of all or some of the electrical appliances (OEE) within the user's network area (apartment, house, or multifamily building). NILM systems enable the understanding of user energy consumption habits, thereby increasing their awareness in this area, which, according to research (International Energy Agency (IEA) World Energy Outlook (WEO), 2022), leads to reduced energy consumption.

NILM systems and devices are designed to monitor only the main circuit that supplies current and voltage to the user's area, such as a home, building, or other location. Based on changes in the signals occurring there, they determine which energy-consuming appliances (EE) within that location are on and what their current operating state is. The main advantage of NILM systems is that there is no need to install multiple measuring devices at each EE (as is the case with ILM (intrusive load monitoring) systems), which should ultimately reduce the cost of developing an energy monitoring system and significantly simplify its installation in existing locations. As a result of the NILM system, the end user can receive a report detailing the individual energy consumption for at least a portion of the EE within that user's area.

Each NILM device comprises an acquisition block, which records current and voltage signal samples, and a sample processing block, where algorithms determine the distribution of circuit load across individual electronic equipment (OEE). NILM devices utilize data processing methods to extract signal characteristics, which are then used by artificial intelligence algorithms to analyze the current operating states of individual OEEs. The result of this analysis is information about which OEEs are currently on and, if applicable, their current operating mode. This information enables the estimation of electrical energy consumption by individual OEEs. Individual OEEs are identified based on a signature (pattern), which can be, for example, a numerical data vector enabling the recognition of OEEs and their operating states.

The devices developed so far using NILM methods do not enable the accurate identification of all OEE used in the user area, such as a household, which was the motivation to develop our own, effective and cheap to implement methods for use in the NILM device.

One of the methods to overcome the known shortcomings of NILM methods is to develop mathematical models (e.g. in the form of state equations of a dynamic system) reflecting the physical properties of some OEE, as in the solution disclosed and published in patent document EP2686937B1 and in other publications, e.g. [Wittmann, F.M.; Lopez, J.C.; Rider, M.J. Nonintrusive Load Monitoring Algorithm Using Mixed-Integer Linear Programming. IEEE Trans. Consum. Electron. 2018, 64, 180-187, doi:10.1109/TCE.2018.2843292].

The methods presented in these disclosures have not been implemented to date, primarily because the models they contain are overly simplified and do not provide satisfactory OEE identification results (especially in real-world use). It should be noted that creating a good and sufficient universal OEE model is an open and difficult task due to the complex physical nature of most OEE.

On the other hand, using mathematical models, which are already complex, requires complex computational methods and significant computing power to use these models to identify OEE. Creating an accurate NILM device using a mathematical OEE model is difficult due to the high production costs.

Due to the shortcomings of existing solutions for detecting potential events mentioned here, it would be beneficial to have a practical solution to the problem of effective detection of potential OEE changes that does not require many hardware resources of the NILM device.

State-of-the-art systems and methods for OEE identification, using e.g. Hidden Markov Models or EMI frequency analysis, or the above-mentioned mathematical models, are difficult to transfer to small and cheap devices that could be installed outside a specialized laboratory in real-world conditions (e.g. in a fuse box in an apartment).

In order to improve the low efficiency of NILM systems, additional sensors are used in prior art solutions, not related to electrical signals coming from the LV power network, as e.g. in the invention disclosed and described under numbers: US11036189B2 and in scientific articles, e.g.: [Zhao, T.; Zhang, C.; Ujeed, T.; Ma, L. Online Methodology for Separating the Power Consumption of Lighting Sockets and Air-Conditioning in Public Buildings Based on an Outdoor Temperature Partition Model and Historical Energy Consumption Data. Appl. Sci. 2021, 11, 1-23, doi:10.3390/app11031031]. Typically, these sensors provide environmental information, such as temperature, lighting, sound, vibration, etc., but also data from other media, such as LANs—for example, patent publications no. US10175276B2 or US9699529B1. Information from such sensors is intended to support NILM device decision-making systems in more accurately identifying changes in OEE operating states. However, the use of multiple additional sensors contradicts the fundamental advantage of NILM systems, which involves using the fewest possible sensors, unlike ILM systems. This, coupled with the questionable effectiveness of NILM systems supported by information from environmental sensors, means that such solutions are also not widely used in practice.

A noted problem with prior art NILM methods is that many disclosed OEE identification methods were developed using only test sets, often not reflecting the conditions in real-world OEE use, e.g., in a home environment. This results in two significant drawbacks. The first is the aforementioned lack of immunity to interference in real-world LV power networks originating from outside the monitored user area (apartment, house, or multifamily building). The second is the insufficient accuracy of identifying low-power OEE when these OEE operate in the presence of other OEE with higher and variable power.

From the perspective of practical implementation of a NILM device, insufficient detection accuracy of potential events often leads to numerous false positives regarding monitored OEE events, which must be identified later in the device's operation. This phenomenon significantly increases the resource consumption of the NILM device, significantly increasing the cost of its practical implementation. Conversely, insufficient sensitivity in potential event detection typically leads to these events being missed during the identification stage, causing the NILM device to miss some changes in the monitored OEE states. Most proposed practical solutions rely on detection of changes in averaged power or current values, e.g., RMS, or changes in phase angles.

The authors of this application noted that this approach is characterized by low detection efficiency of events originating from low-power OEEs on the one hand, and on the other hand, high susceptibility to interference penetrating the monitored power grid circuit from external circuits (this phenomenon is particularly important in networks with a configuration similar to that used in Europe). Additionally, the authors of this application noted that in NEM systems and devices with potential event detection using changes in averaged power or current values, events originating from lower-power OEEs are often masked by multi-state OEEs with higher power and variable power consumption.

There are many disclosed NILM methods, devices, and systems that identify OEE-related events in LV power networks based on OEE patterns consisting of a specified number of signal parameters related to power consumption. Such devices acquire power or energy signals from meters and sensors measuring the fundamental harmonic of the power network voltage. In the case of methods based on higher harmonic parameters, the devices use additional measurement modules that measure current or voltage signals, usually up to the seventh harmonic of the LV power network voltage signal.

The signal parameters that are used in these NILM methods for OEE identification are typically related to the fundamental harmonic of the voltage or current in the LV network, as e.g. in the international application publication no. WO2021176459, and in the methods described in scientific articles, e.g.: [Liu, B.; Luan, W.; Yu, Y. Dynamie Time Warping Based Non-Intrusive Load Transient Identification. Appl. Energy 2017, 195, 634-645, doi:10.1016/j.apenergy.2017.03.010] or parameters related to the harmonics of these signals, as e.g. in the inventions published under numbers: WO2016079229A1, and in the methods described in scientific articles, e.g.: [Agyeman, K.A.; Han, S.; Han, S. Real-Time Recognition Non-Intrusive Electrical Appliance Monitoring Algorithm for a Residential Building Energy Management System. Energies 2015, 8, 9029-9048, doi:10.3390/en8099029].

To improve OEE identification, many early NILM systems and methods utilized Hidden Markov Models (HMMs), as demonstrated in the invention disclosed and published under the number EP2499462B1 and in scientific articles such as [Kolter, J.Z.; Jaakkola, T. Approximate Inference in Additive Factorial HMMs with Application to Energy Disaggregation. J. Mach. Learn. Res. 2012, 22, 1472-1482]. However, these models are computationally complex both during the NILM system testing phase and during the OEE identification phase. Furthermore, such methods and systems typically do not achieve OEE identification performance levels high enough to justify the high computational complexity. Furthermore, models using HMMs face a performance problem of significantly delaying the identification effect with respect to the event associated with the change in the OEE operating state.

Unfortunately, despite many years of research on the NILM group of methods using signals related to the first harmonic of the LV power supply network voltage signal or its harmonics (LF methods, for low frequency), they have not been effectively implemented, primarily due to the imperfect identification in the real-world conditions in which typical LV power supplies are used. LF methods are characterized by low immunity to interference from outside the monitored LV power supply network circuit, low efficiency in identifying operating states of LV power supplies with low rated power, and low efficiency in identifying operating states of LV power supplies with variable power.

Most OEE identification methods require prior detection of potential events related to these OEEs, defined as switching them on, off, or changing their operating state. The commercialization of such OEE identification methods has been hampered by the level of unwanted noise present in the electrical grids of modern buildings. In such conditions, attempts to detect potential events (e.g., changes in device status) prove significantly more difficult than in laboratory conditions, significantly hindering the accuracy of such detection. Despite numerous publications on algorithms for identifying energy consumption by individual OEEs, there are still few studies demonstrating the results of testing these algorithms in the real-world operating environment of NILM systems, such as in real residential buildings. It should be noted that the potential event detection aspect of NILM methods is not emphasized in scientific articles, as these typically focus solely on OEE identification. Authors of scientific publications describe studies of OEE identification algorithms, often conducted solely using publicly available datasets, for example, those described in scientific articles.

Such "tagged" datasets, commonly used to develop and test OEE identification methods, contain data collected from monitored circuits that already contain information from measurement systems about the moments of OEE state changes. Although such datasets are very convenient to use in research on OEE identification algorithms, in practice this approach is insufficient, and an effective event detection method is necessary for the practical application of OEE identification methods in a real-world NILM system.

A partial solution to the identified identification efficiency problems are NILM methods that identify OEE operating states based on changes in the frequency spectrum of voltage or current signals in the high-frequency band (HF methods). Known solutions include those disclosed, for example, in U.S. application no. US20130179124 and methods discussed in scientific articles, e.g., in: [Gupta, S.; Reynolds, M.S.; Patel, S.N. ElectriSense: SinglePoint Sensing Using EMI for Electrical Event Detection and Classification in the Home. In Proceedings of the Proceedings of the 12th ACM international conference on Ubiquitous computing; ACM: New York, NY, USA, September 26, 2010; pp. 139-148]. The methods discussed in these disclosures address some of the OEE identification issues outlined earlier, but these methods result in unacceptable increases in the price of devices and systems utilizing them. This is due to the significant computing power required to determine signal transforms, increased requirements for the analog-to-digital converters used (high sampling rate, high bit resolution), and often significant demands on the telecommunications links connecting the NILM system components.

A disadvantage of most disclosed NILM methods that utilize high-frequency spectral variations is their susceptibility to interference from outside the monitored LV power network circuit, as they rely on voltage measurements in the monitored circuit. This property is less evident in LV power networks, for example, in the USA, due to the different distribution network architecture. In European networks, due to their architecture, in which multiple user areas are supplied within a single distribution network circuit, interference from outside the monitored power network circuit (user area) significantly reduces the effectiveness of NILM devices that utilize voltage variations. Better properties are provided by methods that use the spectrum of the current signal in steady states of OEE operation and in transient states, as well as additionally using an artificially introduced known pulse signal to the monitored circuit, as e.g. in the solutions disclosed and published under the numbers: US20140207398A1, US9104189B2 and discussed in the scientific article [Wójcik, A.; Bilski, P.; Łukaszewski, R.; Dowalla, K.; Kowalik, R. Identification of the State of Electrical Appliances with the Use of a Pulse Signal Generator. Energies 2021, 14, 673, doi:10.3390/en14030673]. However, these methods are difficult to implement and expensive, as they require, for example, the installation of a pulse generator and high-quality analog and analog-to-digital converters.

Both in LF methods using parameters of harmonic signals of currents and voltages of the LV power supply network and in HF methods in the high frequency band, the FFT transform is often used to calculate the parameters of OEE patterns, as e.g. in the inventions disclosed and published under the numbers: US20140200725A1 or in the methods described in scientific articles, e.g.: [Gupta, S.; Reynolds, M.S.; Patel, S.N. ElectriSense: Single-Point Sensing Using EMI for Electrical Event Detection and Classification in the Home. In Proceedings of the Proceedings of the 12th ACM international conference on Ubiquitous computing; ACM: New York, NY, USA, September 26, 2010; pp. 139-148]. In other known solutions, the authors use a wavelet transform instead of the FFT transform, as for example in the publication number: US8983670B2, and in the scientific article e.g.: [Su, Y.C.; Lian, K.L.; Chang, H.H. Feature Selection of Non-Intrusive Load Monitoring System Using STFT and Wavelet Transform. Proc. - 2011 8th IEEE Int. Conf. E-bus. Eng. ICEBE 2011 2011, 293-298, doi:

10.1109/ICEBE.2011.49].

The first method, the FFT method, completely erases information about the time instants relative to the fundamental component of the signal at which significant signals related to changes in the OEE operating mode appear. A better solution is to use appropriate wavelet transforms. However, even in this case, the necessary information is not fully preserved.

NILM methods using OEE signatures have encountered another limitation during their commercialization, significant for potential users of NILM devices. All such methods require either lengthy training of the classifier used, e.g., in the form of a neural network, or the tedious process of training the NILM device through interaction with its user. This problem stems from the properties of the algorithms used, which require providing a number of examples in the process of training the classification of OEE state change signatures. One disclosed approach to training NILM algorithms involves the user repeatedly manually turning OEE on and off while the NILM device is in training mode, as in the solution disclosed in patent application number US20210158186A1. Prior art solutions disclose methods for training, i.e., determining patterns, e.g., in the form of OEE signature vectors. In these solutions, signature values are specific parameters calculated based on measured current or voltage waveforms or directly measured parameters of these signals, such as: active power, reactive power, phase shift angle between the current and voltage signals, peak, average, or effective values of these signals, etc. In the previously described aspect related to the effectiveness of OEE identification methods, it was noted that using such signatures is not sufficient for effective identification of OEE states. Known inventions lack descriptions of methods for training NILM devices that use vectors as OEE signature patterns, which are current signal samples or differences of such vectors, and such patterns are used in the identification method in the present invention.

Other disclosed inventions that create various OEE models describe a method for training these models. The OEE models used require extensive training using multiple data sources. Any suitable techniques may be used to train an OEE model, as described, for example, in U.S. Patent No. US9443195. Training techniques may include, but are not limited to, training neural networks, self-organizing maps, Support Vector Machines (SVMs), decision trees, random forests, logistic regression, Bayesian models, linear and nonlinear regression, and models of mixtures of Gaussian distributions; as, for example, described in WO2011128883A2.

It should be noted that in the disclosed prior art solutions, the aspect of the NILM device training stage is often omitted, as e.g. in the publication of international application number WO2021176459A1. Also, authors of other publications, similarly to the issue of potential event detection, often do not address the problem of training the NILM system (understood as determining real OEE patterns), focusing on the scientific problem of OEE identification, e.g. in: [S. Bao et al., "Feature Selection Method for Nonintrusive Load Monitoring With Balanced Redundancy and Relevancy", in IEEE Transactions on Industry Applications, vol. 58, no. 1, pp. 163-172, Jan.-Feb. 2022, doi:10.1109/TIA.2021.3128469]. In other studies, authors of scientific articles determine OEE patterns from existing, described in detail ("tagged") datasets, e.g. in: [Joao Góis, Lucas Pereira, A novel methodology for identifying appliance usage patterns in buildings based on auto-correlation and probability distribution analysis, Energy and Buildings, Volume 256, 2022, 111618, ISSN 0378-7788, https://doi.org/10.1016Zj.enbuild.2021.111618]. This approach has no practical application in real-world implementations of NILM devices and systems, where the user expects a way to add patterns to the system for new OEE, added to the monitored system after the NILM device is installed in the operating environment of that device or NILM system.

It would be beneficial to provide a practical solution to the problem of training a NILM device, involving a single user interaction for a given OEE in the monitored electrical circuit for the entire duration of the NILM device's use. It would be beneficial to adapt the developed effective training method to NILM devices that use vectors as OEE signature patterns, which are current signal samples or differences of such vectors.

Many disclosed NILM methods and systems, e.g. using Markov models (HMMs), as e.g. in published patent document number: EP2499462B1 or described in scientific articles e.g.: and in scientific articles e.g.: [Kolter, J.Z.; Jaakkola, T. Approximate Inference in Additive Factorial HMMs with Application to Energy Disaggregation. J. Mach. Learn. Res. 2012, 22, 1472-1482] or state models, as e.g. in inventions disclosed and published under numbers: US9739813B2 and in scientific articles e.g.: [Liu, Y.; Liu, C.; Shen, Y.; Zhao, X.; Gao, S.; Huang, X. Non-Intrusive Energy Estimation Using Random Forest Based Multi-Label Classification and Integer Linear Programming. Energy Reports 2021,7, 283-291, doi:10.1016/j.egyr.2021.08.045] cannot report results in real time to end users. This is because such approaches require long periods of observation of full OEE operating cycles to identify their states in the NILM device. As a result, results showing which OEEs were turned on and off and what operating states they were in may only be provided after many hours. In other inventions, data from current or voltage measurements are often transmitted periodically, e.g., every 30 seconds, as in the invention published under the number: WO2021176459A1, e.g., to a central server or cloud services to identify OEEs based on the transmitted data. Only after this data is transmitted does the system decide which OEEs are in the currently defined operating states and then, with a considerable delay, reports the results to the user. Solutions with this drawback are not suitable for widespread implementation in practice. Therefore, solutions that operate in near real-time mode and display changes in OEE operating states with the minimum delay will be beneficial.

The authors of this invention have observed that only appropriate processing of the current signal in the time domain allows for the preservation of all the important information contained in this signal, characteristic for the identification of OEE operating states.

This application is a partial continuation of the Polish invention registered under patent number Pat.243560 on April 24, 2020. The subject of the invention was a device for detecting a change in the operating mode and identifying electrical receivers and a method for detecting a change in the operating mode and identifying electrical receivers.

A device for detecting a change in the operating mode and identifying OEE electrical energy receivers in an electrical power network, in accordance with patent Pat.243560, comprises a voltage path which includes a low-pass filter, the input of which is connected in parallel to the power network and the output through a voltage measurement module and an analog-to-digital converter module and through a phase detection module to the processing module, and a current path which includes a current measurement module, the input of which is connected in series to the power network and the output through an analog-to-digital converter module and a buffer module to the processing module, at the same time the output of the processing module is connected through a cyclic buffer and a filtering module to the input of the change vector determination module, the output of which is connected to the input of the discriminator module, furthermore the output of the OEE standard memory module is connected to the input of the discriminator module.

Patent Pat.243560 also discloses a method for detecting a change in the operating mode and identifying OEE electrical receivers using measurements of instantaneous values of current signals and detection of the voltage phase, which includes:

a) the current measurement stage in the power supply network and the voltage measurement stage in that

b) after the current measurement stage, the current signal is converted into its digital form, which is transmitted to the buffer module, and

c) after the voltage measurement stage, the voltage signal is converted into its digital form, on the basis of which a specific phase value of the voltage signal is detected using the phase detection module, and then

d) using the processing module, appropriate sample values are selected from the buffer module, which are then formed into a matrix in which subsequent columns contain the values of current samples in subsequent voltage periods, starting from the moment the voltage signal reaches a specific phase value, and then

e) using the filtration module, the matrix is filtered by its rows, obtaining a matrix of filtration samples, from which the change vector determination module determines the change vectors that subtract the sample values of the sample matrix after filtration from the period before the change in the state of the OEE receiver from the sample values of the sample matrix after filtration from the period recorded by the change in state, when the OEE receiver that has changed state and will be in a new steady state, and then

f) the change vectors determined using the discriminator module are compared with the reference feature vectors of all OEE receivers in a given electrical circuit stored in the pattern memory module, identifying the OEE receiver based on the most closely matching pattern.

The present invention solves the problem of event detection for identification methods that utilize the time-based form of a signal carrying information about electrical power, preferably for a current signal. Solving this problem overcomes a major limitation of many known event-driven identification methods: the strong dependence of the entire system's performance on event detection. In such solutions, if a single event is missed, the error tends to propagate throughout the system. The present invention allows for increased reliability of event detection in practical embodiments of the present invention, which include a module adapted for efficient event detection with continuous analysis of current signal samples. It should be noted that the event detection method employed is described by advantageously simple operations involving subtraction and comparison of numerical vector values. Therefore, the computational complexity of the event detection employed in the present invention is advantageously very low.

The present invention solves the problem of training identification methods that utilize the temporal form of a signal carrying information about electrical power, preferably for a current signal. According to the invention described, the identification method used therein does not require the development of an OEE model, and therefore, there is no need to train it. Preferably, training the system involves a single (but not limited to, multiple) monitoring and recording of two numerical vectors of current signal sample values measured before and after a change in the monitored OEE state, and calculating and recording the differences in these vector values as a reference for that OEE.

A device for detecting a change in the operating mode or identifying OEE electrical energy receivers in an electrical power network, according to the invention, comprises:

a voltage path that includes a voltage measurement module, the input of which is connected in parallel to the power grid and the output via a phase detection module to the sample selection module; and a current path that includes a current measurement module, the input of which is connected in series to the power grid and the output to the sample selection module;

the output of the sample selector is connected to the matrix forming module, wherein the sample matrix from the matrix forming module comprises a predetermined number of the last K periods of the voltage signal and m samples of the instantaneous current value for each of the last K periods;

at the same time, the output of the matrix formation module is connected via the filtration module to the input of the change vector determination module, wherein the change vector determination module is adapted to determine change vectors, which change vectors are transferred via the output of the change vector determination module to the input of the decision-making module, furthermore, the output of the OEE pattern memory module is connected to the input of the decision-making module;

where the output of the decision-making module provides information on the detection of a change in the operating mode or the identification of OEE electrical energy receivers.

The essence of the invention is that the device for detecting a change in the operating mode or identifying OEE electrical energy receivers in the power network is characterized in that the filtration module is connected to the change vector determination module via a potential event detection module, wherein the potential event detection module is adapted to transmit to the input of the change vector determination module the sample matrix from the matrix forming module and additional information about the number of the period in which a potential change in the OEE operating mode occurred.

It is advantageous when the decision-making module comprises a similarity determination module and a decision module, wherein the input of the similarity determination module is fed with data from the change vector determination module and data from the pattern memory module, the output of the similarity determination module is connected to the first input of the decision module, and the second input of the decision module is connected to the output of the pattern memory module.

It is advantageous if the similarity determination module in the decision module is adapted to compare the values of change vectors with the pattern definition vectors that are stored in the OEE pattern memory module.

It is advantageous if the decision module in the decision module is adapted to identify the OEE receiver and/or identify a change in the state of the OEE receiver based on data from the similarity determination module and the pattern memory module.

It is advantageous when an operating mode selection module is placed between the potential event detection module and the change vector determination module, wherein one of the outputs of the operating mode selection module is connected to the decision module via the pattern vector determination module and the pattern memory module, and the other output of the operating mode selection module is connected to the decision module via the change vector determination module and via the similarity determination module, wherein when the new OEE pattern registration mode is active, the operating mode selection module is adapted to transmit information from the input to the pattern determination module, when the new OEE pattern registration mode is not active, the operating mode selection module is adapted to transmit data only to the change vector determination module.

It is advantageous when the phase detection module is adapted to determine the time instant in which the value of the voltage signal of the electrical network reaches a predefined phase value, preferably when the value of the voltage signal of the electrical network changes from negative to positive.

It is advantageous when the sample selection module is adapted to transfer to the matrix forming module the values of current samples, starting from the sample number determined by the phase detection module for each subsequent period of the voltage signal, wherein the number of transferred samples results from the duration of one period of the voltage signal and the sampling frequency in the current measurement module.

It is advantageous if the matrix forming module is adapted to store the values of current samples for a certain number K of the last periods of the voltage signal, which form a sample matrix that is fed to the input of the filtering module.

It is advantageous when the filtration module is adapted to implement median filtration or averaging filtration or modal filtration.

It is advantageous when the filtration module is adapted to perform filtration for a series of samples taken at moments equidistant in time from the phase determined by the phase detection module.

It is advantageous when the voltage measurement module comprises a voltage transformer to the secondary winding of which a low-pass filter is connected, and the filter output is connected via an amplifier module to an analog-to-digital converter module, the output of which is the output of the voltage measurement module, or the voltage measurement module comprises a voltage transformer to the secondary winding of which an amplifier module is connected, which is connected to an analog-to-digital converter module and further to a digital low-pass filter module, and the output of the digital low-pass filter is the output of the voltage measurement module.

It is advantageous when the current measurement module comprises a current transformer, to the secondary winding of which an amplifier module is connected, which is connected to an analog-to-digital converter module, preferably an anti-aliasing filter is placed between the secondary winding of the current transformer and the amplifier module, wherein the passband of the anti-aliasing filter is related to the sampling frequency of the transducer in the analog-to-digital converter module, or the current measurement module comprises a current transformer, to the secondary winding of which an amplifier module is connected, which is connected to an analog-to-digital converter module, preferably the analog-to-digital converter module is connected to a digital anti-aliasing filter module, wherein the output of the digital anti-aliasing filter or the analog-to-digital converter is the output of the current measurement module.

According to the invention, a method for detecting a change in the operating mode and identifying OEE electrical receivers using measurements of instantaneous values of current signals and detection of the voltage phase comprises:

a stage of measuring the current in the supply network and a stage of measuring the voltage, whereby after the stage of measuring the current, the current signal is converted into its digital form, which is transferred to the sample selection module, and after the stage of measuring the voltage, the voltage signal is converted into its digital form, on the basis of which, using the phase detection module, a specific phase value of the voltage signal is detected, and then, in the stage of forming the matrix, using the matrix forming module, appropriate sample values are selected from the sample selection module, which are further formed into a matrix in which the next m rows contain samples of instantaneous current values in k columns of subsequent voltage periods starting from the moment the voltage signal reaches a specific phase value, and then, in the filtering stage, using the filtration module, the matrix is filtered by its rows, obtaining a matrix of samples after filtration, from which the change vector determination module determines change vectors comprising the differences in the sample values of the sample matrix after filtration from the period before the change in the state of the OEE receiver from the sample values of the sample matrix after filtration from the period recorded after the analyzed change in state, when the OEE receiver that has changed state and will be in a new steady state, and then follows the identification stage, in which the change vectors determined by the module are The decision-making process is compared with the reference feature vectors of OEE receivers in a given electrical circuit stored in the pattern memory module, identifying the OEE receiver based on the best-matching pattern.

The essence of the invention is that the method of detecting a change in the operating mode and identifying OEE electrical receivers using measurements of instantaneous values of current signals and detection of the voltage phase is characterized in that after filtering the sample matrix in the filtration module, a step of detecting potential events takes place, wherein the detection consists in determining in the module of detecting potential events change vectors for at least some of the K voltage periods, which vectors allow for determining a proportionality coefficient proportional to the average of the absolute values of the elements of these vectors, wherein the period k, in which the determined coefficient locally reaches the maximum value, is considered to be the period in which a potential change in the OEE mode or state occurred, then information about the number of the period in which a potential change in the OEE mode or state occurred is transferred to the module for determining change vectors. The change vectors determined in the change vector determination module include the differences in the sample values of the sample matrix after filtering for the period before the period in which a potential change in the OEE mode or state occurred, from the sample values of the sample matrix after filtering for the period after the period in which a potential change in the OEE mode or state occurred, with only those periods for which the proportionality coefficient reached its maximum value being taken into account. Only selected change vectors determined in the change vector determination module are further compared in the decision module and are used to identify OEE based on the best-matching pattern.

It is advantageous when determining the proportionality coefficient in the potential event detection module comprises determining change vectors comprising differences in the sample values of the sample matrix after filtration from the period k-o' before the change in the OEE receiver state from the sample values of the sample matrix after filtration from the period k+o" recorded after the analyzed period k, wherein only selected periods before and after are analyzed for the predefined at least one pair of shift values o' and o", preferably shift o' and o" corresponds to the OEE category and the shift values o' and o" may not be equal, the proportionality coefficient is determined on the basis of the value of the determined difference in the sample values of the matrix for each predefined at least one pair of shifts o' and o", preferably the proportionality coefficient is the average of the absolute value of the determined difference in the sample values.

It is advantageous to use the phase detection module to determine the moment in time at which the voltage value of the electrical network changes from negative to positive.

It is advantageous when, by means of the matrix forming module, the values of current samples are transferred from the sample selection module in the form of a sample matrix, wherein for each period k, starting from the sample determined by the voltage phase detection module, a number m of samples corresponding to each of the voltage periods is transferred, preferably the matrix forming module contains a matrix buffer for the last K voltage periods and transfers to the output a sample matrix for each voltage period, which sample matrix is fed to the input of the filtering module.

It is advantageous when the filtration is performed for a series of samples taken at moments equidistant in time from the phase determined by the phase detection module.

It is advantageous if all steps are performed on time-domain signal samples.

It is advantageous when the current signal sampling moments are correlated in time with the moment when the voltage signal reaches a specific phase.

It is advantageous when the filtration is a median filtration or an averaging filtration or a mode filtration.

Preferably, the method comprises:

- determining in the determination module similarities between all the pattern change vectors taken from the pattern memory module, wherein the pattern memory module contains pattern change vectors characteristic for OEE, and the set of change vectors determined in the change vector determination module, wherein the determined similarities are transferred to the decision module;

- making decisions by the decision module based on similarities determined in the similarity determination module and on the basis of data from the pattern memory module with a threshold value characteristic for each type of OEE and a possible change in the state of this OEE, the pattern of which is stored in the pattern memory module.

It is advantageous when decision-making by the decision module includes:

determining the maximum value of the similarity coefficient for the entire similarity vector, and comparing this maximum value with the threshold value taken from the pattern memory module, then making a decision by the decision module based on this comparison, if the maximum value of the similarity coefficient for the similarity vector exceeds the threshold value, then the decision module decides that the potential event in period k concerns a change in the state of a specific OEE, whereas if the maximum value of the similarity coefficient for the similarity vector does not exceed the threshold value, then the decision module decides that the previously detected potential event is not an event indicating a change in the state of any OEE in the tested power supply circuit of the power network.

It is advantageous when, at the decision-making stage in the decision module, in a situation where the given decision selection criteria are not sufficient, statistical algorithms are used to select the most probable OEE from among the registered patterns, preferably a decision tree algorithm or a random forest, or K-nearest neighbors (KNN), or convolutional neural networks (CNN), or support vector machines (SVM), or deep neural networks (DNN), or Naive Bayes.

It is advantageous that before the decision-making stage in the decision module (19), a step of adding change patterns to the pattern memory module (17) is performed, wherein the step of adding patterns depends on predefined data provided by the user about the category of the device being added and the moment of its connection to the network, the steps of adding patterns include:

setting the operating mode selection module (14) to the pattern addition mode (output duplication mode), which results in supplying a sample matrix to the pattern determination module (16), determining the pattern change vectors, wherein the determination of the pattern change vectors is carried out analogously to the determination of the change vectors, wherein when determining the pattern change vectors, predefined values of shifts o’ and o”, and a predefined threshold value characteristic for the given OEE category are assumed, then storing the pattern change vector, together with the corresponding shift values o’ and o”, and a predefined threshold value characteristic for the given OEE category in the pattern memory module (17), wherein after adding the pattern change vector to the pattern memory module (16), OEE recognition is started as defined in the method according to the invention or another device is added.

The present disclosure improves the prior invention (Pat. 234560), at least by changes in signal processing in the identification stage, but also by supplementing it with new effective solutions regarding the detection stage of potential events related to the monitored OEE and by supplementing it with solutions regarding the device training mode.

The present invention utilizes an identification method that preferably utilizes only the temporal form of signals, preferably average frequency. This method is described in detail later in this paper. A characteristic feature of this identification method is that it preferably considers moments within the fundamental component periods, preferably at the time of occurrence of a specific voltage level (phase of this voltage), in order to obtain information, preferably from a pseudo-periodic signal characteristic of the monitored OEE. An additional characteristic of the identification method used in the invention is a special method of filtering this pseudo-periodic signal, preferably median filtering. The applied filtration method, presented further in the detailed description of the invention, is different from the denoising filtration methods used in other inventions, which, although they facilitate event detection, significantly reduce the effectiveness of OEE identification, as they cause the loss of significant information contained in the current signals of the LV power supply network, as e.g. in the inventions disclosed and published under the numbers: US9104189B2, CN106022645B (WO2017211288A1, US20170351288A1).

It should be noted that the OEE identification method used in the present invention is described by simple operations involving subtraction and comparison of the values of a numerical vector. Therefore, the computational complexity of the event detection used in the invention is advantageously very low.

Moreover, by using a simple similarity measure for OEE classification (where scaling only requires adding new numerical vectors as patterns of new OEEs), OEE classification can be easily scaled to account for new OEEs in the monitored LV power network circuit. However, classification methods used in other inventions, such as LDA, SVM, and neural networks, often require retraining the classifier or adding (e.g., manually) new classifiers to the system. The proposed solution according to the invention is therefore scalable to larger environments with a larger number of OEEs or with OEEs having multiple operating modes.

The applied approach to the OEE identification method is therefore characterized by lower computational complexity and greater scalability compared to other inventions based on, for example, mathematical models, HMM or high-frequency analysis (EMI).

The advantageous effect of using the developed identification method in the device according to the invention is that a very good efficiency of OEE identification was achieved, using cheap means in the practical implementation of the NILM device.

In contrast to inventions employing identification methods using, e.g., H MMs or state models, the invention disclosed herein advantageously is capable of producing OEE classification results in near real time, with a delay of about 1 second or less, after enabling the actual OEE.

In the further part of the description of the invention, a device and a method are described in detail, which implements both the method of processing measurement data and the method of event detection and the method of training the system.

The subject of the invention is shown in an embodiment in the drawing, in which Fig. 1 shows a block diagram of a device according to an embodiment of the invention.

During stable operation (i.e., steady-state) of an electrical energy receiver (EEE) or multiple such receivers, it has been observed that the recorded instantaneous current values are similar at repeatable instants (tm) relative to the time the fundamental component signal of the LV network voltage reaches a specific phase (tm, where k is the number of successive periods of the fundamental component of the network voltage signal). When an additional receiver is turned on, an already-on receiver is turned off, or its status changes, the instantaneous current values undergo a change that is characteristic for the given type of receiver or its state. The concept of the device according to the invention involves identifying the operating state of the electrical energy receiver based on continuous measurements of the total current and supply voltage waveform of the LV network and processing the results of these measurements. The essence of data processing used to identify operating receivers is to determine the difference in the instantaneous current signal values in appropriately selected periods of the fundamental component signal of the LV network voltage measured at instants tm.

The device according to an exemplary embodiment of the invention includes in the voltage path 22 a voltage sensor module, which is, for example, a voltage transformer 2, the input of which is connected in parallel to the power grid, and which transforms the voltage signal of the LV network (low voltage network) to levels acceptable by typical A/A and A/C conversion modules. In the exemplary device, the voltage transformer is a transformer that steps down the voltage level Uzas. At the same time, the voltage transformer Tn provides galvanic separation of the measuring systems from the LV network. As a result, the voltage transformer module 2 transforms the supply voltage signal U zas to the voltage signal U1(t).

The output of voltage transformer 2 is connected via filter module 4 and amplifier module 6 and analog-to-digital converter module 8 to phase detection module 10.

Filter module 4 converts the voltage signal ui(t), received from voltage transformer 2, into a voltage signal ui’(t) by removing components with frequencies higher than the fundamental power grid voltage (50 Hz in Europe) from the voltage signal ui(t). The low-pass filtration used in filter module 4 is intended to facilitate the determination of instant t0. For example, a low-pass filter with a stopband above a frequency of 70 Hz is used for this purpose.

It is possible to implement filter module 4 in a digital manner. In the embodiment, this module is located after the analog-to-digital processing module 8 in the voltage path of the device according to the invention, and the amplifier module 6 is connected directly to the voltage transformer 2. Changing the analog to digital form of filter module 4 does not change the operation of the device according to the invention.

The amplifier module 6 converts the voltage signal ui’(t) to the voltage signal U1”(t) in such a way that the signal ui”(t) meets the requirements of the analog-to-digital processing module 8 in the voltage path of the device according to the embodiment. These requirements relate to the selected input voltage range for the A/D converter used in this signal processing path and depend on the specific implementation of the device. For example, the voltage amplitude U1”(t) is 12V p-p.

In the voltage path of the device, in the example embodiment, the analog-to-digital processing module 8 receives the voltage signal U1"(t), which is processed in the analog-to-digital processing module 8 into a series of samples of this signal u(n). For example, a converter with a bit resolution of 12 bits and a sampling frequency of Fs = 10 kHz was used.

After sampling the voltage signal U1”(t) in the analog-to-digital converter module 8, the digital signal u(n) is subjected to phase detection in the phase detection module 10. The operation of the phase detection module 10 consists in determining the index n0 of the sample of the u(n) series for the instant t0k. The instant t0k represents the time when the signal of the fundamental component of the network voltage Nn (where k are the numbers of the subsequent periods of the fundamental component of the network voltage signal) reaches a specific phase. The index n0 is transferred from the output of the phase detection module to one of the inputs of the current i(n) sample selection module 9 in the current path and enables the appropriate selection of current i(n) samples in this module and further processing of these samples in the manner described below.

The device according to an exemplary embodiment of the invention includes in the current path 32 a current measurement sensor module, which is, for example, a current transformer 1, the input of which is connected in series (current) to the phase conductor of the power network. The operation of the current transformer 1 Tp consists in transforming the current signal of the LV network to a voltage signal with levels acceptable by typical A/A and A/C conversion modules. Changes in the voltage signal u2(izas(t)) at the output of the current transformer 1 proportionally represent changes in the currents supplying the receivers (OEE) connected to the measured circuit of the LV power network. At the same time, the current transformer 1 provides galvanic isolation of the measuring systems from the LV power network. The voltage signal u2(izas(t)) is transmitted from the output of the current transformer 1 to further modules in the current path of the device according to the invention.

The output of the current transformer 1 is connected via the anti-aliasing filter module 3 and the amplifier module 5 and the analog-to-digital converter module 7 to the current i(n) sample selection module 9.

The anti-aliasing filter module 3 is optional and constitutes an aspect of the embodiment, and its absence does not significantly affect the effectiveness of the entire device according to the invention. In the absence of the anti-aliasing filter module 3, the output of the current transformer 1 is connected via the amplifier module 5 and the analog-to-digital converter module 7 to the current i(n) sample selection module 9.

Filter module 3 is a low-pass filter module that acts as an anti-aliasing filter with a cutoff frequency Fic depending on the selected sampling frequency Fis used in the 7-channel current analog-to-digital converter module.

The operation of filter module 3 consists in converting the voltage signal U2(izas(t)) received from current transformer 1 into the voltage signal U2'(izas(t)) by removing from the voltage signal U2(izas(t)) the components with frequencies higher than the limiting frequency Fic.

The operation of the amplifier module 5 consists in further converting the voltage signal U2'(izas(t)) to the voltage signal u2"(izas(t)) in such a way that the signal U2"(izas(t)) meets the requirements of the analog-to-digital converter module 7 in the current path of the device according to the invention. These requirements relate to the selected range of input voltages for the A/D converter used in this signal processing path.

In the current path of the device according to the invention, the analog-to-digital converter module 7 receives the voltage signal U2"(izas(t)), which is processed in the analog-to-digital converter module 7 into a series of samples of this signal i(n). For example, a converter with a bit resolution of 12 bits and a sampling frequency of Fis = 10 kHz was used.

The current i(n) sample selection module 9 of the device in the example embodiment accepts at its inputs a series of i(n) samples from the output of the analog-to-digital converter module 7 of the current measurement module 32 and the values of the n0 indexes of reaching a specific phase of the voltage signal from the phase detection module 10, which were determined on the basis of data from the output of the voltage measurement module 22. The operation of the current i(n) sample selection module 9 consists in selecting samples from the i(n) series starting from the index no for each period of the supply network voltage Nn that has reached a specific phase - for example, the voltage signal has changed its sign (the so-called zero crossing).

The output of the sample selection module 9 of the device according to the invention is connected via the matrix forming module 11 and the filter module 12 to the potential event detection module 13.

The matrix forming module 11 receives as its input the selected samples i(n) from the output of the sample selection module (9). The matrix forming module 11, using for example a circular buffer, transforms the values of the signal samples i(n) into the matrix Ip.

The individual rows of the Ip matrix correspond to the instantaneous current values that occurred at the same time interval relative to instant t0 in periods k from 1 to K. The index k indicates the period number of the fundamental component of the voltage signal in the power network, and the index m indicates the number of the current sample in this period. A single period has M samples, where M is a natural number corresponding to the number of samples recorded during one period (20 ms) of the fundamental component of the power network voltage (50 Hz), for example M = 200 samples. The individual columns of the Ip matrix therefore contain samples that were recorded in the kth period (of the 50 Hz fundamental component). The Ip matrix has Fs/50 rows, where Fs denotes the sampling frequency used in the analog-to-digital converter module 7 in the current measurement module 32 of the device.

Data in the form of a matrix Ip is passed from the input of matrix forming module 11 to the input of filter module 12. Filter module 12 filters the rows of matrix Ip, resulting in matrix I. For example, median filtration of order n = 15 is used. In other aspects of the embodiment, filter module 12 performs averaging filtration or modal filtration. Median filtration of order n involves determining a moving median, for a window of width n, on the sample values along the rows of matrix Ip. Matrix I contains, in each subsequent column, filtered values of current signal samples of subsequent periods of the fundamental component of the supply network voltage Nn. In other aspects of the embodiment, the construction of matrices I and Ip may be different, e.g., rows may be swapped with columns. In still other aspects of the embodiment, matrix forming module 11 may transform the values of signal samples i(n) to the appropriate series of vectors of these sample values. A person skilled in the art may organize the ordering of the signal sample values i(n) and their filtering in a different way, which does not change the essence of the invention.

PL 248440 Β1

The matrix I is transferred from the output of the matrix forming module 11 to the input of the potential event detection module 13. The potential event detection module 13 first determines the vector of changes Alk,o that occurred in the current signal, or more precisely in the columns of matrix I, in the time corresponding to the assumed number of periods of the fundamental component of the network voltage Nn. The assumed number of periods o is predefined and depends on the type of detected OEE; for example, for inductive loads, the value of parameter o assumes the value o = 7. Possible values of the offset parameter can be determined empirically. For the order value of the used median filter n = 15 and for an inductive load, it is advantageous to assume an offset o = 7, which means that the vector of changes Alk,o is determined for current signal samples taken in period k-o and for samples in period k+o, in accordance with formula W.1.

δι _λ , _ο = (ik ₊₀ - lk- ₀ ) (Wi)

The change vectors Alk,o are determined for all periods k in the range from period o+1 to the last period for which samples of the current signal i(n) reduced by o+1 (available in matrix I) can be taken.

Further operation of the potential event detection module 13 consists in calculating the values of the average change in the period SZk for each designated change vector Alk,o, i.e. for each period k, in accordance with formula W.2.

5Z, _C = (meanfaiJjudge/fc ₀ ))) (W.2)

Then, the potential event detection module 13, for all periods k, from the previously defined range) determines the peak values of all mean values of changes in the period SZk.

As a result of its operation, the potential event detection module 13 determines and transmits at its output the numbers of periods k, where period k is the period in which the determined coefficient of average change values in period SZk locally reaches the maximum value for the established o. These periods are considered to be periods in which a potential change in the OEE mode or state for the selected o occurred.

Additionally, the potential event detection module 13 transmits the I matrix at its output in the same form in which this matrix was transmitted to its input from the output of the filter module 12.

The operating mode selection module 14 receives data from the potential event detection module 13 in the form of a matrix I and the numbers of periods k in which a state change event of one of the OEEs could have occurred. The operating mode selection module 14 determines the operation of the subsequent modules of the device according to the settings made by the user of the device – the device either registers a new receiver or recognizes defined OEE types. Because the user determines the operating mode, they can also set appropriate parameters, such as offset o.

In the learning mode, the device in the example embodiment includes modules that allow for determining OEE patterns and storing these patterns in the device's non-volatile memory. In the example embodiment, this memory may be constructed as a database.

In the second stage of the learning mode of the device according to the invention, the operating mode selection module 14 is connected to the pattern memory module 17 via the pattern vector determination module 16.

The pattern vector determination module 16 receives at its input from the output of the operation mode selection module 14 data in the form of matrix I and the numbers of periods k in which an event of a change in the state of one of the OEEs could have occurred.

The pattern vector determination module 16 collects sample values from matrix I from period k’ before the change in the OEE state and sample values of matrix I from period k’ recorded after the analyzed change in state, when OEE has changed state and will already be in a new steady state.

Notations • The value of the m-th sample in the k-th period is denoted as i _m , _k . Then the sample vector corresponding to the k-th period of the current signal can be defined by the formula (W.3).

ffc = [G.fc< - Gn.fc’ «'«.»] (W.3) • X denotes the number of OEE in the monitored circuit of the Nn power supply network, which can be detected using the device according to the invention.

• As x, where: χε N λχ e < 1,X > is the serial number of each OEE in the monitored circuit of the Nn power supply network.

• Sx is the number of OEE operating states with the sequence number x.

PL 248440 Β1 • As ae <1 ,s _x > the number of the previous operating state of the OEE with the sequence number x is denoted.

• As b e <1 ,Sx> the number of the new OEE operating state with the sequence number x is designated.

• As o'(x, _a ,b) is denoted the appropriate number of periods of the signal with frequency 50 Hz before the change of the OEE state with the sequence number x from the state with the number a to another state with the number b, i.e. then the period k'= k-o'(x, _a ,b).

• As o”(x,a,b) we mean the appropriate number of periods of the 50 Hz signal after the change of the state of device x from state number a to another state number b, i.e. then the period k =k+O (x,a,b).

The values o'(x, _a ,b) and o"( _X , _a ,b) are characteristic for OEE with sequence number x. The values o'(x, _a ,b) and o"(x,a,b) should be selected so that if k' is the period occurring before the change of state a to b of OEE with sequence number x, then the period k" is the period after the transition of OEE with sequence number x to the steady state b. The values that o'( _X , _a ,b) and o"( _X , _a ,b) can assume are determined experimentally and stored in the memory of the device according to the invention. Unlike in the potential event detection module 13, in which one value o is used, i.e. o' and o are equal, and these values are selected for specific (user-defined) types of devices, not specific devices. In the change vector determination module 15 and the pattern vector determination module 16, the values o' and o may be different from those used in the potential event detection module (13). The purpose of detection in the potential event detection module 13 is only to detect changes in the signal, the purpose of calculations in modules 15 and 16 is to determine the change vector assuming that in a given period k the OEE changed from state a to state b.

• O is denoted as the set of all o'( _X , _a , b) and o"( _X , _a , b) - for all xe <1 , X> and for all pairs (a, b) e<1 , s _x > iab - in practice, the number of possible pairs (a, b), i.e. the number of different changes of states of an OEE with the order number x, is limited by the technical realization of this OEE. In the example of the implementation of the device according to the invention, a finite set O is assumed.

The operation of the pattern vector determination module 16 consists in determining the OEE feature vector with the order number x when the OEE state changes from state a to state b, marked as:

_Λ .(χ,α,6) ^.ίχ,α,ΰ) “'pattern . The vector ^a 'pattern is determined based on the changes that occurred in the current signal as a result of the change of state za to state b by the OEE with the sequence number x. This vector is determined after receiving the signal about turning on this OEE. This signal comes from the user of the device, who put the device in the state of adding new patterns.

During its operation, the pattern vector determination module 16 determines the feature vector _Λ ιίϊ.“.ί>) ^a 'pattern for OEE with the order number x by subtracting the values of the samples of matrix I from the period k' before the change of the state of OEE no. x forced by the user of the device from the values of the corresponding samples of matrix I from the period k" recorded when OEE no. x is already in the new steady state λ (χ.B,ίϊ)

b. The formula for determining the standard feature vector "'the OEE formula with the ordinal number x takes the form (W.4.).

. ,ίχα,Β) _ , _ r , (W.4) “'formula — ' k+0“ę _x , _a j,·) 'k~o _cxul ,j for ke N Λ ke (o + 1Λ - o^)) Λι ε <1,^0

The pattern memory (database) module 17 receives at its input the output of the pattern vector determination module 16 and stores in its non-volatile memory the determined pattern feature vector . . (χ,Λ,Β) "'OFE pattern with the ordinal number x together with the parameters dT( _X , _a .b), o'( _X , _a ,b) and o"(x,a,b).

The device according to the invention, after determining the standard feature vector ^a ', sends a signal to the user of this device about the need to turn off or change the state of the OEE with the sequence number x.

Multiple use of the device according to the invention in its learning mode enables subsequent determination of Λ I (Λ of the standard feature vectors ^and 'the formula for the number X of all OEE in the electrical circuit covered by the monitoring of the device according to the invention.

.,μ,α,Β)

Determining the standard feature vectors ^a, the pattern for all X OEE in a given electrical circuit allows the creation of a database of patterns of all OEE in the form of a quad:

PL 248440 Β1 f rIT λ' ” 1 Λΐί^' ⁰ ^) t“ _WMT , ai(xab), o ₍ x. _Ł b), ° The determination of each standard feature vector must be performed with OEEs other than OEEs with the order number x switched off.

The operation of the pattern database module 17 consists in permanently recording, storing, and making available to other modules of the device according to the invention patterns for each OEE that the device is able to identify. The pattern of each OEE with the sequence number x is the feature vector _A .(w) ^Q/ pattern consisting of a series of numbers with the cardinality resulting from the ratio of the sampling frequency in the A/D converter of the converter module 7 and the frequency of the network voltage Nn, i.e. for the European network: Fs/50.

The parameters of each pattern feature vector ^a 'pattern are information about which OEE and which of its states a and b this vector refers to. Additionally, in the pattern memory module 17, for each feature vector ^a, pattern, information is stored on how many periods o'( _X , _a , b) and o"( _X , _a , b) of the fundamental component of the supply network voltage Nn elapsed between the subtracted periods when determining this vector. In the pattern data module 17, a parameter specifying the threshold value dT(x. _a . b) is also stored, on the basis of which the decision module 19 decides (in the manner described below) whether the potential event concerns a change in the state of a to b OEE with the sequence number x, or whether it is an event not related to any change in the state of OEE in a given supply circuit of the network Nn. Therefore, each pattern of each change in the state of each OEE, which is stored in the non-volatile memory of the pattern memory module 17 of the device according to the invention, is a quadruple: · dTuaw, o"^)} assigned to a change in a specific operating state a to another specific operating state b OEE with the sequence number x.

In the identification mode path of the device according to the invention, the operating mode selection module 14 is connected via the change vector determination module 15 and via the similarity determination module 181 to the decision module 19.

Additionally, the pattern memory module 17 is connected directly to the decision module 19.

The change vector determination module 15 receives at its input from the output of the operating mode selection module 14 data in the form of matrix I and the numbers of periods k in which an event of a change in the state of one of the OEEs could have occurred.

The change vector determination module 15 takes the sample values of matrix I from the period k’ before the change in the OEE state and the sample values of matrix I from the period k’ recorded after the analyzed change in state, when the OEE has changed state and will be in a new steady state.

The following designations are adopted: X, x, and _m ,k, o'(x, a, b), o"(x, _a , b) identical as in the case of the pattern vector determination module 16 described above. The change vector determination module 15 calculates k', k", zi/mk, lk also in the same way as in the case of the pattern vector determination module 16, therefore the description of the method of these calculations will not be repeated here.

The operation of the change vector determination module 15 consists in determining the change vector κ, which concerns the changes that occurred in the signal as a result of a change in state from a to b by the OEE with the number x. This vector is determined at the moment of receiving a signal about a potential event (i.e., in period k of the supply network voltage Nn) of a change in the state of the OEE. This signal comes from the potential event detection module 13, propagated by the operating mode selection module 14 and is delivered to the input of the change vector determination module 15 in the form of the value of the period number k, in which an event of a change in the state of one of the OEEs could have occurred.

The method of determining the change vector « in the change vector determination module 15 consists in subtracting the values of the I matrix samples from the period before the change in the OEE state from the values of the corresponding I matrix samples from the period recorded after the analyzed change in state, when the OEE that has changed state will already be in a new steady state.

The value of the difference between the values of m-th samples from the period k”=k+o”(x, _a ,b) and the period k'=k-o'(x, _a ,b) is denoted as Ai _m ,ki, the module for determining the change vectors 15 is calculated according to the formula (W.5).

~ (W.5) dlame N Λ me (1,M) Ak ε N Λ ke (1,/C) ^£ N Ao^ _atl) ΕΝΛχξ (1,X)

PL 248440 Β1

Based on the definitions of Ai _m ,k and l _k the change vector determination module 15 calculates the change vector ^nft) according to the formula (W.6).

. __ j __ t ^{J fc+o} (xaJ>) 'kC \x, _a ,by (W.6) for k £ NA k ε + 1.Χ ^_ Λ x £ (Li) (χ ¹ o.

The change vector is also a feature vector, based on which the change in the OEE state with the ordinal number x is identified by comparison with the reference feature vectors

Λ, ^a, the pattern of all OEEs in a given electrical circuit. The device according to the invention performs this comparison in the similarity determination module 181.

The similarity determination module 181 receives at its inputs from the output of the change vector determination module 15 the calculated values of the change vector * and from the output of the pattern memory module 17 the matrix M*zor containing the pattern vectors of all X OEEs, as well as the matrix O containing the parameters o'( _X , _a , b) and o"(x, a, b) appropriate for these pattern vectors. The operation of the similarity determination module 18 consists in comparing the change vector * obtained at its input _with each pattern ^A/ pattern for the appropriate values of the parameter o'( _X , _a , b) and o"(x, a, b). The similarity determination module 18 can use various known similarity measures, for example the mean square measure. The result of the operation of the similarity determination module 18 is the creation of a vector of similarity coefficients d, which contains in its elements the calculated similarity coefficients d< ^{xa b} > for each OEE pattern.

The decision module 19 receives at its inputs a vector of similarity coefficients d from the input of the similarity determination module 18 and a threshold value dT(x, _a , b) from the output of the pattern database module 17, characteristic for each OEE with the number xi for each pair of state changes from a to b of this OEE.

The operation of decision module 19 consists in determining the maximum value of the coefficient d ^(xab) for the entire vector d and comparing this maximum value with the threshold value dT(X, a, b). Decision module 19 then makes a decision based on this comparison. If the maximum value of the coefficient d ^{(xa b)} for vector d exceeds the threshold value dT(X, _a , b), then the system decides that the examined event concerns a change in the OEE state with the sequence number x and a change in its state za to another state b.

If the maximum value of the coefficient d ^(xab) for vector d does not exceed the threshold value dT(x, _a ,b), then the decision module 19 decides that the previously detected potential event is not an event indicating a change in the state of any OEE in the tested LV network supply circuit.

Claims (24)

1. A device for detecting changes in the operating mode or identifying OEE electrical energy receivers in the power grid, which includes: a voltage path which includes a voltage measurement module (22), the input of which is connected in parallel to the power network and the output through a phase detection module (10) to the sample selection module (9), and a current path which includes a current measurement module (32), the input of which is connected in series to the power network and the output to the sample selection module (9), the output of the sample selection module (9) is connected to the matrix forming module (11), wherein the sample matrix from the matrix forming module (11) contains a predetermined number of the last K periods of the voltage signal and m samples of the instantaneous current value for each of the last K periods, at the same time the output of the matrix forming module (11) is connected via a filtering module (12) to the input of the change vector determination module (15), wherein the change vector determination module (15) is adapted to determine change vectors, which change vectors are transferred via the output of the change vector determination module (15) to the input of the decision module (18), furthermore the output of the pattern memory module is connected to the input of the decision module (18), (17) OEE, wherein the output of the decision-making module (18) provides information on detecting a change in the operating mode or identifying OEE electrical energy receivers, characterized in that the filtration module (12) is connected to the change vector determination module (15) via a potential event detection module (13), wherein the potential event detection module (13) is adapted to transmit to the input of the change vector determination module (15) a sample matrix from the matrix forming module (11) and additional information on the number of the period in which a potential change in the OEE operating mode occurred. 2. A device according to claim 1, characterized in that the decision module (18) comprises a similarity determination module (181) and a decision module (19), wherein the input of the similarity determination module (181) is fed with data from the change vector determination module (15) and data from the pattern memory module (17), the output of the similarity determination module (181) is connected to the first input of the decision module (19), and the second input of the decision module (19) is connected to the output of the pattern memory module (17). 3. A device according to claim 2, characterized in that the similarity determination module (181) in the decision module (18) is adapted to compare the values of the change vectors with the pattern definition vectors that are stored in the pattern memory module (17) of the OEE. 4. A device according to claim 2 or 3, characterized in that the decision module (19) in the decision module (18) is adapted to identify the OEE receiver and/or identify a change in the state of the OEE receiver based on data from the similarity determination module (181) and from the pattern memory module (17). 5. A device according to claim 1-4, characterized in that an operation mode selection module (14) is placed between the potential event detection module (13) and the change vector determination module (15), wherein one of the outputs of the operation mode selection module (14) is connected to the decision module (19) via the pattern vector determination module (16) and the pattern memory module (17), and the other output of the operation mode selection module (14) is connected to the decision module (19) via the change vector determination module (15) and via the similarity determination module (181), wherein when the new OEE pattern registration mode is active, the operation mode selection module (14) is adapted to transmit information from the input to the pattern determination module (16), when the new OEE pattern registration mode is not active, the operation mode selection module (14) is adapted to transmit data only to the change vector determination module (15). 6. A device according to claim 1-5, characterized in that the phase detection module (10) is adapted to determine the moment of time at which the value of the voltage signal of the electrical network reaches a predefined phase value, preferably when the value of the voltage signal of the electrical network changes from a negative to a positive value. 7. A device according to claim 1-6, characterized in that the sample selection module (9) is adapted to transfer to the matrix forming module (11) the values of current samples, starting from the sample number determined by the phase detection module (10) for each subsequent period of the voltage signal, wherein the number of transferred samples results from the duration of one period of the voltage signal and the sampling frequency in the current measurement module (32). 8. A device according to claim 1-7, characterized in that the matrix forming module (11) is adapted to store the values of current samples for a determined number K of the last periods of the voltage signal, which form a sample matrix that is fed to the input of the filtering module (12). 9. A device according to claim 8, characterized in that the filtration module (12) is adapted to perform median filtration or average filtration or mode filtration. 10. Devices according to claims 1-9, characterized in that the filtration module (12) is adapted to perform filtration for a series of samples taken at moments equidistant in time from the phase determined by the phase detection module (4). 11. A device according to claim 1-10, characterized in that the voltage measurement module (22) comprises a voltage transformer (2), to the secondary winding of which a low-pass filter (4) is connected, and the output of the filter (4) is connected via an amplifier module (6) to an analog-to-digital converter module (8), the output of which is the output of the voltage measurement module (32), or the voltage measurement module (22) comprises a voltage transformer (2), to the secondary winding of which an amplifier module (6) is connected, which is connected to an analog-to-digital converter module (8) and further to a low-pass digital filter module (42), and the output of the low-pass digital filter (42) is the output of the voltage measurement module (22). 12. Device according to claim A current measuring module (32) comprises a current transformer (1), to the secondary winding of which an amplifier module (5) is connected, which is connected to an analog-to-digital converter module (7), preferably an anti-aliasing filter (3) is placed between the secondary winding of the current transformer (1) and the amplifier module (5), wherein the passband of the anti-aliasing filter (3) is related to the sampling frequency of the transducer in the analog-to-digital converter module (7), or the current measuring module (32) comprises a current transformer (1), to the secondary winding of which an amplifier module (5) is connected, which is connected to an analog-to-digital converter module (7), preferably the analog-to-digital converter module (7) is connected to a digital anti-aliasing filter module (52), wherein the output of the digital anti-aliasing filter (52) or the analog-to-digital converter (7) is the output of the module current measurement (32). 13. A method for detecting a change in the operating mode and identifying OEE electrical receivers using measurements of instantaneous values of current signals and voltage phase detection, comprising: a stage of measuring the current in the supply network and a stage of measuring the voltage, wherein after the current measurement stage the current signal is converted into its digital form, which is transmitted to the sample selection module (9), and after the voltage measurement stage the voltage signal is converted into its digital form, on the basis of which a specific phase value of the voltage signal is detected by means of the phase detection module (10), and then in the matrix formation stage using the matrix formation module (11) the appropriate sample values are selected from the sample selection module (9), which are further formed into a matrix in which the next m rows contain samples of instantaneous current values in k columns of subsequent voltage periods starting from the moment the voltage signal reaches a specific phase value, and then in the filtration stage using the filtration module (12) the matrix is filtered by its rows, obtaining a matrix of samples after filtration, from which the change vector determination module (15) determines the change vectors containing the differences in the sample values of the sample matrix after filtration from the period before a change in the state of the OEE receiver from the values of the samples of the sample matrix after filtration from the period recorded after the analyzed change of state, when the OEE receiver which has changed the state and will be in a new steady state, and an identification stage takes place, in which the change vectors determined by means of the decision-making module (18) are compared with the standard feature vectors of OEE receivers in a given electric circuit stored in the pattern memory module (17), identifying the OEE receiver on the basis of the most matched pattern, characterized in that after filtering the sample matrix in the filtration module (12), a stage of potential event detection takes place, wherein the detection consists in determining in the potential event detection module (13) the vectors of changes for at least some of the K voltage periods, which vectors allow for determining a proportionality coefficient proportional to the average of the absolute values of the elements of these vectors (W.2), wherein the period k in which the determined coefficient locally reaches the maximum value is considered to be the period in which a potential change of the OEE mode or state occurred, then information on the number of the period in which a potential change occurred a change in the OEE mode or state is transferred to the change vector determination module (15), wherein the change vectors determined in the change vector determination module (15) include differences in the sample values of the sample matrix after filtering for the period before the period in which a potential change in the OEE mode or state occurred, from the sample values of the sample matrix after filtering for the period after the period in which a potential change in the OEE mode or state occurred, wherein only those periods for which the proportionality coefficient reached the maximum value are taken into account, only selected change vectors determined in the change vector determination module (15) are further compared in the decision-making module (18) and are used to identify OEE based on the best-matching pattern. 14. A method according to claim 13, characterized in that determining the proportionality coefficient in the potential event detection module (13) comprises determining change vectors comprising differences in sample values of the sample matrix after filtering from the period k-o' before the change in the OEE receiver state from the sample values of the sample matrix after filtering from the period k+o" recorded after the analyzed period k, wherein only selected periods before and after are analyzed for the predefined at least one pair of shift values o' and o", preferably shift o' and o" corresponds to the OEE category and shift values o' and o" may not be equal, the proportionality coefficient is determined on the basis of the value of the determined difference in the sample values of the matrix for each predefined at least one pair of shifts o' and o", preferably the proportionality coefficient is the average of the absolute value of the determined difference in the sample values. 15. A method according to claim 13 or 14, characterized in that the phase detection module (10) is used to determine the time instant in which the voltage value of the electrical network changes from negative to positive. 16. A method according to claim 13-15, characterized in that by means of the matrix forming module (11) the current sample values are transferred from the sample selection module (9) in the form of a sample matrix, wherein for each period k starting from the sample determined by the voltage phase detection module (10), a number m of samples corresponding to each of the voltage periods is transferred, preferably the matrix forming module (11) contains a matrix buffer for the last K voltage periods and outputs a sample matrix for each voltage period, which sample matrix is fed to the input of the filtering module (12). 17. A method according to claim 13-16, characterized in that the filtration is performed for a series of samples taken at moments equidistant in time from the phase determined by the phase detection module (10). 18. A method according to claim 13-17, characterized in that all steps are performed on time-domain signal samples. 19. A method according to claim 13-18, characterized in that the current signal sampling moments are correlated in time with the moment when the voltage signal reaches a specific phase. 20. A method according to claim 13-19, characterized in that the filtration is a median filtration or an average filtration or a mode filtration. 21. A method according to any one of claims 13-20, comprising: determining in the similarity determination module (181) similarities between all the pattern change vectors taken from the pattern memory module (17), wherein the pattern memory module (17) contains pattern change vectors characteristic for OEE, and the set of change vectors determined in the change vector determination module (15), wherein the determined similarities are transferred to the decision module (19); making decisions by the decision module (19) based on the similarities determined in the similarity determination module (181) and on the basis of data from the pattern memory module (17) with a threshold value characteristic for each type of OEE and a possible change in the state of this OEE, the pattern of which is stored in the pattern memory module (17). 22. A method according to claim 13-21, characterized in that the decision-making by the decision module (19) comprises: determining the maximum value of the similarity coefficient for the entire similarity vector, and comparing this maximum value with the threshold value taken from the pattern memory module (17), then making a decision by the decision module (19) based on this comparison, if the maximum value of the similarity coefficient for the similarity vector exceeds the threshold value, then the decision module (19) decides that the potential event in period k concerns a change in the state of a specific OEE, whereas if the maximum value of the similarity coefficient for the similarity vector does not exceed the threshold value, then the decision module (19) decides that the previously detected potential event is not an event indicating a change in the state of any OEE in the tested power supply circuit of the power network. 23. A method according to claim 13-22, characterized in that in the decision-making stage in the decision module (19), in a situation where the given decision selection criteria are insufficient, statistical algorithms are used to select the most probable OEE from among the registered patterns, preferably it is a decision tree algorithm or a random forest, or K-nearest neighbors (KNN), or convolutional neural networks (CNN), or support vector machines (SVM), or deep neural networks (DNN), or Naive Bayes. 24. A method according to claim 13-23, characterized in that before the decision-making step in the decision module (19), a step of adding change patterns to the pattern memory module (17) is performed, wherein the step of adding patterns depends on predefined data provided by the user about the category of the device being added and the moment of its connection to the network, the steps of adding patterns include: setting the operating mode selection module (14) to the pattern addition mode (output duplication mode), which results in supplying a sample matrix to the pattern determination module (16), determining the pattern change vectors, wherein the determination of the pattern change vectors is carried out analogously to the determination of the change vectors, wherein when determining the pattern change vectors, predefined shift values o’ and o”, and a predefined threshold value characteristic for the given OEE category are assumed, then storing the pattern change vector, together with the corresponding shift values o’ and o”, and a predefined threshold value characteristic for the given OEE category in the pattern memory module (17), wherein after adding the pattern change vector to the pattern memory module (16), OEE recognition is started as defined in claim 13 or another device is added.