Enel S.p.A is testing and implementing a number of Internet of Things (IoT) based solutions for the online monitoring of a wide set of variables, which enables for an instant and comprehensive view of the condition of each asset installed in MV/LV Distribution Sub-stations (DSs). MV/LV-DSs could be affected by different faults, depending both on environmental parameters such as humidity and temperature and usage level. Different use cases in the Proof of Concept initially implemented on 100 distribution sub-stations (POC100) are proposed by Enel S.p.A. Two use cases are reviewed and investigated in this thesis, which are Partial Discharge (PD) detection in air insulated switchgears by air decomposition using ozone sensors and jumper interruption detection for overhead transmission line using electrical parameters of the transformer. Five secondary sub-stations across Italy are selected for the analysis in this thesis which are Archimedes, Colletti, Cofano, Ingardia and Torre Razze. It is well known that electrical discharges in air produce ozone. High ozone concentration in air-insulated switchgears has been reported and is always associated with the presence of PDs. Since conventional methods are more time consuming, required downtime, expertise, and usually required off-line monitoring with expensive detection apparatus. Therefore, new approaches are investigated in this thesis. The proposed method is based on chemical detection method, which is cost-effective and compatible with the upcoming patterns especially the Artificial Intelligence (AI), Machine Learning (MI), Deep Learning (DL), Data Analytics and Internet of Things (IoT). Since these technologies are the future trends towards predictive maintenance of the smart grids. The environmental parameter such as ozone, humidity and temperature are continuously monitored by IoT sensors installed inside the MV air insulated switchgear containers. The data are sent to the cloud-based platform where the proposed algorithm analysis the data and extracts the useful information such as rate of change of the ozone concentration in the air inside the MV air insulated switchgear containers to detect the PD activity and also its correlation with other environmental parameters. i.e. humidity and temperature. Based on the analysis and results found via proposed method, threshold levels of the ozone concentration in the air are proposed to alert Distribution System Operators (DSOs) about the PD activity. The continuous monitoring of the environmental parameters based on the proposed algorithm will be visualized via Grafana dashboard. Python programming language is used to implement the proposed approach. This all leads to the ability of Predictive Maintenance, thus reducing the number of outages, improving quality of service and reducing costs. The power distribution overhead line network supplies power to its customers, and the lines exposed outdoors are susceptible to external and environmental interference. When an open-circuit fault occurs on the jumper of the overhead line, the jumper will be interrupted, which will naturally affect the power supply of the customer. Such failures are usually only detected after receiving a customer’s supply interruption report. A fault detection method based on timely analysis of symmetrical voltage components on power lines and other measures has been proposed in this thesis. The proposed method for the jumper interruption detection has been tested via MATLAB software. The approach will help DSOs to easily detect the jumper interruption via visualizing the proposed parameters on dashboard.
Enel SpA sta testando e implementando una serie di soluzioni basate su “Internet of Things” (IoT) per il monitoraggio online di un ampio set di variabili che forniscono una visione “real-time” complessiva delle condizioni di ogni “asset” installato in cabina secondaria (MT/BT). Le cabine secondarie MT/BT potrebbero essere soggette a diversi guasti a seconda sia delle condizioni ambientali, come umidità e temperatura, sia delle condizioni di utilizzo. Diversi “Casi d’Uso” sono stati proposti da Enel SpA in ambito di un PoC (Proof of Concept) effettuato su 100 cabine secondarie (PoC100), due dei quali sono stati analizzati in questa tesi. Il primo riguarda la rilevazione di scariche parziali (PD) in quadri isolati in aria (GIS) tramite sensori di ozono, il secondo analizza l'interruzione del collegamento (Jumper) tra due linee aeree di distribuzione utilizzando i parametri elettrici del trasformatore. Sono state selezionate cinque cabine secondarie per l'analisi del primo caso d’uso: Archimede, Colletti, Cofano, Ingardia e Torre Razze. È noto che le scariche parziali rilascino ozono in ambiente. In presenza di scariche parziali è stata sempre riscontrata un'elevata concentrazione di ozono nei quadri isolati in aria (GIS). Poiché i metodi convenzionali di rilevazione delle scariche parziali richiedono tempi lunghi, periodi di disservizio, esperienza e costosi apparati di rilevamento, in questa tesi vengono analizzati dei nuovi approcci. Il metodo proposto si basa sul sull’analisi della variazione della quantità di zono tramite sensore ad hoc, meno costoso e più compatibile con i modelli di Intelligenza Artificiale (AI), Machine Learning (MI), Deep Learning (DL), Data Analytics e Internet of Things (IoT), che rappresentano il paradigma futuro della manutenzione predittiva delle reti elettriche interconnesse (Smart Grid). I parametri ambientali come ozono, umidità e temperatura sono costantemente monitorati dai sensori IoT installati all'interno degli scomparti dei quadri MT isolati in aria. I dati provenienti dai sensori vengono inviati ad una piattaforma in “cloud” dalla quale l’algoritmo sviluppato estrae i dati, come la velocità di variazione della concentrazione di ozono nell'aria, al fine di rilevare scariche parziali e valutare la correlazione con i parametri ambientali (temperatura e umidità). Sulla base dell'analisi e dei risultati ottenuti tramite lo studio svolto in questa tesi, vengono proposti alcuni valori soglia della concentrazione di ozono nell'aria per allertare gli operatori del distributore (DSO) sulla presenza di scariche parziali. Il monitoraggio “real-time” dei parametri dedotti dall’algoritmo proposto potrà essere visualizzato tramite dashboard Grafana. È stato utilizzato il linguaggio di programmazione Python per sviluppare il modello al fine di effettuare manutenzione predittiva, riducendo così il numero di interruzioni, migliorando la qualità del servizio e diminuendo i costi. Le linee aeree di distribuzione elettrica forniscono energia ai clienti e sono suscettibili alle interferenze esterne e ambientali. Quando si verifica un guasto di “circuito aperto” sul collegamento tra linee aeree (Jumper), l’interruzione influenzerà l'alimentazione dell’utente. Tali guasti vengono generalmente segnalati dal cliente ed estinti solo dopo la ricezione di un reclamo. In questa tesi è stato proposto un metodo di rilevamento dei guasti basato sull'analisi delle componenti simmetriche di tensione e altre misure. L’algoritmo proposto per rilevare l'interruzione del collegamento (Jumper) è stato simulato tramite il software MATLAB. L'approccio potrà aiutare i DSO a rilevare facilmente l'interruzione visualizzando specifici parametri su un’apposita dashboard.
IoT applications in distributed sub-station review and new approaches
Kakar, Malak Gulbadin Khan;THOTA, NAGA DEEP
2019/2020
Abstract
Enel S.p.A is testing and implementing a number of Internet of Things (IoT) based solutions for the online monitoring of a wide set of variables, which enables for an instant and comprehensive view of the condition of each asset installed in MV/LV Distribution Sub-stations (DSs). MV/LV-DSs could be affected by different faults, depending both on environmental parameters such as humidity and temperature and usage level. Different use cases in the Proof of Concept initially implemented on 100 distribution sub-stations (POC100) are proposed by Enel S.p.A. Two use cases are reviewed and investigated in this thesis, which are Partial Discharge (PD) detection in air insulated switchgears by air decomposition using ozone sensors and jumper interruption detection for overhead transmission line using electrical parameters of the transformer. Five secondary sub-stations across Italy are selected for the analysis in this thesis which are Archimedes, Colletti, Cofano, Ingardia and Torre Razze. It is well known that electrical discharges in air produce ozone. High ozone concentration in air-insulated switchgears has been reported and is always associated with the presence of PDs. Since conventional methods are more time consuming, required downtime, expertise, and usually required off-line monitoring with expensive detection apparatus. Therefore, new approaches are investigated in this thesis. The proposed method is based on chemical detection method, which is cost-effective and compatible with the upcoming patterns especially the Artificial Intelligence (AI), Machine Learning (MI), Deep Learning (DL), Data Analytics and Internet of Things (IoT). Since these technologies are the future trends towards predictive maintenance of the smart grids. The environmental parameter such as ozone, humidity and temperature are continuously monitored by IoT sensors installed inside the MV air insulated switchgear containers. The data are sent to the cloud-based platform where the proposed algorithm analysis the data and extracts the useful information such as rate of change of the ozone concentration in the air inside the MV air insulated switchgear containers to detect the PD activity and also its correlation with other environmental parameters. i.e. humidity and temperature. Based on the analysis and results found via proposed method, threshold levels of the ozone concentration in the air are proposed to alert Distribution System Operators (DSOs) about the PD activity. The continuous monitoring of the environmental parameters based on the proposed algorithm will be visualized via Grafana dashboard. Python programming language is used to implement the proposed approach. This all leads to the ability of Predictive Maintenance, thus reducing the number of outages, improving quality of service and reducing costs. The power distribution overhead line network supplies power to its customers, and the lines exposed outdoors are susceptible to external and environmental interference. When an open-circuit fault occurs on the jumper of the overhead line, the jumper will be interrupted, which will naturally affect the power supply of the customer. Such failures are usually only detected after receiving a customer’s supply interruption report. A fault detection method based on timely analysis of symmetrical voltage components on power lines and other measures has been proposed in this thesis. The proposed method for the jumper interruption detection has been tested via MATLAB software. The approach will help DSOs to easily detect the jumper interruption via visualizing the proposed parameters on dashboard.| File | Dimensione | Formato | |
|---|---|---|---|
|
IOT Applications in Distribution Substations Review and new Approaches.pdf
non accessibile
Dimensione
8.68 MB
Formato
Adobe PDF
|
8.68 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/167163