Denominazione: Battery EneRgy managemEnt systems for reNewable and cItizen energy CommunitiEs
Acronimo: BERENICE
Tipologia: progetto PRIN 2022 finanziato da MUR- Next Generation EU; Anno inizio 2023
Responsabile scientifico per Sapienza Università di Roma: Rodolfo Araneo
Ente finanziatore: Missione 4 “Istruzione e Ricerca” - Componente C2 Investimento 1.1
“Fondo per il Programma Nazionale di Ricerca e Progetti di Rilevante Interesse Nazionale (PRIN)” Decreto Direttoriale n. 104 del 02 febbraio 2022 - Avviso pubblico per la presentazione di Progetti di ricerca di Rilevante Interesse Nazionale (PRIN) da finanziare nell’ambito del PNRR
PRIN 2022 P2022R3L83 - CUP MASTER xxxxxxxxxxx - CUP B53D23024050001
Importo finanziato: 299884,27 €
Partners
- Dipartimento di Ingegneria Astronautica, Elettrica, Energetica, SAPIENZA Università di Roma
- Politecnico di Milano
- Università degli Studi di Genova
Finalità
BERENICE (Battery EneRgy managemEnt systems for reNewable and cItizen energy CommunitiE) aims at developing innovative battery and energy management methodologies for stationary battery energy systems to be used in next-generation Energy Communities (ECs) and demonstrating the effectiveness of such methodologies through a full-scale demonstrator to be installed in the Savona Campus grid. While Battery Energy Storage Systems (BESSs) are socially and technically appealing for their propelling role in the smartization and widespread of ECs, the industrial exploitation of their effective impact, primarily when distributed throughout the single grid or throughout connected grids, has been significantly slowed down by the pivoting role exercised by the research in Battery Management System (BMS) for batteries in Electric Vehicles (EVs) and hybrid EVs. BERENICE responds by taking a comprehensive conceptual and manufacturing approach, starting from ab-initio multi-physics modeling, passing through advanced artificial intelligence approaches, and ending in designing and constructing an integrated Battery Energy Management System (BEMS) unit to be tested in a real demonstrator named BASTET (Battery Advanced System - TEsT). This research is expected to develop model-based, flexible, modular BEMS technologies, which allow increasing in BESS operating functionalities at the battery and system levels while paving the way to unprecedented applications downstream, such as harmonic co-operating functionalities in distributed configurations and direct interfacing with breakthrough technologies (i.e., internet of energy, internet of things, blockchains), and, finally, making them commercially competitive in the energy market beyond 2023.
BERENCICE employed a community-based approach, that is, a way of working based on an inclusive partnership with communities of persons of concern that recognizes their resilience, capacities, and resources. BERENICE mobilizes and builds on these to deliver protection, assistance, and solutions while supporting community processes and goals. BERENICE (a woman's name) built an equal and active partnership with women, men, girls, and boys of diverse ages and backgrounds in all areas of our work. BERENICE demands understanding and considering the prevailing context, the receiving population, gender roles, community dynamics, protection risks, community concerns, and priorities, and that we work with people of concern during the various stages of UNHCR’s program cycle. BERENICE recognizes our facilitation role as external actors, our limitations in terms of capacities, resources, and the temporary nature of our presence, as well as the longer-term impact of our interventions.
Up to the present date:
The units continued classifying several batteries. The cycling tests were carried out in particular on lithium batteries at a current regime equal to C/4, and at the same time, charge and discharge tests were performed at different C-rates, in particular C/4, C/2, C, 2C, and 3C. Through the analysis of these tests, an electrochemical characterization was carried out that describes the behavior of the battery by identifying the values of different electrochemical parameters such as internal resistance, exchange current, activation and diffusion overvoltages, and the relationship between SOC and OCV. Furthermore, a data fitting procedure was developed to arrive at a model of the electrochemical behavior of the batteries over time that has proven to be quite accurate so far. The electrolytic resistance seems to remain practically constant during the 166 cycles, while the internal resistance changes over time through the diffusive component. Furthermore, it has been verified that the cathode, over time, changes its electrochemical behavior for a used battery. The equilibrium between the two solid phases inside the material seems interrupted, and the lithium gradually fills or empties the crystal lattice of the cathode material without forming two distinct phases in equilibrium between them. The analysis highlights a difference between the OCV values between charge and discharge equal to about 40 mV. Hence, a hysteresis is present, as reported in the literature; instead, almost all the other parameters are the same for both charge and discharge. They are the same as those obtained for charges at different C-rates because the electrochemical behavior, with some differences, is similar anyway.
The units has focused on the theoretical framework for EMS (WP2) investigating the effectiveness of data-driven forecasting techniques to estimate the proper contribution of BESS, based on their healthy status, to provide optimal dispatching of active power among the different decentralized electrical generation units in a microgrid over a typical daily based horizon. The overall system will be able to combine RES with storage using forecasting data and methods taking into account both RES sources and connected loads, exploiting the WP1 Berenice outputs related to battery status (e.g., SoC, SoH, and RUL estimated parameters). The project's primary computational intelligence technique is based on neural network learning algorithms applied to an EMS.
The units focused moreover:
1) from the theoretical standpoint, the Machine Learning (ML) and Deep Learning (DL) techniques investigated to forecast the Renewable Energy Sources (RES) production have been implemented in the EMS . Moreover, a receding horizon strategy has been set up to check how the effectiveness of the forecasting improves when it is updated every day or every hour and how this reflects on the performances of the EMS in terms of optimizing the overall energy cost of the microgrid/energy community.
2) from the experimental viewpoint, the units have continued the measurements conducted on the University of Genova Savona campus venue Smart Polygeneration Microgrid (SPM). We recall that, in the SPM there is a storage system made of 6 Fiamm Sonick model modules ST523, each consisting of 240 cells. Each module is characterized by a nominal capacity of 23.5 kWh and an output voltage of 620 V. The six modules are electrically connected in parallel. The storage has been connected to the SPM operation since 2014. Different types of charge/discharge tests have been performed. In this third four month’ time slot, focus has been posed on the characterization tests, conducted as follows:
Starting from 100% SOC, the batteries are discharged with a power P (7, 11, 17, 22, 27, 38, 45, 57 kW) down to a SOC equal to 10-15%. The available measurements were downloaded from both the battery management system (NIDEC) and the microgrid's SCADA.
In particular:
• From NIDEC: energy content of each module (measured in Ah) - 1 value every minute; DC voltage of each module - 1 value every minute.
• From SCADA: AC side measurements of the entire storage system (current, phase voltage, frequency, power factor, active power, reactive power, apparent power), DC side measurements of the whole storage system (current, voltage, SOC).
During the 4th four-month period, the BioLogic model BCS-915 Module of 8 channels 15 A with 0 V complete with Electrochemical Impedance Spectroscopy, moted on a BioLogic model 24U cabinet-w/Core, Stop, 20out Powerstrip was delivered to the Rome Unit. The instrumentation was networked using a special computer that allows remote control and monitoring. This made it possible to set up a laboratory that could also be remotely controlled and monitored by the other units in Genoa and Milan. At the same time, an Electrochemical Workstation Squidstat Penta was ordered from Quantum Design doing functions of Single-channel potentiostat/galvanostat complete with EIS frequency range: 10 µHz to 2 MHz, Maximum current: ±5 A, Voltage Scan Range: ±10 V.
With the acquired equipment, a characterization campaign of different types of batteries was carried out, specifically, Li-Ion rechargeable batteries, Lithium Ion Polymer Batteries, HR20/D Accumulators, Lithium Phosphate Rechargeable Batteries, LiFePO4, NiMH, and LiPO. The above batteries were tested at different C-rates (e.g., C/10, C/5, C/2, C), at different temperatures, and with different numbers of successive cycles (e.g., 10, 20, 50) to broadly characterize their range of behavior.
The data made available allowed the units to develop models for the diagnostic and prognostic analysis of batteries based not only on classic parameters established in the literature, such as State of Charge and State of Life but on more advanced techniques for monitoring past life expressed in terms of historical trends in voltages, currents and temperatures. These models based on neural networks allow the classification Of the instantaneous State of the battery and the prediction of its future operation to characterize its performance and predict possible failures. The models also provide valuable support for predicting possible runaway events in modern battery systems.
Therefore, the research activity is completing the first five working packages, slightly later than initially planned due to difficulties in setting up the laboratory, which have now mostly been overcome.
At the same time, close collaboration with industries has enabled the design of an evolved BEMS management system to be routed by running tests on a test BESS system.
The BESS system is a “1-hour system” consisting of No. 1 DC buses connected to a Nidec inverter (Store Control 1); each DC bus is connected to 10 battery strings. The strings (STACK) are made with two types of drawers (BATTERY-PACK), called A and B, which are electrically identical but have mechanically reversed + and - terminals. Each battery pack in the rack houses two type A battery modules or two type B modules connected in series with each other. This allows the electrical connection between battery packs to be simplified by using copper rods. The single stack consists of 17 battery packs (8 type A + 9 type B) plus a control box containing string contactors, string fuses, and stack control electronics (STACK CONTROLLER)
Each half of the battery pack (rack drawer) consists of two basic modules, each consisting of 8 groups of 4 cells connected in series (4P8S configuration); the two modules are connected in parallel. Every battery module is monitored by a BATTERY PACK CONTROLLER (BPC), which communicates with the 4 CELL GROUP CONTROLLER (CGC), each installed on two battery bodies consisting of 4 cells each The CGC transmits information to the BPC regarding the temperatures and voltages of groups of 4 cells and enables cell balancing via special relays. All BPCs communicate via CAN bus with the Stack Controller located in the top left drawer of the rack.
A BEMS has been developed on the system as follows:
- Battery Management System (BMS): This system collects information such as voltage, current, and temperature levels from the batteries, protects them from operating outside the operational range areas, and keeps the charge levels of the different batteries balanced.
- Storage Management System (SMS): This system is responsible for dialoguing with BMS and EMS by controlling the inverters according to the connected strings. The SMS system has been integrated with BMS and EMS using a MESA device/SunSpec protocol.
- Energy Management System (EMS): This system communicates directly with SMS and Power Conversion System (PCS) to ensure efficient coordination of various plant components. It comprises machine-to-machine (M2M) logic and a web portal (Kobold).
The system is open and will allow to test different BEMS techniques that are under development by the research group.
Risultati attesi
Risultati raggiunti