Multiport DC converter and intelligent energy management system

Abstract

To provide a multiport converter designed to supply direct current loads.SOLUTION: A multiport converter comprises one or more energy sources, a battery, a multiport DC-DC converter connected to the energy sources and to a load, and an IoT device for communication and control of the system. The multiport DC-DC converter further comprises an energy management system for selectively controlling the charging or discharging of the battery and the amount of energy directed to the load based on predictions of energy generation from the energy sources, based on a load demand and based on optimized scenarios. The multiport converter is used in green hydrogen generation plants powered by renewable energy sources, such as wind and photovoltaic energy, and a system capable of self-managing and ensuring better use of energy from the sources and better efficiency of green hydrogen generation.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to Brazilian Patent Application No. P1020240094115, filed May 13, 2024, the contents of which are incorporated herein by reference.

[0002] The present invention relates to an electrical energy conversion system with a multi-port converter for powering a DC load based on, but not limited to, solar, wind and battery bank energy. [Background technology]

[0003] The production of hydrogen from mixed renewable sources (solar and wind) is gaining importance for flexible and non-polluting energy directions. Furthermore, wind energy generation is considered complementary to photovoltaic (PV) solar energy, as their generating capacity is complementary depending on weather variability. When solar energy is at its peak generation, wind energy is generally in its valley, and vice versa.

[0004] Furthermore, energy storage systems (ESS) are an excellent complement to intermittent renewable energy generation, as they can store energy during peak generation periods for future energy supply during periods of low renewable generation, improving the stability and reliability of these systems.Green hydrogen (GH2) can be considered a form of energy storage, obtained by electrolysis of water using renewable energy sources, albeit in a long-term chemical form.

[0005] Hybrid microgrids, consisting of batteries, PV power generation systems, and electrolyzer-based ESSs, operate on direct current (DC). On the other hand, wind power systems operate on three-phase alternating current (AC), but require a DC conversion stage to convert the generated energy into grid-compliant energy. Conventional microgrids, which integrate these technologies through an AC bus, have multiple DC / DC, AC / DC, and DC / AC conversion stages, as illustrated in Figure 1, which leads to inefficient processing.

[0006] The collection of different conversion stages in the setup used to produce green hydrogen requires them to be compatible with each other, resulting in a decrease in overall process efficiency and making it difficult to integrate the various devices. Furthermore, this partitioning of energy converters makes the standardization and implementation of microgrids a more complex task, as these devices must be compatible with each other in terms of voltage and operating current. Even after the converters are standardized, they still need to be configured for operation, requiring additional control devices such as programmable logic controllers (PLCs) to orchestrate their operation.

[0007] Current technology Brazilian Patent Application Publication No. 112022012326, entitled "Conversor deenergia multiportas" ("Multiport Energy Converter"), discloses a multiport converter that provides faster dynamic response to load changes than prior art systems, including hybrid energy storage systems (HESS) that allow for a smaller primary energy storage system (ESS) and a longer lifespan of the primary ESS (e.g., energy battery), or maintain the same ESS size while allowing for a wider range or longer lifespan of energy sources. Multiport converters advantageously reduce investment and maintenance costs and can also advantageously provide a path for the input to directly power the load. All of these benefits can be achieved while reducing the number of active switches and overall component count compared to prior art systems.

[0008] Chinese Patent Application No. 212726480, entitled "Grid-connected and off-grid type wind-solar-water hydrogen storage fuel cell direct current interconnection microgrid system," discloses a grid-connected and off-grid type wind-solar-water hydrogen storage fuel cell direct current interconnection microgrid system, which includes a DC bus, a wind energy power generation system, a hydroelectric power generation system, a solar cell energy power generation system, a battery-powered energy storage system, a water electrolysis hydrogen production system, a fuel cell system, an electric load, an oxygen storage system, a hydrogen storage system, and a hydrogen load. The DC bus is connected to the output of the wind energy power generation system, the output of the hydroelectric power generation system, the output of the solar cell energy power generation system, the battery energy storage system, the energy interface of the water electrolysis hydrogen production system, the output of the fuel cell system, and the electric load. The oxygen output of the water electrolysis hydrogen production system is connected to the oxygen storage system. The input of the hydrogen storage system and the output of the hydrogen storage system are connected to the hydrogen input of the hydrogen load and the fuel cell system. This system can effectively improve the energy utilization efficiency of a clean energy microgrid.

[0009] Chinese Patent Application Publication No. 202210811923, titled "Optimized operation control method for flexible DC power distribution network," discloses an optimized operation control method for a flexible DC energy distribution network, suitable for optimizing the operation of a multi-terminal flexible DC energy distribution network under normal operating conditions. It optimizes the operation of the DC energy distribution network based on three time scales: day-ahead forecast optimization, continuous intraday correction, and real-time intraday feedback correction. The method employs strategies such as energy storage charging and discharging, converter power adjustment, flexible load power adjustment, and solar cell output adjustment, thereby improving the consumption capacity of distributed energy sources, reducing operating losses, and ensuring economical and efficient operation of the DC distribution network. This method fully utilizes the flexibility of the flexible DC energy distribution network and combines the potential of adjustable and controllable source-load resources, such as energy storage, load, and energy source, to achieve safe and stable operation of the flexible DC energy distribution network and improve the absorption capacity and utilization rate of distributed energy sources. Summary of the Invention [Problem to be solved by the invention]

[0010] The present invention describes a multi-port DC converter that aims to minimize costs, increase efficiency, and increase the feasibility of implementing local power generation, thereby allowing a microgrid to have a single multi-port energy converter. The multi-port DC converter of the present invention unifies the equipment required to process energy from multiple sources, simplifying and reducing the design costs and difficulties in implementing and controlling a microgrid. [Means for solving the problem]

[0011] Furthermore, the present invention presents an advanced energy management algorithm based on artificial intelligence (AI) for controlling the DC converter. The algorithm is responsible for determining which energy sources will be used to power the load, ideally the GH2 electrolyzer, and aims to keep them operating at their maximum power point, taking into account the characteristics of each source. In this way, the overall efficiency of the microgrid is maximized. The AI ​​uses historical energy generation data and weather information to accurately predict energy generation from PV and wind sources. Based on these predictions, the AI ​​makes strategic decisions about energy storage in the battery bank. Maximum GH2 generation can be guaranteed, taking into account future demand specified by the user. The optimization algorithm plays a key role in planning the optimal time setpoint for the electrolyzer. Based on the energy generation forecast and GH2 generation needs, the algorithm determines the optimal time to direct energy to the electrolyzer, taking into account demand conditions and the characteristics of available energy sources.

[0012] Communication between the multiport converter, the PLC (Programmable Logic Controller) controlling the load, and the artificial intelligence in the cloud can be handled by a device called IoT Manager. This is a communication hub that receives the operating instructions calculated by the AI ​​and directs them to the multiport converter, as well as aggregating data collected from the system (consisting of the load, multiport converter, battery, and auxiliary peripherals) and sending it to a server. It can be connected to the BMS (Battery Management System) modules of the converter and battery via CAN (Controller Area Network) communication with the PLC, which controls the hydrogen plant, for example, via the ModBus protocol, and has Internet connectivity via 2G / 4G mobile networks and Wi-Fi.

[0013] The invention will now be described with reference to a general embodiment thereof and with reference to the accompanying drawings, in which: FIG. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a diagram of a conventional AC microgrid according to the current state of the art. [Figure 2] FIG. 2 is a simplified diagram of a DC microgrid using a multi-port DC converter according to the present invention. [Figure 3] FIG. 3 is a detailed diagram of a DC microgrid using a multi-port DC converter according to the present invention. [Figure 4] FIG. 4 is a diagram of the software architecture of a multi-port DC converter according to the present invention. [Figure 5] FIG. 5 is a diagram of the general architecture of an energy management module according to the present invention. [Figure 6] FIG. 6 shows an architecture diagram of an IoT manager device according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0015] Specific embodiments of the present disclosure are described below. In an effort to provide a concise description of these embodiments, all features of an actual implementation may not be described herein. It should be understood that in the development of any actual implementation, like any engineering or design project, myriad implementation-specific decisions must be made to achieve the developer's specific objectives, including compliance with system-related and business constraints that may vary from implementation to implementation. Moreover, it should be understood that such a development effort, while complex and time-consuming, may represent a routine undertaking of design, fabrication, and manufacture for those of ordinary skill in the art having the benefit of this disclosure.

[0016] The present invention comprises an embodiment of a single energy converter with a multi-port topology, which reduces the equipment required to process energy from multiple sources, so that each source operates from the same converter arm, reducing the cost of power electronics. An example of the proposed topology is shown in Figure 2.

[0017] The present technology has been developed for powering electrolyzers in green hydrogen generation plants (GH2), which are typically powered by renewable energy sources such as wind and solar cells. Therefore, most of the drawings and examples in this description show the electrolyzer as the load and wind and solar photovoltaic sources as the energy supply sources. However, the invention is not limited in this manner. Those skilled in the art will immediately recognize that the present description can be used to power any type of load powered by direct current and with any type of energy source, simply by making the necessary adjustments to the control and power supply modules comprising the multi-port converter.

[0018] The use of a DC grid and the elimination of the energy conversion step allows for a significant increase in the overall efficiency of the system compared to a conventional grid. Furthermore, designing the converter for a specific application (e.g., GH2 generation) allows it to operate at its peak efficiency over the nominal operating range.

[0019] In addition to the concept of increased efficiency, switches such as gallium nitride (GaNFET) and silicon carbide (SiC) transistors are preferably used, which have switching losses an order of magnitude lower when compared to MOSFET or IGBT transistors, allowing the converter to achieve higher switching frequencies and higher efficiencies, and minimizing the size of the device by reducing passive components.

[0020] At least one of the converter arms must have bidirectional energy handling capacity to support the use of a battery bank as one of the supply sources. The battery bank is used to store excess electrical energy generated by renewable sources during peak generation hours for future use during peak energy consumption hours, ensuring stability of energy supply to the load.

[0021] Based on this, Figure 3 shows a block diagram of the proposed multi-port converter, illustrating a preferred application where the power is supplied by wind energy generation and photovoltaic (PV) energy generation and the load is the electrolyzer of a GH2 plant.

[0022] The arm connecting the wind power generation consists of a three-phase rectifier whose function is to convert the three-phase AC voltage of the wind turbine into direct current, preferably followed by a unidirectional buck-boost DC / DC converter operated by a current control system that limits the current supplied to the electrolyzer while ensuring maximum operating efficiency of the generator. These types of maximum power control, also known as MPPT (Maximum Power Point Tracking), are algorithms that aim to maximize energy extraction from systems that exhibit nonlinear power curves. In the case of wind power systems, the majority of MPPT applications rely on measurements of mechanical parameters such as machine speed or blade angle.

[0023] In the present invention, the multi-port converter preferably uses a strategy that is not affected by the mechanical parameters of the wind turbine, which depends on the constructive aspects of the wind turbine. An MPPT algorithm based on reading the voltage and current of the subsequent rectifier should be used, which makes the tracking method more efficient and more flexible for use with different wind turbines.

[0024] Meanwhile, the connection port to the PV module is configured with a unidirectional DC / DC arm. This also uses MPPT control to maximize the output power of the solar array. Solar cell systems and wind power systems have nonlinear power curves, so a tracking algorithm is required. Therefore, the MPPT algorithm used here may be similar to that used in wind power systems, with voltage and current also used as input parameters for the algorithm.

[0025] For the battery energy storage system, a bidirectional DC / DC converter is used, which charges the battery system during periods of maximum energy generation and discharges the battery system during electrolyzer operation, i.e., using the battery system as an energy source during periods of low energy generation from wind and / or PV sources.

[0026] The structure of a multiport converter consists of a hierarchical control system with three control levels: primary control, secondary control, and tertiary control. Each level performs a specific function and has a different response speed. Communication between the hierarchical levels is carried out through an IoT system called the IoT Manager, shown in Figure 6.

[0027] The IoT manager shown in FIG. 6 includes a microcontroller unit (MCU) connected to a memory (e.g., flash memory) and FRAM. The IoT manager communicates with the BMS, the multi-port DC converter, and the PLC via a transceiver for bidirectional data transmission and reception between these components. For the BMS, communication is preferably performed via a CAN (Controller Area Network), while for the PLC, the Modbus protocol is preferably used. The IoT manager also bidirectionally communicates with the cloud, from which weather data or operator control data can be obtained when the operating parameters of the multi-port DC converter are changed. Communication with the cloud can be performed via a WiFi module and / or a mobile network module (2G / 4G). The MCU is further configured to receive an analog signal (setpoint) from the PLC.

[0028] Primary controls are the lowest level in the control hierarchy and, as a result, have a faster response than other levels of control. Primary controls of multiport converters refer to the most basic level of control equipment, i.e., local voltage and current controllers, that directly monitor and respond to local network conditions. This includes controllers for energy storage systems (e.g., batteries) and wind and PV converter ports. These primary controls act to maintain local stability by controlling voltages, currents, and other variables essential to the operation of individual components of the multiport DC converter. Generally, primary-level controls act on the switch-level control of the converters according to control criteria provided by higher-level controls and may operate with a centralized or independent strategy. Thus, their primary function is to control the voltage, current, and power of each converter input, interacting with the maximum power point trackers (MPPTs) of the wind and PV systems, aiming for DC currents with ripples limited to 10% or less of their nominal values, in addition to controlling the bus voltage and current supplied to the electrolyzer.

[0029] Secondary control occupies an intermediate level in the control hierarchy. In the case of the multiport converter of the present invention, secondary control is responsible for limiting voltage, balancing current between devices, and managing battery energy storage. This level of control aims to reduce deviations from the nominal voltage of the converter's intermediate bus using a closed-loop voltage controller (PI, PID, or resonant controllers can be used, techniques already known in the current technology). Control of the bidirectional battery arm is essential to balance this dynamics and ensure energy is available when needed. Another point that may be noted in a battery-focused control strategy relates to the intermittent sources associated with microgrids. Sources such as PV and wind are intermittent depending on weather conditions. Batteries can compensate for this variability by storing energy when there is excess generation and releasing it when power generation is low, ensuring a stable supply. In addition to the safety and reliability of operating as an energy backup, there are also issues of peak demand and optimizing the use of renewable sources.

[0030] Tertiary control is the highest level of control in the hierarchy. Its main purpose is to optimize the operation of multi-port converters. At this level, technologies such as the Internet of Things (IoT), predictive systems, and energy management optimization are applied. Tertiary control acts as an optimizer for the previous control system. Its function is to predict energy generation and consumption and optimize the use of energy resources, with the goal of ensuring that energy is delivered to the loads as planned.

[0031] The interaction of the hierarchical control levels plays a key role in ensuring the stability, efficiency, and safety of the system and is essential for optimal operation. The interaction between the control levels is as follows: The tertiary control, running on the server, determines the percentage of power (relative to the nominal power) to be applied to the load and where this energy originates (PV / wind source or battery) according to the energy management algorithm and weather forecast. Therefore, the tertiary control determines the operating setpoints for the output arm of the multiport converter and the bidirectional arm of the battery. These setpoints are values ​​between 0 and 1 for the output arm and between -1 and 1 for the bidirectional arm. These values ​​are then transmitted to the multiport converter via an internet connection to the IoT manager. The IoT manager receives commands from the server and directs them to the multiport converter via the CAN network. Meanwhile, the second-level control begins in the multiport converter, converting the received setpoint data into current reference values. This is done by multiplying the received value by the nominal current of each converter arm. These new current reference values ​​are used by the current control loops of the converter output arm and the bidirectional battery arm at the primary control level. To protect the converters and stabilize the intermediate bus voltage through energy balance, additional adjustments are made to the current values: the sum of the converter's input and output currents must always be zero, so that all energy is directed to the load or battery. These current references are then transmitted to the primary level control, which runs in the DSP (digital signal processor) of each arm of the converter and applies this reference to a current control loop, which, by means of a PI or PID controller, adjusts the switching of the power switches of the arms of the multiport converter to deliver current at the established level.

[0032] Intelligent Energy Management System The intelligent energy management system means comprises an optimization system and an artificial intelligence (AI) with a prediction and optimization module. This AI is responsible for managing the various energy sources used to generate GH2. The AI ​​is adapted to make automated decisions, with the goal of balancing and optimizing the system. The optimization system includes:

[0033] Data Entry: 1) Solar Power Generation Forecast: Information about the amount of energy generated by a solar system over a given period of time. 2) Current Battery Status: The current charge level of the battery bank. 3) Load Demand: The amount of energy / power required at a given time in the future.

[0034] System Constraints: The system into which the multi-port converter is inserted imposes constraints such as maximum and minimum solar cell power yield, maximum and minimum battery capacity, and the power range that the electrolyzer can accommodate.

[0035] Optimization problem formulation: The objective is to maximize energy availability to the load at all times, which is achieved by adjusting the hourly power to the load and determining when to store or draw energy from the battery.

[0036] Decision variables: Variables such as the nominal power of the load and the power supplied to or extracted from the battery to charge the battery are determined by an optimization algorithm.

[0037] Initial conditions and boundaries: Initial conditions, such as minimum power of the load, are established and bounds are defined to ensure that the solution is within the constraints of the system.

[0038] Solving optimization problems: Optimization algorithms are applied to find ideal values ​​of the decision variables that ensure the load demand is met at any given time, subject to system constraints.

[0039] Algorithm output: The algorithm output gives the ideal power per time to the load and the power used to charge or discharge the battery.

[0040] Below are some example scenarios of application of the present invention taken from validation tests. These scenarios used an electrolyzer from a green hydrogen production plant as a load, with the aim of maximizing the production of green hydrogen. Such application examples should not be construed as limiting and are given here merely as examples. Those skilled in the art will immediately recognize that the present invention is applicable to any type of load.

[0041] Scenario 1 - Charged battery, high solar power forecast: Initial condition: charged battery. Solar power generation forecast: High power generation throughout the day. Hydrogen demand: Moderate. Algorithm behavior: In this scenario, optimization may suggest that most of the energy generated is directed to the electrolyzer, maximizing hydrogen production. The battery is charged, so there is no need to store energy. The electrolyzer operates near its maximum capacity. There is no battery charging, as solar generation meets demand, so there is no need to store energy for periods of low generation.

[0042] Scenario 2 - Discharged battery, high solar power forecast: Initial condition: Fully / partially discharged battery. Solar power generation forecast: High power generation throughout the day. Hydrogen demand: Moderate. Algorithm behavior: May choose to store excess energy in the battery during periods of high power generation for use during periods of low power generation.

[0043] Scenario 3 - Partially charged battery, solar power peak: Initial condition: partially charged battery. Solar power generation forecast: Peak power generation mixed with periods of low power generation. Hydrogen demand: high-constant. Algorithm behavior: Energy from solar peaks is used to charge the battery, allowing later demand to be met and maintaining constant hydrogen production.

[0044] Scenario 4 - Discharged battery, low solar power: Initial condition: Discharged battery. Solar power generation forecast: Low power generation throughout the day. Hydrogen demand: Moderate. Algorithm behavior: Prioritizes hydrogen production using the limited energy available from solar power generation, drawing on stored energy from the battery as needed.

[0045] Scenario 5 - Urgent hydrogen demand, low solar power generation: Initial condition: partially charged battery. Solar power generation forecast: Low power generation throughout the day. Hydrogen demand: urgent and high. Algorithm behavior: Prioritize battery use and focus on hydrogen generation even if it results in a large discharge.

[0046] These are only illustrative examples, and the effectiveness of the algorithm depends on the specific dynamics of each system, such as load characteristics, component efficiencies, and solar power generation patterns. The algorithm is adapted to optimize the balance between hydrogen production and the specific demands of the energy storage and off-grid system.

[0047] The forecasting and optimization module, which also belongs to the AI ​​module of the present invention, operates on the basis of short-term / medium-term weather forecasts and energy generation data, aiming for long-term equilibrium using historical data. Weather data is captured through two possible options: if the microgrid has weather stations connected to the cloud and integrated into a database, or by obtaining data through external APIs such as OpenWeatherMap API, WeatherBit API, Dark Sky API, AccuWeather API, etc.

[0048] Two machine learning algorithms, XGBoost and LSTM (long short-term memory), are used for power generation forecasting. LSTM is a type of recurrent neural network (RNN) architecture designed to handle the complexity of long time sequences. Unlike traditional RNNs, LSTMs have a "gate" mechanism that allows them to remember and forget information over time, which is important for capturing relevant temporal patterns. LSTM architectures consist of cell units, each equipped with three main gates: an input gate, a forget gate, and an output gate. These gates control the flow of information and allow selective retention or forgetting of data. This makes LSTMs effective at modeling long-term dependencies in temporal data. The basic steps in implementing an LSTM model include data preparation, model building, training, and prediction. During training, the model adjusts the weights of connections between cell units to optimize its ability to predict patterns in the input time sequence.

[0049] XGBoost is a machine learning library that excels at building robust, efficient, and accurate decision tree models. Using boosting techniques, XGBoost combines multiple weak decision trees to construct a stronger, more generalizable model. This method is particularly effective at predicting continuous values, making it well suited for temporal prediction problems.

[0050] Building an XGBoost model involves defining hyperparameters such as the maximum tree depth, learning rate, and number of trees. During training, the XGBoost model adjusts these hyperparameters to minimize the loss function, thereby improving its ability to generalize to new data.

[0051] The use of algorithms such as LSTM and XGBoost to forecast energy generation over a 96-hour horizon is an advanced application in the field of time series. When forecasting with XGBoost, the model combines the predictions of each individual tree to generate a final forecast. The interpretability provided by the tree structure used allows for a clearer understanding of the factors affecting the forecast. In data preparation, it is important to properly organize the time sequence to account for the 96-hour horizon. Normalizing the data facilitates model convergence. Both LSTM and XGBoost present specific challenges and considerations when working with a 96-hour horizon. LSTM excels at capturing complex temporal patterns, while XGBoost, due to its interpretability, provides a clearer view of the factors affecting the forecast. Continuous experimentation, tuning to the architecture and hyperparameters, along with cross-validation, are fundamental practices for optimizing the performance of these models in specific scenarios.

[0052] Therefore, using machine learning to predict power generation and use it as input to optimization algorithms allows for intelligent decisions and the autonomous ability to adapt to fluctuations in demand and power generation, ensuring efficient storage management.

[0053] The integration of this intelligence with the multiport converter is performed by the IoT Manager, a communications device specifically developed for the needs of the present invention, as described above and shown in Figure 6. The intelligent control algorithm operates at the third hierarchical control level of the multiport converter and receives information from the other hierarchical levels via IoT functionality. The intelligent control algorithm communicates with the cloud server using a 2G / 4G cellular wireless module to receive the primary and secondary level operating parameters and send the collected data to the cloud database.

[0054] In third-order control, an AI algorithm runs on the cloud and is responsible for determining the operating parameters for each port of the DC / DC converter.

[0055] The third level of hierarchical control also has an intelligent energy management system (EMS), which is divided into two modules: a machine learning module and an optimization module. A cloud architecture allows data to be collected, processed, and stored in a centralized manner. A database keeps historical records of, for example, solar and wind power generation data, hydrogen demand, battery energy storage, and other relevant metrics such as the status of multiport converters, hydrogen plants, and battery banks. An application programming interface (API) enables integration with other systems or devices, such as an IoT manager, providing real-time information and enabling communication between system components.

[0056] The EMS begins by collecting data acquired by sensors from the energy sources, in this case, the solar panels and wind turbines, which provides information about current energy production. This data is sent by the IoT Manager to a cloud database and stored for future analysis. A machine learning module then generates a prediction of energy production. Fed this historical data and weather information, the machine learning module predicts the amount of energy generated by the solar panels and wind turbines over a future time horizon, such as 96 hours in the future. Based on this prediction, the EMS can make a decision on whether to store energy in a battery bank. The overall software architecture is shown in Figure 4 and is described in more detail below.

[0057] The application's web system is designed with a layered architecture, with each layer handling a specific area of ​​operation. The web system uses two databases: one for storing operational data and the other for storing system administration and registration information. This allows for efficient insertion and agile retrieval of IoT Manager data to provide information to the dashboard. The operational database is preferably high-performance and scalable, e.g., MongoDB, a NoSQL database. To ensure communication between applications, a modular backend service is proposed, with each module performing a specific task. Integration between the frontend weather system (e.g., OpenWeatherMap API, WeatherBit API, Dark Sky API, AccuWeather API) and the platform is achieved through APIs in the backend. Communication with the IoT Manager module is achieved through IoT services that pass through the data processing module. The web system has a backend that receives, processes, and stores data from the IoT Manager, interacts with the frontend, provides information to the dashboard, and handles AI algorithms. The frontend allows users to interact with the system and is developed with reusable components using JavaScript and the ReactJS framework. Application state control is managed by Redux, which facilitates the sharing of these states between components.

[0058] The forecasting and optimization module shown in FIG. 5 is responsible for defining load demand based on several forecast scenarios, such as those illustrated above, as well as controlling when the battery should be charged and when it should be used as an energy source. In the example of the present disclosure, the user needs to define the amount of hydrogen production in cubic meters, i.e., the load demand. Based on a forecast of future energy generation, the forecasting and optimization module evaluates whether the amount of energy generated is sufficient to meet the load demand (electrolyzer). If the forecast indicates high future energy generation (greater than consumption), the scheduling forecasting and optimization module can direct a large portion of the energy, for example, more than 50% of the power, or even all of the energy generated at that moment, directly to the electrolyzer and increase the output setpoint. This ensures that a large portion of the immediately available renewable energy is available to power the load, since there is a guarantee of sufficient future energy generation for operation without requiring battery energy. On the other hand, if the forecasted future energy generation is low (less than demand), the system can choose to store energy in the battery bank, lowering the output setpoint and resulting in reduced hydrogen production. This strategy aims to store excess energy, when available, to ensure a stable supply of energy to the electrolyzer, even when the instantaneous renewable generation is lower than the consumption by the electrolyzer. In this way, the forecasting and optimization module uses battery storage as a strategic energy reserve to maximize the energy output setpoint for the electrolyzer, keeping it operating in its region of maximum efficiency, even when renewable generation is low.

[0059] The system communicates with the electrolyzer using APIs and the IoT Manager to provide information on energy availability and hydrogen demand. Based on this data, the electrolyzer can adjust its hydrogen production to meet demand. Figure 5 shows the schematic architecture of the analytics module that controls the forecast of energy generation and manages energy allocation.

[0060] Figure 5 shows the input variables for the optimization algorithm. The first variable is "Solar and Wind Power Forecast," which provides information about the amount of energy expected to be generated by solar and wind sources at a given time in the future. The second input variable is "Battery State of Charge," which indicates the amount of energy stored in the battery. This information is essential for identifying whether there is available capacity to store excess energy or whether energy needs to be used immediately. Based on the current battery state of charge, the system is configured to perform several predictions to achieve optimization through scenarios that ensure maximization of the planned hydrogen production. These simulations consider system constraints such as solar and wind power boundaries as well as the maximum storage capacity of the battery bank. Based on the simulation results, the optimization algorithm selects the best scenario that satisfies the system constraints and determines the "Maximum Hourly Setpoint for the Electrolyzer." This value represents the maximum power that should be directed to the electrolyzer at a given time to maximize hydrogen production.

[0061] Other application examples Renewable energy generation: The system of the present invention is applicable to wind farms and solar plants where multiple renewable energy sources are connected to the electrical grid. The multi-port DC / DC converter allows for the integration and efficient management of these different energy sources, maximizing renewable energy generation.

[0062] Microgrid: The system of the present invention can be used to establish autonomous microgrids in remote or island areas where access to the main energy grid may be limited. Different energy sources such as solar panels, wind turbines and battery storage systems can be integrated and efficiently managed by converters and management systems to provide stable and sustainable energy to the community.

[0063] Hydrogen Fuel Station: With the growing interest in hydrogen fuel cell vehicles, the system of the present invention can be applied to hydrogen fuel stations. Different renewable energy sources, such as panels and wind turbines, can be integrated with multi-port DC / DC converters and management systems to ensure the production of green hydrogen for fueling vehicles.

[0064] Integration with natural gas networks: The green hydrogen produced can be integrated into existing natural gas networks, contributing to the decarbonization of the gas sector. Multi-port DC / DC converters and intelligent IoT energy management systems can be used to manage the production of green hydrogen and its injection into natural gas networks, optimizing the use of renewable sources and ensuring the stability and quality of the hydrogen produced.

[0065] Electrification of transportation: With the increasing demand for electric vehicles, the system of the present invention can be applied to charging stations or electric vehicle refueling stations. Different energy sources, such as solar panels and battery energy storage systems installed on the station's roof, can be integrated and managed to provide clean and renewable energy for charging electric vehicles.

[0066] Industrial infrastructure: The system can be implemented in industrial facilities to optimize energy consumption and reduce operational costs. Different energy sources available at the facility, such as rooftop solar panels, nearby wind turbines, or cogeneration systems, can be integrated and managed by converters and management systems to provide an efficient and sustainable energy supply for industrial activities.

[0067] This deployment contributes to reducing the costs of the renewable hydrogen production chain and involves maximizing plant efficiency through the use of a DC network with a single converter and its intelligent decision-making for energy generation and management. This measure can also support the deployment of distributed generation measures by producing hydrogen close to the site of consumption, whether industrial or mobile, helping to reduce the costs associated with transporting this fuel.

[0068] The combination of technical solutions described herein constitutes a complete energy control and conversion system for operating an electrolysis plant to produce hydrogen locally. The resulting system is robust, scalable, low-cost (compared to conventional multi-port DC converters), easy to implement, and suitable for deployment in various locations with hydrogen demand. Furthermore, it has intelligent monitoring and control features, providing efficient process management.

[0069] Advantageously, the multi-port converter according to the present invention can dynamically adjust energy scheduling, fully utilizing renewable generation and battery storage. This results in optimized operation of the microgrid, enabling efficient hydrogen production and intelligent use of available resources. This innovative approach provides more efficient and sustainable operation and contributes to the production of hydrogen in an environmentally friendly manner.

[0070] While aspects of the present disclosure are susceptible to various modifications and alternative forms, specific embodiments have been shown by way of example in the drawings and are described in detail herein. It should be understood, however, that the invention is not limited to the particular forms disclosed. Rather, the invention is intended to cover all modifications, equivalents, and alternatives falling within the scope of the present invention as defined by the following appended claims.

Claims

1. 1. A multi-port DC converter, comprising: one or more energy sources; A battery, Equipped with The multi-port DC converter comprises: receiving energy from one or more energy sources and / or the battery; charging the battery using at least a portion of the energy from the one or more energy sources; configured to power a load by directing energy from the one or more energy sources and / or the battery; The multi-port DC converter further comprising an energy management system for selectively controlling the charging or discharging of the battery and the amount of energy directed to the load.

2. 10. The multi-port DC converter of claim 1, wherein the one or more energy sources comprise a solar energy source and a wind energy source, and the load comprises a green hydrogen (GH2) electrolyzer.

3. 3. The multi-port DC converter of claim 2, wherein the energy management system provides selective control based on a forecast of energy generation from the one or more energy sources and / or based on a forecast of load demand.

4. 10. An energy management system for use with the multi-port DC converter of claim 1, comprising: a machine learning module for predicting energy generation from each of the one or more energy sources; a battery charge state monitoring module; a communication module including an energy management system and a cloud service that identifies a load demand in communication with the load; a prediction and optimization module configured to determine battery charging using at least a portion of the energy generated by the one or more energy sources and / or battery discharging to supply energy to the load based on a prediction of energy generation of each of the one or more energy sources, the battery state of charge, and load demand, and / or to increase or decrease the energy directed to the load; An energy management system comprising:

Images