According to the present invention, a machine learning-based vessel energy-power control management system provides a structure in which condition information of a vessel power source (a generator, a battery, a fuel cell or the like) and vessel driving equipment (an engine, a motor or the like) is primarily collected by a vessel information collection-management unit provided on a vessel, and then, is transmitted to a cloud server unit for EMS-PMS in a time slot capable of connection to an Internet communication network to be formed into a bigdata database, and the optimal control of the vessel power source and the vessel driving equipment, the detection and response of an abnormal equipment condition and the calculation of an output value by power source based on power usage prediction are performed through a machine learning regression analysis algorithm on a ground control server unit communicating with the cloud server unit for EMS-PMS, thus enabling optimal control in terms of power management/energy management, customized to a current state and intrinsic characteristics of a vessel, to be performed with high reliability and stability. Moreover, the system provides a structure in which a vessel driving equipment control signal and a vessel power source control signal generated by a machine learning regression analysis algorithm are individually calculated in a form of an input value calculation function algorithm for optimal control, a control signal generation algorithm for abnormal equipment condition detection-response, and a power usage prediction-based power source output value calculation algorithm, and then, is finally confirmed through reinforcement learning on a vessel virtual environment simulation module to be delivered to a vessel device control unit while making operation customized with the vessel, thus enabling high-performance/high-efficiency vessel optimal control.
본 발명에 따른 머신러닝 기반 선박 에너지-전력제어 관리시스템은 선박 전력원(발전기, 배터리, 연료전지 등)과 선박 구동기기(엔진, 모터 등)의 상태정보가 1차적으로 선박에 구비되는 선박정보 수집-관리유닛에 수집된 후 인터넷 통신망 접속가능시간 대에 EMS-PMS용 클라우드 서버유닛에 전송되어 빅데이터로 데이터베이스화되고, EMS-PMS용 클라우드 서버유닛과 통신하는 육상관제 서버유닛에서의 머신러닝 회귀분석 알고리즘을 통해 선박 전력원과 선박 구동기기의 최적제어, 기기 이상상태 검출과 대응, 전력사용예측 기반 전력원 별 출력값 산출이 수행되는 구조를 제공함으로써 선박의 자체특성과 현재상태에 맞추어진 전력관리/에너지관리 측면의 최적제어가 높은 신뢰성과 안정성을 가지면서 진행될 수 있고, 머신러닝 회귀분석 알고리즘에 의해 생성되는 선박 구동기기 제어신호와 선박 전력원 제어신호가 최적제어용 입력값 산출함수 알고리즘, 기기 이상상태 검출-대응용 제어신호생성 알고리즘, 전력사용예측 기반 전력원 별 출력값 산출 알고리즘의 형태로 초기 산출된 후 선박 가상환경 시뮬레이션 모듈에서의 강화학습을 통해 최종 확정되어 선박장치 제어유닛으로 전달되면서 해당 선박에 맞춤식으로 동작하게 되는 구조를 제공함으로써 고성능/고효율의 선박 최적제어가 가능해질 수 있는 기술적 특징을 갖는다.
- Machine learning based ship energy-power management system
머신러닝 기반 선박 에너지-전력제어 관리시스템
2021-04-30
Patent
Electronic Resource
Korean
IPC: | B63J Hilfsanlagen auf Schiffen , AUXILIARIES ON VESSELS / B63H Schiffsantrieb oder Steuerung der Schiffe , MARINE PROPULSION OR STEERING / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES |
Hybrid power ship energy management system based on fuzzy control
European Patent Office | 2025
|SHIP POWER ENERGY MANAGEMENT SYSTEM USING NEURAL NETWORK
European Patent Office | 2024
|Pure battery power ship energy management method and system and pure battery power ship
European Patent Office | 2024
|Ship Model for Machine Learning Based Ship Signature Classification
DataCite | 2025
|Multi-energy energy management system of diesel-electric hybrid power ship
European Patent Office | 2020
|