A triad framework for ship carbon reduction: Direct CO2 measurement, multi-intelligence fusion prediction, and Cauchy-enhanced speed optimization

• First integration of direct CO₂ measurement, multi-intelligence fusion prediction, and sailing speed optimization, achieving closed-loop decision-making for micro-level ship operation carbon reduction. • Effective cleansing of raw data achieved by integrating manual expertise, statistical IQR method, and HDBSCAN clustering algorithm. • Multi-intelligence CO₂ prediction model constructed by fusing multiple categories of meta-heuristic algorithms. • Enhanced Equilibrium Optimizer (EEO) algorithm designed by incorporating Cauchy mutation strategy for speed optimization, experimentally verified to improve convergence performance by over 20 %. • Application of the triad framework enables direct carbon emission reduction of 12.27 t during a 6-day offshore voyage Ship carbon emissions have become a global concern, yet significant challenges persist in precise quantification, real-time assessment, and operational optimization. This study proposes an integrated "direct measurement-emission prediction-speed optimization" framework to bridge these gaps. Firstly, a shipboard CO₂ direct measurement system is deployed to collect high-frequency dynamic data. Through synergistic application of domain expertise, statistical methods, and HDBSCAN clustering, erroneous values, anomalies, and outliers are effectively identified and eliminated, yielding a high-confidence fine-grained dataset for microscopic emission analysis and operational decisions. Secondly, four metaheuristic algorithms are incorporated into a dual-adaptive prediction architecture, expanding its mechanistic boundaries to construct a highly reliable and precise CO₂ prediction model. Finally, an Enhanced Equilibrium Optimizer (EEO) with Cauchy mutation strategy demonstrates superior performance in carbon-oriented speed optimization. During a 6-day coastal voyage, EEO algorithm achieves direct carbon reduction of 12.27 tons and 4.11% energy efficiency improvement, realizing synergistic gains in decarbonization and Energy Efficiency Operational Indicator (EEOI). This research provides a holistic technical framework and granular decision support for shipping decarbonization.