Multi-source errors evaluation of machine tools: from research gaps to methodologies and applications

Multi-source errors, as critical obstacles limiting the accuracy retention and machining performance of machine tools, hold fundamental and strategic significance for achieving high-precision, high-efficiency, and high-reliability machining in modern manufacturing systems. However, these errors typically exhibit complex characteristics such as strong coupling, time-variance, and nonlinearity, which challenge traditional methods of error identification, modeling, and compensation in terms of adaptability, real-time capability, and integration. Therefore, it is imperative to establish a systematic and intelligent multi-source error control framework. Firstly, this work systematically reviews typical error sources and their evolution mechanisms, evaluates multi-scale detection technologies including laser interferometry, double ball-bar systems, multi-sensor fusion, and vision-based systems, and constructs an intelligent error identification and evaluation framework. Next, it reviews classical modeling methods such as homogeneous transformation matrices, screw theory, thermal equilibrium models, finite element analysis, and modal analysis, compares physical modeling, data-driven, and hybrid modeling strategies, and develops an integrated multi-source error modeling architecture centered on digital twin technology and artificial intelligence. Furthermore, key technologies, including geometric error mapping and real-time compensation, online thermal error prediction and active temperature control, dynamic error suppression, and adaptive control, are summarized. A multi-level integrated error compensation architecture is proposed by combining physical models, data models, and cyber-physical synchronization. This architecture encompasses core processes such as error traceability and decoupling, dynamic prediction, real-time compensation, and closed-loop optimization, emphasizing engineering implementation mechanisms based on cyber-physical collaboration, multi-physics coupling, and multi-scale fusion, thereby effectively enhancing accuracy stability and control robustness under complex operating conditions. Finally, frontier challenges such as constructing high-fidelity coupled models from heterogeneous multi-source data, edge–cloud collaborative control, and cross-platform interoperability are discussed. The application prospects of multi-source error evaluation are also envisioned, providing theoretical foundations and technical support for the precise management and optimization of the entire lifecycle accuracy of machine tools.