Abstract:With the widespread integration of fiber-optic communication, optoelectronic sensing terminals, and intelligent electronic devices into the power grid dispatching system, the complexity of dispatching operations has been continuously increasing. The accurate and timely execution of cross-level dispatching instructions is crucial for the safe and stable operation of the power grid. Addressing issues such as instruction semantic ambiguity, inconsistent execution status, and delayed closed-loop response in the optoelectronic sensing transmission system for power grid dispatching scenarios, a system for full-process monitoring and cross-level execution verification of dispatching instructions is proposed. This system establishes a hierarchical monitoring framework covering the entire lifecycle of instruction preparation, issuance, execution, and feedback. It utilizes edge computing to achieve preprocessing and real-time transmission of optoelectronic sensing data and on-site equipment operation data, and incorporates blockchain technology to ensure tamper-proof and traceable transmission of dispatching instructions. Furthermore, a bidirectional collaborative verification mechanism is designed, integrating intelligent algorithms such as state estimation, rule engines, Long Short Term Memory Network (LSTM), and Word to Vector (Word2Vec) models, to achieve pre-verification and reverse verification between superior and subordinate dispatching levels, reducing the risk of cross-level semantic ambiguity and misoperations. The technology achieves recognition accuracy rates of 98.0%, 99.0%, and 99.0% for instruction format errors, parameter violations, and equipment state conflicts, respectively, representing a 25% improvement in semantic matching compared to traditional schemes. Under normal operating conditions, the end-to-end delay is as low as 32 ms, and the delay under fault conditions is controlled within 70 ms, a 50% reduction compared to traditional schemes. After correction, the compliance rate is maintained between 98.8% and 99.3%, and the closed-loop recovery rate under fault conditions is improved by 15%. The results indicate that the proposed technology significantly enhances the accuracy, real-time performance, and safety of dispatching instruction execution supported by optoelectronic sensing data, providing effective technical support for full-process monitoring and intelligent verification of cross-level dispatching in the power grid.