Empirical Measurement and Influence Mechanism of Resource Allocation Efficiency in Collaborative R&D of New Energy Technologies Within Distributed Innovation in Electric Power System Operators

Accelerated technological iteration in new energy has prompted electric power system operating entities to engage in cross-organizational and cross-regional collaborative R&D through distributed innovation models. These models not only support knowledge integration and resource synergy but also play a vital role in enhancing system-wide planning and operational scheduling efficiency. However, due to variations in R&D intensity, structural configurations, and knowledge heterogeneity across projects, notable divergences in resource allocation efficiency have emerged, which further impact the responsiveness and flexibility of automated system dispatching.To characterize this divergence and its underlying mechanisms,this study constructs a sample of 187 collaborative R&D projects involving 42 power and new energy enterprises from 2012 to 2022. Input indicators include project-level R&D expenditure, R&D personnel hours, and experimental equipment conversion values, while output indicators comprise authorized patents, high-value patent ratios, prototype systems, and standard proposals, forming a collaborative R&D input-output database.Using the SBM-DEA model to measure resource allocation efficiency at the project level, results indicate an average efficiency value of 0.73 for the sample projects. Approximately 31% of projects operate on the efficient frontier, while around 27% exhibit redundant investment space.Further analysis using a double Bootstrap truncated regression model linked efficiency scores to firms' degree centrality within distributed innovation networks, cross-regional collaboration intensity, technological distance, and knowledge stickiness levels. Findings indicate that each standard deviation increase in network centrality boosts project efficiency by approximately 6.2%.while the interaction term between technological distance and knowledge stickiness index exerts a significant negative impact on efficiency. Projects characterized by high stickiness and high technological distance exhibit efficiency levels approximately 11%–14% lower than those with low stickiness and low distance. This study quantitatively reveals the formation mechanism of resource allocation efficiency in new energy collaborative R&D among electric power system operators within a distributed innovation context, providing methodological support for optimizing project portfolios and partner selection.