Abstract: The scale of clean energy and electric energy substitution is expanding with the rapid development of China's new power system and the steady introduction of “dual carbon” strategic goals. Electric energy signals under the high proportion of renewable energy access and high-power dynamic load applications lead to nonlinear random dynamic changes, often causing serious deviations in electric energy measurements and affecting the fairness and rationality of electric energy trading. This study focuses on the energy economy, in the context of problems in implementing the aforementioned national strategies. Furthermore, this study identifies scientific problems, explores the important characteristics of dynamic loads that cause power metering deviations, and analyzes the local and global features of complex power dynamic load signals to address the challenges in accurately characterizing the global features of high-power dynamic load signals in existing research. Additionally, the method of constructing binary dynamic power testing signal models is explored. First, a discrete mathematical model is constructed for complex dynamic load signals of electrified railways and electric arc furnaces collected onsite. The important features of instantaneous voltage and current amplitudes are analyzed and extracted in the waveform domain, which reflects the approximate stability of voltage signals, the fast random dynamic fluctuation characteristics of current signals, the main characteristics of current amplitudes being an approximately Gaussian distribution, and decreasing autocorrelation coefficients. Second, based on run-length sequence mapping, a binary run-length sequence of complex dynamic load signals is constructed to analyze and extract important features, such as local and global run-length mode changes, modulation depth, and impact strength of current amplitudes on electrified railways and electric arc furnaces in the run-length domain. Compared with the proposed time-, frequency-, and time-frequency domain feature analysis methods proposed, the method suggested in this study has significant advantages in simultaneously extracting the local and global features of complex dynamic electrical-energy signals, characterizing important features such as large-scale fluctuations (large fluctuations), rapid changes over time (fast time-varying), and strong randomness. Finally, constraint conditions are constructed based on the typical characteristics of the run and waveform domains of complex power dynamic load signals. Using feature modeling methods, a binary m-sequence dynamic energy-testing signal model with specific parameters is constructed such that the testing signal reflects the typical features of the dynamic load signal and the most significant factors affecting energy measurement errors and covers the maximum range of feature parameter changes. This can also allow the simultaneous completion of the dynamic error testing of energy meters and the traceability of energy values. A dynamic error-testing system is built for electric-energy measurements, and the dynamic error of the electric-energy meter is tested under binary dynamic electric-energy-testing signal conditions. Experimental verification showed that the test signal reflected the typical characteristics of dynamic loads under the influence of electric-energy measurement errors. The research content of this paper provides a theoretical basis for the analysis of dynamic energy signal characteristics in complex scenarios, the construction of multifeature constraint models, and the dynamic error testing of energy meters.