国网宁夏电力有限公司

This paper proposes a lithium battery state of health(SOH)prediction method that combines a Diffusion model and a Transformer to address the issue of low accuracy in predicting the SOH of lithium batteries under com-plex operating conditions.By integrating the nonlinear auto regression with exogenous input(NARX)mechanism with a structured state space model,a dynamic model consisting of convolutional neural network,Mamba model,Transformer decoder,and conditional spread model is constructed to improve the prediction accuracy and generali-zation performance of lithium battery SOH.Firstly,we stack the multi cycle lithium battery data into a three-dimen-sional tensor and standardize it.A dual path feature input module is used to extract multi-dimensional temporal fea-tures and historical SOH correlation information separately.The main path captures the long-range dependencies of capacity decay through convolutional layers and Mamba state space models,while the auxiliary path uses historical SOH values to enhance temporal features.Then,the Transformer decoder will read the dynamic behavior of the bat-tery based on the NARX method,predict the SOH of the battery at cycle t,and complete the NARX feedback loop to enable the model to merge its past predictions.A conditional diffusion model is designed at the output of the model,which guides the diffusion process through conditional features,effectively suppressing the accumulation of prediction errors and improving robustness to noise.The experimental results show that the proposed model can a-chieve accurate prediction of battery SOH and effectively integrate cross cycle degradation features,achieving high-precision prediction of lithium battery SOH under complex operating conditions.

The continuous increase in the penetration ratio of new energy in active distribution network has led to a sharp increase in operational data.While,coupled with the intermittent nature of new energy sources and other fac-tors,has posed significant challenges to dispatching strategies.Therefore,a real-time optimization scheduling mod-el called VMD-BiLSTM-PPO is proposed.Based on edge computing,this model constructs a multi-region energy autonomous framework,adopts the proximal policy optimization(PPO)algorithm of deep reinforcement learning to minimize operating scheduling costs and achieve optimized scheduling of the distribution network with cloud-edge collaboration.It decentralizes a large amount of computing and data storage tasks to the edge,effectively reducing the computing and data transmission loads at the scheduling center.Furthermore,in order to effectively alleviate the impact of new energy fluctuation,it adopts a new energy output prediction framework based on variational mode decomposition(VMD)decomposition and bi-directional long short-term memory(BiLSTM)in the PPO algorithm.Simulation experiments indicate that the model can improve the utilization rate of new energy and enhance the eco-nomic performance of real-time distribution network scheduling.

The complex structure of distribution network and the cumbersome operation procedures make it easy to cause misoperation accidents during the operation process.In response to this issue,this paper studies the research of distribution network security risk prevention and control system based on multi-source data.The overall frame-work of the system implementation and the specific design of each module are designed.Multiple system data from the same source system,self distribution system,and supply and service system are integrated,and real-time data processing and classification storage are adopted using a subscription message queue.Based on the integration of multi-source data from distribution network security risk prevention and control system,the random forest algorithm is used to judge the abnormal operating status.Through the actual application of a city company,the method and system are verified to meet practical needs in abnormal data detection,response speed,and misoperation prevention unlocking accuracy,thereby ensuring the safe operation of distribution network.

In the new energy grid,the intermittent impact of natural resources such as wind and solar makes the power of the grid stochastic,leading to the problem of multiple objectives needing to be optimized simultaneously.Traditional particle swarm optimization combined with a single model often needs to give up the consideration of load,which is a key factor,to ensure that the single model solution process will not fall into the local optimal solu-tion.In this regard,a research on staged operation optimization of new energy grid resources considering load con-straints is proposed.By using the differential autoregressive moving average(ARIMA)model to predict the low-fre-quency part of the load,and using a deep belief network(DBN)to predict the high-frequency part of the load,the two parts are superimposed and reconstructed to generate the load prediction results for the new energy grid.A two-stage optimization operation model is designed for power grid resources,including day ahead and real-time.Final-ly,a global optimal solution is performed for the two-stage configuration model to achieve optimization of resource operation in the new energy grid through using the adaptive mutation particle swarm optimization algorithm.The ex-perimental results show that the proposed algorithm has a lower amount of abandoned wind and solar power in each time period within 24 hours compared to the comparative algorithm.The new energy consumption rate reaches the highest 97.6%at 3 hours,and the controllable load time shift rate and operating cost are much lower than the com-parative algorithm,improving the automation and intelligence level of the power grid.

In the context of the power and carbon trading markets， addressing the flexibility requirements of power systems with a high penetration of renewable energy while minimizing investments and construction costs is crucial. This paper proposes a collaborative planning method for source-grid-load-storage， incorporating electricity-carbon coupling and flexible supply-demand balance. First， the electricity-carbon coupling relationship is established by considering electricity prices and carbon trading prices. Second， the adjustable potential of flexible load is analyzed， and a price-based demand response model is developed based on adaptive time-sharing tariffs under electricity-carbon coupling. Third， the flexibility supply capacity of various resources within the power system is assessed， leading to the formulation of a flexibility supply-demand equilibrium model. This model is further refined using a full life-cycle theory. Subsequently， a two-layer source-grid-load-storage cooperative planning model is proposed， with the upper layer aimed at minimizing the integrated cost of construction investment using the equal annual value method， while the lower layer seeks to minimize the sum of the annual operational cost of the power system and the annual carbon trading cost. Finally， the improved adaptive genetic algorithm is used to solve the model. The case study results demonstrate that the proposed method provides a valuable reference for the collaborative planning of source-grid-load-storage systems.