Search

The continuous increase in renewables penetration poses a severe challenge to the frequency control of interconnected power grid. Since the conventional automatic generation control (AGC) strategy does not consider the power flow constraints of the network, the traditional approach is to make tentative generator power adjustments based on expert knowledge and experience, which is time consuming. The optimal power flow-based AGC optimization model has a long solution time and convergence issues due to its non-convexity and large size. Deep reinforcement learning has the advantage of “offline training and online end-to-end strategy formation”, which yet cannot ensure the security of artificial intelligence (AI) in power grid applications. A coordinated optimal control method is proposed for active power and frequency control based on safe deep reinforcement learning. First, the method models the frequency control problem as a constrained Markov decision process, and an agent is designed by considering various safety constraints. Then, the agent is trained using the example of East China Power Grid through continuous interactions with the grid. Finally, the effect of the agent and the conventional AGC strategy is compared. The results show that the proposed approach can quickly generate control strategies under various operating conditions, and can assist dispatchers to make decisions online.

Effectively tapping into the response potential of load-side resources is one of the effective means to enhance the control capabilities of load resources. To efficiently perceive the response potential of loads， a method for assessing the response potential of load-side resources is proposed based on SWARA （step-wise weight assessment ratio analysis）， CRITIC （criteria importance through intercriteria correlation）， and MARCOS （measurement of alternatives and ranking according to compromise solution）. Firstly， a multidimensional set of impact factor indicators for load response potential is constructed based on factors from the load， economic， social， and system perspectives. Fuzzy number theory is then employed to assess the weights of these indicators and the response potential of the load. Subsequently， the comprehensive weights of various indicators are calculated by combining SWARA and CRITIC. Utilizing MARCOS， the utility function values of load response potential for each type of load are determined based on the integrated indicator weights， resulting in a ranking of load response potential. Lastly， the effectiveness of the proposed method is validated through a case study using part of the dispatchable user loads in Shanghai.

Abstract The progression of human degenerative and hypoxic/ischemic diseases is accompanied by widespread cell death. One death process linking iron-catalyzed reactive species with lipid peroxidation is ferroptosis, which shows hallmarks of both programmed and necrotic death in vitro. While evidence of ferroptosis in neurodegenerative disease is indicated by iron accumulation and involvement of lipids, a stable marker for ferroptosis has not been identified. Its prevalence is thus undetermined in human pathophysiology, impeding recognition of disease areas and clinical investigations with candidate drugs. Here, we identified ferroptosis marker antigens by analyzing surface protein dynamics and discovered a single protein, Fatty Acid-Binding Protein 5 (FABP5), which was stabilized at the cell surface and specifically elevated in ferroptotic cell death. Ectopic expression and lipidomics assays demonstrated that FABP5 drives redistribution of redox-sensitive lipids and ferroptosis sensitivity in a positive-feedback loop, indicating a role as a functional biomarker. Notably, immunodetection of FABP5 in mouse stroke penumbra and in hypoxic postmortem patients was distinctly associated with hypoxically damaged neurons. Retrospective cell death characterized here by the novel ferroptosis biomarker FABP5 thus provides first evidence for a long-hypothesized intrinsic ferroptosis in hypoxia and inaugurates a means for pathological detection of ferroptosis in tissue.

Abstract The heterogeneity of macrophages influences the response to immune checkpoint inhibitor (ICI) therapy. However, few studies explore the impact of APOE+ macrophages on ICI therapy using single‐cell RNA sequencing (scRNA‐seq) and machine learning methods. The scRNA‐seq and bulk RNA‐seq data are Integrated to construct an M.Sig model for predicting ICI response based on the distinct molecular signatures of macrophage and machine learning algorithms. Comprehensive single‐cell analysis as well as in vivo and in vitro experiments are applied to explore the potential mechanisms of the APOE+ macrophage in affecting ICI response. The M.Sig model shows clear advantages in predicting the efficacy and prognosis of ICI therapy in pan‐cancer patients. The proportion of APOE+ macrophages is higher in ICI non‐responders of triple‐negative breast cancer compared with responders, and the interaction and longer distance between APOE+ macrophages and CD8+ exhausted T (Tex) cells affecting ICI response is confirmed by multiplex immunohistochemistry. In a 4T1 tumor‐bearing mice model, the APOE inhibitor combined with ICI treatment shows the best efficacy. The M.Sig model using real‐world immunotherapy data accurately predicts the ICI response of pan‐cancer, which may be associated with the interaction between APOE+ macrophages and CD8+ Tex cells.

In the field of defect detection in sewers, some researches focus on high accuracy. However, it is challenging for portable on-site devices to provide high performance. This paper proposes a lightweight sewer defect detection model, You Only Look Once (YOLO) v5-Sewer. Firstly, the backbone network of YOLOv5s is replaced with a stacked MobileNetV3 block. Secondly, the C3 module of the neck of YOLOv5s is improved with a C3-Faster module. Thirdly, to compensate for the accuracy loss due to the lightweight network, a channel attention (CA) and convolutional block attention module (CBAM) are added to the proposed method. Finally, the Efficient Intersection over Union (EIOU) is adopted as the localization loss function. Experimental validation on the dataset shows that YOLOv5-Sewer achieves a 1.5% reduction in mean Average Precision (mAP) while reducing floating-point operations by 68%, the number of parameters by 55%, and the model size by 54%, compared to the YOLOv5s model. The detection speed reaches 112 frames per second (FPS) with the GPU (RTX 3070Ti). This model successfully implements a lightweight design while maintaining the detection accuracy, enhancing its functionality on low-performance devices.

Summary: Expansion of the hexanucleotide repeat GGGGCC in the C9orf72 gene is the most common genetic factor in amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). Poly-Gly-Ala (poly-GA), one form of dipeptide repeat proteins (DPRs) produced from GGGGCC repeats, tends to form neurotoxic protein aggregates. The C9orf72 GGGGCC repeats and microglial receptor TREM2 are both associated with risk for ALS/FTD. The role and regulation of TREM2 in C9orf72-ALS/FTD remain unclear. Here, we found that poly-GA proteins activate the microglial NLRP3 inflammasome to produce interleukin-1β (IL-1β), which promotes ADAM10-mediated TREM2 cleavage and inhibits phagocytosis of poly-GA. The inhibitor of the NLRP3 inflammasome, MCC950, reduces the TREM2 cleavage and poly-GA aggregates, resulting in the alleviation of motor deficits in poly-GA mice. Our study identifies a crosstalk between NLRP3 and TREM2 signaling, suggesting that targeting the NLRP3 inflammasome to sustain TREM2 is an approach to treat C9orf72-ALS/FTD.