Intelligent adjustment for power system operation mode based on deep reinforcement learning

Power flow adjustment is a sequential decision problem.The operator makes decisions to ensure that the power flow meets the system's operational constraints,thereby obtaining a typical operating mode power flow.However,this decision-making method relies heavily on human experience,which is inefficient when the system is complex.In addition,the results given by the current evaluation system are difficult to directly guide the intelligent power flow adjustment.In order to improve the efficiency and intelligence of power flow adjustment,this paper proposes a power flow adjustment method based on deep reinforcement learning.Combining deep reinforcement learning theory with traditional power system operation mode analysis,the concept of region mapping is proposed to describe the adjustment process,so as to analyze the process of power flow calculation and manual adjustment.Considering the characteristics of power flow adjustment,a Markov decision process model suitable for power flow adjustment is constructed.On this basis,a double Q network learning method suitable for power flow adjustment is proposed.This method can adjust the power flow according to the set adjustment route,thus improving the intelligent level of power flow adjustment.The method in this paper is tested on China Electric Power Research Institute(CEPRI)test system.

Neurosurgery in Parkinson's disease:Social adjustment, quality of life and coping strategies

Subthalamic nucleus deep brain stimulation has become a standard neurosurgical therapy for ad- vanced Parkinson＇s disease. Subthalamic nucleus deep brain stimulation can dramatically improve the motor symptoms of carefully selected patients with this disease. Surprisingly, some specific dimensions of quality of life, ＂psychological＂ aspects and social adjustment do not always improve, and they could sometimes be even worse. Patients and their families should fully understand that subthalamic nucleus deep brain stimulation can alter the motor status and time is needed to readapt to their new postoperative state and lifestyles. This paper reviews the literatures regarding effects of bilateral subthalamic nucleus deep brain stimulation on social adjustment, quality of life and coping strategies in patients with Parkinson＇s disease. The findings may help to understand the psychoso-cial maladjustment and poor improvement in quality of life in some Parkinson＇s disease patients.

Need for multiple biomarkers to adjust parameters of closed-loop deep brain stimulation for Parkinson's disease

Closed-loop deep brain stimulation（DBS）：DBS has been established as a surgical therapy for movement disorders and select neuropsychiatric disorders.Various efforts to improve the clinical outcomes of the procedure have been previously made.Several factors affect the DBS clinical outcomes such as lead position,programming technique,

Deep Learning-Based Control System Design for Emergency Vehicles through Intersections

This paper addresses the challenge of integrating priority passage for emergency vehicles with optimal intersection control in modern urban traffic. It proposes an innovative strategy based on deep learning to enable emergency vehicles to pass through intersections efficiently and safely. The research aims to develop a deep learning model that utilizes intersection violation monitoring cameras to identify emergency vehicles in real time. This system adjusts traffic signals to ensure the rapid passage of emergency vehicles while simultaneously optimizing the overall efficiency of the traffic system. In this study, OpenCV is used in combination with Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) to jointly complete complex image processing and analysis tasks, to realize the purpose of fast travel of emergency vehicles. At the end of this study, the principle of the You Only Look Once (YOLO) algorithm can be used to design a website and a mobile phone application (app) to enable private vehicles with emergency needs to realize emergency passage through the application, which is also of great significance to improve the overall level of urban traffic management, reduce traffic congestion and promote the development of related technologies.

Enrichment Mechanism and Prospects of Deep Oil and Gas

With the deepening of oil and gas exploration,the importance of depth is increasingly highlighted.The risk of preservation of storage space in deep reservoirs is greater than that in shallow and medium layers.Deep layers mean older strata,more complex structural evolution and more complex hydrocarbon accumulation processes,and even adjustment and transformation of oil and gas reservoirs.This paper systematically investigates the current status and research progress of deep oil and gas exploration around the world and looks forward to the future research focus of deep oil and gas.In the deep,especially the ultra-deep layers,carbonate reservoirs play a more important role than clastic rocks.Karst,fault-karst and dolomite reservoirs are the main types of deep and ultra-deep reservoirs.The common feature of most deep large and medium-sized oil and gas reservoirs is that they formed in the early with shallow depth.Fault activity and evolution of trap highs are the main ways to cause physical adjustment of oil and gas reservoirs.Crude oil cracking and thermochemical sulfate reduction(TSR)are the main chemical modification effects in the reservoir.Large-scale high-quality dolomite reservoirs is the main direction of deep oil and gas exploration.Accurate identification of oil and gas charging,adjustment and reformation processes is the key to understanding deep oil and gas distribution.High-precision detection technology and high-precision dating technology are an important guarantee for deep oil and gas research.

Design of deep-water omnidirectional spirit level

Attitude adjustment is a key link in the installation process of underwater facilities in deep water.To solve this problem,an omnidirectional spirit level for deep water was developed.The sealing principle of the spirit level and the principle of deep-water pressure resistance are analyzed,and the threaded connection strength is checked.The mechanical simulation verifies that the spirit level can withstand the pressure of 2000 m water depth,and the water pressure test is carried out for 30 min in a 20 MPa hyperbaric chamber.After the experiment is completed,the appearance of the spirit level is intact and there is no leakage.The experiment results show that the deep-water omnidirectional spirit level can be used in the deep sea within 2000 m.

基于改进门控循环神经网络的采煤机滚筒调高量预测

采煤机自适应截割技术是实现综采工作面智能化开采的关键技术。针对采煤机在复杂煤层下自动截割精度较低的问题,提出了一种基于改进门控循环神经网络(GRU)的采煤机滚筒调高量预测方法。鉴于截割轨迹纵向及横向相邻数据之间的相关性,采用定长滑动时间窗法对获取的采煤机滚筒高度数据进行预处理,将输入数据划分为连续、大小可调的子序列,同时处理横向、纵向的特征信息。为提高模型预测效率,满足循环截割的实时性要求,提出了一种用因果卷积改进的门控循环神经网络(CC-GRU),对输入数据进行双重特征提取和双重数据过滤。CC-GRU利用因果卷积提前聚焦序列纵向的局部时间特征,以减少计算成本,提高运算速度;利用门控机制对卷积得到的特征进行序列化建模,以捕捉元素之间的长期依赖关系。实验结果表明,采用CC-GRU模型对采煤机滚筒调高量进行预测,平均绝对误差(MAE)为43.80 mm,平均绝对百分比误差(MAPE)为1.90%,均方根误差(RMSE)为50.35 mm,决定系数为0.65,预测时间仅为0.17 s;相比于长短时记忆(LSTM)神经网络、GRU、时域卷积网络(TCN),CC-GRU模型的预测速度较快且预测精度较高,能够更准确地对采煤机调高轨迹进行实时预测,为工作面煤层模型的建立和采煤机调高轨迹的预测提供了依据。

可调弯鞘在回收困难的伞形下腔静脉滤器取出术中的应用

目的探讨可调弯鞘在回收困难的伞形下腔静脉滤器取出术中的应用价值。方法收集2021年7月至2023年2月于南京中医药大学附属无锡医院就诊的24例滤器回收困难患者临床资料,均采用可调弯鞘进行滤器回收,分析所有患者的手术成功率,观察下腔静脉血流通畅情况、管壁情况、造影剂外溢或滞留情况、滤器情况、相关并发症发生情况。结果23例患者手术成功,1例因患有阿尔茨海默病,术中不能配合而放弃手术。回收滤器均结构完整、无折断,无变形,滤器取出后行下腔静脉造影均未发现造影剂滞留或外渗。手术时间25~120 min,平均58 min,未明显增加手术时间及费用,无相关并发症发生。结论应用可调弯鞘可以显著提高回收钩贴壁的下腔静脉滤器回收成功率,安全、有效,具有一定的临床应用价值。

“三改联动”下高中压内缸适应灵活运行改造探讨

为实现“双碳”目标和构建新型电力系统,高中压内缸如何适应灵活性运行成为现役煤电高中压合缸机组“三改联动”时重点关注内容之一。文章通过对高中压内缸结构要素组合、典型结构方案、结合电厂实际运行情况和相应结构有限元分析,得出高中压进汽部分分开的内缸结构作为适应灵活性运行优选结构,并对现役机组高中压内缸改造方案进行探讨。

塔里木盆地塔北地区多期断裂差异匹配控制下超深岩溶缝洞储层成藏特征

超深断控岩溶缝洞型油藏是塔北地区重要油藏类型,探讨多期构造活动与深层油气成藏匹配关系对区域油气勘探具有重要意义。基于哈拉哈塘油田、金跃油田和富满油田连片地震资料,以野外地质露头断裂特征为指导,采用多种地震精细解释手段,分期、分级、分段刻画研究区断裂。基于研究区中寒武统玉尔吐斯烃源岩认识,结合加里东期、海西期和喜马拉雅期3期成藏的特点,将研究区走滑为主的控藏断裂划分为加里东早期、加里东中晚期、海西晚期和喜马拉雅期。进一步分析多期断裂继承性关系、通源特点和调整作用等,结合多种类型岩溶缝洞型储层的开发现状,探讨了研究区走滑断裂控制下的岩溶缝洞型储层成藏差异性。结果表明:(1)研究区油气藏关键因素是油源断裂通源性与后期断裂的调整作用,将加里东早期正断裂系统定义为源内断裂,加里东晚期形成的走滑断裂系统定义为源外断裂,源内断裂利于寒武系烃源岩排烃,源外断裂进一步沟通烃源岩实现油气运移成藏,基于这种认识建立4种通源模式;(2)结合研究区存在加里东晚期、海西期和喜马拉雅期3个主力生烃期的认识,海西晚期部分北西向走滑断裂继承性发育至二叠系,对加里东期油藏有一定的破坏和调整作用,喜马拉雅期部分北东向走滑断裂系统继承性发育至新近系,对早期油气藏起破坏和调整作用,建立了3种调整样式;(3)根据断裂匹配关系,建立6种走滑断裂控藏等级,并将研究区加里东中晚期走滑断裂带逐一划分,叠合开采现状图显示差异控藏断裂与油气生产情况匹配度较高;(4)选择发育断裂与岩溶共同控储的研究区,建立多期断裂系统与多种类型岩溶缝洞油气藏的匹配关系,并将认识成功应用于研究区井位勘探中,取得较好效果,为受控于走滑断层的岩溶缝洞油气藏勘探开发提供一定指导意义。

禁食低代谢场景下基于非接触式生理心理感知的多模态情绪调适研究

目的探索禁食低代谢场景下结合非接触式生理心理检测方法的全流程情绪调适模式,验证非接触式生理及心理检测算法精度,评估面向抑郁、焦虑等负性情绪状态的多模态情绪调适方案效果。方法部署非接触式生理心理检测算法与情绪调适方案,搭建多模态情绪调适系统,采集“绿航星旅Ⅷ”15 d完全禁食人体低代谢实验志愿者生理心理相关数据,并结合指夹式血氧仪与量表结果,验证系统内部署的非接触式生理心理检测算法精度,设计情绪调适实验,设置声音、穴位、磁、组合调适四个组别,比较志愿者调适前后量表评分,验证系统情绪调适效果。结果实验数据显示非接触式心率检测模型的Bland-Altman图差值平均数为﹣0.497 bpm,且95.3%的误差值检测数据处于95%一致性区间内;非接触式心理检测模型对于应激状态、焦虑情绪、抑郁情绪识别准确率均超过80%,疲劳状态、愤怒状态识别准确率超过70%;经情绪调适后,志愿者应激水平得到显著改善(P<0.05),积极情绪得到显著性促进(P=0.005),实时消极情绪评分明显改善。结论非接触式生理心理检测方法能有效识别禁食低代谢场景下志愿者生理及情绪状态,声音、穴位、磁、组合方案能够有效缓解志愿者负面情绪状态,为未来深空探测与地外驻留场景中航天员身心健康管理提供了新的技术途径。

基于BNN-IKAE双层多智能体深度强化学习的大电网母线电压智能自动调整方法

母线电压调整是保证电能质量和电力系统安全稳定运行的重要措施,但目前自动电压控制系统的数据处理及分析效率和正确性仍难以满足愈加复杂的大型电网要求,据此本文提出了一种二值神经网络知识经验融合(BNN-IKAE)双层多智能体深度强化学习算法,在深度强化学习的基础上进行大电网母线电压调整。本文首先介绍了常规的调整流程,并据此搭建了马尔科夫决策过程(MDP)模型;然后针对大电网可调元件参数复杂的问题设计了双层多智能体结构,通过引入二值神经网络(BNN)降低了网络复杂度,解决了模型计算速度慢的问题,并融合了专家经验的知识经验融合(IKAE)模块,通过专家经验池和存量判定机制提高了模型的收敛性和奖励值。最后,在东北地区电网中对提出的基于BNN-IKAE的双层多智能体深度强化学习模型的母线电压调整能力进行了仿真验证,与常规方法相比其调整时间减少了79.331%,调整的成功率增加了19.23%,结果表明基于BNN-IKAE双层多智能体深度强化学习的大电网母线电压智能自动调整方法能够提高计算速度和成功率。

基于OCT原位测量的可调环模激光焊飞溅定量评价

为了快速准确定量评价金属飞溅以优化工艺和保证焊接质量,采用1060铝合金可调环模(variable beam profile,VBP)激光焊为研究对象,搭建了基于光学相干层析成像(opticalc coherence tomography,OCT)的激光焊匙孔深度原位测量系统,并提出了一种1DCNN-BiLSTM复合深度学习模型,该模型利用两种网络单元的特性对匙孔深度信息进行局部和全局时序特征挖掘,实现了飞溅状态的定量评价.结果表明,该模型的飞溅识别准确率达到99.69%,为VBP激光焊工艺优化和质量控制提供了指导依据和闭环反馈.

基于卷积神经网络的滚动轴承故障诊断研究综述

随着机器学习技术的兴起,深度学习被用于故障诊断领域并得到迅速发展,其中,卷积神经网络是具有出色特征提取能力的深度学习模型,因其适用于处理图像数据和高维数据而成为故障诊断研究的热点。针对传统故障诊断方法难以解决轴承振动信号存在的特征提取困难和信号噪声污染的问题,为高效、准确地完成滚动轴承故障诊断工作,首先,对卷积神经网络的结构进行了简单介绍,并研究了近年来经典卷积神经网络模型用于滚动轴承故障诊断的重要进展;然后,从深度特征提取、超参数调整和网络结构优化等角度,对各种优化卷积神经网络的方法原理进行了简单介绍,详细探讨了将卷积神经网络应用于滚动轴承故障诊断的优化途径和已经取得的研究进展;最后,对几种典型优化方法的优势与不足进行了比较,并对不同角度优化卷积神经网络的途径进行了总结。研究结果表明:基于卷积神经网络的滚动轴承故障诊断方法还需要解决数据不平衡、模型特征提取能力不足和泛化性不强的问题,后续研究工作应聚焦于多源数据融合、模型性能优化以及多方技术结合等方向。

Power System Flow Adjustment and Sample Generation Based on Deep Reinforcement Learning

With the increasing complexity of power system structures and the increasing penetration of renewable energy,the number of possible power system operation modes increases dramatically.It is difficult to make manual power flow adjustments to establish an initial convergent power flow that is suitable for operation mode analysis.At present,problems of low efficiency and long time consumption are encountered in the formulation of operation modes,resulting in a very limited number of generated operation modes.In this paper,we propose an intelligent power flow adjustment and generation model based on a deep network and reinforcement learning.First,a discriminator is trained to judge the power flow convergence,and the output of this discriminator is used to construct a value function.Then,the reinforcement learning method is adopted to learn a strategy for power flow convergence adjustment.Finally,a large number of convergent power flow samples are generated using the learned adjustment strategy.Compared with the traditional flow adjustment method,the proposed method has significant advantages that the learning of the power flow adjustment strategy does not depend on the parameters of the power system model.Therefore,this strategy can be automatically learned without manual intervention,which allows a large number of different operation modes to be efficiently formulated.The verification results of a case study show that the proposed method can independently learn a power flow adjustment strategy and generate various convergent power flows.

Distribution rules of remaining oil by bottom water flooding and potential exploitation strategy in fault-controlled fractured-vuggy reservoirs

Based on the tectonic genesis and seismic data of fault-controlled fractured-vuggy reservoirs,the typical fractured-vuggy structure features were analyzed.A 3D large-scale visual physical model of“tree-like”fractured-vuggy structure was designed and made.The experiments of bottom-water flooding and multi-media synergistic oil displacement after bottom-water flooding were conducted with different production rates and different well-reservoir configuration relationships.The formation mechanisms and distribution rules of residual oil during bottom-water flooding under such fractured-vuggy structure were revealed.The producing characteristics of residual oil under different production methods after bottom-water flooding were discovered.The results show that the remaining oil in"tree-like"fractured-vuggy structure after bottom-water flooding mainly include the remaining oil of non-well controlled fault zones and the attic remaining oil at the top of well controlled fault zones.There exists obvious water channeling of bottom-water along the fault at high production rate,but intermittent drainage can effectively weaken the interference effect between fault zones to inhibit water channeling.Compared with the vertical well,horizontal well can reduce the difference in flow conductivity between fault zones and show better resistance to water channeling.The closer the horizontal well locates to the upper part of the“canopy”,the higher the oil recovery is at the bottom-water flooding stage.However,comprehensive consideration of the bottom-water flooding and subsequent gas injection development,the total recovery is higher when the horizontal well locates in the middle part of the“canopy”and drills through a large number of fault zones.After bottom water flooding,the effect of gas huff and puff is better than that of gas flooding,and the effect of gas huff and puff with large slug is better than that of small slug.Because such development method can effectively develop the remaining oil of non-well controlled fault zones and the attic remaining oil at the top of well controlled fault zones transversely connected with oil wells,thus greatly improving the oil recovery.

基于置信度策略优化的SuperGlue口腔特征匹配算法研究

随着深度学习技术的不断进步,特征匹配算法在计算机视觉领域的重要性日益凸显。传统的SuperGlue算法在特征匹配准确度上已经表现出了优异性能,但在处理低光照和纹理复杂的口腔图像时,其效率和准确性仍有提升的空间。针对上述问题,文中提出一种基于置信度策略优化的SuperGlue口腔特征匹配算法。首先,通过引入一个置信度评分机制,可以更准确地评估特征点对之间的匹配可能性,让算法聚焦于更可能正确的匹配点对;其次,提出动态置信度阈值调整策略,根据口腔图像对的特性和特征点分布自动调整阈值,以达到匹配数量与质量平衡的目的。经过一系列实验验证,改进后的算法在效率和准确性方面都取得了显著提升,尤其是在特征点多样性和图像质量不一的情况下,展现了更好的鲁棒性。设计算法的成功实现,为口腔视觉领域中的特征匹配问题提供了一种新的解决思路,具有重要的理论价值和实际应用前景。

颈深部脓肿的临床特征及预后影响因素分析

目的回顾分析颈深部脓肿患者的临床特征及预后影响因素。方法回顾性分析2014年10月至2023年8月中国医科大学附属第一医院耳鼻咽喉科急诊手术治疗的166例颈深部脓肿患者的病历资料,按照预后分为预后良好组和预后不良组,根据临床资料总结其临床特征,结合计算年龄校正Charlson共病指数(aCCI)评分,采用多因素logistic回归分析2组患者各项指标对预后的影响。结果2组性别、年龄、纵隔感染、脓毒症、糖尿病、高血压、C反应蛋白(CRP)、降钙素原(PCT)、aCCI评分比较,有统计学差异(P<0.05),表明以上因素对颈深部脓肿患者预后有显著影响。多因素回归分析结果表明纵隔感染、高血压、aCCI评分是颈深部脓肿患者预后的独立危险因素(P<0.05)。结论纵隔感染、高血压以及aCCI评分是影响颈深部脓肿患者预后的独立危险因素,可初步评估颈深部脓肿患者的预后。

基于DBN-ELM的构网型并网逆变器控制参数自适应调整方法

“双高”电力系统中电网阻抗呈现宽范围时变特性,构网型并网逆变器控制参数缺乏自适应调整能力,存在失稳风险。对此,提出一种基于深度置信网络-极限学习机的构网型并网逆变器控制参数自适应调整方法。建立闭环极点映射模型,利用深层架构对控制参数与关键极点之间的映射关系进行训练;通过训练好的闭环极点映射模型预测得到相应的关键极点,识别出关键极点最接近参考极点时构网型并网逆变器的控制参数;通过自适应调整控制参数,确保系统在电网阻抗变化时跟踪参考极点,实现自适应稳定控制。理论分析和仿真结果均表明,所提方法能够实现控制参数的自适应调整,有效提高构网型并网逆变器对电网阻抗变化的适应性。

基于深度强化学习的移动通信网载波调整算法

为解决如何准确、及时地对移动通信网络扇区进行载波调整的问题,提出了一种基于深度强化学习的扇区扩(减)容算法。采用Model-based强化学习方法,建立了容量指标概率动态模型的多模型组合,利用真实环境的历史数据对模型进行训练,并在此基础上构建了虚拟环境。然后用神经网络构建智能体,并使之与虚拟环境互动,采用短展开技术,产生虚拟样本。最后利用虚拟样本,采用DQN算法对智能体进行策略优化,使其给出扇区扩(减)容操作的建议。实验结果表明,训练后的智能体给出的载波调整建议,达到了较高的正确率。