Table of Content

05 April 2020, Volume 53 Issue 4

Previous Issue    Next Issue

Study on Technical Standard of Distributed Resources Grid Integration

HE Guoqing, WANG Weisheng, LIU Chun, ZHANG Yue, LIANG Zhifeng, SUN Wenwen, LI Guanghui

2020, 53(4):  1-12,176.  DOI: 10.11930/j.issn.1004-9649.202002061

Asbtract ( )   PDF (6370KB) ( )  

References | Related Articles | Metrics

Recently, the rapid growth of distributed resources (DR) has promote the development of standards for DR grid integration. In this paper, the existing DR standards about grid integration in China and abroad were introduced first, and then the representative standards in North America, Europe and China were compared according to power control, grid adaptability, power quality, synchronization, safety and protection, etc. Finally, this paper indicated the main differences between North American, European and Chinese standards, analyzed the development trend of DR grid integration standards in the world and presented reasonable suggestions for the updating of Chinese DR standards. The work in this paper may provide available reference for the development and review of DR grid integration standards in China and enhance the understanding of DR standards for DR developers, as well as equipment manufacturers, so as to promote the development of DR industry in future.

Active Distribution Network Cost Allocation Method Based on Distribution Factor Method

ZHANG Yi, CHEN Qixin, XIA Qing, SHI Lei, YU Le, XIE Xu, ZHANG Jing

2020, 53(4):  13-21.  DOI: 10.11930/j.issn.1004-9649.201906010

Asbtract ( )   PDF (3627KB) ( )  

References | Related Articles | Metrics

Considering the characteristics of active distribution network’s flexible topology and variable power flow, the traditional method of cost allocation is no longer applicable. Therefore, this paper proposes a new cost allocation method for the active distribution network based on the generalized load distribution factor (GLDF). By calculating the distribution factor of nodes and topology, the proposed method can dynamically adapt to the changing operation mode of active distribution network. Through network loss modification and analysis of the full load level utilization, the DC power flow model kernel of the distribution factor method is improved. This method not only has good adaptability to active distribution network, but also can be used to objectively measure the usage of different load nodes in different operation modes. Moreover, this method is helpful to guide the market users to adjust their electricity consumption behavior. The feasibility and rationality of the method is verified by an example.

Calculation Method for Critical Conditions of Distribution Network Considering Distributed Generation “Plug and Play”

JIANG Zhen, YU Li, HUANG Yanlu, YE Linhao, ZHENG Shuang, GENG Guangfei, TANG Wei

2020, 53(4):  22-31.  DOI: 10.11930/j.issn.1004-9649.201906028

Asbtract ( )   PDF (3848KB) ( )  

References | Related Articles | Metrics

In order to provide a theoretical basis for the realization of “plug and play” for distributed generation (DG) and other elements in the distribution network, from the macroscopic point of view, the first and second synergy systems of smart distribution networks that consider “plug and play” elements are established. Take distributed generation as an example, considering the constraints of load and DG power changes, voltage, line capacity, power flow reversal and DG power factor, a mathematical model for calculating the “plug and play” critical condition of distributed generation is proposed. And based on the consideration of load and DG power fluctuation, a critical condition execution strategy is proposed. Furthermore, through theoretical analysis, the constraint conditions in the critical condition model are equivalently simplified, and a fast estimation method for solving the “plug and play” critical conditions of DG is obtained. Taking the IEEE 33 nodes as an example, the proposed calculation method and simplified algorithm are compared and analyzed. The results show that the proposed simplified algorithm is feasible and can quickly estimate the DG “plug and play” critical conditions.

A SNOP Capacity Planning Method Considering Network-loss Optimization Benefit and Investment Cost

HE Yingjing, SHEN Shuyi, HE Dong, ZHANG Jingrui, LI Fan, ZHU Keping, WANG Songyan

2020, 53(4):  32-40.  DOI: 10.11930/j.issn.1004-9649.201911044

Asbtract ( )   PDF (4020KB) ( )  

References | Related Articles | Metrics

Soft normally open point (SNOP) has strong ability in distribution network power flow control, but its application in distribution network is limited due to its quite high investment costs. An optimal planning method is thus proposed with consideration of the network-loss reduction benefit and SNOP investment cost. Firstly, the operation scenarios of the distribution network are generated according to time-series based on the life cycle of SNOP. Based on the restrains of SNOP and system operation states, the network-loss reduction benefits of SNOP are quantitatively analyzed under different SNOP capacities. And then, the SNOP cost is analyzed according to the topologic structure of modular multilevel converter, and the relationship between SNOP capacity and cost is linearly fitted. At last, a planning scheme is proposed for optimal SNOP installation location and capacity allocation considering the balance between network-loss reduction benefit and SNOP investment cost in daily operations. Case study shows that, limited by the high investment and operation costs of SNOP, it is feasible to install one SNOP in a distribution network under most conditions.

Investment Decision-making Method for Regional Distribution Network Planning Considering Distributed Generations

KONG Xiangyu, YAO Jingtao, CUI Kai, SUN Fangyuan, YANG Jingang

2020, 53(4):  41-48.  DOI: 10.11930/j.issn.1004-9649.202002096

Asbtract ( )   PDF (3757KB) ( )  

References | Related Articles | Metrics

As an important power generation mode of renewable energy, the distributed generation (DG) is increasingly integrated into distribution networks. Aiming at the uncertainties in investment structure and investment benefits of the regional distribution network with DG, an investment decision-making method is proposed for distribution network planning considering DG integration. Firstly, based on analysis of the DG impact on the distribution network, an association analytic method is given for investment benefits of the distribution network, and intermediate variables, such as distribution network renovation measures, affected operating parameters and investment benefit indicators, are introduced to associate the investment with benefits. Secondly, by taking the optimal comprehensive benefits as the objective, an investment decision-making method is proposed for distribution network planning, and the main processes and key steps are given. At last, the effectiveness of the proposed method is verified through case study of a power supply area in a northern China province.

Impact of Dispersed Wind Farm Integration on Distributed Network in Different Scenarios

LIAN Yiqing, YUAN Zhiyong, LEI Jinyong, MA Xiyuan, TAO Siyu

2020, 53(4):  49-58.  DOI: 10.11930/j.issn.1004-9649.201910118

Asbtract ( )   PDF (4725KB) ( )  

References | Related Articles | Metrics

In order to study the impact of dispersed wind farm (DWF) integration on distribution network in different scenarios, a simulation model is established for the distribution network with DWFs, and a comparative analysis is made. Firstly, based on the basic model of distribution network, the paper analyzes the impact mechanism of DWF access on node voltage, power flow distribution and network loss. Then, based on a case of a distribution network in China Southern Power Grid, the paper studies the impacts of DWF access with different penetration rates on the distribution network in different scenarios. Finally, according to the simulation results, the paper summarizes the impact of dispersed wind power access on the distribution network. The general law of voltage and network loss distribution in distribution network is obtained under different access location, access capacity and access form of DWFs. The simulation results show that different access modes of DWFs, including accessing at the head, middle and end terminals of distribution network feeder, or access of single or multiple wind farms, have different impacts on voltage and network loss distributions. It is therefore important in practice to reasonably determine the access capacity and location of DWFs.

Reliability Evaluation of Cyber-physical Distribution Network Considering the Impact of Weather Conditions

WU Yingjun, FAN Tingting, XU Hao, NI Ming

2020, 53(4):  59-68.  DOI: 10.11930/j.issn.1004-9649.201912037

Asbtract ( )   PDF (4321KB) ( )  

References | Related Articles | Metrics

The deep integration of power and communication is the development trend of distribution network, and weather is one of the most important factors affecting the reliability of the distribution network. Therefore, it is necessary to consider the influence of weather in studying and evaluating the reliability of the distribution cyber-physical system. First, the weather factors that have the greatest impact on power component faults and communication component faults are selected, and divided into intervals to build a weather state model suitable for analyzing power component faults and communication component faults. Second, based on the historical fault data and historical weather data, the power component failure probability model and communication component failure probability model are established with weather conditions as variables. Third, for the consequence evaluation of monitoring and control failure caused by communication component faults, a load loss quantification analysis method is given. And then, considering the faults of power components and communication components, a reliability assessment method is proposed for the distribution cyber-physical system based on Monte Carlo method. Finally, simulations were performed on the improved IEEE 33-node test system, which has verified the feasibility and effectiveness of the proposed method.

Reactive Power Optimization of Distribution Network Based on Random Matrix and Historical Scenario Matching

AN Ran, WU Junyong, SHI Chen, ZHU Xiaowen, SHAO Meiyang, HUANG Xing, CAI Rong

2020, 53(4):  69-78.  DOI: 10.11930/j.issn.1004-9649.201903019

Asbtract ( )   PDF (3926KB) ( )  

References | Related Articles | Metrics

Reactive power optimization and voltage management of distribution networks is very important for the optimal operation of distribution networks. By introducing big data theory into reactive power optimization of distribution networks, a reactive power optimization method is proposed based on random matrix and historical scenario matching, which does not need the model and parameters of the distribution network, and can directly use the data generated during the operation of the distribution network to construct seven high-dimensional random matrices and extract 57 characteristic indicators. The extracted characteristic indicators are reduced in dimension to match the existing scenarios in the historical database, and the scenarios closest to the statistical characteristics of the current system are found. The control strategy under the matching scenarios is adopted as the reactive power optimization control strategy in the current period to reduce the active power loss and node voltage deviation. Finally, the method is verified on the modified IEEE-37 node distribution network model, where the model of random loads, including the distributed generations such as photovoltaic/wind power, and electric vehicles are added. The results show that the proposed method can effectively optimize the reactive power of the distribution network without the need of its model and parameters, and the online decision-making speed is fast.

Research on Economic Optimization Scheduling of Park Integrated Energy System with Electricity-heat-cool-gas Load

SUN Qiang, XIE Dian, NIE Qingyun, ZHANG Lihui, CHEN Qian, CHEN Jiejun

2020, 53(4):  79-88.  DOI: 10.11930/j.issn.1004-9649.201912038

Asbtract ( )   PDF (3922KB) ( )  

References | Related Articles | Metrics

Integrated energy system is an important way to achieve coordinated optimization between different energy systems. This paper firstly analyzes the four subsystems of electricity-heat-cool-gas, and constructs an economic optimization scheduling model for the park integrated energy system with the total cost as objective. Secondly, the non-linear method is used to transform the constructed scheduling model into a mixed integer programming problem. Finally, based on two scenarios of independent operation and coupled scheduling of subsystems, and combined with the time-of-use electricity price, the mixed integer programming software is used to solve the model. The result shows that, compared to the independent operation mode, the coupling scheduling mode can flexibly combine various energy conversion devices, greatly reducing the system's external energy purchase and the total system costs.

Research on the Transaction Mechanism and Benefit of Promoting New Energy Consumption

YANG Jianhua, WANG Xiongfei, XIAO Daqiang, ZHAO Jiawei, LIU Dunnan

2020, 53(4):  89-95.  DOI: 10.11930/j.issn.1004-9649.201904024

Asbtract ( )   PDF (3476KB) ( )  

References | Related Articles | Metrics

Based on the problems for promoting high proportion of new energy consumption, this paper firstly constructs a contribution model of market mechanism to new energy consumption, aiming at the three market mechanisms of power generation rights trading, standby ancillary services trading and time of use (TOU) electricity price trading of self-owned power plants. Secondly, based on the market mechanism of promoting new energy consumption, the paper analyses the environmental benefits, benefits of saving electricity purchasing cost, and benefits of improving power supply reliability for promoting new energy consumption, and constructs a quantitative model to quantify the benefits of market mechanism. Finally, the data of northwest and southeast coastal areas are selected to calculate the benefits. The results show that the effective market mechanism can promote the high proportion of new energy consumption, while bringing considerable benefits.

Optimal Configuration of BESS in Commercial Area Considering Virtual Energy Storage Characteristics of Load

GAO Wenhao, ZHAO Haibing, YIN Shuangrui, CUI Demin, TIAN Huaiyuan, GE Yang, LI Xiaobo, SUN Haitao

2020, 53(4):  96-104.  DOI: 10.11930/j.issn.1004-9649.201809067

Asbtract ( )   PDF (3900KB) ( )  

References | Related Articles | Metrics

The rapid development of energy storage technology has gradually increased the economic efficiency and flexibility of power systems, while the virtual energy storage of load(VES) based on demand side management has received extensive attention as a kind of generalized energy storage. We take the commercial area prosumers as the research objects and establish the VES model of air conditioning load with time decoupling characteristics based on human comfort range. On this basis, considering the seasonal and climatic differences between load and photovoltaic (PV) power generation and the life cycle cost, a bi-level optimal configuration model for battery energy storage system (BESS) considering the VES characteristics is established, and the Gurobi solver combined with the modified differential evolution (MDE) is adopted. Finally, the specific example shows that the participation of VES can effectively improve the economic efficiency of BESS configuration in commercial area, and the impact of energy storage system on PV consumption is analyzed.

Probability Prediction of Low-voltage Tripping Failures in Distribution Transformer Station Areas Based on XGBoost Algorithm

WU Qiong, YU Wencheng, HONG Haisheng, YU Lei, DUAN Lian, SHANG Mingyuan, LIU Zhe

2020, 53(4):  105-113.  DOI: 10.11930/j.issn.1004-9649.201908140

Asbtract ( )   PDF (3834KB) ( )  

References | Related Articles | Metrics

In view of the frequent low-voltage tripping faults in the distribution transformer station areas during summer season, a machine learning-based warning model is proposed for low-voltage tripping in this paper. Firstly, a combination method based on isolated forest and synthetic minority oversampling technique-nominal continuous (SMOTE-NC) algorithm is proposed in data pre-processing phase. The aim of this combination method is to oversample the representative trip fault samples to adjust imbalance rate of dataset based on outlier separation. Secondly, the extreme gradient boosting (XGBoost) classifier model is trained using the optimized data sets to predict the probability of low-voltage tripping faults in the target station area. Finally, the proposed algorithm is validated by the measured data in an area of Guangzhou, and the results show it has good application effects.

Research and Analysis of Short-term Load Forecasting Based on Frequency Domain Decomposition

MA Yuan, ZHANG Qian, LI Guoli, MA Jinhui, DING Jinjin

2020, 53(4):  114-121.  DOI: 10.11930/j.issn.1004-9649.201809089

Asbtract ( )   PDF (4513KB) ( )  

References | Related Articles | Metrics

To study the effect of frequency domain component prediction on the accuracy of short-term load forecasting. This paper proposes to decompose the original load data based on the frequency domain decomposition algorithm,and decompose the data into four parts: daily cycle, weekly cycle, low frequency and high frequency components. Among them, the daily and weekly cycle components are forecasted by Elman neural network; the low-frequency components are forecasted by random forest; the high-frequency components are secondarily decomposed using the Mallat algorithm to obtain the low-frequency part and the high-frequency part respectively. The low-frequency part is selected as the training sample and is combined with the Elman neural network to predict the high-frequency components. The results of each frequency domain component are recombined to achieve high-precision prediction of power load. The actual load data of a certain city in Anhui Province is taken as an example. Compared with the forecasting results of Elman neural network, random forest method and frequency domain component forecasting method, the proposed method can effectively improve the accuracy and reduce the degree of dispersion of predicted and true values.

Coordinated Charging Guiding Strategy for Electric Vehicles in Temporal-spatial Dimension Considering User Satisfaction Degree

JIANG Yijing, YU Aiqing, HUANG Minli

2020, 53(4):  122-130.  DOI: 10.11930/j.issn.1004-9649.201807075

Asbtract ( )   PDF (5983KB) ( )  

References | Related Articles | Metrics

The charging behavior of large-scale electric vehicles (EVs) is random and uncertain in temporal and spatial dimensions. An architecture for coordinated charging guiding model is thus built based on the internet of vehicles system with integration of multi-networks. In order to reflect the diversity of charging time choices for users accessing the system, the EVs are divided into receiving/not receiving system scheduling clusters according to the user’s charging wishes. Then, the satisfaction function under different dimensions is established to mobilize the participation of EV users, and a coordinated charging guidance strategy is proposed on this basis. In temporal level, the load curve is adjusted via directing the charging time of EV users. In spatial level, the selection of charging station is planned for various vehicles. Based on an IEEE 33-node distribution grid with integration of four charging stations, a model simulation is made using the optimization software LINGO11. The simulation results proves that the proposed guidance strategy has good controlling effects, and can improve the load curve of grid and the equipment utilization of charging stations.

Intelligent Charging Navigation for Electric Vehicles Based on Real-time Electricity Price

WU Rui, ZHOU Liangsong, YAO Zhandong

2020, 53(4):  131-138,146.  DOI: 10.11930/j.issn.1004-9649.201904106

Asbtract ( )   PDF (9066KB) ( )  

References | Related Articles | Metrics

A large number of electric vehicles (EVs) connected to the grid will affect the planning and operation of the power system. Hence it is of great significance to guide the charging behavior of electric vehicles in an orderly manner. Regarding this kind of problem, an electric vehicle intelligent charging navigation strategy based on real-time electricity price is proposed, where the PSO-GA algorithm is used to solve the optimal scheme. The motivation is to guide the EV charging toward an orderly way by virtue of the real-time electricity price mechanism. Furthermore, according to the different requirements of users, three charging navigation strategies are designed for different objectives, i.e., the minimization of the user's time cost, expense cost or overall cost respectively. Finally, the IEEE-33 bus system model is established in Matlab. By utilizing the Monte Carlo method, the road condition and EV space-time distribution information is simulated in a real traffic area on the map so as to verify the effectiveness of the method. The simulation results show that the method can reduce the travel cost of EV users and improve the node voltage quality of the power distribution system.

Bi-uncertainty Network Frame Planning Method for Distribution Network with Electric Vehicles

SHEN Yiting, ZHANG Jing, WU Peng, LIU Lu, YANG Jianlin

2020, 53(4):  139-146.  DOI: 10.11930/j.issn.1004-9649.201905117

Asbtract ( )   PDF (5956KB) ( )  

References | Related Articles | Metrics

Considering the stochastic uncertainty of charging load of electric vehicle (EV) and the fuzzy uncertainty of regular load in the distribution system, a hybrid chance-constrained distribution grid planning model including both fuzzy variables and random variables is established on the basis of the theory of uncertain planning. In this paper, the objective function is to minimize the total cost of fixed investment and fuzzy random network loss in the planning period, with the hybrid chance-constraint programming introduced to process the transmission power constraint and node voltage constraint. In addition, the balance can be well maintained between the investment cost and operation risk by setting two different confidence level parameter values. Finally, a genetic algorithm based on hybrid simulation is proposed to solve the model built in this paper, and a 25-node example is used to verify the correctness and validity of the proposed method.

Research on Electric Vehicle Smart Charging Strategy on Carbon Emission Minimization

ZHANG Yu, ZHANG Yue, HAN Xinyang, HAN Bing, LU Shaofeng, XUE Fei

2020, 53(4):  147-154.  DOI: 10.11930/j.issn.1004-9649.201812092

Asbtract ( )   PDF (6164KB) ( )  

References | Related Articles | Metrics

At the beginning, this paper built a generator marginal output power incremental model according to the carbon emission characteristics of the clean energy source and traditional generation devices, geographical location of clean energy source, base load distributions and historical operation data. Then, a carbon emission model based on the scenario of out-of-order electric vehicles charging is built by taking the spatial and temporal driving and battery charging characteristics of electric vehicles into account. After that, in order to take more advantages of clean energy by reducing the marginal carbon emission increase from electric vehicle charging, a smart charging strategy between electric vehicles charging and renewable energy in carbon market is proposed in the model. Finally, carbon emission evaluation indexes are proposed in this paper on the basis of carbon cost calculation model. Results in the case studies show that, the proposed smart charging strategy could effectively reduce the carbon emission.

Safety Supervision Method for Power Operation Site Based on Machine Learning and Image Recognition

CHANG Zhengwei, PENG Qian, CHEN Ying

2020, 53(4):  155-160.  DOI: 10.11930/j.issn.1004-9649.201811057

Asbtract ( )   PDF (2560KB) ( )  

References | Related Articles | Metrics

Aiming at the safety supervision of power operation site, a research is carried out on personnel intrusion detection. It is proposed to use the histogram of oriented gradient (HOG) and support vector machine (SVM) for full frame personnel detection, and then the OpenCV-based image processing technology is used to determine any personnel intrusion into the alert area. Through the images of the work site captured by video monitoring equipment, the above-said methods are used to identify the on-site personnel and their dangerous behaviors in real time, and to issue an alarm signal. The experimental results show that the accuracy of the detecting results reaches 92%, and the automatic safety supervision of power operation site is realized, which greatly improves the safety level of the power operation site.

Design and Application of UAV Intelligent Inspection System for Transmission Lines Based on Cloud and Fog-edge Heterogeneous Collaborative Computing Architecture

HUANG Zheng, WANG Yongqiang, WANG Hongxing, GAO Chao, BAI Cang

2020, 53(4):  161-168.  DOI: 10.11930/j.issn.1004-9649.201907167

Asbtract ( )   PDF (2961KB) ( )  

References | Related Articles | Metrics

To solve the problems of the low computing capacity and low integration degree of the data processing platform of the existing unmanned aerial vehicle (UAV) inspection systems for transmission lines, a new UAV inspection system is proposed based on cloud and fog-edge heterogeneous collaborative computing architecture, and a detail introduction is made to the system architecture, network architecture and operation process. Through integration of the advanced artificial intelligence technology and the specific functions of UAV patrol inspection, an overall UAV inspection solution scheme is given based on the artificial intelligence technology, with such three main functions as the independent path planning, multi-UAV multi-task synergy and intelligent recognition being displayed. The proposed system can comprehensively upgrade the intelligent level and independent running ability of UAVs, and improve the efficiency and quality, of power transmission line inspection and reduce the operational cost.

Performance Analysis of Three Kinds of Integrated Trough Solar Energy Combined Cycle Systems

ZHANG Nan, DUAN Liqiang, DING Zeyu, WANG Jianhua, QU Jie, LUO Na

2020, 53(4):  169-176.  DOI: 10.11930/j.issn.1004-9649.201812088

Asbtract ( )   PDF (5123KB) ( )  

References | Related Articles | Metrics

Integrated solar combined cycle (ISCC) power generation system has been proved to be a high efficient and clean way for power generation, which can reduce not only the cost of solar power generation, but also the fuel consumption of gas steam combined cycle system. In this paper, three kinds of trough ISCC systems with different heat transfer mediums including the heat conducting oil, direct steam generation (DSG) and compressed air are investigated and compared in depth. The performance differences between them are analyzed, especially in the aspects of typical day performance, year performance and economic performance. From the results the ISCC system integrated with DSG has demonstrated the best thermal and economic performance among them.

Energy Efficiency Assessment of Large Capacity Coal-fired Boiler Based on Exergy Analysis

WANG Weiliang, WANG Yuzhao, LYU Junfu, SHAO Wenjun, CHI Zhongjun, ZHANG Ji

2020, 53(4):  177-185.  DOI: 10.11930/j.issn.1004-9649.201812118

Asbtract ( )   PDF (4606KB) ( )  

References | Related Articles | Metrics

In order to study the efficiency in energy conversion and transformation process in coal-fired power generation, and to take advantage of the energy conservation potential, on the basis of the analysis and comparison of the heat balance and exergy balance, the intrinsic relationship between the heat loss and exergy loss is disclosed such that the breakdown of energy conservation potential in the entire process of coal-fired power generation were assessed. Accordingly, the approaches such as increasing the hot air temperature, using oxy combustion, decreasing the excess air coefficient and improving the steam conditions are investigated in depth to evaluate their effects on exergy losses in combustion process, steam heat transfer, air preheating heat transfer, and so on. It is found out that, increasing the hot air temperature, application of oxy combustion, improving the steam temperature and pressure could reduce the net coal consumption by 6~ 9, 6 ~ 10, 11, and 5 ~ 12 g/(kW·h) respectively. These conclusions could provide the guidance for the direction of further development of energy conservation technologies in coal-fired power industry.

Energy-Saving Optimization of Steam Turbine DEH under Speed Control Mode

XU Zhangfu, DENG Tongtian, ZHONG Jingliang, WANG Suobin, ZHANG Shihai

2020, 53(4):  186-192.  DOI: 10.11930/j.issn.1004-9649.201908014

Asbtract ( )   PDF (4813KB) ( )  

References | Related Articles | Metrics

Currently thermal power industry calls for the operation cost reduction, quality and efficiency improvement as well as energy-saving and emission reduction urgently. In this paper, the problems that have been existing in digital electric-hydraulic control system (DEH) speed control mode widely are discussed in terms of long start-up duration, inadequate warming-up process, difficult control over cylinder temperature rise rate, the complex operation, unstable speed and unsuccessful re-connecting to the grid in a short time after load rejection. Starting with the regulating valve and control mechanism of steam turbine, the heating process and dynamic process after load rejection are analyzed. By comparing the pros and cons of the existing techniques, the unreasonable settings of valve position function during the start-up process and the mismatch between DEH speed control mode and bypass valve characteristics after load rejection are concluded as the root causes of the above problems. At last, an optimization method of variable valve position function in steam turbine DEH speed control mode is proposed, which can realize automatic warming-up and fast grid-connection of the unit after load rejection. Therefore it can improve the security of the unit while maintaining its economic performance. In addition, the operators' workload is greatly reduced and their stress at work is relieved.