[go: up one dir, main page]

CN105974818B - A data mining method for cloud-visualized machine-network coordinated control response characteristics - Google Patents

A data mining method for cloud-visualized machine-network coordinated control response characteristics Download PDF

Info

Publication number
CN105974818B
CN105974818B CN201610286711.9A CN201610286711A CN105974818B CN 105974818 B CN105974818 B CN 105974818B CN 201610286711 A CN201610286711 A CN 201610286711A CN 105974818 B CN105974818 B CN 105974818B
Authority
CN
China
Prior art keywords
data
machine
temperature
thermal performance
control
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201610286711.9A
Other languages
Chinese (zh)
Other versions
CN105974818A (en
Inventor
马瑞
范辉
彭钢
侯倩
徐欣航
金飞
高志存
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Hebei Energy Technology Service Co Ltd
Original Assignee
State Grid Hebei Energy Technology Service Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by State Grid Hebei Energy Technology Service Co Ltd filed Critical State Grid Hebei Energy Technology Service Co Ltd
Priority to CN201610286711.9A priority Critical patent/CN105974818B/en
Publication of CN105974818A publication Critical patent/CN105974818A/en
Application granted granted Critical
Publication of CN105974818B publication Critical patent/CN105974818B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B17/00—Systems involving the use of models or simulators of said systems
    • G05B17/02—Systems involving the use of models or simulators of said systems electric

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a kind of clouds to visualize machine net coordinated control response characteristic data digging method, belong to the field of cloud calculation of smart grid, on the basis of its online thermal performance data check processing analysis platform, the thermal performance data of acquisition unit and the verification that standardizes to it in real time, it establishes mechanism simulation model and thermal performance precision checking is carried out to it, sample is obtained from optimal mechanism simulation model, establish machine net coordinated control response characteristic equation, it determines raw sample data matrix and normalization is standardized to each value therein, matrix after solving standardization normalization, after seeking the non-negative characteristic value of the matrix and arranging from big to small, it asks and meets the principal component number that cumulative proportion in ANOVA is greater than 80%, original principal component is substituted with the principal component for meeting above-mentioned condition, principal component scatter plot of data is drawn as base Plinth, according to principal component characteristic value correlation degree rendering enginer net coordinated control response characteristic data backbone degree Visualization Model figure.

Description

A kind of cloud visualization machine net coordinated control response characteristic data digging method
Technical field
The invention belongs to the field of cloud calculation of smart grid, and in particular to a kind of cloud visualization machine net coordinated control response is special Property data digging method.
Background technique
The first, the operation data o'clock of fired power generating unit DCS system real-time control is at 10000 points or more, for all the time all In the real time data of generation, have been able to reach mass data rank.In the control process of conventional, the real-time number of magnanimity According to or according to traditional approach and modern control theory, being used in the relationship map of 2D plane to operation picture Wherein limited one group of key operation data point is supplied to operations staff and DCS as the input and output of operation and monitoring scheme Unit is reasonably controlled；Traditional method can not provide the more abundant three-dimensional stereo data interactive relationship of operations staff, It can only be according to production process micro-judgment.
The second, fired power generating unit process control is the Process Control System of a complicated multivariable, crucial operation number According to, for example Stream temperature and main vapour pressure can just occur to change accordingly often after an obvious lag time, And affected to practical control performance, it is especially true to the unit of circulating fluidized bed boiler.It after all, is due to controlled pair As caused by self-information deficiency；And traditional method, it is all based on thermodynamic equilibrium relationship or transmission function relationship obtains Analysis model is that can not find profound data relationship based on priori knowledge and local knowledge；
Third, fired power generating unit information are complicated, and technical threshold is high, before dispatching of power netwoks personnel can only by dispatch command come pair Power plant carries out empirical operation, the inside concrete condition of generating set can not be understood in depth, to make to power grid overall performance Optimal scheduling decision, net side and source lack intuitive effective communication mechanism.
Summary of the invention
Technical problem to be solved by the invention is to provide one kind to meet the actual production of unit net source coordination, is able to reflect The cloud of power grid optimal scheduling decision visualizes machine net coordinated control response characteristic data digging method.
To solve the above problems, the technical solution used in the present invention is:
A kind of cloud visualization machine net coordinated control response characteristic data digging method, key technology are using following step It is rapid:
Step 1, online thermal performance data check handle analysis platform on the basis of, in real time acquire unit thermal performance Data simultaneously carry out standardization verification to the thermal performance data of acquired unit；
Step 2, realization mechanism Building of Simulation Model:
It is divided into different functional groups in unit modeling process, submodel, the son are established to each functional group Model passes through model combination after building up, and builds entire unit model；
Step 3 carries out thermal performance precision checking to mechanism simulation model:
By the thermal performance data input of the online thermal performance data check processing collected real-time unit of analysis platform Thermal performance index calculated value is calculated by mechanism simulation model, to heating power in mechanism model simulation model in step 2 Energy index calculated value and optimal thermal performance curve acquire deviation A, to the thermal performance index inputted in mechanism simulation model Ideal value and optimal thermal performance curve acquire deviation B, are compared, sentence for deviation A corresponding to same power and deviation B Break and use thermal performance index corresponding to smaller value in deviation A and deviation B, obtains optimal mechanism simulation model；
Step 4, according to machine net coordinated control principle, determine that machine net coordinated control response characteristic is response time ki；Machine net The variation of coordinate responses time is formed by stacking by the caused response rate of change of the independent variation of each Con trolling index, from step 3 Optimal mechanism simulation model in obtain sample, it is following (1.1) to establish following machine net coordinated control response characteristic equation:
Wherein, kiCoordinate the AGC response time for machine net；
xiTo influence the rate of change that machine net coordinates the Con trolling index of AGC response time；
N is the number for acquiring data sample；
ai、bi、ciFor xiSpecific gravity impact factor coefficient；
Step 5, the raw sample data matrix R for determining machine net coordinate responses characteristicn×p, wherein footnote p is influence machine net association Adjust the number of the Con trolling index of AGC response time；To raw sample data matrix Rn×pIn each value be standardized at normalizing The calculation formula (1.2) of reason is as follows, eliminates influence of the data difference magnitude to calculating；
Wherein, xijFor raw sample data matrix Rn×pIn the i-th sample j-th of Con trolling index initial data；
For the value after standardization normalization；
E(xij) it is raw sample data matrix Rn×pIn j-th of Con trolling index raw sample data average value；
For j-th of Con trolling index raw sample data variance；
Raw sample data matrix Rn×pIn each value standardization normalization after obtain matrix R*；
Step 6, on labview platform, use matrix and cluster tool box, solution matrix R*, obtain the coordinated control of machine net Response characteristic y1,y2,...,ypA principal component, while obtaining the non-negative characteristic values of p, by the p non-negative characteristic values from big λ is classified as to float1,λ2,...,λp, while obtain respectively with λ1,λ2,...,λpCorresponding feature vector u1,u2,...,up；
Step 7 utilizes the eigenvalue λ arranged from big to small in step 61,λ2,...,λp, calculate and meet cumulative variance contribution Rate α (k) be greater than 80% corresponding to k value, calculation formula (1.3) is as follows；
After acquiring k value, λ is taken1,λ2,...,λpIn principal component corresponding to preceding k characteristic value replace original p principal component；
It is flat in R language based on step 8, the principal component corresponding to the k characteristic value before machine net coordinated control response characteristic Platform carries out cluster visual analyzing to the simulation model data of step 3, using k-medoids clustering method, according to step 3 Simulation model data, rendering enginer net coordinated control response characteristic principal component scatter plot of data；
Step 9, based on the coordinated control response characteristic principal component scatter plot of data of machine net described in step 8, according to master Composition characteristics value correlation degree rendering enginer net coordinated control response characteristic data backbone degree Visualization Model figure.Further, institute Stating functional group different in step 2 includes air and flue system, main steam system, high pressure cylinder and high collateral line system, Gao Jia and steam bleeding system.
Further, machine net is influenced in the step 4 to coordinate the Con trolling index of AGC response time to include boiler implosion index With turbine control index.
Further, the boiler implosion index includes；The boiler implosion index includes main steam pressure, main steam temperature Degree, pulverizer outlet temperature, flue gas exhaust gas temperature, an air temperature, Secondary Air temperature, combustion chamber draft, reheater inlet temperature, Reheater outlet temperature, economizer inlet temperature, economizer exit temperature, flue gas oxygen content, main feed temperature and the degree of superheat.
Further, the turbine control index includes that first stage pressure, governing stage temperature, #1 high add exhaust steam pressure, #1 Height plus exhaust temperature, #2 high add exhaust steam pressure, #2 high that exhaust temperature, #3 high is added to add exhaust steam pressure, #3 high that exhaust temperature, #4 is added to remove Oxygen device exhaust steam pressure, #4 deaerator exhaust temperature, #5 are low to add exhaust steam pressure, #5 are low to add exhaust temperature, #6 low plus exhaust steam pressure, #6 Low plus exhaust temperature, #7 are low to add exhaust steam pressure, #7 low plus exhaust temperature, condensing water temperature and condensed water vacuum degree.
Further, labview language is used at CSV format, to make all data conversions for R language in the step 9 With.
The present invention the beneficial effects of adopting the technical scheme are that
This method realizes imitative according to the cloud visual numeric simulation of operating unit response net side energy requirement scheduling in netting True technology and emulation platform carry out real-time Data Transmission, and emulation platform is modulated to actual set and power grid characteristic and is approached, is tested It demonstrate,proves advanced optimization algorithm and establishes net source security margin evaluation system, be dispatching of power netwoks service.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, invention is carried out combined with specific embodiments below Clear, complete description.
Embodiment is the net source coordination visual numeric simulation analysis method of progress by taking 600MW supercritical unit as an example, tool Steps are as follows for body:
Step 1, online thermal performance data check handle analysis platform on the basis of, in real time acquire unit thermal performance Data simultaneously carry out standardization verification to the thermal performance data of acquired unit；
Step 2, realization mechanism Building of Simulation Model:
It is divided into different functional groups in unit modeling process, submodel, the son are established to each functional group Model passes through model combination after building up, and builds entire unit model；
Step 3 carries out thermal performance precision checking to mechanism simulation model:
By the thermal performance data input of the online thermal performance data check processing collected real-time unit of analysis platform Thermal performance index calculated value is calculated by mechanism simulation model, to heating power in mechanism model simulation model in step 2 Energy index calculated value and optimal thermal performance curve acquire deviation A, to the thermal performance index inputted in mechanism simulation model Ideal value and optimal thermal performance curve acquire deviation B, are compared, sentence for deviation A corresponding to same power and deviation B Break and use thermal performance index corresponding to smaller value in deviation A and deviation B, obtains optimal mechanism simulation model；
Step 4, according to machine net coordinated control principle, determine that machine net coordinated control response characteristic is response time ki；Machine net The variation of coordinate responses time is formed by stacking by the caused response rate of change of the independent variation of each Con trolling index, from step 3 Optimal mechanism simulation model in obtain sample, it is following (1.1) to establish following machine net coordinated control response characteristic equation:
Wherein, kiCoordinate the AGC response time for machine net；
xiTo influence the rate of change that machine net coordinates the Con trolling index of AGC response time；
N is the number for acquiring data sample；
ai、bi、ciFor xiSpecific gravity impact factor coefficient；
Step 5, the raw sample data matrix R for determining machine net coordinate responses characteristicn×p, wherein footnote p is influence machine net association Adjust the number of the Con trolling index of AGC response time；To raw sample data matrix Rn×pIn each value be standardized at normalizing The calculation formula (1.2) of reason is as follows, eliminates influence of the data difference magnitude to calculating；
Wherein, xijFor raw sample data matrix Rn×pIn the i-th sample j-th of Con trolling index initial data；
For the value after standardization normalization；
E(xij) it is raw sample data matrix Rn×pIn j-th of Con trolling index raw sample data average value；
For j-th of Con trolling index raw sample data variance；
Raw sample data matrix Rn×pIn each value standardization normalization after obtain matrix R*；
Step 6, on labview platform, use matrix and cluster tool box, solution matrix R*, obtain the coordinated control of machine net Response characteristic y1,y2,...,ypA principal component, while obtaining the non-negative characteristic values of p, by the p non-negative characteristic values from big λ is classified as to float1,λ2,...,λp, while obtain respectively with λ1,λ2,...,λpCorresponding feature vector u1,u2,...,up；
Step 7 utilizes the eigenvalue λ arranged from big to small in step 61,λ2,...,λp, calculate and meet cumulative variance contribution Rate α (k) be greater than 80% corresponding to k value, calculation formula (1.3) is as follows；
After acquiring k value, λ is taken1,λ2,...,λpIn principal component corresponding to preceding k characteristic value replace original p principal component；
It is flat in R language based on step 8, the principal component corresponding to the k characteristic value before machine net coordinated control response characteristic Platform carries out cluster visual analyzing to the simulation model data of step 3, using k-medoids clustering method, according to step 3 Simulation model data, rendering enginer net coordinated control response characteristic principal component scatter plot of data；
Step 9, based on the coordinated control response characteristic principal component scatter plot of data of machine net described in step 8, according to master Composition characteristics value correlation degree rendering enginer net coordinated control response characteristic data backbone degree Visualization Model figure.Further, institute Stating functional group different in step 2 includes air and flue system, main steam system, high pressure cylinder and high collateral line system, Gao Jia and steam bleeding system.
Further, machine net is influenced in the step 4 to coordinate the Con trolling index of AGC response time to include boiler implosion index With turbine control index.
Further, the boiler implosion index includes；The boiler implosion index includes main steam pressure, main steam temperature Degree, pulverizer outlet temperature, flue gas exhaust gas temperature, an air temperature, Secondary Air temperature, combustion chamber draft, reheater inlet temperature, Reheater outlet temperature, economizer inlet temperature, economizer exit temperature, flue gas oxygen content, main feed temperature and the degree of superheat.
Further, the turbine control index includes that first stage pressure, governing stage temperature, #1 high add exhaust steam pressure, #1 Height plus exhaust temperature, #2 high add exhaust steam pressure, #2 high that exhaust temperature, #3 high is added to add exhaust steam pressure, #3 high that exhaust temperature, #4 is added to remove Oxygen device exhaust steam pressure, #4 deaerator exhaust temperature, #5 are low to add exhaust steam pressure, #5 are low to add exhaust temperature, #6 low plus exhaust steam pressure, #6 Low plus exhaust temperature, #7 are low to add exhaust steam pressure, #7 low plus exhaust temperature, condensing water temperature and condensed water vacuum degree.
Further, labview language is used at CSV format, to make all data conversions for R language in the step 9 With.
Online unit data and simulation model are docked, the emulation that actual set characteristic similarity meets required precision is established The specific method is as follows for model:
If online unit data are divided into dry cooling condition prerequisite according to operating condition, static number is run first, in accordance with any operating condition Design of Simulation verification is carried out according to unit, so that the design conditions of simulation model meet the required precision of actual production test, than It is only required such as original simulation model and meets simulation operations precision prescribed in 50% and 100% load condition, then in order to full The requirement of sufficient l-G simulation test just further segments replicating machine operating condition, is gear with 10%, be respectively set 10%, 30%, 50%, the load conditions condition such as 70%, 100% needs for unit online data to be modified simulation model, improves model Accuracy and availability；Secondly, being directed to the dynamic characteristic of the simulation model, exported using online unit data and simulation model The deviation of data, the optimal simulation parameters numerical value of iteration optimization complete actual set characteristic similarity and meet required precision Simulation model.
It is pretreated to data progress that the specific method is as follows:
According to the relationship of net source energy balance, all original variables in relation to net source coordination data of simulation model are carried out Standardization；Input raw data matrix or correlation matrix arrange in principal () and fa () function, are calculating Before ensure there is no missing values in data.
On Xinhua's OC6000E emulation platform, the 600MW supercritical thermal power unit that is identical based on actual set characteristic Simulation model counts the magnanimity that net source coordination control loop generates using R language platform for net source coordination control loop in real time According to principal component analysis and k-medoids clustering is carried out, it is finally established based on the net source coordination data cloud visualization of backbone degree Depth data method for digging.
The Executive Meeting of the State Council held on June 24 in 2015 has passed through " " internet+" action instruction " and (has called in the following text " opinion ")." internet+" this new industry mode formally becomes the national plan of action of China.On July 1st, 2015 state affairs Institute's publication clearly proposes " internet+" wisdom energy " about the instruction for actively pushing forward " internet+" action ", here+ Be the wisdom energy, rather than be only limited to existing smart grid, the existing concept in energy internet, there is more profound meaning, Contain the core competitiveness of China's energy innovation.Therefore carry out cloud computing technology research in intelligent power generation technical field to have carved Do not allow to delay, the data of current system can be provided in the case where guaranteeing that existing electric system hardware infrastructure is basically unchanged Source and processor resource are integrated, and net interior unit to the energy of extra-high voltage grid real-time response and advanced analysis to greatly improve Power provides effective support for the development of intelligent power grid technology.
The operation data o'clock of fired power generating unit DCS system real-time control is at 10000 points or more, for all generating all the time Real time data, have been able to reach mass data rank.Traditional method can not provide the more abundant three-dimensional of operations staff Stereo data interactive relationship, can only be according to production process micro-judgment；Fired power generating unit process control is a complicated multivariable Process Control System, traditional method is to be with priori knowledge and local knowledge since controlled device self-information is insufficient Basis, it can not find profound data relationship；Power plant can only be carried out by dispatch command before dispatching of power netwoks personnel Empirical operation, can not understand the inside concrete condition of generating set in depth, to make the tune optimal to power grid overall performance Decision is spent, net side and source lack intuitive effective communication mechanism.
This, in Southern Hebei Network, the actual number of Southern Hebei Network main force generating set is realized using achievement and method at fruit product It is docked according to scheduling simulation model, it is organic closely to connect production reality with scientific research, in combination with when operation of power networks It reliability and solves the problems, such as the various preliminary analysis encountered, and has carried out the optimisation strategy architectural study of seriation.Utilize net source Coordination data backbone degree visual analysis method establishes Hebei South Power Network net source energy balance emulation dispatch platform, formulates Electricity net safety stable nargin appraisement system.It is finally reached to produce the purpose that bring benefit pushes science and technology research and development, further The hatching for promoting scientific achievement opens the research and development mode of the benign cycle of " establishing production with scientific research, promote scientific research to produce ". By integrate Advanced Control Strategies iteration optimization work, reduce debug time and increase test safety, while achieve compared with Good economic benefit.
Net source coordination data backbone degree visual analysis method, is improved to original unit emulation platform can satisfy electricity The test simulation platform of net scheduling verifying, establishes net source coordination simulation model for units all in Hebei South Power Network, makes Emulation platform becomes the technical platform that can be run with real service in dispatching of power netwoks；
Increase net source coordination data backbone degree visual analysis method to dispatching of power netwoks using the supercritical unit based on R language After test divides boundary progress simulation modeling assessment, interior unit responsive electricity grid scheduling Mean Speed will be netted and improve 7%, reduce machine Group operating cost 5%,；Electricity net safety stable nargin improves 11%, reduces 81% to unit and power grid impact amplitude.
Certainly, those skilled in the art in the art are it should be appreciated that above-described embodiment is intended merely to illustrate this hair It is bright, and be not used as limitation of the invention, if in spirit of the invention, to the variation of above-described embodiment, Modification etc. will all be fallen within the scope of the claims.

Claims (6)

1.一种云可视化机网协调控制响应特性数据挖掘方法，其特征在于：其包括步骤如下：1. a cloud visualization machine network coordination control response characteristic data mining method, is characterized in that: it comprises the steps as follows: 步骤1、在在线热力性能数据校验处理分析平台基础上，实时采集机组的热力性能数据并对所采集机组的热力性能数据进行规范化校验；Step 1. On the basis of the online thermal performance data verification processing and analysis platform, collect the thermal performance data of the unit in real time and conduct standardized verification of the collected thermal performance data of the unit; 步骤2、实现机理仿真模型建立：Step 2. Realize the establishment of the mechanism simulation model: 在机组建模过程中将其划分为不同的功能组，对每个功能组建立子模型，所述子模型建好后通过模型合并，搭建出整个机组模型；In the process of unit modeling, it is divided into different functional groups, and a sub-model is established for each functional group. After the sub-model is built, the model is merged to build the entire unit model; 步骤3、对机理仿真模型进行热力性能精度校验：Step 3. Verify the thermal performance accuracy of the mechanism simulation model: 将在线热力性能数据校验处理分析平台采集到的实时机组的热力性能数据输入步骤2中的机理模型仿真模型，通过机理仿真模型计算得到热力性能指标计算值，对热力性能指标计算值与最优热力性能曲线求得偏差A，对机理仿真模型中已输入的热力性能指标理想值与最优热力性能曲线求得偏差B，针对同一功率所对应的偏差A和偏差B进行比较，判断并采用偏差A和偏差B中较小值所对应的热力性能指标，得到最优机理仿真模型；Input the real-time unit thermal performance data collected by the online thermal performance data verification processing and analysis platform into the mechanism model simulation model in step 2, and obtain the calculated value of the thermal performance index through the mechanism simulation model. The deviation A is obtained from the thermal performance curve, and the deviation B is obtained from the ideal value of the thermal performance index input in the mechanism simulation model and the optimal thermal performance curve, and the deviation A and the deviation B corresponding to the same power are compared, and the deviation is judged and adopted. The thermal performance index corresponding to the smaller value of A and deviation B is obtained, and the optimal mechanism simulation model is obtained; 步骤4、根据机网协调控制原理，确定机网协调控制响应特性为响应时间ki；机网协调响应时间的变化由各控制指标的独立变化所引起的响应变化速率叠加而成，从步骤3的最优机理仿真模型中获取样本，建立如下机网协调控制响应特性方程如下(1.1)：Step 4. According to the principle of machine-network coordinated control, determine that the response characteristic of machine-network coordinated control is the response time k i ; the change of machine-network coordination response time is formed by superposition of the response rate of change caused by the independent change of each control index, from step 3 The sample is obtained from the optimal mechanism simulation model of , and the response characteristic equation of the following machine-network coordinated control is established as follows (1.1): 其中，ki为机网协调AGC响应时间；Among them, k i is the response time of the machine-network coordination AGC; xi为影响机网协调AGC响应时间的控制指标的变化速率；x i is the rate of change of the control index that affects the response time of the machine-network coordination AGC; n为采集数据样本的个数；n is the number of collected data samples; ai、bi、ci为xi的比重影响因子系数；a i , b i , c i are the weighting factor coefficients of xi ; 步骤5、确定机网协调响应特性的原始样本数据矩阵Rn×p，其中脚标p为影响机网协调AGC响应时间的控制指标的个数；对原始样本数据矩阵Rn×p中的每个值进行标准化归一处理的计算公式(1.2)如下，消除数据不同量级对计算的影响；Step 5. Determine the original sample data matrix R n×p of the machine-network coordination response characteristics, where the subscript p is the number of control indicators that affect the machine-network coordination AGC response time; The calculation formula (1.2) for standardizing and normalizing each value is as follows, to eliminate the influence of different magnitudes of data on the calculation; 其中，xij为原始样本数据矩阵Rn×p中第i样本的第j个控制指标原始数据；Among them, x ij is the original data of the j-th control index of the i-th sample in the original sample data matrix R n×p ; 为标准化归一处理后的值； is the value after normalization and normalization; E(xij)为原始样本数据矩阵Rn×p中第j个控制指标原始样本数据的平均值；E(x ij ) is the average value of the original sample data of the jth control index in the original sample data matrix R n×p ; 为第j个控制指标原始样本数据方差； is the original sample data variance of the jth control indicator; 原始样本数据矩阵Rn×p中的每个值标准化归一处理后得到矩阵R*；Each value in the original sample data matrix R n×p is standardized and normalized to obtain a matrix R * ; 步骤6、在labview平台上，使用矩阵和簇工具箱，求解矩阵R*，获得机网协调控制响应特性y1,y2,...,yp个主成分，同时得到p个非负的特征值，将所述p个非负的特征值从大到小排列为λ1,λ2,...,λp，同时得到分别与λ1,λ2,...,λp对应的特征向量u1,u2,...,up；Step 6. On the labview platform, use the matrix and cluster toolbox to solve the matrix R * to obtain the response characteristics y 1 , y 2 , ..., y p principal components of the machine-network coordinated control, and at the same time obtain p non-negative Eigenvalues, arrange the p non-negative eigenvalues as λ 1 ,λ 2 ,...,λ p in descending order, and obtain the corresponding λ 1 ,λ 2 ,...,λ p respectively. eigenvectors u 1 , u 2 ,...,up ; 步骤7、利用步骤6中从大到小排列的特征值λ1,λ2,...,λp，计算满足累积方差贡献率α(k)大于80％所对应的k值，计算公式(1.3)如下；Step 7. Using the eigenvalues λ 1 , λ 2 ,...,λ p arranged from large to small in step 6, calculate the k value corresponding to the cumulative variance contribution rate α(k) greater than 80%, and the calculation formula ( 1.3) as follows; 求得k值后，取λ1,λ2,...,λp中前k个特征值所对应的主成分代替原来p个主成分；After obtaining the k value, take the principal components corresponding to the first k eigenvalues in λ 1 , λ 2 ,...,λ p to replace the original p principal components; 步骤8、以机网协调控制响应特性前k个特征值所对应的主成分为基础，在R语言平台，对步骤3的仿真模型数据进行聚类可视化分析，采用k-medoids聚类方法，根据步骤3的仿真模型数据，绘制机网协调控制响应特性主成分数据散点图；Step 8. Based on the principal components corresponding to the first k eigenvalues of the machine-network coordinated control response characteristic, on the R language platform, perform a clustering visualization analysis on the simulation model data in step 3, using the k-medoids clustering method, according to For the simulation model data of step 3, draw a scatter plot of principal component data of the response characteristic of the machine-network coordinated control; 步骤9、以步骤8中所述机网协调控制响应特性主成分数据散点图为基础，根据主成分特征值关联程度绘制机网协调控制响应特性数据骨干度可视化模型图。Step 9. Based on the principal component data scatter plot of the response characteristic of the machine-network coordinated control described in step 8, draw a visual model diagram of the backbone degree of the machine-network coordinated control response characteristic data according to the correlation degree of the eigenvalues of the principal components. 2.根据权利要求1所述一种云可视化机网协调控制响应特性数据挖掘方法，其特征在于：所述步骤2中不同的功能组包括风烟系统、主蒸汽系统、高压缸及高旁系统、高加及抽汽系统。2. A method for data mining of cloud visualization machine-network coordination control response characteristics according to claim 1, characterized in that: in the step 2, different functional groups include a wind smoke system, a main steam system, a high-pressure cylinder and a high-bypass system , High pressure and extraction system. 3.根据权利要求1所述一种云可视化机网协调控制响应特性数据挖掘方法，其特征在于：所述步骤4中影响机网协调AGC响应时间的控制指标包括锅炉控制指标和汽机控制指标。3 . The method for data mining of cloud visualization machine-network coordination control response characteristics according to claim 1 , wherein the control indexes that affect machine-network coordination AGC response time in the step 4 include boiler control indexes and turbine control indexes. 4 . 4.根据权利要求3所述一种云可视化机网协调控制响应特性数据挖掘方法，其特征在于：所述锅炉控制指标包括主蒸汽压力、主蒸汽温度、磨煤机出口温度、烟气排烟温度、一次风温度、二次风温度、炉膛负压、再热器入口温度、再热器出口温度、省煤器入口温度、省煤器出口温度、烟气含氧量、主给水温度和过热度。4. The data mining method of cloud visualization machine-network coordination control response characteristic according to claim 3, characterized in that: the boiler control index comprises main steam pressure, main steam temperature, coal mill outlet temperature, flue gas exhaust temperature, primary air temperature, secondary air temperature, furnace negative pressure, reheater inlet temperature, reheater outlet temperature, economizer inlet temperature, economizer outlet temperature, flue gas oxygen content, main feed water temperature and heat. 5.根据权利要求3所述一种云可视化机网协调控制响应特性数据挖掘方法，其特征在于：所述汽机控制指标包括调节级压力、调节级温度、#1高加排汽压力、#1高加排汽温度、#2高加排汽压力、#2高加排汽温度、#3高加排汽压力、#3高加排汽温度、#4除氧器排汽压力、#4除氧器排汽温度、#5低加排汽压力、#5低加排汽温度、#6低加排汽压力、#6低加排汽温度、#7低加排汽压力、#7低加排汽温度、凝结水温度和凝结水真空度。5. The method for data mining of cloud visualization machine-network coordination control response characteristics according to claim 3, wherein the steam turbine control index comprises regulation stage pressure, regulation stage temperature, #1 high pressure plus exhaust steam, #1 High add-exhaust temperature, #2 high add-exhaust pressure, #2 high add-exhaust temperature, #3 high add-exhaust pressure, #3 high add-exhaust temperature, #4 deaerator exhaust pressure, #4 deaerator Oxygen exhaust steam temperature, #5 low add and exhaust steam pressure, #5 low add and exhaust steam temperature, #6 low add and exhaust steam pressure, #6 low add and exhaust steam temperature, #7 low add and exhaust steam pressure, #7 low add and exhaust steam pressure Exhaust steam temperature, condensate temperature and condensate vacuum. 6.根据权利要求3所述一种云可视化机网协调控制响应特性数据挖掘方法，其特征在于：所述步骤9中采用labview语言将所有数据转换成CSV格式，供R语言使用。6. A kind of cloud visualization machine-network coordination control response characteristic data mining method according to claim 3, is characterized in that: in described step 9, adopt labview language to convert all data into CSV format, for R language use.
CN201610286711.9A 2016-05-03 2016-05-03 A data mining method for cloud-visualized machine-network coordinated control response characteristics Active CN105974818B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610286711.9A CN105974818B (en) 2016-05-03 2016-05-03 A data mining method for cloud-visualized machine-network coordinated control response characteristics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610286711.9A CN105974818B (en) 2016-05-03 2016-05-03 A data mining method for cloud-visualized machine-network coordinated control response characteristics

Publications (2)

Publication Number Publication Date
CN105974818A CN105974818A (en) 2016-09-28
CN105974818B true CN105974818B (en) 2019-01-25

Family

ID=56993540

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610286711.9A Active CN105974818B (en) 2016-05-03 2016-05-03 A data mining method for cloud-visualized machine-network coordinated control response characteristics

Country Status (1)

Country Link
CN (1) CN105974818B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109100411A (en) * 2018-06-14 2018-12-28 中国矿业大学 A kind of chemometrics application method for coal Soluble Organic Matter
CN111103789B (en) * 2019-11-20 2023-04-18 国网河北省电力有限公司电力科学研究院 Source network load comprehensive energy scheduling analysis method, system and terminal equipment
CN118777860B (en) * 2024-09-12 2024-12-06 格伊(杭州)电气有限公司 A method and system for detecting circuit faults of low voltage circuit breaker
CN120610472B (en) * 2025-08-12 2025-10-21 国家能源集团江苏电力有限公司 Multi-target energy consumption monitoring and optimizing method and system for secondary reheating unit

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012128609A8 (en) * 2011-03-23 2012-10-18 Petroliam Nasional Berhad (Petronas) An improved fuel blending system and method
EP2645329A1 (en) * 2012-03-27 2013-10-02 Westfälische Wilhelms-Universität Münster Method and system for image segmentation
CN103995985A (en) * 2014-06-10 2014-08-20 华东理工大学 Fault detection method based on Daubechies wavelet transform and elastic network
CN104242306A (en) * 2014-09-12 2014-12-24 清华大学 Power system self-adaption partition method based on principal component analysis method
CN104834938A (en) * 2015-04-30 2015-08-12 北京环境特性研究所 Hyper-spectral information extraction method based on main component and cluster analysis
CN105303468A (en) * 2015-11-20 2016-02-03 国网天津市电力公司 Comprehensive evaluation method of smart power grid construction based on principal component cluster analysis
CN105512388A (en) * 2015-12-04 2016-04-20 河北省电力建设调整试验所 Sub-critical thermal power generating unit enhancing stimulation and simulation modeling method based on LABVIEW

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012128609A8 (en) * 2011-03-23 2012-10-18 Petroliam Nasional Berhad (Petronas) An improved fuel blending system and method
EP2645329A1 (en) * 2012-03-27 2013-10-02 Westfälische Wilhelms-Universität Münster Method and system for image segmentation
CN103995985A (en) * 2014-06-10 2014-08-20 华东理工大学 Fault detection method based on Daubechies wavelet transform and elastic network
CN104242306A (en) * 2014-09-12 2014-12-24 清华大学 Power system self-adaption partition method based on principal component analysis method
CN104834938A (en) * 2015-04-30 2015-08-12 北京环境特性研究所 Hyper-spectral information extraction method based on main component and cluster analysis
CN105303468A (en) * 2015-11-20 2016-02-03 国网天津市电力公司 Comprehensive evaluation method of smart power grid construction based on principal component cluster analysis
CN105512388A (en) * 2015-12-04 2016-04-20 河北省电力建设调整试验所 Sub-critical thermal power generating unit enhancing stimulation and simulation modeling method based on LABVIEW

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
电力系统动态响应的数据挖掘;张鹏飞;《万方学术论文》;20061208;全文

Also Published As

Publication number Publication date
CN105974818A (en) 2016-09-28

Similar Documents

Publication Publication Date Title
Guo et al. A multivariable hybrid prediction system of wind power based on outlier test and innovative multi-objective optimization
CN103440528B (en) Thermal power unit operation optimization method and device based on power consumption analysis
CN105974818B (en) A data mining method for cloud-visualized machine-network coordinated control response characteristics
CN101498459B (en) A Modeling Method for Boiler Combustion Optimization
CN103968367B (en) Based on the method for the general predictive control that Fuzzy Neural PID controls
CN106936627B (en) Thermal power equipment performance monitoring method based on big data analysis and mining
CN103345546B (en) The governor parameter discrimination method that frequency locus combines with particle cluster algorithm
CN103488869A (en) A method for short-term load forecasting of wind power generation based on least squares support vector machine
CN108876163B (en) A fast evaluation method for transient power angle stability based on comprehensive causal analysis and machine learning
CN113722656B (en) A real-time health evaluation method and system for thermal power generating units
CN110262450B (en) A Fault Prediction Method for Cooperative Analysis of Multiple Fault Characteristics for Steam Turbines
CN109272258B (en) Regional wind and solar power generation resource evaluation method based on K-means clustering
CN103886405A (en) Boiler combustion condition identification method based on information entropy characteristics and probability nerve network
CN108667069A (en) A kind of short-term wind power forecast method returned based on Partial Least Squares
CN109471420A (en) Control performance monitoring method of air preheater for large coal-fired generator sets in smart power plants based on CVA-SFA
CN108182553A (en) A kind of coal-fired boiler combustion efficiency On-line Measuring Method
CN111950118B (en) A gas-steam combined power generation unit operation optimization method and equipment
CN106771847B (en) A kind of 35kV distribution network transmission line thunderbolt risk prediction technique
CN111262280A (en) Modeling and analyzing method for primary frequency modulation of pressurized water reactor nuclear power unit
Rui et al. A novel approach for modelling of an injector powered transonic wind tunnel
CN102661974A (en) Thermal-operating-parameters-based pulverized coal heat value on-line identification method
Wu et al. A hybrid least square support vector machine for boiler efficiency prediction
CN113379093B (en) Energy consumption analysis and optimization method for oil-gas gathering and transportation system
CN105158007B (en) A kind of coal consumption sequence test method based on leakage vapour rate of passing a bridge
CN117060390A (en) An evaluation method for online static voltage stability margin of power systems

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
CB02 Change of applicant information

Address after: 050091 Hebei high tech District of Shijiazhuang City Xinshi North Road, No. 368

Applicant after: National network Hebei Energy Technology Service Co., Ltd.

Address before: 050091 Hebei high tech District of Shijiazhuang City Xinshi North Road, No. 368

Applicant before: Hebei Electric Power Construction & Adjustment Research Institute

CB02 Change of applicant information
GR01 Patent grant
GR01 Patent grant