CN101752866A - Automatic heavy-load equipment early warning implementation method based on decision tree - Google Patents
Automatic heavy-load equipment early warning implementation method based on decision tree Download PDFInfo
- Publication number
- CN101752866A CN101752866A CN200810204336A CN200810204336A CN101752866A CN 101752866 A CN101752866 A CN 101752866A CN 200810204336 A CN200810204336 A CN 200810204336A CN 200810204336 A CN200810204336 A CN 200810204336A CN 101752866 A CN101752866 A CN 101752866A
- Authority
- CN
- China
- Prior art keywords
- load
- decision tree
- value
- heavy
- early warning
- 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.)
- Pending
Links
Images
Classifications
-
- Y02E40/76—
-
- Y04S10/54—
-
- Y04S10/545—
Abstract
The invention discloses an automatic heavy-load equipment early warning implementation method based on decision tree in automatic monitoring technical field, and provides the early warning decision of the heavy-load equipment; a decision tree is created in each time period of each monitoring group to realize the early warning of the heavy-load equipment, the operation experience and relevant history information are sufficiently utilized to figure out the early warning scheme for the current power grid; when the heavy-load equipment really occurs overload situation, an early warning signal is transmitted in advance, the out-of-limit time is given to the dispatching staff so that the dispatching staff can quickly judge the early warning locations, the early warning property and the severity, determine the affected area, also take correct measures to reduce the affected range, avoid accidents and reduce the possibly-caused load loss. The invention greatly improves the safety of the power supply equipment and also improves the automation degree of the automatic power dispatching system.
Description
Technical field
The present invention relates to a kind of method of automatically-monitored technical field, particularly relate to a kind of automatic heavy-load equipment early warning implementation method based on decision tree.
Background technology
Security and stability and the economy of pursuing power system operation are the main targets of system's operation always.Along with electric system reform deeply and the fast development of power construction, the power transmission and transforming equipment of the transmission line of some big capacity units, long distance and high pressure, ultra high voltage puts into operation in succession, cause some important equipment, for example the service conditions of transformer, circuit is more and more abominable, and higher load factor even overload will be more and more frequent.Operation conditions to these visual plants monitors, has higher requirement, and requires can give warning in advance when overload takes place, and allows the dispatcher have the sufficient time to handle overload equipment, changes disadvantageous operation conditions.
At present, electric power scheduling automatization system generally adopts both at home and abroad: the pattern that definite value-early warning is set.This pattern does not adopt historical load variations information, only consider current load variations situation, there are the following problems: 1) wrong report: in load variations time period faster, even some apparatus of load reaches the definite value of setting, but owing to be varying loading, with very fast falling, will report by mistake according to the definite value that is provided with; 2) early warning is untimely: in the load dull fast rise period, adopt definite value to trigger early warning, it is few to leave dispatcher's processing time for, handling untimely some automatics that will cause moves, occur losing the load phenomenon, serious may lead to a disaster, and causes heavy economic losses.
The traditional decision-tree of data mining is to utilize the information gain in the information theory to seek the attribute field that has the maximum information gain in the illustrative data base, sets up the node of decision tree, sets up the branch of tree again according to the different values of this attribute field; Concentrate at each branch and to repeat to set up next node and the ramifying of tree.
Decision tree is a tree structure that is similar to flow chart, and wherein each internal node is illustrated in a test on the attribute, and each branch is represented a test output, and each leaf nodes is represented class.The top-most node of tree is a root node, and for the sample classification to the unknown, the property value of sample is tested on decision tree.The path by root to leaf node.
Summary of the invention
At the defective that exists in the present prior art, technical problem to be solved by this invention provides a kind of solution dispatching automation of electric power systems to the automatic heavy-load equipment early warning implementation method, can make full use of the electric network data library information, provide the early warning of heave-load device to judge, satisfy the monitoring requirement of operations staff heave-load device.
In order to solve the problems of the technologies described above, a kind of automatic heavy-load equipment early warning implementation method provided by the present invention based on decision tree, the early warning decision of heave-load device is provided, adopts each time period of each supervision group to set up a decision tree to realize the early warning of heave-load device.
The present invention includes following concrete steps:
The first step, set up the supervision group according to the operation of power networks situation, each supervision group is responsible for gathering the information on load of some equipment, and information on load comprises: effective power, electric current, voltage etc., heavily loaded operating value AV, the heavy load starting value SV of each supervision group are set, start resetting value RV;
Wherein, heavily loaded operating value AV (action set-value) is meant after the load heavy duty rate of heavily loaded group of equipment reaches certain value needs the early warning amount; Heavy load starting value SV (start set-value) is meant that the load heavy duty rate of heavily loaded group of equipment reaches the startup value that begins heavily loaded class value is carried out integration behind the certain value; After startup resetting value RV (restore set-value) was meant that heavily loaded group of equipment early warning starts, the heavily loaded rate of loading fell after rise less than certain value, and there is following relation: AV＞SV＞RV in the three.
In second step,, be divided into 4-6 time period according to peak valley is flat according to network load situation time division section;
The 3rd step, set up database, be responsible for data message, markers that storage supervision group collects, retrieve fast for decision tree;
In the 4th step, the data that the first step is obtained according to time sampling uniformly-spaced, comprise following information to each sample: time marking T, heavy duty action sign FA, heavy load starting sign FS, heavily loaded involution sign FR, heavily loaded rate LR in each time period;
Wherein, heavily loaded rate LR (load rate) refers to heave-load device load and heavy duty action ratio extremely; Heavy duty action sign FA (flag action) is that the heavily loaded load heavy duty rate of organizing of true respresentation is greater than heavily loaded operating value AV; Heavy load starting sign FS (flag start) is that the heavily loaded rate of the load of the heavily loaded group of true respresentation is greater than heavy load starting value SV; Heavy duty involution sign FR (flag restore) be the heavily loaded rate of the load of true respresentation heavy duty group greater than heavily loaded resetting value RV, FR and FA, AV mutual exclusion among the three.
The 5th step, the data that obtain according to the 4th step generate following data: during heavy load starting load to the single order load derivative Δ P of time, heavy load starting to heavy duty section operate time load to time integral amount Δ S, current time to heavy duty difference operate time Δ T, wherein, Δ P is non-linear, Δ P, Δ S, Δ T data all are saved in the database of the 3rd step foundation, when decision tree is set up in formation, retrieve necessary sample information with this, determine nodal community and the property value of decision tree;
In the 6th step, Δ P, Δ S are carried out discretization handle;
The 7th step, extract the information of preserving in the 5th step, determine root node, internal node, leaf node, with certain centrifugal pump among the Δ S of discretization as root node, with the Δ p of other Δs S value of discretization, discretization as internal node, with operate time Δ T as leaf node;
The 8th step, most of sample informations according to some days, use greed (ID3) algorithm that the data of decision tree are classified, and correction node attribute values, each individual path is an output of node attribute values, adjust Δ S centrifugal pump, Δ p centrifugal pump repeatedly, make the degree of depth minimum of decision tree, final formation and time period the same number of decision tree;
The 9th step, according to the cutting that sample information is set the decision-making branch of redundancy, finish the quantitative analysis of decision tree nodes property value, obtain the attribute correction value of each internal node, generate optimum decision tree;
The tenth step, according to all the other sample informations, finish the check analysis of decision tree nodes property value, node of decision tree and property value are kept at database, obtain decision tree from database during real time execution, predicting the outcome of decision tree judged as the prediction of equipment operation operating mode, used for the heave-load device early warning.
In the 6th step, describedly Δ P, Δ S are carried out discretization handle, its discrete size of counting equals T * 60/t, and T is the affiliated time period hourage of decision tree, and t is the time that gives warning in advance;
In the 5th step, thus operate time Δ T, if its value is 0, represent that this heavy load starting can not cause early warning action; If its value is non-0 value, represent that this heavy load starting can cause the early warning action, and be Δ T operate time;
In the 8th step, described greed (ID3) algorithm, its with comentropy as the separate targets evaluation function, it constructs decision tree in top-down mode of defeating in detail, find the part in whole spaces, it guarantees that decision tree foundation is the simplest, and each test data of doing is minimum, the decision tree mean depth of ID3 algorithm construction is less, and classification speed is very fast.
In the 9th step, the described cutting that the decision-making branch of redundancy is set is meant and adopts minimum cost complexity beta pruning method, crops redundant decision-making branch.In the construction process of decision tree, many branches may reflect is noise or isolated point in the training data, and it uses statistical measures minimum cost complexity beta pruning method, detects and reject this interfering data, cuts off the least reliable branch.This will cause classifying faster, improve the ability of the correct classification of test data, thereby improve the accuracy of classifying on unknown data.
When the inventive method really is about to transship at heave-load device, to send early warning signal in advance, and provide the out-of-limit time that is about to, make the dispatcher can judge place, early warning character and the order of severity that early warning takes place apace, determine the influence area, and in time adopt right measures and dwindle coverage, avoid accident, reduce the load loss that may cause.
The present invention has realized a kind of method for supervising, the method can make full use of operating experience and history-sensitive information, excavating the early warning scheme that is fit to current electrical network, make the fail safe of power supply unit obtain bigger raising, also is the embodiment that the electric power scheduling automatization system automatization level improves.The present invention simultaneously is based on the data method of decision tree, and its data mining mode is irrelevant with application, has realized the cross-platform application of system, supports the operating system platform of UNIX, the various family releases of WINDOWS.
Compare present early warning system, following advantage arranged based on the automatic heavy-load equipment early warning method of decision tree:
1) do not need to analyze concrete operational mode, the early warning Knowledge Source is in the data of history;
2) can set the time that gives warning in advance;
3) can correctly discern early warning and load normal variation, not report by mistake;
4) adopt the method for decision tree when optimizing, directly to browse decision tree, optimize sooner, finally can find the solution of an optimum or suboptimum.
Description of drawings
Fig. 1 is the flow chart of the embodiment of the invention based on the automatic heavy-load equipment early warning method of decision tree;
Fig. 2 is the decision tree growth schematic diagram of a supervision group of the embodiment of the invention;
Fig. 3 is heavily loaded operating value AV of the embodiment of the invention, heavy load starting value SV, starts resetting value RV and concerns schematic diagram.
Embodiment
Below in conjunction with description of drawings embodiments of the invention are described in further detail, but present embodiment is not limited to the present invention, every employing similarity method of the present invention and similar variation thereof all should be listed protection scope of the present invention in.
As shown in Figure 1, present embodiment comprises following concrete steps:
The first step is set up the supervision group, and heavily loaded operating value AV, the heavy load starting value SV of each supervision group is set, and starts resetting value RV;
Wherein, heavily loaded operating value AV (action set-value) is meant after the load heavy duty rate of heavily loaded group of equipment reaches certain value needs the early warning amount; Heavy load starting value SV (start set-value) is meant that the load heavy duty rate of heavily loaded group of equipment reaches the startup value that begins heavily loaded class value is carried out integration behind the certain value; After starting resetting value RV (restore set-value) and being meant that heavily loaded group of equipment early warning starts, the heavily loaded rate of loading falls after rise less than certain value, the relation between the three as shown in Figure 3, that is: AV＞SV＞RV.
Second step being divided into 4 time periods according to peak valley is flat, and 0～7 is that time period 1,7～11 are that time period 2,11～13 are that time period 3,13～17 are that incorporate into the time period 1 time period 4,17～24, is divided into 4 decision trees;
The 3rd step, set up database, be responsible for data message, markers that storage supervision group collects, retrieve fast for decision tree;
The 4th step, data to first step acquisition, in each time period according to the sampling of 3 seconds time uniformly-spaced, there are 28800 samples every day, each sample is comprised following information: time marking T, heavy duty action sign FA, heavy load starting sign FS, heavily loaded involution sign FR, heavily loaded rate LR, each supervision group is preserved 144000 data messages every day.
Wherein, heavily loaded rate LR (load rate) refers to heave-load device load and heavy duty action ratio extremely; Heavy duty action sign FA (flag action) is that the heavily loaded load heavy duty rate of organizing of true respresentation is greater than heavily loaded operating value AV; Heavy load starting sign FS (flag start) is that the heavily loaded rate of the load of the heavily loaded group of true respresentation is greater than heavy load starting value SV; Heavy duty involution sign FR (flag restore) be the heavily loaded rate of the load of true respresentation heavy duty group greater than heavily loaded resetting value RV, FR and FA, AV mutual exclusion among the three.
The 5th step, the data that obtain according to the 4th step generate following data: during heavy load starting load to the single order load derivative Δ P of time, heavy load starting to heavy duty section operate time load to time integral amount Δ S, current time to heavy duty difference operate time Δ T, wherein, Δ P is non-linear, Δ P, Δ S, Δ T data all are saved in the database of the 3rd step foundation, when decision tree is set up in formation, retrieve necessary sample information with this, determine nodal community and the property value of decision tree;
The 6th step, Δ P, Δ S to be carried out discretization handle, its discrete size of counting equals T * 60/t, T is the affiliated time period hourage of decision tree, and t is the time that gives warning in advance, to the T=7 of time period 1, the time t=5 that gives warning in advance minute, then discrete counting was 7 * 60/5=84;
The 7th step, extract the information of preserving in the 5th step, determine root node, internal node, leaf node, with certain centrifugal pump among the Δ S of discretization as root node, with the Δ p of other Δs S value of discretization, discretization as all the other nodes, with operate time Δ T as leaf node;
The 8th step, according to n days n * 28800 * 80% sample, use greed (ID3) algorithm that the data of decision tree are classified, and correction node attribute values, each individual path is an output of node attribute values, adjust Δ S centrifugal pump, Δ p centrifugal pump repeatedly, make the degree of depth minimum of decision tree, final formation and time period the same number of decision tree;
As shown in Figure 2, be the decision tree growth schematic diagram of a supervision group in the present embodiment
Wherein: node 1 is a Δ S centrifugal pump 1,
Node 21: Δ S centrifugal pump 2 (the conjunction rule: greater than)
Node 22: Δ S centrifugal pump 3 (the conjunction rule: less than)
Node 31: Δ P centrifugal pump 1 (the conjunction rule: greater than)
Node 32: Δ P centrifugal pump 2 (the conjunction rule: less than)
Node 41: Δ P centrifugal pump 3 (the conjunction rule: greater than)
Node 42: Δ P centrifugal pump 4 (the conjunction rule: less than)
Leaf node ● Δ T1: non-0, will transship behind the Δ T1, need early warning
Leaf node zero Δ T0:0, early warning need not early warning with involution after starting
The 9th step, according to the cutting that sample information is set the decision-making branch of redundancy, finish the quantitative analysis of decision tree nodes property value, obtain the attribute correction value of each internal node, generate optimum decision tree;
The tenth step, according to all the other n * 28800 * 20% sample informations, finish the check analysis of decision tree nodes property value, node of decision tree and property value are kept at database, obtain decision tree from database during real time execution, predicting the outcome of decision tree judged as the prediction of equipment operation operating mode, used for the heave-load device early warning.
Claims (5)
1. the automatic heavy-load equipment early warning implementation method based on decision tree is characterized in that, comprises the steps:
The first step, set up the supervision group according to the operation of power networks situation, each supervision group is responsible for gathering the information on load of some equipment, and information on load comprises: effective power, electric current, voltage, heavily loaded operating value AV, the heavy load starting value SV of each supervision group are set, start resetting value RV;
Wherein, heavily loaded operating value AV is meant after the load heavy duty rate of heavily loaded group of equipment reaches certain value needs the early warning amount; Heavy load starting value SV is meant that the load heavy duty rate of heavily loaded group of equipment reaches the startup value that begins heavily loaded class value is carried out integration behind the certain value; After startup resetting value RV was meant that heavily loaded group of equipment early warning starts, the heavily loaded rate of loading fell after rise less than certain value, and there is following relation: AV＞SV＞RV in the three;
In second step,, be divided into 4-6 time period according to peak valley is flat according to network load situation time division section;
The 3rd step, set up database, be responsible for data message, markers that storage supervision group collects, retrieve fast for decision tree;
The 4th step, the data that the first step is obtained, according to time sampling uniformly-spaced, each sample comprises following information: time marking T, heavy duty action sign FA, heavy load starting sign FS, heavily loaded involution sign FR, heavily loaded rate LR in each time period;
Wherein, heavy duty rate LR is meant heave-load device load and heavy duty action ratio extremely, heavy duty action sign FA is that the heavily loaded load heavy duty rate of organizing of true respresentation is greater than heavily loaded operating value AV, heavy load starting sign FS is that the heavily loaded rate of the load of the heavily loaded group of true respresentation is greater than heavy load starting value SV, heavy duty involution sign FR be the load heavy duty rate of true respresentation heavy duty group greater than heavily loaded resetting value RV, FR and FA, AV mutual exclusion among the three;
The 5th step, the data that obtain according to the 4th step generate following data: during heavy load starting load to the single order load derivative Δ P of time, heavy load starting to heavy duty section operate time load to time integral amount Δ S, current time to heavy duty difference operate time Δ T, wherein, Δ P is non-linear, Δ P, Δ S, Δ T data all are saved in the database of the 3rd step foundation, when decision tree is set up in formation, retrieve necessary sample information with this, determine nodal community and the property value of decision tree;
In the 6th step, Δ P, Δ S are carried out discretization handle;
The 7th step, extract the information of preserving in the 5th step, determine root node, internal node, leaf node, with certain centrifugal pump among the Δ S of discretization as root node, with the Δ p of other Δs S value of discretization, discretization as all the other internal nodes, with operate time Δ T as leaf node;
The 8th step, part sample information according to some days, use greedy algorithm that the data of decision tree are classified, and correction node attribute values, each individual path is an output of node attribute values, adjust Δ S centrifugal pump, Δ p centrifugal pump repeatedly, make the degree of depth minimum of decision tree, final formation and time period the same number of decision tree;
The 9th step, according to the cutting that sample information is set the decision-making branch of redundancy, finish the quantitative analysis of decision tree nodes property value, obtain the attribute correction value of each internal node, generate optimum decision tree;
The tenth step, according to all the other sample informations, finish the check analysis of decision tree nodes property value, node of decision tree and property value are kept at database, obtain decision tree from database during real time execution, predicting the outcome of decision tree judged as the prediction of equipment operation operating mode, used for the heave-load device early warning.
2. the automatic heavy-load equipment early warning implementation method based on decision tree according to claim 1 is characterized in that, in the 6th step, describedly Δ P, Δ S are carried out discretization handle, its discrete size of counting equals T * 60/t, and T is the affiliated time period hourage of decision tree, and t is the time that gives warning in advance.
3. the automatic heavy-load equipment early warning implementation method based on decision tree according to claim 1 is characterized in that, in the 5th step, so operate time Δ T, if its value is 0, represent that this heavy load starting can not cause early warning action; If its value is non-0 value, represent that this heavy load starting can cause the early warning action, and be Δ T operate time.
4. the automatic heavy-load equipment early warning implementation method based on decision tree according to claim 1, it is characterized in that, in the 8th step, described greedy algorithm, it as the separate targets evaluation function, constructs decision tree in top-down mode of defeating in detail with comentropy, finds the part in whole spaces, guarantee that decision tree foundation is the simplest, each test data of doing is minimum.
5. the automatic heavy-load equipment early warning implementation method based on decision tree according to claim 1, it is characterized in that, in the 9th step, the described cutting that the decision-making branch of redundancy is set, be meant the minimum cost complexity beta pruning method that adopts, crop redundant decision-making branch, in the construction process of decision tree, many branches may reflect is noise or isolated point in the training data, it uses statistical measures minimum cost complexity beta pruning method, detection is also rejected this interfering data, cuts off the least reliable branch.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200810204336A CN101752866A (en) | 2008-12-10 | 2008-12-10 | Automatic heavy-load equipment early warning implementation method based on decision tree |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200810204336A CN101752866A (en) | 2008-12-10 | 2008-12-10 | Automatic heavy-load equipment early warning implementation method based on decision tree |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101752866A true CN101752866A (en) | 2010-06-23 |
Family
ID=42479413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200810204336A Pending CN101752866A (en) | 2008-12-10 | 2008-12-10 | Automatic heavy-load equipment early warning implementation method based on decision tree |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101752866A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102902623A (en) * | 2012-09-27 | 2013-01-30 | 华北电力大学 | Implementation method for test optimization of complex system |
| CN103337012A (en) * | 2013-07-08 | 2013-10-02 | 国家电网公司 | Power grid equipment monitoring oriented multi-subject intelligent integrated alarm analysis method |
| CN104376397A (en) * | 2014-10-15 | 2015-02-25 | 中国农业银行股份有限公司 | Data real-time analysis method and device |
| CN104506340A (en) * | 2014-11-21 | 2015-04-08 | 河南中烟工业有限责任公司 | Creation method of decision tree in industrial Ethernet fault diagnosis method |
| CN105024380A (en) * | 2014-04-28 | 2015-11-04 | 苏州华天国科电力科技有限公司 | Energy saving control method for regulating DC voltage of active filter at fixed time |
| CN106642061A (en) * | 2016-12-28 | 2017-05-10 | 湖南坤宇网络科技有限公司 | Early warning method for boiler separate cylinder failures based on decision tree system |
| CN106682833A (en) * | 2016-12-28 | 2017-05-17 | 湖南坤宇网络科技有限公司 | Boiler cotton filter obstruction early warning method based on decision tree system |
| CN106682777A (en) * | 2016-12-28 | 2017-05-17 | 湖南坤宇网络科技有限公司 | Boiler flue gas tube blockage early warning method based on decision tree system |
| CN106709657A (en) * | 2016-12-28 | 2017-05-24 | 湖南坤宇网络科技有限公司 | Decision tree system-based boiler intake pump failure early warning method |
| CN106779234A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of coal-saving apparatus for boiler early warning method for failure based on decision tree system |
| CN106779235A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of boiler feed early warning method for failure based on decision tree system |
| CN106779233A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of heat-insulated fault rupture method for early warning of boiler based on decision tree system |
| CN106779450A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of boiler water softener efficiency lost resin method for early warning based on decision tree system |
| CN106779236A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of boiler drum deformation method for early warning based on decision tree system |
| CN106845689A (en) * | 2016-12-28 | 2017-06-13 | 湖南坤宇网络科技有限公司 | A kind of boiler feed temperature method for early warning based on decision tree system |
| CN106845690A (en) * | 2016-12-28 | 2017-06-13 | 湖南坤宇网络科技有限公司 | A kind of boiler water level method for early warning based on decision tree system |
| CN110824586A (en) * | 2019-10-23 | 2020-02-21 | 上海理工大学 | Rainfall prediction method based on improved decision tree algorithm |
| WO2021165811A1 (en) * | 2020-02-21 | 2021-08-26 | International Business Machines Corporation | Optimal interpretable decision trees using integer programming techniques |
-
2008
- 2008-12-10 CN CN200810204336A patent/CN101752866A/en active Pending
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102902623A (en) * | 2012-09-27 | 2013-01-30 | 华北电力大学 | Implementation method for test optimization of complex system |
| CN103337012A (en) * | 2013-07-08 | 2013-10-02 | 国家电网公司 | Power grid equipment monitoring oriented multi-subject intelligent integrated alarm analysis method |
| CN103337012B (en) * | 2013-07-08 | 2016-04-13 | 国家电网公司 | Towards the multi-threaded intelligent comprehensive alert analysis method of grid equipment monitoring |
| CN105024380A (en) * | 2014-04-28 | 2015-11-04 | 苏州华天国科电力科技有限公司 | Energy saving control method for regulating DC voltage of active filter at fixed time |
| CN104376397A (en) * | 2014-10-15 | 2015-02-25 | 中国农业银行股份有限公司 | Data real-time analysis method and device |
| CN104506340A (en) * | 2014-11-21 | 2015-04-08 | 河南中烟工业有限责任公司 | Creation method of decision tree in industrial Ethernet fault diagnosis method |
| CN106779234A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of coal-saving apparatus for boiler early warning method for failure based on decision tree system |
| CN106779450A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of boiler water softener efficiency lost resin method for early warning based on decision tree system |
| CN106682777A (en) * | 2016-12-28 | 2017-05-17 | 湖南坤宇网络科技有限公司 | Boiler flue gas tube blockage early warning method based on decision tree system |
| CN106709657A (en) * | 2016-12-28 | 2017-05-24 | 湖南坤宇网络科技有限公司 | Decision tree system-based boiler intake pump failure early warning method |
| CN106642061A (en) * | 2016-12-28 | 2017-05-10 | 湖南坤宇网络科技有限公司 | Early warning method for boiler separate cylinder failures based on decision tree system |
| CN106779235A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of boiler feed early warning method for failure based on decision tree system |
| CN106779233A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of heat-insulated fault rupture method for early warning of boiler based on decision tree system |
| CN106682833A (en) * | 2016-12-28 | 2017-05-17 | 湖南坤宇网络科技有限公司 | Boiler cotton filter obstruction early warning method based on decision tree system |
| CN106779236A (en) * | 2016-12-28 | 2017-05-31 | 湖南坤宇网络科技有限公司 | A kind of boiler drum deformation method for early warning based on decision tree system |
| CN106845689A (en) * | 2016-12-28 | 2017-06-13 | 湖南坤宇网络科技有限公司 | A kind of boiler feed temperature method for early warning based on decision tree system |
| CN106845690A (en) * | 2016-12-28 | 2017-06-13 | 湖南坤宇网络科技有限公司 | A kind of boiler water level method for early warning based on decision tree system |
| CN110824586A (en) * | 2019-10-23 | 2020-02-21 | 上海理工大学 | Rainfall prediction method based on improved decision tree algorithm |
| WO2021165811A1 (en) * | 2020-02-21 | 2021-08-26 | International Business Machines Corporation | Optimal interpretable decision trees using integer programming techniques |
| GB2608322A (en) * | 2020-02-21 | 2022-12-28 | Ibm | Optimal interpretable decision trees using integer programming techniques |
| US11676039B2 (en) | 2020-02-21 | 2023-06-13 | International Business Machines Corporation | Optimal interpretable decision trees using integer linear programming techniques |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101752866A (en) | Automatic heavy-load equipment early warning implementation method based on decision tree | |
| CN110687473B (en) | Fault positioning method and system for relay protection test of intelligent substation | |
| CN106446016A (en) | Parallel association rule mining-based method for predicting running reliability of power distribution network | |
| CN105245185A (en) | Regional distributed photovoltaic fault diagnosis system and method for access power distribution network | |
| CN109885021A (en) | A kind of method and system that acquisition equipment is monitored based on real-time data collection | |
| CN114793018A (en) | Electrical intelligent data processing device for offshore power grid | |
| CN104616092A (en) | Distributed log analysis based distributed mode handling method | |
| CN114295940A (en) | Distribution network fault state monitoring system and method based on smart city | |
| CN108710099A (en) | Capacitance type potential transformer monitoring alarm method and system | |
| CN101609987B (en) | Joint coordination method for master-station automatic voltage control and plant-station voltage reactive power control | |
| CN110866616A (en) | Fault early warning method and device for secondary equipment of transformer substation | |
| CN110556920A (en) | Distribution automation monitoring method, system, terminal and storage medium | |
| CN111669123A (en) | Method and device for fault diagnosis of photovoltaic string | |
| CN107453354A (en) | A kind of weak link recognition methods of power distribution network | |
| CN103675354A (en) | Anemometer fault detection method and system | |
| CN113806420A (en) | Power grid data monitoring method and device | |
| CN108764544A (en) | Circuit hidden danger prediction technique and device | |
| CN111830342A (en) | Multi-system-fused intelligent substation relay protection dynamic scheduled inspection method | |
| CN108596450B (en) | Power grid risk early warning method and system | |
| CN110070266B (en) | Method and system for identifying major event of power grid cascading failure | |
| CN110502257B (en) | Transformer substation monitoring information management, control and checking method | |
| CN104680010A (en) | Method for screening steady-state operation data of turbine unit | |
| CN110908956A (en) | Information protection main station system and fault information filing method thereof | |
| CN112737120B (en) | Regional power grid control report generation method and device and computer equipment | |
| CN105005939A (en) | Point to point fold ratio method based grid jump operation data discrimination and correction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100623 |