A bus duct system intelligent control method based on big data

Abstract

The present invention discloses an intelligent control method for a bus duct system based on big data, and belongs to the technical field of bus duct data control. The present invention includes: obtaining data room load change data in a big database, the big database refers to a construction database, storing historical operation data of each data room plugging in a bus duct, and the historical operation data contains each load change of the data room and the control area area of ​​the data room; based on the data room load change data and the control area area of ​​the data room, constructing a prediction growth model for the data room, and analyzing the future current load change trend of the data room; obtaining the current carrying data under the extreme temperature rise value of the bus duct under test conditions, and outputting the rated current of different bus duct types under different extreme temperature rise values; based on the future current load change trend, the control system intelligently selects the bus duct type for recommendation, and feeds back the recommendation information to the administrator port to realize the intelligent control of the bus duct.

Description

Technical Field

[0001] The present invention relates to the technical field of bus duct data control, and in particular to an intelligent control method for a bus duct system based on big data. Background Art

[0002] The main culprits of electrical fire accidents include: cables, wires, high and low voltage equipment, transformers, bus ducts, electrical components, etc.

[0003] Most bus duct accidents are caused by long-term temperature rise and heating, which leads to aging of insulation materials and short circuits, causing fire accidents. The limit temperature rise value is the most critical safety technical parameter of the bus duct. When carrying different rated currents, the temperature rise is completely different. When the limit temperature rise is 70K and 90K respectively, the current carrying capacity of the same bus duct differs by more than 15%. At the same time, when facing the construction of different data centers, considering the addition of power load equipment at a later stage, the selection of the limit temperature rise value of the bus duct often exceeds a certain limit, resulting in increased power loss and accelerated aging of the insulation materials of the bus duct. Summary of the invention

[0004] The purpose of the present invention is to provide a bus duct system intelligent control method based on big data to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a bus duct system intelligent control method based on big data, the method comprising the following steps:

[0006] S1. Obtain data on load changes in a data center from a large database, wherein the large database refers to a construction database, and stores historical operation data on plugging bus ducts in each data center, wherein the historical operation data includes each load change in the data center and the control area of ​​the data center;

[0007] S2. Based on the data center load change data and the control area of ​​the data center, a prediction growth model for the data center is constructed to analyze the future current load change trend of the data center;

[0008] S3. Obtain the current carrying data of the bus duct under the extreme temperature rise value under the test conditions, and output the rated current of different bus duct types under different extreme temperature rise values; the extreme temperature rise test refers to the detection of the temperature rise of the incoming line section, conductor, plug interface conductor, connector, shell, etc. of the bus duct. After running at full load current, the highest temperature that stabilizes is subtracted from the ambient temperature to obtain the temperature rise value. The conductor specifications of other currents are calculated according to the current carrying capacity per square millimeter when the test sample is qualified. The estimated current is preferably smaller than the sample current;

[0009] S4. Based on the future current load change trend, the control system intelligently selects the bus duct type for recommendation and feeds back the recommendation information to the administrator port.

[0010] According to the above technical solution, the construction of the forecast growth model of the data center includes:

[0011] Construct the control area area set of the data room;

[0012] Forming coordinate scatter points;

[0013] Build a data center based on the coordinate scatter points, and use the data center to process the control area of ​​the data room;

[0014] To achieve the predicted growth of future current load changes in data rooms under different control area areas of data rooms.

[0015] According to the above technical solution, forming coordinate scatter points includes:

[0016] The initial current load value and the final current load value of the load change data of the data room of group L are obtained, and the initial current load value is used as the abscissa and the final current load value is used as the ordinate to form coordinate scatter points.

[0017] According to the above technical solution, the construction of a data center based on scattered coordinate points includes:

[0018] Randomly select H data from L groups of coordinate scatter points as the first-level cluster centers, where H is a positive integer greater than 1 and less than L;

[0019] Calculate the difference value between each coordinate scatter point and each first-level cluster center, and assign each coordinate scatter point to the first-level cluster center corresponding to the minimum difference value. The specific calculation includes:

[0020]

[0021] Where F represents the differentiation value; x i ,y i represents the initial value and final value of the current load of the i-th coordinate scatter point; m j 、n j The initial value and final value of the current load representing the jth first-level cluster center; the control area is divided in this way, because in the actual calculation process, the actual control area of ​​the data room is closely related to the change and growth of the current load, so there is an inevitable connection between the maximum current carrying value of the entire data room and the actual control area, and the data in the large database are mixed, and it is inevitable that there are different control areas of the data room, which will lead to different data and cannot be accurately predicted and processed.

[0022] The H data are processed continuously, and a new first-level cluster center is selected in each group. The selection method of the new first-level cluster center includes:

[0023] S2-1. In the same group, calculate the sum of the difference values ​​between any coordinate scatter point and all other coordinate scatter points, record it as the selection value, take the coordinate scatter point corresponding to the minimum selection value as the new first-level cluster center, and regroup based on the original L groups of coordinate scatter points, and calculate the new first-level cluster center again after the allocation, and set a calibration threshold number at the same time. If the number of coordinate scatter points in the formed group is lower than the calibration threshold number, delete the corresponding first-level cluster center and reallocate the coordinate scatter points to other groups;

[0024] S2-2, repeat step S2-1 until the entire function converges, where function convergence means that the data within the group does not change during two adjacent allocation processes; at this time, take the corresponding group and the first-level clustering center within the group for output, and each output first-level clustering center is recorded as a data center.

[0025] According to the above technical solution, the control area processing of the data room using the data center includes:

[0026] The control area of ​​the data room corresponding to the coordinate scatter point in each data center is obtained, and special values ​​are screened out for the control area of ​​the data room in each data center. The special value refers to the difference between the mode of the control area of ​​the data room in the data center and the mode exceeds the system preset value. The control area of ​​the remaining data room after screening out the special value is taken, and the minimum and maximum values ​​are used as the two ends of the range value to form a data interval of the control area area as the output of the corresponding data center. The special value here is to remove some unexpected values ​​and enhance the accuracy of the system.

[0027] Get the output of all data centers and divide them on the linear axis according to the size of the data interval. If there are adjacent data intervals that intersect, divide them by the midpoint of the intersecting part; if there are adjacent intervals that do not intersect, divide them by the midpoint of the blank part that does not intersect, and fill in the adjacent intervals to form a new data interval.

[0028] Finally, all processed data intervals are output.

[0029] According to the above technical solution, based on each data interval, obtaining a forecast growth model of the data center in the same data interval includes:

[0030] Obtain the historical data in the same data interval, each time the load of the data room changes, form a load change data group, each load change data group forms a curve;

[0031] Select any one of the curves as the control group curve, calculate the reference degree between the control group curve and other curves, and calculate the sum of the reference degrees as the selection value of the control group curve, obtain the selection values ​​of all curves, and select the prediction model corresponding to the control group curve with the largest value as the prediction growth model of the data center under the selected data interval.

[0032] According to the above technical solution, the reference degree includes:

[0033] The two groups of curves are regarded as two sequences. Each point on the sequence corresponds to the current load change data. The distance matrix between the points of the two sequences is calculated. Among all the paths from the upper left corner to the lower right corner of the matrix, the element on the path and the path with the minimum corresponding path are recorded as the reference degree.

[0034] Set the distance matrix to G, and the shortest path length from the upper left corner to any point (r, s) to L min (r, s); where r and s refer to any point in the two sequences respectively, then:

[0035] L min (r, s) = d(r, s)

[0036] +min{L min (r-1, s-1), L min (r, s-1), L min (r-1, s)}

[0037] When (r, s) is at the lower right corner of the matrix, L min (r, s) is the reference degree.

[0038] According to the above technical solution, the prediction model corresponding to the control group curve with the largest selection value includes:

[0039] The load change data of the control group curve corresponding to the largest selection value is obtained, and a regression curve is formed by linear regression, which is output as a prediction model.

[0040] According to the above technical solution, the control area of ​​the data room is obtained in real time, the corresponding data interval is selected, and the future current load change trend is output in real time according to the prediction model corresponding to the data interval. The administrator inputs the time period node, and the control system intelligently selects the bus duct type for recommendation based on the time period node, and feeds back the recommended information to the administrator port.

[0041] Compared with the prior art, the beneficial effects achieved by the present invention are: the present invention can realize intelligent safety monitoring of the bus duct, and select the bus duct type based on the limit temperature rise value, so as to ensure that the construction party takes into account the subsequent current compliance development and changes when installing the bus duct, and prevents accidents before they happen. At the same time, when facing the construction of different data rooms, it can quickly add power load equipment, reduce power loss, reduce the workload of replacing the bus duct, and save energy and protect the environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0043] Figure 1 It is a flow chart of a bus duct system intelligent control method based on big data in the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 The present invention provides a technical solution: a bus duct system intelligent control method based on big data, the method comprising the following steps:

[0046] S1. Obtain data on load changes in a data center from a large database, wherein the large database refers to a construction database, and stores historical operation data on plugging bus ducts in each data center, wherein the historical operation data includes each load change in the data center and the control area of ​​the data center;

[0047] S2. Based on the data center load change data and the control area of ​​the data center, a prediction growth model for the data center is constructed to analyze the future current load change trend of the data center;

[0048] S3. Obtain the current carrying data of the bus duct under the extreme temperature rise value under the test conditions, and output the rated current of different bus duct types under different extreme temperature rise values;

[0049] S4. Based on the future current load change trend, the control system intelligently selects the bus duct type for recommendation and feeds back the recommendation information to the administrator port.

[0050] The forecast growth model for building a data center includes:

[0051] Construct the control area area set of the data room;

[0052] Forming coordinate scatter points;

[0053] Build a data center based on the coordinate scatter points, and use the data center to process the control area of ​​the data room;

[0054] To achieve the predicted growth of future current load changes in data rooms under different control area areas of data rooms.

[0055] The forming of coordinate scatter points comprises:

[0056] The initial current load value and the final current load value of the load change data of the data room of group L are obtained, and the initial current load value is used as the abscissa and the final current load value is used as the ordinate to form coordinate scatter points.

[0057] The method of constructing a data center based on coordinate scatter points includes:

[0058] Randomly select H data from L groups of coordinate scatter points as the first-level cluster centers, where H is a positive integer greater than 1 and less than L;

[0059] Calculate the difference value between each coordinate scatter point and each first-level cluster center, and assign each coordinate scatter point to the first-level cluster center corresponding to the minimum difference value. The specific calculation includes:

[0060]

[0061] Where F represents the differentiation value; x i ,y i represents the initial value and final value of the current load of the i-th coordinate scatter point; m j 、n j represents the initial value and final value of the current load of the jth first-level cluster center;

[0062] The H data are processed continuously, and a new first-level cluster center is selected in each group. The selection method of the new first-level cluster center includes:

[0063] S2-1. In the same group, calculate the sum of the difference values ​​between any coordinate scatter point and all other coordinate scatter points, record it as the selection value, take the coordinate scatter point corresponding to the minimum selection value as the new first-level cluster center, and regroup based on the original L groups of coordinate scatter points, and calculate the new first-level cluster center again after the allocation, and set a calibration threshold number at the same time. If the number of coordinate scatter points in the formed group is lower than the calibration threshold number, delete the corresponding first-level cluster center and reallocate the coordinate scatter points to other groups;

[0064] S2-2, repeat step S2-1 until the entire function converges, where function convergence means that the data within the group does not change during two adjacent allocation processes; at this time, take the corresponding group and the first-level clustering center within the group for output, and each output first-level clustering center is recorded as a data center.

[0065] The control area processing of the data room using the data center includes:

[0066] Obtain the control area area of ​​the data room corresponding to the coordinate scatter point in each data center, and screen out special values ​​for the control area area of ​​the data room in each data center, where the difference between the special value and the mode of the control area area of ​​the data room in the data center exceeds the system preset value, and take the control area area of ​​the remaining data room after screening out the special value, with the minimum value and the maximum value as the two ends of the range value, to form a data interval of the control area area as the output of the corresponding data center;

[0067] Get the output of all data centers and divide them on the linear axis according to the size of the data interval. If there are adjacent data intervals that intersect, divide them by the midpoint of the intersecting part; if there are adjacent intervals that do not intersect, divide them by the midpoint of the blank part that does not intersect, and fill in the adjacent intervals to form a new data interval.

[0068] Finally, all processed data intervals are output.

[0069] Based on each data interval, the forecast growth model of the data center in the same data interval is obtained, including:

[0070] Obtain the historical data in the same data interval, each time the load of the data room changes, form a load change data group, each load change data group forms a curve;

[0071] Select any one of the curves as the control group curve, calculate the reference degree between the control group curve and other curves, and calculate the sum of the reference degrees as the selection value of the control group curve, obtain the selection values ​​of all curves, and select the prediction model corresponding to the control group curve with the largest value as the prediction growth model of the data center under the selected data interval.

[0072] The references include:

[0073] The two groups of curves are regarded as two sequences. Each point on the sequence corresponds to the current load change data. The distance matrix between the points of the two sequences is calculated. Among all the paths from the upper left corner to the lower right corner of the matrix, the element on the path and the path with the minimum corresponding path are recorded as the reference degree.

[0074] Set the distance matrix to G, and the shortest path length from the upper left corner to any point (r, s) to Lmin (r, s); where r and s refer to any point in the two sequences respectively, then:

[0075] L min (r, s) = d(r, s)

[0076] +min{L min (r-1, s-1), L min (r, s-1), L min (r-1, s)}

[0077] When (r, s) is at the lower right corner of the matrix, L min (r, s) is the reference degree.

[0078] The prediction model corresponding to the control group curve with the largest selection value includes:

[0079] The load change data of the control group curve corresponding to the largest selection value is obtained, and a regression curve is formed by linear regression, which is output as a prediction model.

[0080] The control area of ​​the data room is obtained in real time, the corresponding data interval is selected, and the future current load change trend is output in real time according to the prediction model corresponding to the data interval. The administrator inputs the time period node, and the control system intelligently selects the bus duct type for recommendation based on the time period node and feeds back the recommendation information to the administrator port.

[0081] In this embodiment:

[0082] Use a large database (i.e., construction database) to obtain data on load changes in the data center;

[0083] The big database stores the historical operation data of each data room plug-in bus duct, and the historical operation data includes each load change of the data room and the control area of ​​the data room;

[0084] Based on the data center load change data and the control area of ​​the data center, a forecast growth model for the data center is constructed to analyze the future current load change trend of the data center;

[0085] The above include:

[0086] Construct the control area area set of the data room;

[0087] Forming coordinate scatter points;

[0088] Build a data center based on the coordinate scatter points, and use the data center to process the control area of ​​the data room;

[0089] To achieve the predicted growth of future current load changes in data rooms under different control area areas of data rooms.

[0090] The forming of coordinate scatter points comprises:

[0091] Get the initial current load value and the final current load value of the load change data of the computer room of group L, use the initial current load value as the horizontal coordinate and the final current load value as the vertical coordinate to form a coordinate scatter point. For example, if the initial current load value is 10 and the final current load value is 100, then the coordinate (10, 100) is formed.

[0092] The method of constructing a data center based on coordinate scatter points includes:

[0093] Randomly select H data from L groups of coordinate scatter points as the first-level cluster centers, where H is a positive integer greater than 1 and less than L;

[0094] Calculate the difference value between each coordinate scatter point and each first-level cluster center, and assign each coordinate scatter point to the first-level cluster center corresponding to the minimum difference value. The specific calculation includes:

[0095]

[0096] Where F represents the differentiation value; x i ,y i represents the initial value and final value of the current load of the i-th coordinate scatter point; m j 、n j represents the initial value and final value of the current load of the jth first-level cluster center;

[0097] For example, take any two coordinate scatter points (20, 120) and (10, 100); then

[0098]

[0099] The H data are processed continuously, and a new first-level cluster center is selected in each group. The selection method of the new first-level cluster center includes:

[0100] S2-1. In the same group, the sum of the difference values ​​between any coordinate scatter point and all other coordinate scatter points is calculated, recorded as the selection value, and the coordinate scatter point corresponding to the minimum selection value is taken as the new first-level cluster center, and regrouped based on the original L groups of coordinate scatter points, and the new first-level cluster center is calculated again after the allocation, and a calibration threshold number is set at the same time. If the number of coordinate scatter points in the formed group is lower than the calibration threshold number, the corresponding first-level cluster center is deleted and the coordinate scatter points are reallocated to other groups; for example, if the sum of the difference values ​​between any coordinate scatter point (10, 100) and other coordinate scatter points in the data group is the smallest, then (10, 100) is selected as the new first-level cluster center, replacing one of the H data randomly specified initially;

[0101] S2-2, repeat step S2-1 until the entire function converges, where function convergence means that the data within the group does not change during two adjacent allocation processes; at this time, take the corresponding group and the first-level clustering center within the group for output, and each output first-level clustering center is recorded as a data center.

[0102] The control area processing of the data room using the data center includes:

[0103] Obtain the control area area of ​​the data room corresponding to the coordinate scatter point in each data center, and screen out special values ​​for the control area area of ​​the data room in each data center, where the difference between the special value and the mode of the control area area of ​​the data room in the data center exceeds the system preset value, and take the control area area of ​​the remaining data room after screening out the special value, with the minimum value and the maximum value as the two ends of the range value, to form a data interval of the control area area as the output of the corresponding data center;

[0104] Get the output of all data centers, divide them on the linear axis according to the size of the data interval. If there are adjacent data intervals that intersect, take the midpoint of the intersecting part for division; if there are adjacent intervals that do not intersect, take the midpoint of the blank part that does not intersect for division, and fill in the adjacent intervals to form new data intervals; for example, if there are two intervals (60, 80) and (70, 90), take the midpoint 75 of the intersecting part (70, 80) for division to form two new intervals (60, 75) and (75, 90). In this application, all data interval endpoints are set to include the front but not the back;

[0105] Finally, all processed data intervals are output.

[0106] Based on each data interval, the forecast growth model of the data center in the same data interval is obtained, including:

[0107] Obtain the historical data in the same data interval, each time the load of the data room changes, form a load change data group, each load change data group forms a curve;

[0108] Select any one of the curves as the control group curve, calculate the reference degree between the control group curve and other curves, and calculate the sum of the reference degrees as the selection value of the control group curve, obtain the selection values ​​of all curves, and select the prediction model corresponding to the control group curve with the largest value as the prediction growth model of the data center under the selected data interval.

[0109] The references include:

[0110] The two groups of curves are regarded as two sequences. Each point on the sequence corresponds to the current load change data. The distance matrix between the points of the two sequences is calculated. Among all the paths from the upper left corner to the lower right corner of the matrix, the element on the path and the path with the minimum corresponding path are recorded as the reference degree.

[0111] Define a cumulative distance, start matching the two sequences from point (0, 0), and when you reach a point, the distances calculated from all previous points are accumulated. When you reach the end point (r, s), this cumulative distance is the final total distance, which is the similarity of the sequences.

[0112] Set the distance matrix to G, and the shortest path length from the upper left corner to any point (r, s) to L min (r, s); where r and s refer to any point in the two sequences respectively, then:

[0113] L min (r, s) = d(r, s)

[0114] +min{L min (r-1, s-1), L min (r, s-1), L min (r-1, s)}

[0115] When (r, s) is at the lower right corner of the matrix, the accumulated distance L min (r, s) is the sum of the Euclidean distance (similarity) between points r and s and the cumulative distance of the smallest neighboring element that can reach the point, so L min (r, s) is the reference degree.

[0116] The prediction model corresponding to the control group curve with the largest selection value includes:

[0117] The load change data of the control group curve corresponding to the largest selection value is obtained, and a regression curve is formed by linear regression, which is output as a prediction model.

[0118] Obtain the current carrying data of the bus duct at the extreme temperature rise value under the test conditions, and output the rated current of different bus duct types at different extreme temperature rise values;

[0119] The control area of ​​the data room is obtained in real time, the corresponding data interval is selected, and the future current load change trend is output in real time according to the prediction model corresponding to the data interval. The administrator inputs the time period node, and the control system intelligently selects the bus duct type for recommendation based on the time period node and feeds back the recommendation information to the administrator port.

[0120] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0121] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A bus duct system intelligent control method based on big data, characterized by: The method comprises the following steps: S1. Obtain data on load changes in a data center from a large database, wherein the large database refers to a construction database, and stores historical operation data on plugging bus ducts in each data center, wherein the historical operation data includes each load change in the data center and the control area of ​​the data center; S2. Based on the load change data of the data room and the control area of ​​the data room, a prediction growth model of the data room is constructed to analyze the future current load change trend of the data room; S3. Obtain the current carrying data of the bus duct under the extreme temperature rise value under the test conditions, and output the rated current of different bus duct types under different extreme temperature rise values; S4. Based on the future current load change trend, the control system intelligently selects the bus duct type for recommendation and feeds back the recommendation information to the administrator port; The control area of ​​the data room is obtained in real time, the corresponding data interval is selected, and the future current load change trend is output in real time according to the prediction model corresponding to the data interval. The administrator inputs the time period node, and the control system intelligently selects the bus duct type for recommendation based on the time period node, and feeds back the recommended information to the administrator port; The forecast growth model for building a data center includes: Construct the control area area set of the data room; Forming coordinate scatter points; Build a data center based on the coordinate scatter points, and use the data center to process the control area of ​​the data room; To achieve the prediction growth of future current load changes in data rooms under different control area of ​​data rooms; The forming of coordinate scatter points comprises: Obtain the current load initial value and the current load final value of the load change data of the data room of group L, use the current load initial value as the abscissa and the current load final value as the ordinate to form coordinate scatter points; The method of constructing a data center based on coordinate scatter points includes: Randomly select H data from L groups of coordinate scatter points as the first-level cluster centers, where H is a positive integer greater than 1 and less than L; Calculate the difference value between each coordinate scatter point and each first-level cluster center, and assign each coordinate scatter point to the first-level cluster center corresponding to the minimum difference value. The specific calculation includes: ； in, represents the differentiated value; , represents the initial value and final value of the current load of the i-th coordinate scatter point; , represents the initial value and final value of the current load of the jth first-level cluster center; The H data are processed continuously, and a new first-level cluster center is selected in each group. The selection method of the new first-level cluster center includes: S2-1. In the same group, calculate the sum of the difference values ​​between any coordinate scatter point and all other coordinate scatter points, record it as the selection value, take the coordinate scatter point corresponding to the minimum selection value as the new first-level cluster center, and regroup based on the original L groups of coordinate scatter points, and calculate the new first-level cluster center again after the allocation, and set a calibration threshold number at the same time. If the number of coordinate scatter points in the formed group is lower than the calibration threshold number, delete the corresponding first-level cluster center and reallocate the coordinate scatter points to other groups; S2-2, repeat step S2-1 until the entire function converges, where function convergence means that the data within the group does not change during two adjacent allocation processes; at this time, take the corresponding group and the first-level clustering center within the group for output, and each output first-level clustering center is recorded as a data center.

2. According to the big data based intelligent control method of bus duct system according to claim 1, it is characterized by: The control area processing of the data room using the data center includes: Obtain the control area area of ​​the data room corresponding to the coordinate scatter point in each data center, and screen out special values ​​for the control area area of ​​the data room in each data center, where the difference between the special value and the mode of the control area area of ​​the data room in the data center exceeds the system preset value, and take the control area area of ​​the remaining data room after screening out the special value, with the minimum value and the maximum value as the two ends of the range value, to form a data interval of the control area area as the output of the corresponding data center; Get the output of all data centers and divide them on the linear axis according to the size of the data interval. If there are adjacent data intervals that intersect, divide them by the midpoint of the intersecting part; if there are adjacent intervals that do not intersect, divide them by the midpoint of the blank part that does not intersect, and fill in the adjacent intervals to form a new data interval. Finally, all processed data intervals are output.

3. According to the big data based intelligent control method of bus duct system according to claim 2, it is characterized by: Based on each data interval, the forecast growth model of the data center in the same data interval is obtained, including: Obtain the historical data in the same data interval, each time the load of the data room changes, form a load change data group, each load change data group forms a curve; Select any one of the curves as the control group curve, calculate the reference degree between the control group curve and other curves, and calculate the sum of the reference degrees as the selection value of the control group curve, obtain the selection values ​​of all curves, and select the prediction model corresponding to the control group curve with the largest value as the prediction growth model of the data center under the selected data interval.

4. According to the big data-based intelligent control method for bus duct system of claim 3, it is characterized by: The references include: The two groups of curves are regarded as two sequences. Each point on the sequence corresponds to the current load change data. The distance matrix between the points of the two sequences is calculated. Among all the paths from the upper left corner to the lower right corner of the matrix, the element on the path and the path with the minimum corresponding path are recorded as the reference degree. Set the distance matrix to G, and the shortest path length from the upper left corner to any point (r, s) to L min (r, s); where r and s refer to any point in the two sequences respectively, then: ； When (r, s) is at the lower right corner of the matrix, That is the reference degree.

5. According to the big data-based intelligent control method for bus duct system of claim 4, it is characterized by: The prediction model corresponding to the control group curve with the largest selection value includes: The load change data of the control group curve corresponding to the largest selection value is obtained, and a regression curve is formed by linear regression, which is output as a prediction model.

Images