An energy management platform-based device state data analysis system and method

Abstract

The present application relates to the technical field of device state data management, in particular to a device state data analysis system and method based on an energy management platform, which comprises collecting historical maintenance records of a target device and historical data and real-time monitoring data of sensors, performing similarity comparison on the real-time monitoring data, judging whether the real-time monitoring data is similar to the data monitored before failure, predicting whether the device will fail, and according to the similarity of the data, combining different early warning thresholds to expand the early warning grading judgment of each real-time monitoring data. The judgment data generated in each early warning grading judgment process is combined with the early warning data generated in the corresponding real-time monitoring data collection period to adaptively adjust the related early warning thresholds in the early warning grading judgment process, and output the early warning grading judgment result of the next device state monitoring period.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of device state data management, in particular to a device state data analysis system and method based on an energy management platform. BACKGROUND

[0002] Energy is the basis and driving force for the progress of human civilization, is related to the national economy and people's livelihood and national security, and is related to human survival and development. It is of great importance to promote economic and social development and improve people's well-being. Energy equipment is related to energy development, energy use and energy saving. The operation state of energy equipment is directly related to the safety of people's life and property and the safety of the ecological environment.

[0003] Therefore, it is crucial to be able to early warn of possible failures of energy equipment and reduce the losses caused by energy equipment failures. However, in the prior art, the historical data of energy equipment need to be manually summarized to correct the early warning threshold. Since the early warning threshold cannot be flexibly adjusted according to the actual operation, false alarms and false warnings of normally operating energy equipment may occur. The early warning data cannot reflect the actual operation of the energy equipment, which is not conducive to the judgment of the device operation by the device maintenance personnel. SUMMARY

[0004] The purpose of the present application is to provide a device state data analysis system and method based on an energy management platform to solve the problems raised in the background.

[0005] In order to solve the above technical problems, the present application provides the following technical scheme: a device state data analysis method based on an energy management platform, the method comprising:

[0006] Step S100: collecting the historical maintenance records of the target device, constructing a maintenance database of the target device, collecting the historical monitoring data of the sensors monitoring the operation and maintenance parameters of the target device, and constructing a first historical operation and maintenance database of the target device, wherein each monitoring sensor corresponds to a monitoring operation and maintenance parameter of a target device;

[0007] Step S200: setting a real-time monitoring data collection period T2, collecting and screening the real-time monitoring data of the sensors, and obtaining a real-time monitoring data set corresponding to each real-time monitoring data collection period T2;

[0008] Step S300: comparing the data features extracted from each real-time monitoring data set with the data features extracted from the corresponding first historical operation and maintenance database, and obtaining a data feature similarity comparison result corresponding to each real-time monitoring data collection period T2;

[0009] Step S400: The data feature similarity comparison results corresponding to each real-time monitoring data collection period T2 are combined with different early warning thresholds to develop early warning classification judgments for each real-time monitoring data collection period T2.

[0010] Step S500: The judgment data generated in each early warning classification judgment process is combined with the early warning data generated in the corresponding real-time monitoring data collection period T2 to adaptively adjust the related early warning thresholds in the early warning classification judgment process.

[0011] Step S600: The adjusted related early warning thresholds are used in the early warning classification judgment process of the next equipment state monitoring period, and the early warning classification judgment result of the next equipment state monitoring period is output.

[0012] Further, step S100 includes:

[0013] Step S101: Extract the historical maintenance records in the first historical operation and maintenance database, extract the record generation time t corresponding to each historical maintenance record, and collect the various equipment operation state monitoring data exhibited by the target monitoring sensor energy equipment in the previous historical monitoring data collection period T1 before the record generation time t;

[0014] Step S102: Collect all the collected various equipment operation state monitoring data to obtain the first historical operation and maintenance database C1, C1 = {X1, X2, X3, …, XN}, wherein X1, X2, X3, …, XN represent the historical state operation monitoring data sequences corresponding to the 1st, 2nd, 3rd, …, Nth operation and maintenance parameter items in the time period T1. N} respectively. N

[0015] Further, step S200 includes:

[0016] Step S201: Set the real-time operation parameter collection period T2, collect the real-time state operation monitoring data corresponding to each operation parameter in the T2 time period, and arrange each real-time monitoring data according to the collection time to obtain N first real-time monitoring data sequences;

[0017] Step S202: In each first real-time monitoring data sequence, sequentially select two adjacent state operation monitoring data g j and g j+1 , satisfying the condition 1≤j≤M-1, wherein M represents the total amount of state operation monitoring data in each first real-time monitoring data sequence;

[0018] Step S203: Denote the collection time interval of g j and g j+1 as ΔT, and denote g j and g​j+1 The absolute value of the difference is ΔE; set a decision threshold P, when the condition is met... At that time, among them, At that time, discard g j+1 ;

[0019] Step S204: Collect the status operation monitoring data retained after data filtering of each first real-time monitoring data sequence to obtain the second real-time monitoring data sequence corresponding to each first real-time monitoring data sequence;

[0020] By filtering real-time monitoring data and removing data whose variation range within a time interval is less than a certain threshold, the computational workload of the system in calculating the similarity between real-time monitoring data and historical detection data is reduced, the time to obtain similarity comparison results is reduced, and the comparison speed in subsequent steps is improved.

[0021] Furthermore, step S300 includes:

[0022] Step S301: Extract the data features of the second real-time monitoring data sequence and the corresponding historical state operation monitoring data sequence for each operation and maintenance parameter item. Let the data feature of the second real-time monitoring data sequence for each operation and maintenance parameter item be S, and let the data feature of the historical state operation monitoring data sequence corresponding to the second real-time monitoring data sequence for each operation and maintenance parameter item be S'. Based on the formula: F'= max(S, S') / min(S, S'), obtain the similarity F' between S and S'.

[0023] Furthermore, the smaller the similarity value F' between S and S', the greater the similarity between the two sets of data; the larger the value F', the less similar the two sets of data.

[0024] Step S302: Collect the similarity scores F' obtained after comparing the similarity scores of the N operation and maintenance parameter items to obtain the data feature similarity set V = {v1, v2, v3, ..., v2} for each real-time monitoring data collection period T2. N}, where v1, v2, v3, ..., v N These represent the similarity of the 1st, 2nd, 3rd, ..., Nth operation and maintenance parameter items in each real-time monitoring data collection period T2.

[0025] Furthermore, step S400 includes:

[0026] Step S401: Set alarm threshold F1 and attention threshold F2 for each operation and maintenance parameter item, satisfying F2 > F1. Then, determine the similarity results of each operation and maintenance parameter collected from each real-time monitoring data collection period T2.

[0027] More preferably, the one-sided F-test value at a confidence level of 95% is selected as the alarm threshold F1;

[0028] Step S402: alarm the operation and maintenance data item in the data feature similarity set V of each real-time monitoring data collection period T2 that is higher than the data alarm threshold F1;

[0029] Step S403: frequency statistics is performed on the operation and maintenance data item in the data feature similarity set V of each real-time monitoring data collection period T2 that is higher than the data attention threshold F2.

[0030] Further, the system corrects the value of the attention threshold F2 by the correction value β, and the specific correction method is as follows:

[0031] Step S501: record the alarm frequency A of an operation and maintenance parameter in a plurality of real-time monitoring data collection periods T2, and the attention frequency B of the data feature similarity less than the data attention threshold F2;

[0032] Step S502: the correction value β is the proportion of A to B, that is, β=A / B×100%;

[0033] Step S503: set the decision threshold ρ, θ, which satisfies the condition 0<2ρ<θ≤1<2θ, when ρ<β≤θ, the existing F2 is kept unchanged, when β≤ρ, the correction step S504 is executed, and when β>θ, the correction step S505 is executed;

[0034] Step S504: when β≤ρ, the correction value F2 * of F2 is ;

[0035] Step S505: when β>θ, the correction value F2 * of F2 is ;

[0036] Step S506: after the decision of a plurality of real-time monitoring data collection periods T2, the correction value F2 * is calculated, and the correction value F2 * replaces the corresponding F2 value as the decision threshold of the operation and maintenance parameter attention threshold;

[0037] According to the generated early warning data, the related early warning threshold is adjusted, the proportion of the actual number of alarms exceeding the first early warning threshold to the number of alarms of the first early warning threshold in the device operation process is used to flexibly adjust the early warning rule, and the early warning content can be made to conform to the actual operation of the device.

[0038] Further, the system comprises a historical monitoring data acquisition module, a real-time monitoring data acquisition and screening module, a similarity comparison module, and a pre-warning grading judgment module.

[0039] Further, the real-time monitoring data acquisition and screening module comprises a monitoring data acquisition sorting unit and a monitoring data selection unit.

[0040] Further, the pre-warning grading judgment module comprises a numerical judgment unit, a data statistics unit, and a threshold correction unit.

[0041] Compared with the prior art, the present application has the following beneficial effects: by monitoring the operation parameters of energy equipment and comparing the data appearing before the historical fault records with the data, the type of the fault occurring in the energy equipment is pre-warned, by counting the frequency of the energy equipment approaching the fault state in a cycle, the equipment with high frequency of approaching the fault state is paid attention to, and by analyzing the pre-warning frequency change, the pre-warning threshold is optimized, so that the pre-warning content is flexibly adjusted according to the actual operation condition of the energy equipment. BRIEF DESCRIPTION OF DRAWINGS

[0042] The accompanying drawings are included to provide a further understanding of the present application, constitute a part of the specification and are included to explain the present application together with the embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0043] Fig. 1 is a structural schematic diagram of a device state data analysis system based on an energy management platform according to the present application;

[0044] Fig. 2 is a flow schematic diagram of a device state data analysis method based on an energy management platform according to the present application. DETAILED DESCRIPTION

[0045] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of the present application.

[0046] Please refer to Figs. 1-2 The present application provides technical solutions:

[0047] Step S100: collecting historical maintenance records of a target device, constructing a maintenance database of the target device, collecting historical monitoring data of sensors monitoring various operation and maintenance parameters of the target device, and constructing a first historical operation and maintenance database of the target device; wherein each monitoring sensor corresponds to an operation and maintenance parameter of a target device;

[0048] The step S100 includes:

[0049] Step S101: extracting historical maintenance records in the first historical operation and maintenance database, extracting the record generation time t corresponding to each historical maintenance record, and collecting various device operation state monitoring data exhibited by the target monitoring sensor energy device within a historical monitoring data collection period T1 before the record generation time t;

[0050] Step S102: collecting all collected various device operation state monitoring data to obtain a first historical operation and maintenance database C1, C1 = {X1, X2, X3, …, XN}, wherein X1, X2, X3, …, XN respectively represent the historical state operation monitoring data sequence corresponding to the 1st, 2nd, 3rd, …, Nth operation and maintenance parameter item within the time period T1. N}, wherein X1, X2, X3, …, XN respectively represent the historical state operation monitoring data sequence corresponding to the 1st, 2nd, 3rd, …, Nth operation and maintenance parameter item within the time period T1. N

[0051] Step S200: setting a real-time monitoring data collection period T2, collecting real-time monitoring data of the sensor, and Screening obtaining a real-time monitoring data set corresponding to each real-time monitoring data collection period T2;

[0052] The step S200 includes:

[0053] Step S201: setting a real-time operation and maintenance parameter collection period T2, collecting real-time state operation monitoring data of each operation and maintenance parameter within the T2 time period, and arranging each real-time monitoring data in order of collection time to obtain N first real-time monitoring data sequences;

[0054] Step S202: in each first real-time monitoring data sequence, sequentially selecting two adjacent state operation monitoring data as g j ​and g j+1 The condition 1≤j≤M-1 is satisfied, where M represents the total amount of status operation monitoring data in each first real-time monitoring data sequence;

[0055] Step S203: Record g j and g j+1 Let the sampling time interval be ΔT, and let g be denoted as g. j and g j+1 The absolute value of the difference is ΔE; set a decision threshold P, when the condition is met... At that time, among them, At that time, discard g j+1 ;

[0056] For example, g1=25.5, g2=28.3, g3=28.2, and the time interval between g1 and g2. The time interval between g2 and g3 is 1 unit of time. The time interval is 5 units, and the real-time operation and maintenance parameter collection cycle is 50 units. A total of 10 real-time monitoring data points are collected within the 50 units. , Therefore, g2 is retained. , Therefore, g3 is discarded;

[0057] Step S204: Collect the status operation monitoring data retained after data filtering of each first real-time monitoring data sequence to obtain the second real-time monitoring data sequence corresponding to each first real-time monitoring data sequence.

[0058] Step S300: Compare the similarity of the data features extracted from each real-time monitoring dataset with the data features extracted from the corresponding first historical operation and maintenance database; obtain the data feature similarity comparison results for each real-time monitoring data collection period T2 respectively;

[0059] Step S300 includes:

[0060] Step S301: Extract the data features of the second real-time monitoring data sequence and the corresponding historical state operation monitoring data sequence for each operation and maintenance parameter item. Let the data feature of the second real-time monitoring data sequence for each operation and maintenance parameter item be S, and let the data feature of the historical state operation monitoring data sequence corresponding to the second real-time monitoring data sequence for each operation and maintenance parameter item be S'. Based on the formula: F'= max(S, S') / min(S, S'), obtain the similarity F' between S and S'.

[0061] Step S302: Collect the similarity scores F' obtained after comparing the similarity scores of the N operation and maintenance parameter items to obtain the data feature similarity set V = {v1, v2, v3, ..., v2} for each real-time monitoring data collection period T2. N}, where v1, v2, v3, ..., v N These represent the similarity of the 1st, 2nd, 3rd, ..., Nth operation and maintenance parameter items in each real-time monitoring data collection period T2.

[0062] Step S400: Based on the data feature similarity comparison results corresponding to each real-time monitoring data collection period T2, and combined with different warning thresholds, make a warning classification judgment for each real-time monitoring data collection period T2.

[0063] Step S400 includes:

[0064] Step S401: Set alarm threshold F1 and attention threshold F2 for each operation and maintenance parameter item respectively. If F2>F1, make a judgment on the similarity results of each operation and maintenance parameter collected from each real-time monitoring data collection period T2.

[0065] Step S402: Issue an alarm for maintenance data items in the data feature similarity set V of each real-time monitoring data acquisition period T2 that are higher than the alarm threshold F1 for that data item;

[0066] Step S403: Perform frequency statistics on the maintenance data items in the data feature similarity set V of each real-time monitoring data collection period T2 that are higher than the data attention threshold F2.

[0067] Step S500: Combine the judgment data generated in each early warning classification judgment process with the early warning data generated within the corresponding real-time monitoring data collection period T2, and make adaptive adjustments to the relevant early warning thresholds in the early warning classification judgment process;

[0068] The system adjusts the value of the attention threshold F2 using a correction value β. The specific correction method is as follows:

[0069] Step S501: Record the number of times an operation and maintenance parameter has been alarmed in several real-time monitoring data collection cycles T2 as alarm frequency A, and the number of times the data feature similarity is less than the attention threshold F2 of that data as attention frequency B;

[0070] Step S502: The correction value β is the proportion of A to B, i.e., β = A / B × 100%;

[0071] Step S503: Set decision threshold ρ, θ, meet the condition 0 < 2ρ < θ ≤ 1 < 2θ, when ρ < β ≤ θ, keep the existing F2 unchanged, when β ≤ ρ, execute correction step S504, when β > θ, execute correction step S505;

[0072] Step S504: When β ≤ ρ, the correction value F2 * is ;

[0073] Step S505: When β > θ, the correction value F2 * is ;

[0074] Step S506: Execute the correction value F2 * after the decision of several real-time monitoring data collection periods T2, replace the corresponding F2 value with the correction value F2 * as the decision threshold of the operation and maintenance parameter attention threshold;

[0075] Step S600: The adjusted related early warning threshold is used for the early warning grading judgment process of the next device state monitoring period, and the early warning grading judgment result of the next device state monitoring period is output;

[0076] For example, the alarm threshold F1 = 5 and the attention threshold F2 = 12, the alarm frequency of a certain operation and maintenance parameter in several real-time monitoring data collection periods is A = 6, the number of times that the data feature similarity is less than the attention threshold F2 of this data is B = 8, the decision threshold ρ = 25%, θ = 60%, β = 6 / 8 × 100% = 75% > θ, , F2 * = 15.5 is used to replace the attention threshold F2 = 12 as the related early warning threshold in the early warning grading judgment process of the next device state monitoring period.

[0077] Among them, the system includes a historical monitoring data collection module, a real-time monitoring data collection and screening module, a similarity comparison module, and an early warning grading judgment module. The historical monitoring data collection module is used to collect the historical monitoring data of the sensors of each operation and maintenance parameter of the monitoring target device. The real-time monitoring data collection and screening module is used to collect and screen the real-time monitoring data of the sensors. The similarity comparison module is used to compare the similarity of the data features extracted from each real-time monitoring data set with the data features extracted from the first historical operation and maintenance database. The early warning grading judgment module is used to make early warning grading judgments on each real-time monitoring data collection period T2 according to the data feature similarity comparison results and in combination with different early warning thresholds.

[0078] The real-time monitoring data collection and screening module comprises a monitoring data collection and sorting unit and a monitoring data selection unit.

[0079] The early warning grading judgment module comprises a numerical judgment unit, a data statistics unit and a threshold correction unit.

[0080] Compared with the prior art, the present application has the following beneficial effects: the type of failure of the energy equipment is warned by comparing the data before the historical failure record with the energy equipment operation parameters, the energy equipment with a high frequency of approaching failure state is paid attention to by counting the frequency of the energy equipment approaching the failure state in a cycle, and the early warning threshold is optimized by analyzing the change of the early warning frequency, so that the early warning content is flexibly adjusted according to the actual operation of the energy equipment.

[0081] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual such relationship or order between these entities or operations. Moreover, the terms "comprise", "contain" 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 inherent to such a process, method, article or device.

[0082] Finally, it should be noted that: the above only describes the preferred embodiments of the present application, and is not used to limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, for those skilled in the art, the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced equivalently. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for analyzing device status data based on an energy management platform, characterized in that: The method comprises: Step S100: collecting historical maintenance records of a target device, constructing a maintenance database of the target device, collecting historical monitoring data of sensors monitoring various operation and maintenance parameters of the target device, and constructing a first historical operation and maintenance database of the target device; wherein each monitoring sensor corresponds to monitoring an operation and maintenance parameter of a target device; Step S200: setting a real-time monitoring data collection period T2, collecting and screening real-time monitoring data of the sensors, and obtaining a real-time monitoring data set corresponding to each real-time monitoring data collection period T2; Step S300: comparing the data features extracted from each real-time monitoring data set with the data features extracted from the first historical operation and maintenance database; and obtaining a data feature similarity comparison result corresponding to each real-time monitoring data collection period T2; Step S400: combining the data feature similarity comparison results corresponding to each real-time monitoring data collection period T2 with different early warning thresholds to perform early warning classification judgment on each real-time monitoring data collection period T2; Step S400 comprises: Step S401: setting an alarm threshold F1 and an attention threshold F2 for each operation and maintenance parameter, respectively, satisfying F2>F1, and determining the similarity results of each operation and maintenance parameter collected from each real-time monitoring data collection period T2; Step S402: alarming the operation and maintenance data items in the data feature similarity set V of each real-time monitoring data collection period T2 that are higher than the data alarm threshold F1 of the items; Step S403: performing frequency statistics on the operation and maintenance data items in the data feature similarity set V of each real-time monitoring data collection period T2 that are higher than the data attention threshold F2 of the items; Step S500: adaptively adjusting the related early warning thresholds in the early warning classification judgment process by combining the generated judgment data in each early warning classification judgment process with the early warning data generated in the corresponding real-time monitoring data collection period T2; Step S500 comprises: Step S501: recording the alarm frequency A of an operation and maintenance parameter in a plurality of real-time monitoring data collection periods T2, and the attention frequency B of the data feature similarity being less than the data attention threshold F2 of the item; Step S502: the correction value β is the proportion of A to B, that is, β=A / B×100%; Step S503: setting a judgment threshold ρ and θ, satisfying the condition 0<2ρ<θ≤1<2θ, keeping the existing F2 unchanged when ρ<β≤θ, executing the correction step S504 when β≤ρ, and executing the correction step S505 when β>θ; Step S504: When β ≤ ρ, the correction value F2 of F2 is F2 = F2 - β * is ; Step S505: When β > θ, the correction value F2 of F2 is F2 = F2 - β * (θ - β) / θ * is ; Step S506: After the decision of a number of real-time monitoring data collection periods T2, calculate the correction value F2 * , replace the corresponding F2 value as the decision threshold of the operation and maintenance parameter attention threshold * ​ Step S600: the adjusted related early warning thresholds are used for the early warning classification judgment process of the next device state monitoring period, and the early warning classification judgment result of the next device state monitoring period is output.

2. The device state data analysis method based on an energy management platform according to claim 1, characterized in that: The step S100 comprises: Step S101: extracting the historical maintenance records in the first historical operation and maintenance database, extracting the record generation time t corresponding to each historical maintenance record, collecting each item of device operation state monitoring data presented by the target monitoring sensor energy equipment within a historical monitoring data collection period T1 before the record generation time t; Step S102: Collect all collected equipment operation status monitoring data to obtain the first historical operation and maintenance database C1, C1 = {X1, X2, X3, ..., X...} N }, where X1, X2, X3, ..., X N These represent the historical status operation monitoring data sequences corresponding to the 1st, 2nd, 3rd, ..., Nth operation and maintenance parameter items within the time period T1.

3. The device state data analysis method based on an energy management platform according to claim 1, characterized in that: The step S200 includes: Step S201: setting a real-time operation and maintenance parameter collection period T2, collecting real-time state operation monitoring data corresponding to each item of operation and maintenance parameter in the T2 time period, and arranging each item of real-time monitoring data according to the collection time to obtain N first real-time monitoring data sequences; Step S202: In each first real-time monitoring data sequence, two adjacent state operation monitoring data are sequentially selected and recorded as g j and g j+1 , satisfying the condition 1≤j≤M-1; wherein M represents the total amount of state operation monitoring data in the first real-time monitoring data sequence. Step S203: Record g j and g j+1 Let the sampling time interval be ΔT, and let g be denoted as g. j and g j+1 The absolute value of the difference is ΔE; set a decision threshold P, when the condition is met... At that time, among them, At that time, discard g j+1 ; Step S204: collecting the state operation monitoring data retained after data screening of each first real-time monitoring data sequence respectively, to obtain a second real-time monitoring data sequence corresponding to each first real-time monitoring data sequence.

4. The device state data analysis method based on an energy management platform according to claim 1, characterized in that: The step S300 includes: Step S301: extracting the data features of the second real-time monitoring data sequence of each operation and maintenance parameter item and the data features of the corresponding historical state operation monitoring data sequence, setting the data features of the second real-time monitoring data sequence of each operation and maintenance parameter item as S, and setting the data features of the historical state operation monitoring data sequence corresponding to the second real-time monitoring data sequence of each operation and maintenance parameter item as S'; based on the formula: F'=max(S, S') / min(S, S'), the similarity F' of S and S' is obtained; Step S302: respectively collect the similarities F' obtained after the similarity comparison of the N operation and maintenance parameter items, to obtain a data feature similarity set V = {v1, v2, v3, …, vN} corresponding to each real-time monitoring data collection period T2. N} wherein v1, v2, v3, …, vN respectively represent the similarities corresponding to the 1st, 2nd, 3rd, …, Nth operation and maintenance parameter items in each real-time monitoring data collection period T2. N ​ 5. A device state data analysis system for implementing the device state data analysis method based on the energy management platform according to any one of claims 1 to 4, characterized by, The system includes a historical monitoring data collection module, a real-time monitoring data collection and screening module, a similarity comparison module, and a warning grading judgment module. The historical monitoring data collection module is used to collect historical monitoring data of sensors of each operation and maintenance parameter of the monitoring target device. The real-time monitoring data collection and screening module is used to collect and screen real-time monitoring data of sensors. The similarity comparison module is used to compare the data features extracted from each real-time monitoring data set with the data features extracted from the first historical operation and maintenance database. The warning grading judgment module is used to make warning grading judgments for each real-time monitoring data collection period T2 according to the data feature similarity comparison results and in combination with different warning thresholds.

6. The equipment state data analysis system of claim 5, wherein, The real-time monitoring data collection and screening module includes a monitoring data collection and sorting unit and a monitoring data selection unit. The monitoring data collection and sorting unit is used to arrange each item of real-time monitoring data according to the collection time. The monitoring data selection unit is used to select real-time monitoring data that meets the judgment requirements.

7. The equipment state data analysis system of claim 5, wherein, The warning grading judgment module includes a numerical judgment unit, a data statistics unit, and a threshold correction unit. The numerical judgment unit is used to determine the threshold interval in which the similarity value is located. The data statistics unit is used to count the number of judgment results in the threshold interval. The threshold correction unit is used to correct the threshold.

Images