Reservoir early warning method and system based on operation and maintenance data

Abstract

The invention belongs to the technical field of reservoir operation and maintenance management, and provides a reservoir early-warning method and system based on operation and maintenance data, and the method comprises the steps: constructing an early-warning database, carrying out the processing and analysis of the operation parameters of a reservoir in a current monitoring period, recognizing the abnormal operation parameters of the reservoir in the current monitoring period, comparing with an early-warning data set, and carrying out the early-warning of the reservoir. And outputting the operation early warning type of the reservoir in the current monitoring period, identifying the maintenance exceeding period through the maintenance frequency of the reservoir in the historical monitoring period, and judging whether maintenance early warning needs to be performed on the reservoir according to the continuity and frequency of the maintenance exceeding period and the change of the operation parameters in the maintenance exceeding period in combination with analysis. The situation that although reservoir operation parameters do not change remarkably, the maintenance frequency exceeds the standard due to other conditions such as reservoir equipment is prevented, and the comprehensiveness and accuracy of reservoir early warning are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of reservoir operation and maintenance management, and particularly relates to a reservoir early warning method and system based on operation and maintenance data. BACKGROUND

[0002] As important infrastructure, the stable operation and safety of a reservoir are of great significance, but there are still some deficiencies in the early warning of the reservoir, for example, the reservoir early warning is not accurate and comprehensive, and therefore it is of great significance to research a reservoir early warning method and system based on operation and maintenance data.

[0003] In the prior art, when the safety of a reservoir is warned, a corresponding operation threshold is usually set, and when the operation parameter of the reservoir exceeds the operation threshold, a warning signal is generated. On the one hand, the prior art has the problem of not timely warning, and the prior art does not identify abnormal operation parameters according to the prediction result, and compare the operation parameters with the past reservoir early warning data to determine the operation warning type of the reservoir according to the comparison result, which results in low timeliness and accuracy of the reservoir early warning. On the other hand, the prior art lacks processing and analysis of the maintenance frequency of the reservoir to determine whether the reservoir needs to be maintained. Since in some cases, the operation parameter of the reservoir may not change significantly, but the maintenance frequency of the reservoir changes due to other factors, if the analysis and warning of the maintenance are lacking, the reservoir early warning will not be comprehensive, and thus the safety and quality of the reservoir cannot be guaranteed.

[0004] Therefore, the application provides a reservoir early warning method and system based on operation and maintenance data. SUMMARY

[0005] In order to make up for the deficiencies of the prior art and solve at least one technical problem proposed in the background.

[0006] The technical scheme adopted by the application to solve the technical problems is: In a first aspect, the application provides a reservoir early warning method based on operation and maintenance data, comprising: Step 1: integrating and processing the early warning data of the reservoir in multiple historical monitoring periods to obtain an early warning database; Step 2: in the current monitoring period, real-time acquisition of the operation parameter of the reservoir, processing and analysis of the operation parameter, prediction of the abnormal operation parameter of the reservoir in the current monitoring period according to the analysis result; Step 3: comparison and analysis of the abnormal operation parameter and the early warning database, processing and outputting the operation warning type of the reservoir in the current monitoring period according to the comparison and analysis result; Step four: comparing and analyzing the maintenance data of the reservoir in multiple historical monitoring periods, identifying the maintenance exceeding standard period in the historical monitoring period, obtaining the operation parameters in the maintenance exceeding standard period and analyzing the continuous performance of the maintenance exceeding standard period, and outputting the continuous state value; Step five: judging whether maintenance warning of the reservoir is needed according to the continuous state value, and if so, generating a maintenance warning signal.

[0007] In a second aspect, the present application provides a reservoir warning system based on operation and maintenance data, comprising: The database construction module integrates and processes the warning data of the reservoir in multiple historical monitoring periods to obtain a warning database; The operation data identification module obtains the operation parameters of the reservoir in real time in the current monitoring period, processes and analyzes the operation parameters, and predicts the abnormal operation parameters of the reservoir in the current monitoring period according to the analysis result; The operation warning module compares and analyzes the abnormal operation parameters and the warning database, and processes and outputs the operation warning type of the reservoir in the current monitoring period according to the comparison and analysis result; The maintenance data acquisition module compares and analyzes the maintenance data of the reservoir in multiple historical monitoring periods, identifies the maintenance exceeding standard period in the historical monitoring period, obtains the operation parameters in the maintenance exceeding standard period, and analyzes the continuous performance of the maintenance exceeding standard period to output the continuous state value; The maintenance warning module judges whether maintenance warning of the reservoir is needed according to the continuous state value, and if so, generates a maintenance warning signal; The reservoir operation and maintenance platform is used for receiving the operation warning type and the maintenance warning signal of the reservoir.

[0008] The beneficial effects of the present application are as follows: 1. The present application constructs a warning database by using the warning data of the reservoir in multiple historical monitoring periods, processes and analyzes the operation parameters of the reservoir in the current monitoring period, identifies the abnormal operation parameters of the reservoir in the current monitoring period, integrates to obtain an abnormal operation group, and outputs the operation warning type of the reservoir in the current monitoring period through the comparison result of the abnormal operation group and the warning data group in the warning database. The present application combines the historical warning data of the reservoir and the prediction of the operation parameters of the reservoir, identifies the operation warning type, and is beneficial to improving the timeliness and accuracy of the reservoir warning.

[0009] 2. Identify the maintenance exceeding period by the number of maintenance of the reservoir in the historical monitoring period, and according to the continuity, number of times of the maintenance exceeding period and the change of the operation parameter in the maintenance exceeding period, combined with analysis, determine whether the reservoir needs to be maintained and warned, prevent the situation that although the operation parameter of the reservoir does not change significantly, the number of maintenance exceeds due to other conditions such as reservoir equipment, and improve the comprehensiveness of the reservoir warning. BRIEF DESCRIPTION OF DRAWINGS

[0010] The application will be further described below with reference to the drawings.

[0011] Figure 1 is a step flow chart of a reservoir warning method based on operation and maintenance data according to an embodiment of the application; Figure 2 is a program block diagram of a reservoir warning system based on operation and maintenance data according to an embodiment of the application. DETAILED DESCRIPTION

[0012] In order to make the technical means, creative features, purposes and effects of the application easy to understand, the application will be further described below with reference to the specific embodiments. EMBODIMENT

[0013] As shown in Figure 1 , a reservoir warning method based on operation and maintenance data according to an embodiment of the application includes the following steps: Step 1: Integrate and process the warning data of the reservoir in multiple historical monitoring periods to obtain a warning database; In some embodiments, the reservoir monitoring report is obtained through the operation and maintenance management platform of the reservoir, and the warning data of the reservoir in multiple historical monitoring periods is obtained through the reservoir monitoring report, wherein the warning data includes the operation warning type and abnormal operation parameter of the reservoir when the warning occurs in the historical monitoring period, and the abnormal operation parameter includes but is not limited to water level, precipitation, flow, etc. The operation warning type and abnormal operation parameter of each time the reservoir occurs warning are correspondingly integrated to obtain a warning data group; For example, it is assumed that in a historical monitoring period, the reservoir occurs flood warning, and at this time the abnormal operation parameters are water level and rainfall, and the integrated warning data group is (flood warning-water level, rainfall); All integrated warning data groups are summarized to construct a warning database; It should be noted that the abnormal operation parameter means that the operation parameter of the reservoir exceeds the corresponding preset threshold value, for example, the water level exceeds the water level threshold value; Step two: in the current monitoring period, the operation parameters of the reservoir are obtained in real time, the operation parameters are processed and analyzed, and the abnormal operation parameters of the reservoir in the current monitoring period are predicted according to the analysis results; Specifically, in the current monitoring period, the operation parameters of the reservoir are obtained in real time, wherein the operation parameters include but are not limited to water level, precipitation, flow and the like; Based on any one operation parameter; According to the real-time monitored operation parameters, the operation change curve is drawn in the constructed rectangular coordinate system, wherein the X axis of the rectangular coordinate system is time, and the Y axis is the operation parameter value; The preset threshold value corresponding to the operation parameter is taken as the first reference value, and the warning threshold value corresponding to the operation parameter is taken as the second reference value. Straight lines parallel to the X axis are drawn through the corresponding points of the first reference value and the second reference value on the Y axis, respectively, to obtain the threshold value line and the warning threshold value line, wherein the threshold value line is obtained from the first reference value, the warning threshold value line is obtained from the second reference value, and the preset threshold value is greater than the warning threshold value; It should be noted that the warning threshold value is a warning value set by the person skilled in the art, and the purpose is to prevent the operation parameter from exceeding the preset threshold value, so that the warning value is set in advance; The operation change curve is monitored in real time. If the operation change curve exceeds the warning threshold value line, the distance between the change end of the operation change curve and the threshold value line is measured, and the ratio processing is performed with the warning distance to obtain the warning proximity value; Wherein, the warning distance is the shortest vertical distance between the threshold value line and the warning threshold value line; It should be noted that the change end of the operation change curve represents one end of the operation change curve changing with time; The warning proximity value is compared with the warning proximity threshold value; If the warning proximity value is greater than or equal to the warning proximity threshold value, an analysis signal is generated; If the warning proximity value is less than the warning proximity threshold value, real-time monitoring is continued; It should be noted that the warning proximity value reflects the proximity of the current operation parameter to the preset threshold value after exceeding the warning threshold value, and the greater the warning proximity value, the closer the operation parameter to the corresponding preset threshold value; Based on the analysis signal, the part of the operation change curve exceeding the warning threshold value line when the analysis signal appears is marked as a warning change line; The time period corresponding to the warning change line on the X axis is measured to obtain a warning time period, the time length corresponding to the warning time period is compared with the time length corresponding to the current monitoring period to obtain a warning time length value, and the warning time length value is marked as SC; The alert period is divided into a plurality of alert segments which are continuous and have equal time lengths, operation parameters at different time points in the alert segment are obtained based on the alert change line, the operation parameters at the different time points in the alert segment are fitted by using a least square method to obtain a fitting model in the alert segment; The fitting models corresponding to adjacent alert segments are integrated into a fitting model group, and an absolute change amount of a regression parameter is obtained by performing difference processing on the regression parameter of the fitting model group; The regression parameter includes a slope and an intercept, and the absolute change amount of the regression parameter is obtained by a slope difference absolute value and an intercept difference absolute value of two fitting models in the fitting model group; For example, assuming that the fitting models in the fitting model group are Y1=K1X+b1 and Y2=K2X+b2, where Y is an operation parameter, X is time, K is a slope, and b is an intercept, and the calculation method of the absolute change amount JL of the regression parameter is: ; The absolute change amount of the regression parameter is compared with a normal fluctuation range; If the absolute change amount of the regression parameter exceeds the normal fluctuation range, the fitting model group is marked as a critical change group; The obtained critical change groups are integrated into a critical change group set (JZ1, JZ2, JZ3,..., JZn); n wherein JZ n represents the nth critical change group, and n is the number of the critical change groups; If the absolute change amount of the regression parameter does not exceed the normal fluctuation range, the fitting model group is marked as a non-critical change group; If the slope of a latter fitting model is greater than the slope of a former fitting model in the critical change group, the critical change group is marked as an increasing critical change group; The obtained increasing critical change groups are integrated into an increasing critical change group set (JK1, JK2, JK3,..., JKm); m wherein JK m represents the mth increasing critical change group, and m is the number of the increasing critical change groups; If the slope of a latter fitting model is less than or equal to the slope of a former fitting model in the critical change group, the critical change group is marked as a non-increasing critical change group; The critical change value JB is obtained by the formula: wherein s1 and s2 are both preset proportion coefficients, s1 is 1.24, s2 is 1.36, and v represents the number of the fitting model groups; It should be noted that the meaning and purpose of the burst value JB are as follows: in the case where the running parameter approaches the preset threshold, the time length of the running parameter exceeding the warning threshold to approaching the preset threshold is analyzed, and whether the running parameter is in a growing burst state after exceeding the warning threshold is determined according to the growth change performance in all time lengths, so as to predict and identify whether the running parameter will become an abnormal running parameter according to the growing burst state of the running parameter, thereby facilitating subsequent reservoir early warning. The burst value JB is compared with the burst threshold value; If the burst value JB is greater than or equal to the burst threshold value, it indicates that the running parameter corresponding to the burst value JB has a sharp change in a short time, exceeds the warning threshold and approaches the preset threshold, and the running parameter is marked as an abnormal running parameter; If the burst value JB is less than the burst threshold value, it indicates that the running parameter corresponding to the burst value JB does not have a sharp change in a short time, and a detailed signal is generated; Based on the detailed signal, the running parameters at different time points in the warning period are obtained based on the warning change line and integrated into a running parameter set; a plurality of models are used to fit the running parameter set to identify the best matching model of the running parameter set; The models include but are not limited to linear regression, nonlinear regression, classification model, etc., and the specific identification process is as follows: S1, the running parameter set is randomly divided into K disjoint subsets (or called “folds”), where K can be 10, 15 or 20; S2, each time one of the subsets is selected as the test set, and the remaining K-1 subsets are used as the training set; S3, the model is trained on the training set, and the performance of the model is evaluated on the test set; S4, repeat K times, each time select a different subset as the test set, until each subset has and only one time as the test set; S5, calculate the average value of the performance indicators obtained in K iterations, wherein the performance indicators include but are not limited to accuracy, precision and recall, etc.; S6, compare the average values of the performance indicators of the running parameter set under different model fitting, and find the best matching model of the running parameter set according to the comparison result; The preset threshold value corresponding to the running parameter is input into the best matching model as a result, and the time point at which the running parameter reaches the preset threshold value, i.e., the threshold reaching time point, is obtained; The threshold reaching time point is compared with the end time point of the current monitoring period; If the threshold reaching time point is before the end time point of the current monitoring period, it indicates that the running parameter will exceed the preset threshold before the end of the current monitoring period, and the running parameter is marked as an abnormal running parameter; If the threshold time point is after the end time point of the current monitoring period, it means that the operation parameter will not exceed the preset threshold before the end of the current monitoring period, and no operation is performed. Step three: comparing and analyzing the abnormal operation parameters and the early warning database, and outputting the operation early warning type of the reservoir in the current monitoring period according to the comparison and analysis result; Specifically, the obtained abnormal operation parameters are integrated into an abnormal operation group, and the abnormal operation group is compared with each early warning data group in the early warning database; based on any comparison, the specific process is as follows: The number of abnormal operation parameters of the same type in the abnormal operation group and the early warning data group is counted, and a ratio processing is performed on the number of abnormal operation parameters in the early warning data group, to obtain an abnormal type same value, which is marked as XT; The number of abnormal operation parameters in the abnormal operation group and the number of abnormal operation parameters in the early warning data group are counted, and a difference processing is performed, to obtain an abnormal operation parameter difference, and a ratio processing is performed on the abnormal operation parameter difference and the number of abnormal operation parameters in the early warning data group, to obtain an abnormal number difference value, which is marked as CY; The obtained abnormal type same value XT and the abnormal number difference value CY are subjected to data processing, through the formula: An comparison matching value BP is obtained; wherein, f1 and f2 are both preset proportion coefficients, wherein, f1 takes a value of 1.379, and f2 takes a value of 1.286; The comparison matching value BP is compared with a comparison matching threshold value; If the comparison matching value BP is greater than the comparison matching threshold value, it means that the matching of the abnormal operation group and the early warning data group is high, and the early warning data group is marked as a matching data group; If the comparison matching value BP is less than or equal to the comparison matching threshold value, it means that the matching between the abnormal operation group and the early warning data group is low, and no operation is performed; The operation early warning type of the reservoir in the current monitoring period is output according to the operation early warning type corresponding to the matching data group; The technical scheme of the embodiment of the application is as follows: the early warning database is constructed by using the early warning data of the reservoir in a plurality of historical monitoring periods, the operation parameters of the reservoir in the current monitoring period are processed and analyzed, the abnormal operation parameters of the reservoir in the current monitoring period are recognized, and the abnormal operation group is integrated, the operation early warning type of the reservoir in the current monitoring period is output through the comparison result of the abnormal operation group and the early warning data group in the early warning database, the historical early warning data of the reservoir and the prediction of the operation parameters of the reservoir are combined, the operation early warning type is recognized in advance, and the timeliness and accuracy of the reservoir early warning are improved.

[0014] Embodiment 2 As Figure 1As shown, based on the basis of example 1, the reservoir early warning method based on operation and maintenance data described in the embodiment of the application comprises the following steps: Step four: comparing and analyzing the maintenance data of the reservoir in multiple historical monitoring periods, identifying the maintenance exceeding standard period in the historical monitoring period, obtaining the operation parameters in the maintenance exceeding standard period and analyzing the continuous performance of the maintenance exceeding standard period, and outputting to obtain the continuous state value; The maintenance data includes the number of maintenance times. Specifically, based on any one historical monitoring period; The number of maintenance times of the reservoir in the historical monitoring period is compared with the preset number of maintenance times; If the number of maintenance times of the reservoir is greater than the preset number of maintenance times, the historical monitoring period is marked as a maintenance exceeding standard period; If the number of maintenance times of the reservoir is equal to the preset number of maintenance times, the historical monitoring period is marked as a normal maintenance period; It should be noted that the preset number of maintenance times represents the standard number of normal maintenance of the reservoir; The operation parameters in the maintenance exceeding standard period are obtained, and the corresponding difference value processing is performed on the operation parameters in the normal maintenance period, and the absolute value of the difference value is taken to obtain the operation parameter absolute difference. The operation parameter absolute difference is processed by ratio to obtain the operation parameter difference ratio. The sum of all operation parameter difference ratios is taken to obtain the operation deviation value, and is marked as YX; Exemplarily, the rainfall and water level in the maintenance exceeding standard period are respectively compared with the rainfall and water level in the normal maintenance period, and the corresponding difference value processing is performed, and the absolute value of the difference value is taken to obtain the rainfall absolute difference and the water level absolute difference. The rainfall absolute difference and the water level absolute difference are respectively processed by ratio to obtain the rainfall difference ratio and the water level difference ratio. The sum is obtained to obtain the operation deviation value; The number of maintenance exceeding standard periods in the historical monitoring period is counted, and the ratio processing is performed on the total number of the historical monitoring period to obtain the maintenance exceeding standard value, and is marked as CB; The number of normal maintenance periods between every two adjacent maintenance exceeding standard periods is counted, and the sum is obtained to obtain the interval period number. The interval period number is processed by ratio to obtain the interval period mean value. The interval period mean value is processed by ratio to obtain the maintenance interval value, and is marked as WJ; It should be noted that the interval is counted once between every two adjacent maintenance exceeding standard periods. The operation deviation value YX, the maintenance exceeding standard value CB and the maintenance interval value WJ are processed by data, and the formula is: A continuous state value LX is obtained, wherein w1, w2 and w3 are preset proportion coefficients, w1 is 1.02, w2 is 1.374, and w3 is 1.496; It should be noted that the meaning and purpose of the continuous state value LX are to analyze whether the reservoir needs to be maintained by analyzing the number of times, continuity and change of the reservoir operation parameter in the maintenance exceeding period, and the purpose of obtaining the continuous state value LX is to identify the case that the number of maintenance times increases while the reservoir operation parameter does not change significantly, and the significance is that although the reservoir operation parameter does not change significantly, the number of reservoir maintenance times may change due to abnormal reservoir equipment and other conditions, and timely analysis and early warning can improve the comprehensiveness of reservoir early warning. Step five: determining whether the reservoir needs to be maintained according to the continuous state value, and if so, generating a maintenance early warning signal; In some embodiments, the continuous state value is compared with a continuous state threshold value; If the continuous state value is greater than or equal to the continuous state threshold value, a maintenance early warning signal is generated and sent to the reservoir operation and maintenance platform; If the continuous state value is less than the continuous state threshold value, no operation is performed; The technical scheme of the embodiment of the present application is that the maintenance exceeding period is identified by the number of maintenance times of the reservoir in the historical monitoring period, and whether the reservoir needs to be maintained is analyzed and judged according to the continuity, number of times and change of the operation parameter in the maintenance exceeding period, so as to prevent the case that the number of maintenance times exceeds due to abnormal reservoir equipment and other conditions although the reservoir operation parameter does not change significantly, and improve the comprehensiveness of reservoir early warning.

[0015] Embodiment 3 As shown in Figure 2 A reservoir early warning system based on operation and maintenance data according to the embodiment of the present application comprises: A database construction module integrates and processes early warning data of the reservoir in a plurality of historical monitoring periods to obtain an early warning database; An operation data identification module obtains the operation parameter of the reservoir in a current monitoring period in real time, processes and analyzes the operation parameter, and predicts the abnormal operation parameter of the reservoir in the current monitoring period according to the analysis result; An operation early warning module compares and analyzes the abnormal operation parameter and the early warning database, and processes and outputs the operation early warning type of the reservoir in the current monitoring period according to the comparison and analysis result; Maintenance data acquisition: comparing and analyzing the maintenance data of the reservoir in multiple historical monitoring periods, identifying the maintenance exceeding standard period in the historical monitoring period, obtaining the operation parameters in the maintenance exceeding standard period and analyzing the continuous performance of the maintenance exceeding standard period, and outputting the continuous state value; Maintenance early warning module: judging whether maintenance early warning of the reservoir is needed according to the continuous state value, and if yes, generating a maintenance early warning signal; Reservoir operation and maintenance platform: used for receiving the operation early warning type and the maintenance early warning signal of the reservoir.

[0016] The basic principle, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above examples, and the above examples and descriptions in the specification are only to illustrate the principle of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A reservoir early warning method based on operation and maintenance data, characterized in that: The method comprises the following steps: Step 1: integrating and processing early warning data of the reservoir in multiple historical monitoring periods to obtain an early warning database; Step 2: in the current monitoring period, real-time acquisition of the operation parameters of the reservoir, processing and analysis of the operation parameters, prediction of abnormal operation parameters of the reservoir in the current monitoring period according to the analysis result; Step 3: comparison and analysis of the abnormal operation parameters and the early warning database, processing and output of the operation early warning type of the reservoir in the current monitoring period according to the comparison and analysis result; Step 4: comparison and analysis of the maintenance data of the reservoir in multiple historical monitoring periods, identification of a maintenance over-standard period in the historical monitoring period, acquisition of the operation parameters in the maintenance over-standard period and analysis combined with the continuous performance of the maintenance over-standard period to output a continuous state value; Step 5: judgment of whether the maintenance early warning of the reservoir is needed according to the continuous state value, and if yes, a maintenance early warning signal is generated. 2.The reservoir early warning method based on operation and maintenance data according to claim 1, characterized in that: The process of processing and analyzing the operation parameters comprises the following steps: based on any one operation parameter; drawing an operation change curve according to the real-time monitored operation parameter, calculating an early warning proximity value if the operation change curve exceeds the early warning threshold line, and the early warning proximity value representing the proximity degree of the current operation parameter to the preset threshold value after exceeding the early warning threshold; generating an analysis signal if the early warning proximity value is greater than or equal to an early warning proximity threshold value; marking the part of the operation change curve exceeding the early warning threshold line as an early warning change line when the analysis signal appears; obtaining a sudden change value JB based on the processing and analysis of the early warning change line; if the sudden change value JB is greater than or equal to a sudden change threshold value, marking the operation parameter as an abnormal operation parameter, otherwise, generating a detailed analysis signal.

3. The reservoir early warning method based on operation and maintenance data according to claim 2, characterized in that: The early warning proximity value is obtained in the following manner: if the operation change curve exceeds the early warning threshold line, measuring the distance between the change end of the operation change curve and the threshold line, and performing ratio processing on the distance and an early warning distance to obtain the early warning proximity value; wherein the early warning distance is the shortest vertical distance between the threshold line and the early warning threshold line.

4. The reservoir early warning method based on operation and maintenance data according to claim 2, characterized in that: The sudden change value JB is obtained in the following manner: measuring the time period corresponding to the early warning change line on the X axis to obtain an early warning time period, performing ratio processing on the time length corresponding to the early warning time period and the time length corresponding to the current monitoring period to obtain an early warning time length value and mark it as SC; dividing the early warning time period into a plurality of early warning segments each having continuous time and equal time length, obtaining operation parameters at different time points in the early warning segment based on the early warning change line, fitting the operation parameters at different time points in the early warning segment by using the least square method to obtain a fitting model in the early warning segment; integrating the fitting models corresponding to adjacent early warning segments into a fitting model group, obtaining an absolute change amount of a regression parameter by performing difference processing on the regression parameter of the fitting model group; if the absolute change amount of the regression parameter exceeds a normal fluctuation range, marking the fitting model group as a sudden change group; The obtained catastrophic groups are integrated into a catastrophic group set (JZ1, JZ2, JZ3...JZ). n ); among them, JZ n This indicates the nth mutation group, where n is the number of the mutation group. If the slope of the next fitting model is greater than the slope of the previous fitting model in the excitation group, the excitation group is marked as an excitation increasing group; the obtained excitation increasing groups are integrated into an excitation increasing group set (JK1, JK2, JK3... JK m ); wherein, JK m represents the mth excitation increasing group, and m is the number of the excitation increasing group. The jump value JB is obtained by the formula: wherein s1 and s2 are both preset proportion coefficients, and v represents the number of the fitting model groups.

5. The reservoir early warning method based on operation and maintenance data according to claim 2, characterized in that: if the detailed analysis signal is generated, the analysis is continued, specifically: based on the early warning change line, obtaining operation parameters at different time points in the early warning time period and integrating them into an operation parameter set; fitting the operation parameter set by using a plurality of models to identify the best matching model of the operation parameter set; S1, randomly dividing the operation parameter set into K non-intersecting subsets; S2, each time one of the subsets is selected as the test set, and the remaining K-1 subsets are selected as the training set; S3, training the model on the training set and evaluating the performance of the model on the test set; S4, repeating K times, each time selecting a different subset as the test set, until each subset has and only has one time as the test set; S5, calculating the average value of the performance indicators obtained in K iterations, wherein the performance indicators include accuracy, precision and recall; S6, comparing the average values of the performance indicators of the running parameter set under different model fitting, and finding the best matching model of the running parameter set according to the comparison result; The preset threshold value corresponding to the running parameter is input into the best matching model as the result, and the time point when the running parameter reaches the preset threshold value, that is, the threshold time point, is obtained; If the threshold time point is before the end time point of the current monitoring period, the running parameter is marked as an abnormal running parameter.

6. The reservoir early warning method based on operation and maintenance data according to claim 1, characterized in that: The process of comparing and analyzing the abnormal running parameter and the early warning database is: The abnormal running parameters are integrated into an abnormal running group, and the abnormal running group is compared with each early warning data group in the early warning database; based on any comparison, the same value XT of the abnormal type and the difference value CY of the abnormal quantity are obtained; The alignment matching value BP is obtained by the formula: wherein f1 and f2 are both preset proportion coefficients. If the comparison matching value BP is greater than the comparison matching threshold value, it indicates that the matching degree of the abnormal running group and the early warning data group is high, and the early warning data group is marked as a matching data group; According to the running early warning type corresponding to the matching data group, the running early warning type of the reservoir in the current monitoring period is output.

7. The reservoir early warning method based on operation and maintenance data according to claim 6, characterized in that: The acquisition method of the same value XT of the abnormal type and the difference value CY of the abnormal quantity is: The number of abnormal running parameters of the same type in the abnormal running group and the early warning data group is counted, and a ratio processing is performed with the number of abnormal running parameters in the early warning data group, to obtain the same value XT of the abnormal type, and mark it as XT; The number of abnormal running parameters in the abnormal running group and the number of abnormal running parameters in the early warning data group are counted and difference processed to obtain the abnormal running parameter difference, and the abnormal running parameter difference is ratio processed with the number of abnormal running parameters in the early warning data group to obtain the difference value of the abnormal quantity, and mark it as CY.

8. The reservoir early warning method based on operation and maintenance data according to claim 1, characterized in that: The acquisition method of the continuous state value is: Based on any one historical monitoring period; If the number of maintenance of the reservoir is greater than the preset number of maintenance, the historical monitoring period is marked as a maintenance exceeding period; Based on the processing and analysis of the maintenance exceeding period, the running deviation value YX, the maintenance exceeding value CB and the maintenance interval value WJ are obtained; The continuous state value LX is obtained by the formula: where w1, w2, and w3 are preset proportionality coefficients.

9. The reservoir early warning method based on operation and maintenance data according to claim 8, characterized in that: The acquisition method of the running deviation value YX, the maintenance exceeding value CB and the maintenance interval value WJ is: The running parameters in the maintenance exceeding period are obtained, and a corresponding difference value processing is performed with the running parameters in the normal maintenance period, and the difference value is taken as an absolute value to obtain the running parameter absolute difference, the running parameter absolute difference is ratio processed with the corresponding running parameter to obtain the running parameter difference ratio, the sum of all running parameter difference ratios is taken as an average value to obtain the running deviation value, and marked as YX; The number of maintenance exceeding periods in the historical monitoring period is counted, and a ratio processing is performed with the total number of the historical monitoring period to obtain the maintenance exceeding value, and marked as CB; The number of normal maintenance periods between every two adjacent maintenance exceeding periods is counted and summed to obtain the interval period number, the interval period number is processed by ratio with the interval number between the maintenance exceeding periods to obtain the interval period mean, the interval period mean is processed by ratio with the total number of the historical monitoring periods to obtain the maintenance interval value, and is marked as WJ.

10. A reservoir early warning system based on operation and maintenance data, characterized in that: Comprise: A database construction module: integrating and processing early warning data of the reservoir in multiple historical monitoring periods to obtain an early warning database; An operation data identification module: in the current monitoring period, real-time acquisition of operation parameters of the reservoir, processing and analysis of the operation parameters, prediction of abnormal operation parameters of the reservoir in the current monitoring period according to the analysis result; An operation early warning module: comparison and analysis of the abnormal operation parameters and the early warning database, processing and output of an operation early warning type of the reservoir in the current monitoring period according to the comparison and analysis result; Maintenance data acquisition: in multiple historical monitoring periods, comparison and analysis of maintenance data of the reservoir, identification of a maintenance exceeding period in the historical monitoring period, acquisition of operation parameters in the maintenance exceeding period and analysis combined with continuous performance of the maintenance exceeding period to output a continuous state value; A maintenance early warning module: judging whether maintenance early warning of the reservoir is needed according to the continuous state value, and if yes, generating a maintenance early warning signal.

Images