A multi-modal drainage monitoring data analysis system and method

Abstract

The application discloses a kind of multimodal drainage monitoring data analysis system and method, it is related to drainage monitoring technical field, and its technical solution key points include the following steps: in monitoring period, the multimodal monitoring data of target drainage system is collected;According to the risk of target drainage system is evaluated to obtain initial comprehensive risk assessment value according to multimodal monitoring data;In multimodal monitoring data, the key characteristic parameter of influence target drainage system operating state is extracted, and the abnormal risk correlation coefficient of different key characteristic parameters and target drainage system is extracted according to historical database, and the abnormal risk value caused by single multimodal monitoring data is predicted based on abnormal risk correlation coefficient and key characteristic parameter to obtain risk prediction value;According to historical database, the coupling influence coefficient between multimodal characteristic parameters is extracted;Effect is to make risk assessment result more close to actual situation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of drainage monitoring, more particularly, it relates to a multi-modal drainage monitoring data analysis system and method. BACKGROUND

[0002] In the operation and maintenance management of urban drainage systems, traditional monitoring and analysis methods rely on single modal data, such as focusing only on flow data or only on water quality indicators. However, the operation of urban drainage systems is a complex process, including water flow movement, water quality change, pipe structure state and other dimensions, each link is interwoven and synergistic. For example, only according to the flow data to judge whether the drainage system is overloaded, but ignoring the pipe corrosion caused by high concentration of pollutants in water quality, because pipe corrosion and damage will also cause poor drainage risk, so the early one-sided evaluation will miss the case, so as to not accurately early warning.

[0003] At the same time, the urban drainage system has accumulated a large amount of historical database during long-term operation, which contains various data under different weather, different time periods and different working conditions. However, the traditional method lacks effective multi-modal data fusion analysis means, that is, it cannot deeply mine the coupling effect between parameters. The frequent fluctuation of flow in the drainage system will affect the pipe pressure, and then accelerate the pipe aging, and the pipe aging will change the flow state of water, thereby affecting the sedimentation effect of water quality, so that the risk early warning is out of touch with the actual system operation state, and cannot provide reliable support for operation and maintenance decision, that is, hinders the development of urban drainage system operation and maintenance towards intelligent and fine direction. SUMMARY

[0004] In view of the deficiencies in the prior art, the purpose of the present application is to provide a multi-modal drainage monitoring data analysis system and method.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0006] A multi-modal drainage monitoring data analysis method, the method comprising the following steps:

[0007] Collecting multi-modal monitoring data of the target drainage system in the monitoring period;

[0008] Evaluating the risk of the target drainage system according to the multi-modal monitoring data to obtain an initial comprehensive risk evaluation value;

[0009] Extracting key characteristic parameters affecting the operation state of the target drainage system from the multi-modal monitoring data, extracting abnormal risk correlation coefficients of different key characteristic parameters and the target drainage system from the historical database, and predicting abnormal risk values caused by single multi-modal monitoring data based on the abnormal risk correlation coefficients and the key characteristic parameters to obtain risk prediction values;

[0010] According to the historical database, a coupling influence coefficient between the multi-modal feature parameters is extracted;

[0011] If it is detected that the target drainage system is not affected by the sudden interference factor, the initial comprehensive risk assessment value is corrected according to the abnormal risk correlation coefficient, the risk prediction value and the coupling influence coefficient to obtain a comprehensive risk assessment value one;

[0012] If it is detected that the target drainage system is affected by the sudden interference factor, the variation amplitude of each modal feature parameter under the action of the interference factor is predicted according to the sudden interference factor intensity, the action time length and the influence weight of each multi-modal monitoring data, and the initial comprehensive risk assessment value is corrected according to the variation amplitude, the coupling influence coefficient, the abnormal risk correlation coefficient and the risk prediction value to obtain a comprehensive risk assessment value two;

[0013] According to the comprehensive risk assessment value one or the comprehensive risk assessment value two, an abnormal early warning result of the target drainage system is output.

[0014] Preferably, the multi-modal monitoring data includes flow monitoring data, water quality monitoring data, pipeline state monitoring data and environmental influence data.

[0015] Preferably, key feature parameters affecting the operation state of the target drainage system are extracted from the multi-modal monitoring data, and the specific steps include the following steps:

[0016] In the flow monitoring data, the peak flow, the average flow and the flow fluctuation coefficient affecting the operation state of the target drainage system are extracted;

[0017] In the water quality monitoring data, the pH value, the suspended solids concentration and the chemical oxygen demand affecting the operation state of the target drainage system are extracted;

[0018] In the pipeline state monitoring data, the pipeline pressure, the corrosion degree and the leakage index affecting the operation state of the target drainage system are extracted;

[0019] In the environmental influence data, the rainfall and the soil moisture content affecting the operation state of the target drainage system are extracted.

[0020] Preferably, the abnormal risk value caused by the single multi-modal monitoring data is predicted based on the abnormal risk correlation coefficient and the key feature parameters to obtain a risk prediction value, and the specific steps include the following steps:

[0021] The key feature parameters are subtracted from the normal operation threshold to obtain a feature parameter deviation value;

[0022] The abnormal risk value caused by the single multi-modal monitoring data is predicted according to the abnormal risk correlation coefficient and the feature parameter deviation value to obtain a risk prediction value.

[0023] Preferably, the coupling influence coefficient between the multi-modal characteristic parameters is extracted according to the historical database, and specifically includes the following steps:

[0024] Risk synergistic action data when the multi-modal characteristic parameters jointly act is extracted from the historical database;

[0025] The mutual influence degree between different modal characteristic parameters is calculated to obtain the coupling influence coefficient between the multi-modal characteristic parameters.

[0026] Preferably, if it is detected that the target drainage system is not affected by a sudden interference factor, the initial comprehensive risk assessment value is corrected to obtain a comprehensive risk assessment value one according to the abnormal risk correlation coefficient, the risk prediction value and the coupling influence coefficient, and specifically includes the following steps:

[0027] Parameter characteristics of a stable operation period are proposed from historical operation state data of each mode of the drainage system, and the normal operation reference data and the parameter normal fluctuation range of each modal parameter characteristic are determined through the parameter characteristics of the stable operation period;

[0028] Real-time operation state data of the target drainage system are collected;

[0029] A parameter deviation fluctuation value is calculated by calculating the deviation range of the real-time operation state data and the normal operation reference data;

[0030] If the parameter deviation fluctuation value is within the parameter normal fluctuation range, it is detected that the target drainage system is not affected by a sudden interference factor;

[0031] The weight proportion of each risk prediction value is determined according to the abnormal risk correlation coefficient;

[0032] The coupling influence coefficient is combined to obtain a superposition result by superimposing calculation of each risk prediction value;

[0033] The initial comprehensive risk assessment value is weighted and corrected according to the weight proportion and the superposition result to obtain the comprehensive risk assessment value one.

[0034] Preferably, if it is detected that the target drainage system is affected by a sudden interference factor, the variation amplitude of each modal characteristic parameter under the action of the interference factor is predicted according to the intensity of the sudden interference factor, the action duration and the influence weight of the sudden interference factor on each multi-modal monitoring data, and specifically includes the following steps:

[0035] If the parameter deviation fluctuation value is not within the parameter normal fluctuation range, it is detected that the target drainage system is affected by a sudden interference factor;

[0036] The intensity of the sudden interference factor and the action duration are determined according to the type of the sudden interference factor;

[0037] The influence weight of the interference factor on each multi-modal monitoring data is calculated;

[0038] According to the influence weight, intensity level and action duration, the variation amplitude of each modal characteristic parameter under the action of the interference factor is predicted.

[0039] Preferably, the initial comprehensive risk assessment value is corrected to obtain a second comprehensive risk assessment value according to the variation amplitude, the coupling influence coefficient, the abnormal risk correlation coefficient and the risk prediction value, and the specific steps include the following steps:

[0040] If the sudden interference factor has a synergistic enhancement effect on the multi-modal monitoring data, an enhancement coefficient of the sudden interference factor on the multi-modal monitoring data is extracted from the historical database, and the dynamic coupling coefficient is obtained by multiplying the coupling influence coefficient by the enhancement coefficient;

[0041] If the sudden interference factor has an antagonistic effect on the multi-modal monitoring data, an attenuation coefficient of the sudden interference factor on the multi-modal monitoring data is extracted from the historical database, and the dynamic coupling coefficient is obtained by multiplying the coupling influence coefficient by the attenuation coefficient;

[0042] The initial comprehensive risk assessment value is corrected to obtain a second comprehensive risk assessment value according to the dynamic coupling coefficient, the variation amplitude, the risk prediction value and the abnormal risk correlation coefficient.

[0043] Preferably, the target drainage system abnormal early warning result is output according to the first comprehensive risk assessment value or the second comprehensive risk assessment value, and the specific steps include the following steps:

[0044] The first comprehensive risk assessment value or the second comprehensive risk assessment value is plotted according to the time sequence to obtain a risk change trend curve;

[0045] According to the curve characteristics, the risk value starting period and the risk value sustained high period are marked, and the target drainage system abnormal early warning result is output.

[0046] A multi-modal drainage monitoring data analysis system, comprising:

[0047] The acquisition module acquires multi-modal monitoring data of the target drainage system within a monitoring period;

[0048] The evaluation module evaluates the risk of the target drainage system according to the multi-modal monitoring data to obtain an initial comprehensive risk assessment value;

[0049] The prediction module extracts key characteristic parameters affecting the operation state of the target drainage system from the multi-modal monitoring data, extracts abnormal risk correlation coefficients of different key characteristic parameters and the target drainage system from the historical database, and predicts abnormal risk values caused by single multi-modal monitoring data based on the abnormal risk correlation coefficients and the key characteristic parameters to obtain risk prediction values;

[0050] The extraction module extracts coupling influence coefficients between multi-modal characteristic parameters from the historical database;

[0051] The first correction module corrects the initial comprehensive risk assessment value according to the abnormal risk correlation coefficient, the risk prediction value and the coupling influence coefficient to obtain a comprehensive risk assessment value one if it is detected that the target drainage system is not affected by a sudden interference factor.

[0052] The second correction module corrects the initial comprehensive risk assessment value according to the variation amplitude, the coupling influence coefficient, the abnormal risk correlation coefficient and the risk prediction value to obtain a comprehensive risk assessment value two if it is detected that the target drainage system is affected by a sudden interference factor.

[0053] The warning module outputs the target drainage system abnormal warning result according to the comprehensive risk assessment value one or the comprehensive risk assessment value two.

[0054] Compared with the prior art, the present application has the following beneficial effects:

[0055] The present application quantifies the correlation between parameter abnormalities and system risks by extracting key characteristic parameters and correlating abnormal risk correlation coefficients in the historical database. The abnormal risk value caused by single multi-modal monitoring data is predicted based on the abnormal risk correlation coefficient and the key characteristic parameter to obtain a risk prediction value, which improves the accuracy of risk tracing. When the target drainage system is not affected by a sudden interference factor, the initial comprehensive risk assessment value is corrected according to the abnormal risk correlation coefficient, the risk prediction value and the coupling influence coefficient to obtain a comprehensive risk assessment value one, so that the risk assessment under the conventional operation scene fully considers the interaction between parameters. When the target drainage system is affected by a sudden interference factor, the initial comprehensive risk assessment value is corrected according to the variation amplitude, the coupling influence coefficient, the abnormal risk correlation coefficient and the risk prediction value to obtain a comprehensive risk assessment value two, so that the risk assessment result is more in line with the actual situation, thereby greatly improving the adaptability to variable scenes and effectively avoiding the risk assessment deviation caused by ignoring parameter coupling or sudden interference. The target drainage system abnormal warning result is output according to the comprehensive risk assessment value one or the comprehensive risk assessment value two, so that the operation and maintenance personnel can clearly know the system risk level and risk evolution trend, and then take targeted operation and maintenance measures. BRIEF DESCRIPTION OF DRAWINGS

[0056] Fig. 1 A schematic diagram of a multi-modal drainage monitoring data analysis method is provided for the present application;

[0057] Fig. 2 A schematic diagram of the steps for obtaining a comprehensive risk assessment value two in a multi-modal drainage monitoring data analysis method is provided for the present application;

[0058] Fig. 3A module schematic diagram of a multi-modal drainage monitoring data analysis system is provided for the present application. DETAILED DESCRIPTION

[0059] In order to make the above objectives, features and advantages of the present application more apparent, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0060] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced without the specific details presented, that in other instances, well-known methods have not been described in detail in order to avoid obscuring aspects of the present application. Thus, the specific embodiments of the present application are presented for purposes of illustration and description.

[0061] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. The "in one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an embodiment that is independent of or mutually exclusive with other embodiments.

[0062] Reference Figs. 1-3 is shown.

[0063] Embodiments further illustrate a multi-modal drainage monitoring data analysis system and method provided by the present application.

[0064] A multi-modal drainage monitoring data analysis method, the method comprising the following steps:

[0065] Collecting multi-modal monitoring data of the target drainage system within a monitoring period;

[0066] Evaluating the risk of the target drainage system according to the multi-modal monitoring data to obtain an initial comprehensive risk evaluation value;

[0067] Extracting key characteristic parameters affecting the operating state of the target drainage system from the multi-modal monitoring data, extracting correlation coefficients between different key characteristic parameters and the target drainage system from the historical database, and predicting abnormal risk values caused by single multi-modal monitoring data based on the correlation coefficients and the key characteristic parameters to obtain risk prediction values;

[0068] Extracting coupling influence coefficients between multi-modal characteristic parameters from the historical database;

[0069] If no sudden interference factor is detected in the target drainage system, the initial comprehensive risk evaluation value is corrected according to the correlation coefficients, the risk prediction values and the coupling influence coefficients to obtain a comprehensive risk evaluation value one;

[0070] If it is detected that the target drainage system is affected by a sudden interference factor, the variation range of each modal characteristic parameter under the action of the interference factor is predicted according to the intensity, duration and influence weight of the sudden interference factor on each multi-modal monitoring data, and the initial comprehensive risk assessment value is corrected to obtain a comprehensive risk assessment value two according to the variation range, coupling influence coefficient, abnormal risk correlation coefficient and risk prediction value.

[0071] The abnormal early warning result of the target drainage system is output according to the comprehensive risk assessment value one or the comprehensive risk assessment value two.

[0072] The multi-modal monitoring data includes flow monitoring data, water quality monitoring data, pipeline state monitoring data and environmental influence data.

[0073] The key characteristic parameters affecting the operation state of the target drainage system are extracted from the multi-modal monitoring data, specifically including the following steps:

[0074] The peak flow, average flow and flow fluctuation coefficient affecting the operation state of the target drainage system are extracted from the flow monitoring data.

[0075] The pH value, suspended solids concentration and chemical oxygen demand affecting the operation state of the target drainage system are extracted from the water quality monitoring data.

[0076] The pipeline pressure, corrosion degree and leakage index affecting the operation state of the target drainage system are extracted from the pipeline state monitoring data.

[0077] The rainfall and soil moisture content affecting the operation state of the target drainage system are extracted from the environmental influence data.

[0078] The multi-modal monitoring data of the target drainage system is collected in a set monitoring period, including flow monitoring data, water quality monitoring data, pipeline state monitoring data and environmental influence data, and the current risk status of the target drainage system is determined based on the collected multi-modal monitoring data to obtain an initial comprehensive risk assessment value.

[0079] The flow monitoring data includes peak flow, average flow and flow fluctuation coefficient. For example, the peak flow of a city drainage system during a rainstorm period reaches 1.2 times of the design load, which is out of the normal range, indicating that the drainage pressure increases dramatically in a short time, so there is a risk of pipeline overload. If the average flow is continuously higher than the historical average, it reflects that the system is running at high load for a long time, so there is a risk of accelerating the wear of equipment. If the flow fluctuation coefficient is too large, it means that the flow is unstable and is easy to cause hydraulic impact of the pipeline. For example, the flow fluctuation coefficient of a certain industrial area reaches 0.8, which is higher than the threshold of 0.3. Compare these parameters with the system design standard and historical normal operation interval, and quantify the risk value of the flow dimension, such as setting the peak flow to be 10%-20% higher than the design value, then the flow risk score is 3 out of 5, and a flow risk assessment sub-model is constructed.

[0080] The water quality monitoring data includes pH value, suspended solids concentration and chemical oxygen demand. If the pH value drops to 5.0, the normal range of pH value is 6.5-8.5, and the acid is too strong to corrode the inner wall of the pipeline; the suspended solids concentration exceeds the standard, for example, the suspended solids concentration reaches 200 mg / L, and the normal standard of suspended solids concentration is 80 mg / L, so this situation is easy to accumulate in the pipe elbow and inspection well, thereby blocking the drainage channel; high chemical oxygen demand indicates high organic pollutant load, for example, the measured chemical oxygen demand is 600 mg / L, which is much higher than the standard of 300 mg / L, so it increases the difficulty of sewage treatment and environmental risk. According to the correlation between water quality parameters, pipeline corrosion, blockage probability and impact on receiving water, a water quality risk assessment sub-model is established to convert the abnormal degree of parameters into risk score, such as 1 times of chemical oxygen demand exceeding the standard, the water quality risk score is 4.

[0081] The pipeline state monitoring data includes pipeline pressure, corrosion degree and leakage index. For example, the corrosion degree detection shows that the pipe wall thickness loss is 30%, and reaches the corrosion warning threshold, which increases the risk of pipeline damage and leakage; the pipeline pressure appears abnormal fluctuation, such as local pressure drop of 20%, which indicates that there may be hidden leakage points in the pipeline; the leakage index exceeds the set threshold, for example, the leakage index of a certain section of pipeline reaches 15, and the normal leakage index should be less than 10, which means that the water leakage has affected the normal drainage efficiency of the system. Based on the historical data of pipeline operation and maintenance, damage and leakage accident cases, a pipeline state risk assessment sub-model is constructed to convert the pipeline state parameters into risk quantitative value, such as 3.5 points for pipeline risk score when the corrosion degree reaches 30% loss.

[0082] The environmental impact data includes rainfall and soil moisture content. For example, if the rainfall accumulates to 150 mm in 24 hours during continuous heavy rain weather, which exceeds the rainfall of 100 mm corresponding to the design recurrence period of the drainage system, the drainage system faces the risk of external water backflow and waterlogging; if the soil moisture content is too high, such as 80%, and the saturated moisture content is 85%, the soil's ability to regulate and store groundwater will be weakened, and the hydrological environment around the drainage system will be deteriorated. The environmental impact risk assessment sub-model is established by combining the urban hydrological model and historical waterlogging events, and the environmental parameter anomaly degree is converted into a risk score, such as 4 points when the rainfall exceeds the design value by 50%.

[0083] The risk weights of the flow monitoring data, water quality monitoring data, pipeline state monitoring data, and environmental impact data are determined through historical conditions, such as 30% for flow monitoring data, 25% for water quality monitoring data, 30% for pipeline state monitoring data, and 15% for environmental impact data. The risk scores output by each sub-model are weighted and summed according to the weights, such as 3 points x 30% + 4 points x 25% + 3.5 points x 30% + 4 points x 15% = 3.55 points, and the initial comprehensive risk assessment value is finally obtained.

[0084] The key feature parameters that play a key role in the operation state of the drainage system are selected from the multi-modal monitoring data, such as peak flow, average flow, and flow fluctuation coefficient in the flow data, pH value, suspended solids concentration, and chemical oxygen demand in the water quality data. The abnormal changes of these parameters will directly affect the system stability. At the same time, the historical database is called to mine the correlation between different key feature parameters and the abnormal risk of the drainage system, thereby forming an abnormal risk correlation coefficient, which quantifies the risk condition of the system caused by the abnormality of a single parameter. Based on the abnormal risk correlation coefficient and the key feature parameters, the abnormal risk value caused by a single multi-modal monitoring data is predicted to obtain a risk prediction value.

[0085] The coupling influence coefficient between multi-modal feature parameters is extracted from the historical database, which reflects the synergistic or offsetting effect of different modal parameters on the system risk when they interact with each other, such as the increase of rainfall changing the flow and thereby affecting the pipeline pressure.

[0086] If there is no sudden disturbance to the drainage system, it means that the system is running in a normal state. At this time, the weights of each risk prediction value are determined according to the abnormal risk correlation coefficient, and the interaction between parameters is considered by combining the coupling influence coefficient. The initial comprehensive risk assessment value is modified by superimposing and weighting each risk prediction value to obtain a comprehensive risk assessment value, which makes the risk assessment more consistent with the actual action relationship between parameters.

[0087] If a burst interference factor is detected, such as extreme weather and unexpected construction, first determine the intensity, duration and influence weight of the interference factor on each modal monitoring data, so as to predict the variation range of each modal characteristic parameter under the interference. Combine the coupling influence coefficient, abnormal risk correlation coefficient and risk prediction value to adjust the initial comprehensive risk assessment value, so as to obtain the comprehensive risk assessment value two.

[0088] The coupling influence coefficient needs to rely on historical database and quantitative analysis. The core is to mine the interaction relationship between different modal characteristic parameters. The specific process is as follows:

[0089] The risk synergistic action data of the multi-modal characteristic parameters is extracted from the historical database. The historical database stores a large amount of operation data of the drainage system under different working conditions, weather, time periods, including multi-modal monitoring records of flow, water quality, pipeline state, environmental influence, and corresponding system failure or abnormal event cases. When extracting, the scene data of multiple parameters acting at the same time need to be selected, for example, the associated data that the increase of rainfall under heavy rain weather leads to the increase of flow peak value, and then causes the abnormality of pipeline pressure, or the synergistic record that the increase of chemical oxygen demand during industrial pollution and the acceleration of pipeline corrosion degree. These data need to include the specific values of each modal characteristic parameter, the action time span, and the final induced system risk performance (such as blockage, leakage, overload), so as to ensure that the data can fully reflect the mutual influence process between parameters.

[0090] Based on the extracted risk synergistic action data, the mutual influence degree between different modal characteristic parameters is quantitatively calculated, and finally the coupling influence coefficient is obtained. This process needs to combine statistical methods and drainage system operation rules. First, the data is preprocessed to remove outliers and complete missing data to ensure data reliability; then the correlation analysis and regression analysis are used to establish the quantitative correlation model between parameters. For example, the mutual influence between flow peak value and pipeline pressure is analyzed through linear regression analysis of the change relationship between the two in historical data. If multiple records show that the flow peak value increases by 10% and the pipeline pressure increases by 15% on average, and the correlation exists stably in different scenes, the influence strength of the two can be quantified. If indirect action (such as water quality pH value) affects pipeline pressure by corroding pipeline, a structural equation model is used to sort out the multi-path action relationship, calculate the influence coefficient of each path and integrate the weighted values. Finally, the mutual influence degree is converted into a coupling influence coefficient in the range of 0-1 or 0-2. The closer the coupling influence coefficient is to 2, the stronger the synergistic action is, and the closer the coupling influence coefficient is to 0, the weaker the mutual influence is. Negative values can represent antagonistic action, for example, the coupling influence coefficient of flow peak value and pipeline pressure is set to 0.8, and the coupling influence coefficient of pH value and pipeline corrosion degree is set to 1.2, so as to accurately quantify the coupling effect between different modal parameters.

[0091] According to the comprehensive risk assessment value one or the comprehensive risk assessment value two, an abnormal early warning result of the target drainage system is output. By comparing the modified risk value with a preset risk threshold value, if the threshold value is exceeded, a corresponding level of early warning is triggered.

[0092] Based on the abnormal risk correlation coefficient and the key characteristic parameter, an abnormal risk value caused by single multi-modal monitoring data is predicted to obtain a risk prediction value, specifically including the following steps:

[0093] The key characteristic parameter is calculated by difference with the normal operation threshold value to obtain a characteristic parameter deviation value.

[0094] According to the abnormal risk correlation coefficient and the characteristic parameter deviation value, an abnormal risk value caused by single multi-modal monitoring data is predicted to obtain a risk prediction value.

[0095] For each type of multi-modal monitoring data corresponding to the key characteristic parameter, its normal operation threshold value is determined. For example, the normal threshold value of the peak flow in the flow monitoring data is determined according to the design bearing capacity of the drainage pipeline, and it is assumed that the peak flow normal threshold value of the drainage pipeline in a certain area is 500 m³ / h; the pH value normal threshold value in the water quality monitoring data is between 6.5-8.5.

[0096] The deviation value of the key characteristic parameter and the normal operation threshold value is calculated. Taking the peak flow of the flow monitoring data as an example, if the currently monitored peak flow is 600 m³ / h, and the normal threshold value is 500 m³ / h, then the deviation value of the peak flow is 600-500=100 m³ / h; if the pH value in the water quality monitoring is 5.8, and the lower limit of the normal threshold value is 6.5, then the deviation value is 5.8-6.5=-0.7, and the negative deviation indicates that it is lower than the normal threshold value, and the same reflects the parameter abnormality. The deviation value quantifies the degree of deviation of the key characteristic parameter from the normal state, and the larger the deviation value, the more significant the parameter abnormality.

[0097] The abnormal risk correlation coefficient is introduced from the historical database, which is obtained through the backtracking analysis of a large number of drainage system failure cases and abnormal events. It reflects the close degree of the specific key characteristic parameter abnormality and the abnormal risk of the drainage system. For example, according to the historical data, the abnormal risk correlation coefficient of the peak flow is 0.8; and the abnormal risk correlation coefficient of the water quality pH value is 0.6, because the triggering mechanism of the pH value abnormality to the system risk is relatively complex, therefore the correlation degree is slightly weak.

[0098] The abnormal risk value caused by single multi-modal monitoring data is predicted by combining the abnormal risk correlation coefficient and the characteristic parameter deviation value. The risk prediction value is obtained by multiplying the two. Taking peak flow as an example, if the deviation value is 100 m³ / h and the correlation coefficient is 0.8, then the risk prediction value is 100 x 0.8 = 80; if the pH value deviation value is -0.7 and the correlation coefficient is 0.6, then the risk prediction value may be |-0.7| x 0.6 = 0.42. The abnormal degree of the key characteristic parameters is converted into an abnormal risk value.

[0099] The coupling influence coefficient between multi-modal characteristic parameters is extracted from the historical database, including the following steps:

[0100] Risk synergy data when multi-modal characteristic parameters act together is extracted from the historical database.

[0101] The coupling influence coefficient between multi-modal characteristic parameters is calculated by calculating the mutual influence degree between different modal characteristic parameters.

[0102] Risk synergy data when multi-modal characteristic parameters act together is extracted from the historical database. The historical database stores a large amount of data of the past operation of the drainage system, including parameter records during normal operation and abnormal failure. Scene data of different modal parameters appearing abnormal or specific combination is screened out. For example, during a heavy rainfall, the peak flow monitoring shows that the peak flow exceeds the standard, at the same time, the suspended solids concentration in the water quality increases, and the pipe state shows that the pressure is abnormal. Under the joint action of these parameters, the drainage system has a record of pipe blockage and overflow failure. This kind of data is risk synergy data.

[0103] The coupling influence coefficient is calculated by calculating the mutual influence degree between different modal characteristic parameters. Based on the extracted risk synergy data, the correlation analysis is used to quantify the mutual influence between different modal parameters. For example, the relationship between peak flow exceeding the standard and pipe pressure abnormality is determined. If the peak flow increases by 10% and the pipe pressure rises by 15% on average, and the stable correlation is shown in multiple data sets, then the mutual influence degree is high. If the pH value decreases by 1, the pipe corrosion degree is aggravated by 20% on average within half a year. By constructing a mathematical model for multiple combinations of different modal parameters, the coupling influence coefficient reflecting the mutual influence degree is calculated.

[0104] Taking the structural equation model as an example, the peak flow has a direct impact on the pipe pressure, and the suspended solids concentration of water quality will affect the pipe corrosion and then indirectly act on the pipe pressure. At this time, the path coefficients of each path need to be integrated and weighted. Assuming that the path coefficient of the direct impact of the peak flow on the pipe pressure in the structural equation model is 0.6, and the path coefficient of the indirect impact of the suspended solids concentration of water quality on the pipe pressure through affecting the pipe corrosion is 0.2, these path coefficients respectively quantify the influence degree of parameters under different action paths. Then, according to the importance of each path in the overall action, the corresponding weight is given, for example, the weight of direct impact is set to 0.7, and the weight of indirect impact is set to 0.3. The weight is determined through historical data, and the sum of the path coefficients multiplied by the corresponding weight is 0.6×0.7+0.2×0.3=0.48, that is, the overall coupling influence coefficient of the multi-modal parameters is 0.48. The greater the coefficient value is, the stronger the interaction between the corresponding modal parameters is, and the comprehensive reflection of the synergistic influence degree of multi-modal parameters on the drainage system risk under the complex action of multiple parameters and multiple paths.

[0105] If it is detected that the target drainage system is not affected by sudden interference factors, the initial comprehensive risk assessment value is corrected to obtain a comprehensive risk assessment value one according to the abnormal risk correlation coefficient, the risk prediction value and the coupling influence coefficient, which specifically includes the following steps:

[0106] From the parameter characteristics of the stable operation period of the historical operation state data of each mode of the drainage system, the normal operation reference data and the parameter normal fluctuation range of each modal parameter characteristic are determined through the parameter characteristics of the stable operation period;

[0107] Collect real-time operation state data of the target drainage system;

[0108] Calculate the deviation range of the real-time operation state data and the normal operation reference data to obtain the parameter deviation fluctuation value;

[0109] If the parameter deviation fluctuation value is within the parameter normal fluctuation range, it means that the target drainage system is not affected by sudden interference factors:

[0110] Taking the flow parameter as an example, the parameter characteristics of the stable operation period are obtained from the flow historical operation state data of the drainage system, and the normal operation reference data of the flow parameter characteristics is 120 cubic meters per hour, and the normal fluctuation range is ±15 cubic meters per hour.

[0111] With the help of the flow sensor, the real-time operation state data of the target drainage system is collected, and the flow situation of the target drainage system is mastered in real time.

[0112] The parameter deviation fluctuation value is calculated by comparing the collected real-time running state data with the normal running benchmark data. Assuming that the real-time running state data of the flow rate collected in real time is 128 cubic meters per hour, the parameter deviation fluctuation value is calculated as 8 cubic meters per hour by comparing the flow rate real-time running state data with the normal running benchmark of 120 cubic meters.

[0113] The parameter deviation fluctuation value is compared with the parameter normal fluctuation range. It can be judged that the flow rate real-time running state data of 8 cubic meters per hour is within the parameter normal fluctuation range, indicating that the current running state of the target drainage system conforms to the normal situation, i.e., there is no sudden interference factor of flow rate surge acting on the target drainage system.

[0114] According to the abnormal risk correlation coefficient, the weight proportion of each risk prediction value is determined;

[0115] The coupling influence coefficient is combined to perform superposition calculation on each risk prediction value to obtain a superposition result;

[0116] According to the weight proportion and the superposition result, the initial comprehensive risk assessment value is weighted and corrected to obtain a comprehensive risk assessment value one.

[0117] The abnormal risk correlation coefficient reflects the close degree of correlation between the abnormality of the multi-modal characteristic parameter and the risk of the drainage system. For example, the abnormal risk correlation coefficient of the flow rate peak value is 0.8, and the correlation coefficient of the water quality pH value is 0.6.

[0118] When calculating the weight, the abnormal risk correlation coefficients of each parameter are first normalized. Assuming that only two parameters, the flow rate peak value (C1=0.8) and the water quality pH value (C2=0.6), are considered, the total correlation coefficient sum is 0.8+0.6=1.4, then the weight of the flow rate peak value =0.8÷1.4≈0.57, and the weight of the water quality pH value =0.6÷1.4≈0.43. In this way, the abnormal risk correlation coefficient is converted into the weight proportion of each risk prediction value.

[0119] The coupling influence coefficient reflects the synergistic effect between multi-modal parameters. For example, a too high flow rate peak value can exacerbate the pipe pressure abnormality, and the coupling influence coefficient of the two is 0.7. Assuming that the risk value predicted based on the flow rate peak value is =6, and the risk value predicted based on the water quality pH value is =4. When performing superposition calculation, the coupling effect between parameters needs to be considered. The superposition result of the single parameter risk is calculated as + =10, and then the coupling influence coefficient is introduced for adjustment. If the coupling effect increases the risk synergy, the superposition formula can be set as , where is the coupling influence coefficient, and this =0.7, then =6+4+0.7×6×4=6+4+16.8=26.8. This method allows the superposition result to include both single-parameter risks and the amplification or reduction of risks through the synergistic effect of multiple parameters, better reflecting the interaction patterns of parameters in actual drainage system operation.

[0120] For example, the initial comprehensive risk assessment value =20, and a weighted correction formula is constructed based on the weight ratio and the superposition result during the correction.

[0121] The corrected formula can be set as follows: ,in, The correction coefficient reflects the degree to which the superposition result adjusts the initial value. Let these be the weighting coefficients for the coupling effect, assuming... =0.6、 =0.3, then =15.908. Through such weighted correction, the parameter weights, coupling effects and initial risk values ​​are integrated to obtain a comprehensive risk assessment value of 15.908. This allows the risk assessment to reflect the true risk level under the synergistic effect of multiple parameters in the drainage system, providing a reliable basis for subsequent anomaly warning.

[0122] Correction coefficient and the weighting coefficient of the coupling effect This process involves fusing statistical data to ensure that these two types of coefficients reflect the actual relationships between various factors in the drainage system. It begins with the collection and processing of large-scale historical data. This data includes operational records of the drainage system under different seasons, weather conditions, flow loads, and pipe aging levels. This data includes multi-dimensional information such as flow rate curves, water quality index fluctuations, pipe pressure monitoring results, and the frequency and type of faults. It also needs to be correlated with corresponding environmental parameters, such as rainfall, temperature, and water usage patterns in the surrounding area. This data undergoes rigorous preprocessing, including outlier removal, missing value imputation, and time series alignment, to ensure the reliability of the analytical basis.

[0123] After data preparation, in-depth data analysis was conducted using various statistical methods. (Regarding the correction coefficient...) Its main purpose is to adjust for the bias of a single factor in risk assessment. Therefore, regression analysis is often used to establish a quantitative relationship between various basic parameters and actual risk results. For example, by using multiple linear regression or nonlinear regression models, the deviation between a specific parameter, the degree of aging of pipeline material, and the actual failure rate can be analyzed, thereby determining... The initial range of values. This process, combined with hypothesis testing, verifies the significance of the parameter's influence, ensuring... The direction and magnitude of the adjustment conform to the actual data patterns.

[0124] Coupling effect weight coefficient It is necessary to reflect the synergistic or antagonistic effects among multimodal parameters. Structural equation modeling is typically used to construct the interaction network between parameters. For example, to determine the cumulative impact of rainfall, pipeline flow rate, and water pollution level on blockage risk, the influence path coefficients of each parameter combination are calculated to determine the impact under different coupling scenarios. Weight allocation. Machine learning algorithms (such as random forests or neural networks) are introduced to train the model and identify higher-order interactions between parameters, further optimizing the process. The precision of the value.

[0125] Correction coefficient By comparing with the initial risk value product term This enables linear correction of biases in single-factor assessments, such as when there are systematic errors in the monitoring data of a certain parameter. The value of this will be increased accordingly to weaken its impact on the result. Coupling effect weighting coefficient Then through integrated coupling risk and weighted sum of modal parameters The product of these parameters transforms the synergistic effects among multiple parameters into quantifiable increases or decreases in risk. For example, when peak flow occurs simultaneously with high pollutant concentrations, The value of is higher than that of a single parameter, to reflect the amplified effect of the cumulative risk. This quantification method allows the risk assessment results to overcome the limitations of single-parameter analysis and more comprehensively reflect the complex operating state of the drainage system.

[0126] If a sudden disturbance is detected in the target drainage system, the variation amplitude of each modal characteristic parameter under the influence of the disturbance is predicted based on the intensity, duration, and impact weight of the disturbance on each multimodal monitoring data. This includes the following steps:

[0127] If the parameter deviation fluctuation value is not within the normal fluctuation range of the parameter, it indicates that a sudden interference factor has been detected in the target drainage system.

[0128] Taking flow parameters as an example, the parameter characteristics of the stable operating period are extracted from the historical operating status data of the drainage system. Based on the parameter characteristics of the stable operating period, the normal operating benchmark data of the flow parameter characteristics is determined to be 120 cubic meters per hour, and the normal fluctuation range is ±15 cubic meters per hour.

[0129] By using flow sensors to collect real-time operating status data of the target drainage system, the flow rate of the target drainage system can be monitored in real time.

[0130] The parameter deviation fluctuation value is calculated by comparing the collected real-time running state data with the normal running benchmark data. Assuming that the real-time running state data of the flow rate collected in real time is 160 cubic meters per hour, the parameter deviation fluctuation value is calculated as 40 cubic meters per hour by comparing the real-time running state data of the flow rate with the normal running benchmark of 120 cubic meters.

[0131] The parameter deviation fluctuation value is compared with the parameter normal fluctuation range. It can be judged that the real-time running state data of the flow rate of 40 cubic meters per hour is not within the parameter normal fluctuation range, indicating that the current running state of the target drainage system does not conform to the normal situation, i.e., the sudden interference factor of flow rate surge acts on the target drainage system.

[0132] The intensity and duration of the sudden interference factor are determined according to the type of the sudden interference factor.

[0133] The influence weight of the interference factor on each multi-modal monitoring data is calculated.

[0134] The variation amplitude of each modal characteristic parameter under the action of the interference factor is predicted according to the influence weight, intensity level, and duration.

[0135] When the drainage system encounters a sudden interference, first, the intensity and duration of the interference are determined according to the type of the interference factor, such as the type of the interference factor is the unexpected discharge of high-concentration wastewater by the factory, combined with historical similar events or industry standards. For example, it is determined that the chemical oxygen demand (COD) is 5000 mg / L by detecting the concentration of wastewater at the sewage outlet, i.e., 500 mg / L higher than the normal discharge, and the intensity level is determined to be severe pollution discharge; at the same time, it is determined that the sewage lasts for 4 hours according to the progress of the factory stop-discharge rectification.

[0136] The influence weight of the interference on the multi-modal monitoring data is calculated. The multi-modal monitoring data includes flow rate monitoring data, water quality monitoring data, pipeline state monitoring data, and environmental impact data: water quality data is directly affected by sewage discharge, so the influence weight is high; the flow rate data may be temporarily increased due to the injection of wastewater, but the proportion of sewage discharge to the system design flow rate is small, so the influence weight is second; the pipeline state data (such as the degree of corrosion) is indirectly affected by the change of water quality, so the influence weight is relatively low. The frequency of system failure caused by abnormality of each modal data is statistically calculated based on the data of historical industrial sewage events. Assuming that in the past 10 times of similar sewage, 7 times were caused by water quality abnormality, 2 times were caused by temporary fluctuation of flow rate, and 1 time was caused by water quality corrosion, the influence weights of water quality, flow rate, and pipeline state are about 0.7, 0.2, and 0.1 respectively after normalization processing (7 / 10, 2 / 10, 1 / 10), which quantifies the influence degree of the interference on different modal data.

[0137] The variation range of each modal characteristic parameter is predicted based on the influence weight, intensity level and action duration. Taking water quality data as an example, the normal average COD of water quality is 30 mg / L, the relationship model of the COD concentration increment corresponding to the disturbance intensity and the action duration is known, such as the concentration increment = intensity coefficient x action duration, the intensity coefficient is 1000 mg / L h, and the water quality influence weight is 0.7; the normal average flow of the flow data is 500 m³ / h, the disturbance causes the flow increment to be associated with the pollution discharge, such as the pollution discharge is 200 m³ / h, the flow influence weight is 0.2, and then the flow variation range is 200 m³ / h x 0.2 = 40 m³ / h, and it is predicted that the flow is increased to 540 m³ / h; the normal corrosion rate of the pipeline state data is 0.1 mm / year, the corrosion acceleration caused by the change of water quality is associated with the COD increment, such as the corrosion rate is increased by 0.05 mm / year for each 1000 mg / L increase of the COD, and the corrosion rate variation range is calculated to be (2800-30) mg / L ÷ 1000 mg / L x 0.05 mm / year x 0.1 = 0.014 mm / year, that is, the corrosion rate is predicted to be increased to 0.114 mm / year.

[0138] The initial comprehensive risk assessment value is corrected to obtain the comprehensive risk assessment value two according to the variation range, the coupling influence coefficient, the abnormal risk correlation coefficient and the risk prediction value, and the specific steps include the following steps.

[0139] If the sudden disturbance factor has a synergistic enhancement effect on the multi-modal monitoring data, the enhancement coefficient of the sudden disturbance factor on the multi-modal monitoring data is extracted from the historical database, the coupling influence coefficient is multiplied by the enhancement coefficient to obtain the dynamic coupling coefficient;

[0140] If the sudden disturbance factor has an antagonistic effect on the multi-modal monitoring data, the attenuation coefficient of the sudden disturbance factor on the multi-modal monitoring data is extracted from the historical database, the coupling influence coefficient is multiplied by the attenuation coefficient to obtain the dynamic coupling coefficient;

[0141] The initial comprehensive risk assessment value is corrected to obtain the comprehensive risk assessment value two according to the dynamic coupling coefficient, the variation range, the risk prediction value and the abnormal risk correlation coefficient.

[0142] When the drainage system encounters a sudden disturbance, if the sudden disturbance is an extreme rainstorm, firstly, the type of the effect of the disturbance on the multi-modal monitoring data is judged. If it is a synergistic enhancement effect, that is, the rainstorm causes the flow to surge, and at the same time, the surface pollutants are washed to increase the water quality suspended solids concentration, the flow and the water quality anomalies are mutually intensified to aggravate the system risk, then the corresponding enhancement coefficient is extracted from the historical database. For example, the synergistic enhancement coefficient of the flow and the water quality parameters in the historical extreme rainstorm event is 1.5. If the coupling coefficient of the flow peak value and the suspended solids concentration is 0.6, the dynamic coupling coefficient is obtained by multiplying the coupling influence coefficient by the enhancement coefficient, which is 0.9, which reflects the intensification of the synergistic effect of the parameters under the disturbance.

[0143] If the disturbance produces antagonistic effects, such as sudden construction disturbance reduces the groundwater level by construction dewatering, although it may alleviate the risk of pipeline leakage, but the vibration of construction machinery may exacerbate the damage of pipeline, the two effects offset each other, then the attenuation coefficient is extracted. For example, the antagonistic attenuation coefficient of pipeline pressure and leakage risk in historical construction disturbance is 0.7, and the dynamic coupling coefficient is 0.56, which is obtained by multiplying the original coupling effect coefficient by the attenuation coefficient, reflecting the offset effect of parameter action under disturbance.

[0144] If extreme rainstorm makes the fluctuation range of flow peak value reach 80%, and the fluctuation range of suspended solids concentration reach 50%, the abnormal risk correlation coefficient of flow peak value is 0.8, and that of suspended solids concentration is 0.6, the predicted value of flow abnormal risk is 6, and that of water quality abnormal risk is 4, the initial comprehensive risk assessment value is corrected. When calculating, the weight of each risk prediction value is determined according to the abnormal risk correlation coefficient (flow weight = 0.8 ÷ (0.8 + 0.6) ≈ 0.57, water quality weight ≈ 0.43), and then the risk prediction value is superimposed according to the weight and coupling coefficient by considering the synergistic effect of parameters (6 × 0.57 × 0.9 + 4 × 0.43 × 0.9 = 3.078 + 1.548 = 4.626) combined with dynamic coupling coefficient. Finally, the initial comprehensive risk assessment value (such as initial value is 8) is corrected by integrating multiple dimensional parameters such as fluctuation range and risk prediction value. For example, the influence factor of fluctuation range (flow fluctuation range 80% corresponds to influence factor 1.2, water quality fluctuation range 50% corresponds to influence factor 1.1), and the comprehensive risk assessment value two = initial value × (1 + superposition result × influence factor average), that is, 8 × (1 + 4.626 × (1.2 + 1.1) ÷ 2) ≈ 42.56, which dynamically reflects the change of system risk under sudden disturbance.

[0145] In the comprehensive risk assessment value two = initial value × (1 + superposition result × influence factor average), the average of influence factor is selected for calculation. From the practical application point of view, the influence factor is often affected by many complex and dynamic factors, such as the uncertainty of the environment where the drainage system is located and the fluctuation of monitoring data, which leads to the fact that the influence factor itself is not a fixed value, but shows certain variability and randomness. If the specific value of the influence factor at a certain moment or under certain conditions is directly used, it is likely that the calculated comprehensive risk assessment value two will deviate greatly due to accidental factors or local conditions, which cannot accurately reflect the overall and long-term risk status of the drainage system. By calculating the average of the influence factor, the information of the influence factor under different times and different conditions can be integrated, so as to more stably and comprehensively reflect the average effect of the influence factor on the comprehensive risk, and make the calculation result of the comprehensive risk assessment value two more representative and reliable, which helps to more accurately assess the risk level of the drainage system and provide a more solid basis for subsequent abnormal warning and other work.

[0146] The changes of the multi-modal characteristic parameters (flow, water quality, pipe state, etc.) of the drainage system under different working conditions (such as encountering different intensity rainstorms, different scale sudden sewage, and various types of sudden interference scenes) and the corresponding actual risk change data need to be collected. Then, the risk changes when each factor acts alone are compared with the risk changes when they act together: if the risk change under the joint action is greater than the superposition sum of the risk change of each factor acting alone, it is determined to be synergistically enhanced; if the risk change under the joint action is less than the superposition sum of the risk change of each factor acting alone, it is determined to be antagonistically offset. For example, the interaction of each factor is analyzed by using multiple linear regression.

[0147] If the risk change under the joint action is greater than the superposition sum of the risk change of each factor acting alone, it is determined to be synergistically enhanced. For example, considering the two factors of rainstorm and urban ground hardening, it is assumed that rainstorm alone will increase the risk of the drainage system by 40%, and urban ground hardening alone will increase the risk by 30%. According to the superposition calculation, the total of the risk change of the two factors acting alone should be 70%. But in reality, the risk of the drainage system under the joint action of rainstorm and urban ground hardening is 90%. This indicates that there is a synergistic effect between rainstorm and urban ground hardening. Ground hardening leads to reduced rainwater infiltration and increased surface runoff, which makes the drainage system face greater water pressure during rainstorms, thus greatly increasing the risk beyond the superposition effect of the two factors acting alone, further amplifying the risk of the drainage system.

[0148] If the risk change under the joint action is less than the superposition sum of the risk change of each factor acting alone, it is determined to be antagonistically offset. For example, considering the two factors of urban rainwater permeable pavement paving and drainage pipe network dredging and maintenance, it is assumed that urban rainwater permeable pavement paving alone will reduce the risk of the drainage system by 22%, and drainage pipe network dredging and maintenance alone will reduce the risk by 18%. According to the superposition calculation, the total of the risk change of the two factors acting alone should be 40%. But in reality, the comprehensive risk of the drainage system under the joint action of urban rainwater permeable pavement paving and drainage pipe network dredging and maintenance is reduced by 32%. This indicates that there is an antagonistic effect between rainstorm and urban ground hardening. The reduced effect of rainwater permeable pavement will weaken the efficiency of pipe network dredging and maintenance. The reduced amount of rainwater entering the pipe network due to the permeable pavement will reduce the actual flow pressure even if the pipe network is not deeply dredged, thus weakening the risk reduction effect of the joint action of the two factors. It is these mutually restrictive and overlapping factors that make the risk reduction amount when rainwater permeable pavement paving and drainage pipe network dredging and maintenance act together less than the superposition sum of the risk change of each factor acting alone, forming an antagonistic offset phenomenon.

[0149] For the acquisition of dynamic coupling coefficient, the variation range of related factors and the change rule of such relationship risk in history are combined to calculate. Taking synergistic enhancement as an example, the change proportion relationship between risk and variation range of related factors in the synergistic enhancement scenario is mined from historical data to determine the calculation method of enhancement coefficient (such as linear relationship, exponential relationship between enhancement coefficient and variation range), and then the dynamic coupling coefficient is obtained by multiplying the enhancement coefficient and the original coupling influence coefficient; if it is antagonistic, the calculation method of attenuation coefficient is determined from the historical data to obtain the dynamic coupling coefficient.

[0150] According to the comprehensive risk assessment value one or the comprehensive risk assessment value two, an abnormal early warning result of the target drainage system is output, specifically including the following steps:

[0151] The comprehensive risk assessment value one or the comprehensive risk assessment value two is plotted according to time sequence to obtain a risk change trend curve;

[0152] According to the curve characteristics, a risk value starting period and a risk value continuously high period are marked, and an abnormal early warning result of the target drainage system is output.

[0153] The comprehensive risk assessment value one or the comprehensive risk assessment value two is sorted according to time sequence. For example, the comprehensive risk assessment values calculated at different times are arranged in sequence to form a time sequence at a fixed time interval. A risk change trend curve is plotted with time as the horizontal coordinate and the comprehensive risk assessment value as the vertical coordinate. The curve intuitively shows how the risk of the drainage system fluctuates in the monitoring period.

[0154] The trend of the curve is observed to determine the risk value starting period, that is, the time node at which the curve deviates from the normal risk level and shows an upward trend. This period marks the beginning of abnormal changes in the risk of the drainage system. At the same time, the risk value continuously high period is identified, that is, the period in which the risk value in the curve is maintained at a high level for a long time and exceeds the normal range.

[0155] According to the marked risk value starting period and continuously high period, the abnormal early warning result of the target drainage system is output in combination with the risk early warning rules of the drainage system.

[0156] A multi-modal drainage monitoring data analysis system, comprising:

[0157] The acquisition module acquires multi-modal monitoring data of the target drainage system in a monitoring period;

[0158] The evaluation module evaluates the risk of the target drainage system according to the multi-modal monitoring data to obtain an initial comprehensive risk assessment value;

[0159] The prediction module: extracts key characteristic parameters affecting the operation state of the target drainage system in the multi-modal monitoring data, extracts the abnormal risk correlation coefficients of different key characteristic parameters and the target drainage system according to the historical database, and predicts the abnormal risk values caused by single multi-modal monitoring data based on the abnormal risk correlation coefficients and the key characteristic parameters to obtain risk prediction values;

[0160] The extraction module: extracts the coupling influence coefficients between the multi-modal characteristic parameters according to the historical database;

[0161] The first correction module: if it is detected that the target drainage system is not affected by a sudden interference factor, the initial comprehensive risk assessment value is corrected to obtain a comprehensive risk assessment value one according to the abnormal risk correlation coefficients, the risk prediction values and the coupling influence coefficients;

[0162] The second correction module: if it is detected that the target drainage system is affected by a sudden interference factor, the variation amplitudes of the multi-modal characteristic parameters under the action of the interference factor are predicted according to the intensity of the sudden interference factor, the action time length and the influence weight of each multi-modal monitoring data, and the initial comprehensive risk assessment value is corrected to obtain a comprehensive risk assessment value two according to the variation amplitudes, the coupling influence coefficients, the abnormal risk correlation coefficients and the risk prediction values;

[0163] The early warning module: outputs the abnormal early warning result of the target drainage system according to the comprehensive risk assessment value one or the comprehensive risk assessment value two.

[0164] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and the necessary general hardware platform, and of course, it can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the method described in each embodiment or some parts of the embodiment.

[0165] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A multi-modal drainage monitoring data analysis method, characterized in that, The method comprises the following steps: Collecting multi-modal monitoring data of the target drainage system in a monitoring period, wherein the multi-modal monitoring data comprises flow monitoring data, water quality monitoring data, pipeline state monitoring data and environmental impact data; Evaluating the risk of the target drainage system according to the multi-modal monitoring data to obtain an initial comprehensive risk evaluation value; Extracting key characteristic parameters affecting the operation state of the target drainage system from the multi-modal monitoring data, specifically comprising the following steps: Extracting peak flow, average flow and flow fluctuation coefficient affecting the operation state of the target drainage system from the flow monitoring data; Extracting pH value, suspended solids concentration and chemical oxygen demand affecting the operation state of the target drainage system from the water quality monitoring data; Extracting pipeline pressure, corrosion degree and leakage index affecting the operation state of the target drainage system from the pipeline state monitoring data; Extracting rainfall and soil moisture content affecting the operation state of the target drainage system from the environmental impact data; Extracting abnormal risk correlation coefficients of different key characteristic parameters and the target drainage system from the historical database, and predicting an abnormal risk value caused by single multi-modal monitoring data based on the abnormal risk correlation coefficients and the key characteristic parameters to obtain a risk prediction value; Extracting coupling influence coefficients between multi-modal characteristic parameters from the historical database; If no sudden interference factor is detected in the target drainage system, correcting the initial comprehensive risk evaluation value according to the abnormal risk correlation coefficients, the risk prediction value and the coupling influence coefficients to obtain a comprehensive risk evaluation value one; If a sudden interference factor is detected in the target drainage system, predicting the variation amplitude of each modal characteristic parameter under the action of the interference factor according to the intensity, action time and influence weight of the sudden interference factor on each multi-modal monitoring data, and correcting the initial comprehensive risk evaluation value according to the variation amplitude, the coupling influence coefficients, the abnormal risk correlation coefficients and the risk prediction value to obtain a comprehensive risk evaluation value two; Outputting an abnormal early warning result of the target drainage system according to the comprehensive risk evaluation value one or the comprehensive risk evaluation value two.

2. The multi-modal drainage monitoring data analysis method of claim 1, wherein, Predicting an abnormal risk value caused by single multi-modal monitoring data based on the abnormal risk correlation coefficients and the key characteristic parameters to obtain a risk prediction value, specifically comprising the following steps: Calculating the deviation value of the characteristic parameters by subtracting the normal operation threshold value from the key characteristic parameters; Predicting the abnormal risk value caused by single multi-modal monitoring data according to the abnormal risk correlation coefficients and the characteristic parameter deviation value to obtain a risk prediction value.

3. The multi-modal drainage monitoring data analysis method of claim 1, wherein, Extracting coupling influence coefficients between multi-modal characteristic parameters from the historical database, specifically comprising the following steps: Extracting risk synergistic action data of the multi-modal characteristic parameters acting together from the historical database; Calculating the mutual influence degree between different modal characteristic parameters to obtain the coupling influence coefficients between the multi-modal characteristic parameters.

4. The multi-modal drainage monitoring data analysis method of claim 1, wherein, If no sudden interference factor is detected in the target drainage system, correcting the initial comprehensive risk evaluation value according to the abnormal risk correlation coefficients, the risk prediction value and the coupling influence coefficients to obtain a comprehensive risk evaluation value one, specifically comprising the following steps: The parameter characteristics of the stable operation period are proposed from the historical operation state data of each mode of the drainage system, and the normal operation reference data and the parameter normal fluctuation range of each mode parameter characteristic are determined through the parameter characteristics of the stable operation period; Real-time operation state data of the target drainage system is collected; The deviation range of the real-time operation state data and the normal operation reference data is calculated to obtain the parameter deviation fluctuation value; If the parameter deviation fluctuation value is within the parameter normal fluctuation range, it indicates that no sudden interference factor is detected in the target drainage system; The weight proportion of each risk prediction value is determined according to the abnormal risk correlation coefficient; The superposition result is obtained by superimposing each risk prediction value combined with the coupling influence coefficient. If the target drainage system is detected to have a sudden interference factor, the variation amplitude of each mode characteristic parameter under the action of the interference factor is predicted according to the intensity, duration and influence weight of the sudden interference factor, which includes the following steps:

5. The multi-modal drainage monitoring data analysis method of claim 4, wherein, If the parameter deviation fluctuation value is not within the parameter normal fluctuation range, it indicates that the target drainage system has a sudden interference factor; The intensity and duration of the sudden interference factor are determined according to the type of the sudden interference factor; The influence weight of the interference factor on each multi-modal monitoring data is calculated; The variation amplitude of each mode characteristic parameter under the action of the interference factor is predicted according to the influence weight, intensity level and duration. The initial comprehensive risk assessment value is corrected to obtain the comprehensive risk assessment value two according to the variation amplitude, coupling influence coefficient, abnormal risk correlation coefficient and risk prediction value, which includes the following steps:

6. The multi-modal drainage monitoring data analysis method of claim 1, wherein, If the sudden interference factor has a synergistic enhancing effect on the multi-modal monitoring data, the enhancing coefficient of the sudden interference factor on the multi-modal monitoring data is extracted from the historical database, and the dynamic coupling coefficient is obtained by multiplying the coupling influence coefficient by the enhancing coefficient; If the sudden interference factor has an antagonistic effect on the multi-modal monitoring data, the attenuation coefficient of the sudden interference factor on the multi-modal monitoring data is extracted from the historical database, and the dynamic coupling coefficient is obtained by multiplying the coupling influence coefficient by the attenuation coefficient; The initial comprehensive risk assessment value is corrected to obtain the comprehensive risk assessment value two according to the dynamic coupling coefficient, variation amplitude, risk prediction value and abnormal risk correlation coefficient. The target drainage system abnormal early warning result is output according to the comprehensive risk assessment value one or the comprehensive risk assessment value two, which includes the following steps:

7. The multi-modal drainage monitoring data analysis method of claim 1, wherein, The risk change trend curve is obtained by drawing a curve according to the time sequence of the comprehensive risk assessment value one or the comprehensive risk assessment value two; The risk value starting period and the risk value sustained high period are marked according to the curve characteristics, and the target drainage system abnormal early warning result is output. It includes:

8. A multi-modal drainage monitoring data analysis system applied to the multi-modal drainage monitoring data analysis method of any one of claims 1-7, characterized in that, The acquisition module collects multi-modal monitoring data of the target drainage system in the monitoring period; The evaluation module evaluates the risk of the target drainage system according to the multi-modal monitoring data to obtain the initial comprehensive risk assessment value; ​ The prediction module is configured to extract key characteristic parameters affecting the operation state of the target drainage system from the multi-modal monitoring data, extract abnormal risk correlation coefficients of different key characteristic parameters and the target drainage system according to the historical database, and predict abnormal risk values caused by single multi-modal monitoring data based on the abnormal risk correlation coefficients and the key characteristic parameters to obtain risk prediction values. The extraction module is configured to extract coupling influence coefficients between the multi-modal characteristic parameters according to the historical database. The first correction module is configured to correct the initial comprehensive risk assessment value to obtain a first comprehensive risk assessment value if no sudden interference factor is detected in the target drainage system according to the abnormal risk correlation coefficients, the risk prediction values and the coupling influence coefficients. The second correction module is configured to predict the variation range of each modal characteristic parameter under the action of the interference factor according to the intensity, the action time and the influence weight of the interference factor on each multi-modal monitoring data, and correct the initial comprehensive risk assessment value to obtain a second comprehensive risk assessment value according to the variation range, the coupling influence coefficients, the abnormal risk correlation coefficients and the risk prediction values. The early warning module is configured to output an abnormal early warning result of the target drainage system according to the first comprehensive risk assessment value or the second comprehensive risk assessment value.

Images