Intelligent drainage monitoring method and system based on IoT perception

Abstract

The present invention provides a drainage intelligent monitoring method and system based on Internet of Things perception, which relates to the field of drainage monitoring technology. The method includes: obtaining facility water level data of multiple drainage facilities through Internet of Things perception technology; obtaining facility parameters of multiple drainage facilities; inputting facility water level data and volume data into a trained facility flow prediction model to obtain predicted facility flow data; obtaining multiple precipitation data; inputting precipitation data, predicted facility flow data, days of use and facility water level data into a trained overflow probability prediction model to obtain predicted overflow probability data; obtaining the harmful substance content of sewage and the surrounding population density of multiple drainage facilities; determining overflow hazard scores; setting hierarchical weights; and determining drainage facilities to be controlled first. According to the present invention, multiple drainage facilities can be hierarchically controlled based on the predicted overflow probability data and overflow hazard scores to reduce the probability and hazard of overflow.

Description

Technical Field

[0001] The present invention relates to the technical field of drainage monitoring, and in particular to an intelligent drainage monitoring method and system based on Internet of Things perception. Background Art

[0002] Among the related technologies, CN118761643A belongs to the field of urban stormwater management technology, specifically an urban stormwater monitoring and control system based on the concept of sponge city, including a server, a rainfall monitoring module, a water level monitoring module, a stormwater combined analysis module, an intelligent control module and a control performance evaluation module; this solution performs real-time processing and analysis of rainfall data and water level data through the stormwater combined analysis module, and the intelligent control module determines whether corresponding adjustments are needed based on the analysis results of the stormwater combined analysis module. If corresponding adjustments are needed, the corresponding urban drainage facilities are controlled based on the control strategy, thereby realizing real-time monitoring, early warning and forecasting, and intelligent adjustment of urban stormwater, significantly improving the efficiency and accuracy of urban stormwater management, reducing the risk of urban waterlogging, and being able to accurately feedback the control performance of drainage facilities to ensure subsequent control performance to improve stormwater management effects, with a high degree of intelligence.

[0003] CN118570015A provides a drainage area risk monitoring method, system, medium and electronic device. The drainage area risk monitoring method comprises: obtaining the standard drainage volume of the drainage area; based on the water flow monitoring data of the water flow monitoring points in the drainage area, obtaining the daily average water flow monitored by the water flow monitoring points, the water flow monitoring data is the water flow monitoring data of the drainage area on several sunny days, the water flow monitoring points include several upstream water flow monitoring points and several downstream water flow monitoring points in the drainage area; based on the standard drainage volume and the daily average water flow, obtaining the drainage area risk monitoring result corresponding to the pipe network risk index. The drainage area risk monitoring method is highly efficient and can achieve real-time monitoring with the help of corresponding hardware equipment.

[0004] Therefore, although drainage monitoring can be achieved in the relevant technology, the relevant technology does not take into account the predicted overflow probability and overflow hazard of the drainage facilities, and the impact on the control of the drainage facilities. That is, it is impossible to perform hierarchical control of multiple drainage facilities based on the predicted overflow probability data and overflow hazard scores.

[0005] The information disclosed in the background technology section of this application is only intended to deepen the understanding of the general background technology of this application, and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0006] The present invention provides a drainage intelligent monitoring method and system based on Internet of Things perception, which can solve the technical problem that related technologies are unable to perform hierarchical control of multiple drainage facilities according to predicted overflow probability data and overflow hazard scores.

[0007] According to a first aspect of the present invention, there is provided a drainage intelligent monitoring method based on Internet of Things sensing, comprising: obtaining facility water level data of multiple drainage facilities at multiple moments in the current monitoring period through Internet of Things sensing technology; obtaining facility parameters of multiple drainage facilities, wherein the facility parameters include volume data and usage days; inputting the facility water level data and the volume data into a trained facility flow prediction model to obtain predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period; obtaining precipitation data of the areas where the multiple drainage facilities are located at multiple moments in the current monitoring period; and inputting the precipitation data, the predicted facility flow data, and the training data into a trained facility flow prediction model. The number of days of use and the water level data of the facilities are input into a trained overflow probability prediction model to obtain the predicted overflow probability data of the multiple drainage facilities at multiple moments in the current monitoring period; the harmful substance content of the sewage of the multiple drainage facilities at multiple moments in the current monitoring period and the surrounding population density of the multiple drainage facilities are obtained; according to the harmful substance content of the sewage and the surrounding population density, the overflow hazard scores of the multiple drainage facilities in the current monitoring period are determined; hierarchical weights are set for the overflow hazard scores and the predicted overflow probability data; according to the hierarchical weights, the overflow hazard scores and the predicted overflow probability data, the drainage facilities to be controlled first are determined.

[0008] Furthermore, the training step of the facility flow prediction model includes: obtaining historical facility water level data and historical facility flow data of multiple drainage facilities at multiple times in multiple historical monitoring periods; processing the historical facility water level data and the volume data through the facility flow prediction model to obtain the historical predicted facility flow data of multiple drainage facilities at multiple times in multiple historical monitoring periods; determining the loss function of the facility flow prediction model based on the historical facility water level data, the volume data, the historical facility flow data and the historical predicted facility flow data; training the facility flow prediction model based on the loss function of the facility flow prediction model to obtain the trained facility flow prediction model.

[0009] Further, according to the historical facility water level data, the volume data, the historical facility flow data and the historical predicted facility flow data, determining the loss function of the facility flow prediction model includes: according to the formula Determine the loss function of the facility flow prediction model ,in, is the volume data of the i-th drainage facility, is the standard volume data of drainage facilities, is the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the historical facility water level data of the ith drainage facility at the first moment of the hth historical monitoring cycle, is the historical facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, The historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, N is the number of drainage facilities, H is the number of historical monitoring periods, M is the number of moments in the monitoring period, i≤N, h≤H, k≤M, and i, h, k, N, H and M are all positive integers.

[0010] Furthermore, the training step of the overflow probability prediction model includes: obtaining historical precipitation data of the areas where multiple drainage facilities are located at multiple times in multiple historical monitoring periods; obtaining historical facility water level data, historical predicted facility flow data and historical overflow probability data of multiple drainage facilities at multiple times in multiple historical monitoring periods; determining the historical usage days of multiple drainage facilities in multiple historical monitoring periods; processing the historical precipitation data, the historical predicted facility flow data, the historical usage days and the historical facility water level data through the overflow probability prediction model to obtain the historical predicted overflow probability data of multiple drainage facilities at multiple times in multiple historical monitoring periods; determining the loss function of the overflow probability prediction model based on the historical precipitation data, the historical predicted facility flow data, the historical usage days, the historical facility water level data, the historical overflow probability data and the historical predicted overflow probability data; and training the overflow probability prediction model based on the loss function of the overflow probability prediction model to obtain the trained overflow probability prediction model.

[0011] Further, according to the historical precipitation data, the historical predicted facility flow data, the historical usage days, the historical facility water level data, the historical overflow probability data and the historical predicted overflow probability data, the loss function of the overflow probability prediction model is determined, including: according to the formula Determine the loss function of the spillover probability prediction model ,in, is the historical usage days of the ith drainage facility in the hth historical monitoring period, is the standard use days threshold of the i-th drainage facility, is the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the standard facility water level data of the i-th drainage facility, is the bottom area of ​​the i-th drainage facility, is the historical precipitation data of the area where the i-th drainage facility is located at the k-th moment in the h-th historical monitoring period, To monitor the adjacent time intervals of the cycle, is the historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, Increase the water volume for drainage facility standards, is the historical overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the historical predicted overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period, N is the number of drainage facilities, H is the number of historical monitoring periods, M is the number of moments in the monitoring period, i≤N, h≤H, k≤M, and i, h, k, N, H and M are all positive integers.

[0012] Furthermore, the overflow hazard scores of multiple drainage facilities in the current monitoring period are determined based on the harmful substance content of the sewage and the surrounding population density, including: averaging the harmful substance content of the sewage of each drainage facility at multiple moments in the current monitoring period to obtain the average value of the content of each harmful substance at multiple moments in the current monitoring period of each drainage facility; and determining the overflow hazard scores of multiple drainage facilities in the current monitoring period based on the average value of the harmful substance content and the surrounding population density.

[0013] Further, according to the average value of the harmful substance content and the surrounding population density, the overflow hazard scores of multiple drainage facilities in the current monitoring period are determined, including: according to the formula Determine the overflow hazard score of the i-th drainage facility in the current monitoring period ,in, is the population density around the ith drainage facility, is the surrounding population density threshold, is the average value of the jth harmful substance content in the ith drainage facility during the current monitoring period, is the j-th hazardous substance content change rate threshold, and is the preset weight, E is the number of types of harmful substances in sewage, j≤E, and j and E are both positive integers.

[0014] According to a second aspect of the present invention, there is provided a drainage intelligent monitoring system based on Internet of Things perception, comprising: a facility water level data module, for obtaining facility water level data of multiple drainage facilities at multiple moments in the current monitoring period through Internet of Things perception technology; a facility parameter module, for obtaining facility parameters of multiple drainage facilities, wherein the facility parameters include volume data and usage days; a predicted facility flow data module, for inputting the facility water level data and the volume data into a trained facility flow prediction model to obtain predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period; a precipitation data module, for obtaining precipitation data of the areas where multiple drainage facilities are located at multiple moments in the current monitoring period; a predicted overflow probability data module, for inputting the precipitation data, the predicted facility flow data into a trained facility flow prediction model, and obtaining predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period. The data, the number of days of use and the water level data of the facilities are input into the trained overflow probability prediction model to obtain the predicted overflow probability data of multiple drainage facilities at multiple moments in the current monitoring period; a harmful substance content and surrounding population density module is used to obtain the harmful substance content of the sewage of multiple drainage facilities at multiple moments in the current monitoring period and the surrounding population density of multiple drainage facilities; an overflow hazard scoring module is used to determine the overflow hazard scores of multiple drainage facilities in the current monitoring period according to the harmful substance content of the sewage and the surrounding population density; a hierarchical weight module is used to set hierarchical weights for the overflow hazard scores and the predicted overflow probability data; a priority control module is used to determine the drainage facilities to be controlled first according to the hierarchical weights, the overflow hazard scores and the predicted overflow probability data.

[0015] Technical effect: According to the present invention, through the Internet of Things sensing technology, the key information such as the facility water level data and precipitation data of multiple drainage facilities can be obtained in real time, which can realize the real-time monitoring and remote control of the drainage facilities and improve the accuracy and timeliness of monitoring. Through the trained facility flow prediction model and overflow probability prediction model, the facility flow data and overflow probability data of the drainage facilities are predicted, so that according to the predicted overflow probability data and overflow hazard score, multiple drainage facilities can be graded and controlled to reduce the probability and hazard of overflow. When determining the loss function of the facility flow prediction model, the influence of historical facility water level data and volume data on the facility flow can be used to determine the influence of the above data on the error of historical predicted facility flow data, and then set weights based on the influence and the relative difference between the historical facility flow data and the historical predicted facility flow data, and based on the characteristics that the shorter the time interval with the current monitoring cycle, the higher the accuracy, and the shorter the time interval with the first moment, the lower the accuracy, so as to perform weighted summation on the errors output by the facility flow prediction model of multiple drainage facilities at each moment in each historical monitoring cycle to obtain the loss function, so as to improve the design accuracy and objectivity of the loss function, thereby improving the training efficiency during the training process and improving the accuracy of the facility flow prediction model. When determining the loss function of the overflow probability prediction model, the influence of historical precipitation data, historical prediction facility flow data, historical facility water level data and historical usage days on the overflow probability can be used to determine the influence of the above data on the error of the historical prediction overflow probability data, and then set weights based on the influence and the relative difference between the historical overflow probability data and the historical prediction overflow probability data, and based on the characteristics that the shorter the time interval with the current monitoring cycle, the higher the accuracy, and the shorter the time interval with the first moment, the lower the accuracy, and then perform weighted summation on the errors output by the overflow probability prediction model of multiple drainage facilities at each moment in each historical monitoring cycle to obtain the loss function, so as to improve the design accuracy and objectivity of the loss function, thereby improving the training efficiency during the training process and improving the accuracy of the overflow probability prediction model. When determining the spillover hazard scores of multiple drainage facilities in the current monitoring period, the surrounding population density and the rate of change of harmful substance content can be used to determine the spillover hazard scores of multiple drainage facilities in the current monitoring period. By evaluating the spillover hazards of drainage facilities from two aspects, the potential high-risk drainage facilities can be discovered in time, so as to take preventive measures or carry out necessary repair work, which will help reduce the occurrence of spillover incidents, protect the surrounding environment and the health of residents, and improve the comprehensiveness and accuracy of the spillover hazard score.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only and do not limit the present invention. Other features and aspects of the present invention will become more apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other embodiments can be obtained based on these drawings without creative work.

[0018] Figure 1 The flowchart of the drainage intelligent monitoring method based on IoT perception according to an embodiment of the present invention is exemplarily shown;

[0019] Figure 2 A block diagram of an intelligent drainage monitoring system based on IoT perception according to an embodiment of the present invention is exemplarily shown. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0022] Figure 1 The flowchart of the drainage intelligent monitoring method based on IoT perception according to an embodiment of the present invention is exemplarily shown, and the method includes:

[0023] Step S101, obtaining facility water level data of multiple drainage facilities through IoT sensing technology at multiple moments in the current monitoring cycle;

[0024] Step S102, obtaining facility parameters of a plurality of drainage facilities, wherein the facility parameters include volume data and usage days;

[0025] Step S103, inputting the facility water level data and the volume data into the trained facility flow prediction model to obtain predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period;

[0026] Step S104, obtaining precipitation data of the areas where the multiple drainage facilities are located at multiple times during the current monitoring period;

[0027] Step S105, inputting the precipitation data, the predicted facility flow data, the usage days and the facility water level data into the trained overflow probability prediction model to obtain the predicted overflow probability data of multiple drainage facilities at multiple moments in the current monitoring period;

[0028] Step S106, obtaining the harmful substance content of sewage from multiple drainage facilities at multiple moments in the current monitoring period and the population density around the multiple drainage facilities;

[0029] Step S107, determining overflow hazard scores of multiple drainage facilities in the current monitoring period according to the harmful substance content of the sewage and the surrounding population density;

[0030] Step S108, setting a grading weight for the spillover hazard score and the predicted spillover probability data;

[0031] Step S109, determining the drainage facilities to be controlled with priority according to the grading weights, the spillover hazard scores and the predicted spillover probability data.

[0032] According to the intelligent drainage monitoring method based on IoT sensing of the embodiment of the present invention, key information such as facility water level data and precipitation data of multiple drainage facilities can be obtained in real time through IoT sensing technology, so as to realize real-time monitoring and remote control of drainage facilities and improve the accuracy and timeliness of monitoring. The facility flow data and overflow probability data of drainage facilities are predicted through the trained facility flow prediction model and overflow probability prediction model, so that multiple drainage facilities can be hierarchically controlled according to the predicted overflow probability data and overflow hazard score, so as to reduce the probability and hazard of overflow.

[0033] According to one embodiment of the present invention, in step S101, each monitoring cycle can be set to 30 minutes, 60 minutes, etc., and the interval between adjacent moments can be set to 2 minutes, 5 minutes, etc., and the present invention does not limit this. The water level data of multiple drainage facilities are monitored using IoT sensing technology, where the drainage facilities can be storage tanks, pump station pits, etc. IoT sensing technology refers to the technology that uses various sensors, communication technologies and data processing technologies to realize the perception, recognition and understanding of the physical world, and can monitor the relevant data of drainage facilities in real time.

[0034] According to one embodiment of the present invention, in step S102, the volume data is the amount of water that the drainage facility can accommodate, which can reflect the drainage capacity of the drainage facility, and the number of days in use is the number of days the drainage facility has been in use since its completion, which can reflect the age of the drainage facility.

[0035] According to one embodiment of the present invention, in step S103, the facility water level data and volume data are input into a trained facility flow prediction model. The facility flow prediction model may be a neural network model, which is trained based on a large amount of historical data through machine learning or deep learning technology, and can analyze the predicted facility flow data of multiple drainage facilities at multiple times in the current monitoring period.

[0036] According to one embodiment of the present invention, the training step of the facility flow prediction model includes: obtaining historical facility water level data and historical facility flow data of multiple drainage facilities at multiple times in multiple historical monitoring periods; processing the historical facility water level data and the volume data through the facility flow prediction model to obtain the historical predicted facility flow data of multiple drainage facilities at multiple times in multiple historical monitoring periods; determining the loss function of the facility flow prediction model based on the historical facility water level data, the volume data, the historical facility flow data and the historical predicted facility flow data; training the facility flow prediction model based on the loss function of the facility flow prediction model to obtain the trained facility flow prediction model.

[0037] According to one embodiment of the present invention, the facility flow rate is the amount of water discharged by the drainage facility per unit time, for example, cubic meters discharged per second (m³ / s). The facility flow rate can be calculated by real-time monitoring of the water level of the drainage facility. For example, when the water level rises, it means that the amount of water discharged is reduced, and when the water level drops, it means that the amount of water discharged is increased. The larger the volume of the drainage facility, the more water is discharged under the same water level change. The facility flow prediction model can predict the predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring cycle based on the relationship between the above-mentioned facility water level data, volume data and the facility flow of the drainage facility. The loss function is determined based on the relative difference between the historical predicted facility flow data and the historical facility flow data of the historical monitoring record. By feedback-adjusting the loss function, a trained facility flow prediction model is obtained.

[0038] According to one embodiment of the present invention, the loss function of the facility flow prediction model is determined based on the historical facility water level data, the volume data, the historical facility flow data and the historical predicted facility flow data, including: determining the loss function of the facility flow prediction model according to formula (1): ,

[0039] (1),

[0040] in, is the volume data of the i-th drainage facility, is the standard volume data of drainage facilities, is the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the historical facility water level data of the ith drainage facility at the first moment of the hth historical monitoring cycle, is the historical facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, The historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, N is the number of drainage facilities, H is the number of historical monitoring periods, M is the number of moments in the monitoring period, i≤N, h≤H, k≤M, and i, h, k, N, H and M are all positive integers.

[0041] According to one embodiment of the present invention, in formula (1), It is the relative error between the historical facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period and the historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period. It is the relative difference between the volume data of the i-th drainage facility and the standard volume data of the drainage facility. The larger the relative difference is, the larger the volume of the i-th drainage facility is and the larger the facility flow of the drainage facility is. It is the ratio of the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring cycle to the historical facility water level data of the ith drainage facility at the 1st moment in the hth historical monitoring cycle. The smaller the ratio is, the greater the water level drop and the greater the facility flow of the drainage facility. It indicates that the volume data of drainage facilities is positively correlated with the facility flow, and the historical facility water level data is negatively correlated with the facility flow. For example, when the volume data of drainage facilities is larger, the amount of water discharged is relatively more under the same water level change, and the facility flow of drainage facilities is larger. The smaller the historical facility water level data is, the larger the facility flow of drainage facilities will be when the water level drops. Therefore, the volume data related data is placed in the numerator position, indicating that the volume data of drainage facilities is larger relative to the standard volume data, that is, The larger the value, the greater the impact on the error of the historical prediction facility flow data. The historical facility water level data related data are placed in the denominator, indicating that the historical facility water level data at a certain moment in a historical monitoring period is smaller than the historical facility water level data at the first moment, that is, The smaller the value, the greater the impact on the error of historical prediction facility flow data. is the weight at the kth moment, which is used to reasonably weight the relative errors at different moments in the loss function. The accuracy of the historical predicted facility flow data at the k+1th moment output by the facility flow prediction model is usually higher than the accuracy of the historical predicted facility flow data at the kth moment. That is, the shorter the time interval between a certain moment and the first moment, the less accurate the prediction result. In order to improve the training efficiency, the higher the weight is set, and vice versa, the more accurate the prediction result is, the lower the weight is. is the weight of the hth historical monitoring period, which is used to reasonably weight the relative errors of different historical monitoring periods in the loss function. The accuracy of the historical predicted facility flow data of the hth historical monitoring period output by the facility flow prediction model is usually higher than the accuracy of the historical predicted facility flow data of the h+1th historical monitoring period. That is, the longer the time interval between a historical monitoring period and the current monitoring period, the less accurate the prediction result. In order to improve the training efficiency, the higher the weight is set. Conversely, the more accurate the prediction result is, the lower the weight is.

[0042] According to one embodiment of the present invention, using , and The relative errors of the historical predicted facility flow data of multiple drainage facilities at each moment in each historical monitoring period are weighted averaged to obtain the training loss function. In the process of training the facility flow prediction model, the loss function is back-propagated and some parameters within the model are adjusted to reduce the value of the loss function of the facility flow prediction model, thereby improving the accuracy of the facility flow prediction model and obtaining the trained facility flow prediction model.

[0043] In this way, the influence of historical facility water level data and volume data on the facility flow can be used to determine the impact of the above data on the error of historical predicted facility flow data, and thus set weights based on the influence and the relative difference between the historical facility flow data and the historical predicted facility flow data, and based on the characteristics that the shorter the time interval with the current monitoring cycle, the higher the accuracy, and the shorter the time interval with the first moment, the lower the accuracy, so as to perform weighted summation on the errors output by the facility flow prediction model of multiple drainage facilities at each moment in each historical monitoring cycle, and obtain a loss function to improve the design accuracy and objectivity of the loss function, thereby improving the training efficiency during the training process and improving the accuracy of the facility flow prediction model.

[0044] According to an embodiment of the present invention, in step S104, precipitation data is obtained through meteorological observation stations, radar monitoring, satellite remote sensing and other methods according to the precipitation conditions of drainage facilities distributed in different geographical areas.

[0045] According to one embodiment of the present invention, in step S105, precipitation data, predicted facility flow data, usage days and facility water level data are input into a trained overflow probability prediction model. The overflow probability prediction model can be a neural network model, which is trained based on a large amount of historical data through machine learning or deep learning technology, and can analyze the predicted overflow probability data of multiple drainage facilities at multiple times in the current monitoring period.

[0046] According to one embodiment of the present invention, the training step of the overflow probability prediction model includes: obtaining historical precipitation data of the area where multiple drainage facilities are located at multiple times in multiple historical monitoring periods; obtaining historical facility water level data, historical predicted facility flow data and historical overflow probability data of multiple drainage facilities at multiple times in multiple historical monitoring periods; determining the historical usage days of multiple drainage facilities in multiple historical monitoring periods; processing the historical precipitation data, the historical predicted facility flow data, the historical usage days and the historical facility water level data through the overflow probability prediction model to obtain the historical predicted overflow probability data of multiple drainage facilities at multiple times in multiple historical monitoring periods; determining the loss function of the overflow probability prediction model based on the historical precipitation data, the historical predicted facility flow data, the historical usage days, the historical facility water level data, the historical overflow probability data and the historical predicted overflow probability data; training the overflow probability prediction model based on the loss function of the overflow probability prediction model to obtain the trained overflow probability prediction model.

[0047] According to one embodiment of the present invention, when the rainfall is too much and exceeds the facility flow of the drainage facility, that is, the water inflow of the drainage facility is greater than the water outflow of the drainage facility, the water volume easily exceeds the volume of the drainage facility, and the overflow probability of the drainage facility is higher. The more days the drainage facility is used and the older the drainage facility is, the higher the overflow probability is. The higher the water level of the drainage facility, the higher the overflow probability. The overflow probability prediction model can predict the predicted overflow probability data of multiple drainage facilities at multiple moments in the current monitoring period based on the relationship between the above-mentioned precipitation data, predicted facility flow data, days of use and facility water level data and the overflow probability of the drainage facility. The loss function is determined based on the relative difference between the historical predicted overflow probability data and the historical overflow probability data of the historical monitoring records. By feedback-adjusting the loss function, a trained overflow probability prediction model is obtained.

[0048] According to one embodiment of the present invention, the loss function of the spillover probability prediction model is determined based on the historical precipitation data, the historical predicted facility flow data, the historical usage days, the historical facility water level data, the historical spillover probability data and the historical predicted spillover probability data, including: determining the loss function of the spillover probability prediction model according to formula (2): ,

[0049] (2),

[0050] in, is the historical usage days of the ith drainage facility in the hth historical monitoring period, is the standard use days threshold of the i-th drainage facility, is the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the standard facility water level data of the i-th drainage facility, is the bottom area of ​​the i-th drainage facility, is the historical precipitation data of the area where the i-th drainage facility is located at the k-th moment in the h-th historical monitoring period, To monitor the adjacent time intervals of the cycle, is the historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, Increase the water volume for drainage facility standards, is the historical overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the historical predicted overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period, N is the number of drainage facilities, H is the number of historical monitoring periods, M is the number of moments in the monitoring period, i≤N, h≤H, k≤M, and i, h, k, N, H and M are all positive integers.

[0051] According to one embodiment of the present invention, in formula (2), It is the relative error between the historical overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period and the historical predicted overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period. It is the product of the bottom area of ​​the ith drainage facility and the historical precipitation data of the area where the ith drainage facility is located at the kth moment in the hth historical monitoring period, indicating the water inflow of the drainage facility during the time period between the k-1th moment and the kth moment in the hth historical monitoring period. For example, the bottom area of ​​the drainage facility is 1m 2 , the rainfall data measured in the area where the drainage facility is located at the kth moment in the hth historical monitoring period is 5mm, then the water inflow of the drainage facility is 0.005m³. It is the product of the adjacent time intervals of the monitoring period and the historical predicted facility flow data of the ith drainage facility at the kth moment of the hth historical monitoring period, indicating the water discharge of the drainage facility in the time period between the k-1th moment and the kth moment of the hth historical monitoring period. It is the difference between the water inflow of the drainage facility and the water outflow of the drainage facility in the time period between the k-1th moment and the kth moment of the hth historical monitoring period, and the ratio of the standard growth water volume of the drainage facility. The larger the ratio is, the greater the growth water volume of the drainage facility is, and the greater the overflow probability of the drainage facility is. It is the ratio of the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period to the standard facility water level data of the ith drainage facility. The larger the ratio is, the higher the water level is and the greater the probability of overflow of the drainage facility is. It is the ratio of the historical usage days of the ith drainage facility in the hth historical monitoring period to the standard usage days threshold of the ith drainage facility. The larger the ratio is, the older the drainage facility is and the greater the probability of overflow of the drainage facility is. It indicates that the growth of drainage facilities, historical facility water level data, and historical use days are positively correlated with the overflow probability. For example, the greater the historical precipitation data in the area where the drainage facilities are located, the more water will flow into the drainage facilities. The greater the historical predicted facility flow data of the drainage facilities, the more water will flow out of the drainage facilities. When the water inflow of the drainage facilities is greater than the water outflow, the greater the growth of drainage facilities, the greater the overflow probability of the drainage facilities. The higher the water level of the drainage facilities, the greater the overflow probability of the drainage facilities. The more days the drainage facilities are used, the older the drainage facilities, and the greater the overflow probability of the drainage facilities. Therefore, the greater the growth of drainage facilities relative to the standard growth of drainage facilities, that is, The larger the value of , the greater the impact on the error of the historical prediction spillover probability data, and the larger the historical facility water level data is relative to the standard facility water level data, that is, The larger the value of , the greater the impact on the error of the historical prediction spillover probability data, and the larger the historical usage days relative to the standard usage days threshold, that is, The larger the value of , the greater the impact on the error of historical prediction spillover probability data. is the weight at the kth moment, which is used to reasonably weight the relative errors at different moments in the loss function. The accuracy of the historical predicted spillover probability data at the k+1th moment output by the spillover probability prediction model is usually higher than the accuracy of the historical predicted spillover probability data at the kth moment. That is, the shorter the time interval between a certain moment and the first moment, the less accurate the prediction result. In order to improve the training efficiency, the higher the weight is set, and vice versa, the more accurate the prediction result is, the lower the weight is. is the weight of the h-th historical monitoring period, which is used to reasonably weight the relative errors of different historical monitoring periods in the loss function. The accuracy of the historical predicted spillover probability data of the h-th historical monitoring period output by the spillover probability prediction model is usually higher than the accuracy of the historical predicted spillover probability data of the h+1-th historical monitoring period. That is, the longer the time interval between a historical monitoring period and the current monitoring period, the less accurate the prediction result. In order to improve the training efficiency, the higher the weight is set. Conversely, the more accurate the prediction result is, the lower the weight is.

[0052] According to one embodiment of the present invention, using , and The relative errors of the historical predicted overflow probability data of multiple drainage facilities at each moment in each historical monitoring period are weighted averaged to obtain the training loss function. In the process of training the overflow probability prediction model, the loss function is back-propagated and some parameters within the model are adjusted to reduce the value of the loss function of the overflow probability prediction model, thereby improving the accuracy of the overflow probability prediction model and obtaining the trained overflow probability prediction model.

[0053] In this way, the influence of historical precipitation data, historical predicted facility flow data, historical facility water level data and historical usage days on the overflow probability can be used to determine the influence of the above data on the error of historical predicted overflow probability data, and then set weights based on the influence and the relative difference between the historical overflow probability data and the historical predicted overflow probability data, and based on the characteristics that the shorter the time interval with the current monitoring cycle, the higher the accuracy, and the shorter the time interval with the first moment, the lower the accuracy, so as to perform weighted summation on the errors output by the overflow probability prediction model of multiple drainage facilities at each moment in each historical monitoring cycle, and obtain the loss function, so as to improve the design accuracy and objectivity of the loss function, thereby improving the training efficiency during the training process and improving the accuracy of the overflow probability prediction model.

[0054] According to an embodiment of the present invention, in step S106, multiple drainage facilities are sampled at multiple moments in the current monitoring cycle, and the harmful substance content of the sewage is measured using spectral detection technology, reflecting the degree of pollution of the sewage discharged by the drainage facilities. The surrounding population density is the population distribution in the surrounding area of ​​the drainage facility, expressed as the number of people per square kilometer, and the population distribution within a range of 50 kilometers or 100 kilometers around the midpoint of the drainage facility as a circle can be used as the surrounding population density, and the present invention does not limit this.

[0055] According to an embodiment of the present invention, in step S107, the overflow hazard scores of multiple drainage facilities in the current monitoring period are determined based on the harmful substance content of the sewage and the surrounding population density.

[0056] According to one embodiment of the present invention, step S107 includes: averaging the harmful substance content of sewage at multiple moments in the current monitoring period of each drainage facility to obtain the average value of the content of each harmful substance at multiple moments in the current monitoring period of each drainage facility; and determining the spillover hazard scores of multiple drainage facilities in the current monitoring period based on the average value of the harmful substance content and the surrounding population density.

[0057] According to one embodiment of the present invention, for each harmful substance in sewage, the average content thereof in the current monitoring period is calculated, that is, the content of the harmful substance at all times is summed up and divided by the number of monitoring times, so as to obtain the average content of each harmful substance in each drainage facility in the current monitoring period. Based on the average content of the harmful substance and the surrounding population density, the spillover hazard scores of multiple drainage facilities in the current monitoring period are determined.

[0058] According to an embodiment of the present invention, the spillover hazard scores of multiple drainage facilities in the current monitoring period are determined according to the average value of the harmful substance content and the surrounding population density, including: determining the spillover hazard score of the i-th drainage facility in the current monitoring period according to formula (3): ,

[0059] (3),

[0060] in, is the population density around the ith drainage facility, is the surrounding population density threshold, is the average value of the jth harmful substance content in the ith drainage facility during the current monitoring period, is the j-th hazardous substance content change rate threshold, and is the preset weight, E is the number of types of harmful substances in sewage, j≤E, and j and E are both positive integers.

[0061] According to one embodiment of the present invention, in formula (3), It is the ratio of the surrounding population density of the i-th drainage facility to the surrounding population density threshold. The larger the ratio is, the more people are around the i-th drainage facility, and the greater the spillover hazard of the drainage facility. It is the relative difference between the average value of the j-th harmful substance content in the ith drainage facility during the current monitoring period and the threshold value of the change rate of the j-th harmful substance content. The larger the relative difference, the greater the j-th harmful substance content in the ith drainage facility, the easier it is to affect the soil environment around the drainage facility, and the greater the spillover hazard of the drainage facility. It means that the relative differences of the contents of multiple harmful substances in the i-th drainage facility are averaged. The larger the average value, the greater the overflow hazard of the drainage facility. and The value of is weighted and summed to obtain the spillover hazard score of the i-th drainage facility in the current monitoring period. The larger the spillover hazard score, the greater the spillover hazard of the drainage facility.

[0062] In this way, the spillover hazard scores of multiple drainage facilities in the current monitoring period can be determined based on the average value of the harmful substance content and the surrounding population density. The spillover hazards of drainage facilities can be evaluated from two aspects: the average value of the harmful substance content and the surrounding population density. Potential high-risk drainage facilities can be discovered in time, so that preventive measures can be taken or necessary repair work can be carried out, which will help reduce the occurrence of spillover incidents, protect the surrounding environment and the health of residents, and improve the comprehensiveness and accuracy of the spillover hazard score.

[0063] According to one embodiment of the present invention, in step S108, the spillover hazard score and the predicted spillover probability data respectively represent the degree of hazard caused by the spillover event and the possibility of the spillover event. Therefore, the hierarchical weights of the spillover hazard score and the predicted spillover probability data can be set according to the spillover hazard and spillover probability of the drainage facilities. For example, the hierarchical weight of the spillover hazard score is low and the hierarchical weight of the predicted spillover probability data is high, so as to better reflect the spillover hazard and urgency of multiple drainage facilities in the comprehensive evaluation.

[0064] According to one embodiment of the present invention, in step S109, the maximum value of the predicted spillover probability data of the drainage facility at multiple moments in the current monitoring period is taken, and the spillover hazard score and the maximum value of the predicted spillover probability data are weighted and summed using the above-mentioned hierarchical weights to obtain a comprehensive risk score for the drainage facility. The comprehensive risk scores of each drainage facility are arranged in descending order to determine the drainage facilities that are to be controlled first, that is, the drainage facility with the largest comprehensive risk score is the drainage facility that is to be controlled first, thereby reducing the probability and hazard of spillover.

[0065] According to the intelligent drainage monitoring method based on IoT sensing of the embodiment of the present invention, key information such as facility water level data and precipitation data of multiple drainage facilities can be obtained in real time through IoT sensing technology, so as to realize real-time monitoring and remote control of drainage facilities and improve the accuracy and timeliness of monitoring. The facility flow data and overflow probability data of drainage facilities are predicted through the trained facility flow prediction model and overflow probability prediction model, so that multiple drainage facilities can be hierarchically controlled according to the predicted overflow probability data and overflow hazard score, so as to reduce the probability and hazard of overflow. When determining the loss function of the facility flow prediction model, the influence of historical facility water level data and volume data on the facility flow can be used to determine the influence of the above data on the error of historical predicted facility flow data, and then set weights based on the influence and the relative difference between the historical facility flow data and the historical predicted facility flow data, and based on the characteristics that the shorter the time interval with the current monitoring cycle, the higher the accuracy, and the shorter the time interval with the first moment, the lower the accuracy, so as to perform weighted summation on the errors output by the facility flow prediction model of multiple drainage facilities at each moment in each historical monitoring cycle to obtain the loss function, so as to improve the design accuracy and objectivity of the loss function, thereby improving the training efficiency during the training process and improving the accuracy of the facility flow prediction model. When determining the loss function of the overflow probability prediction model, the influence of historical precipitation data, historical prediction facility flow data, historical facility water level data and historical usage days on the overflow probability can be used to determine the influence of the above data on the error of the historical prediction overflow probability data, and then set weights based on the influence and the relative difference between the historical overflow probability data and the historical prediction overflow probability data, and based on the characteristics that the shorter the time interval with the current monitoring cycle, the higher the accuracy, and the shorter the time interval with the first moment, the lower the accuracy, and then perform weighted summation on the errors output by the overflow probability prediction model of multiple drainage facilities at each moment in each historical monitoring cycle to obtain the loss function, so as to improve the design accuracy and objectivity of the loss function, thereby improving the training efficiency during the training process and improving the accuracy of the overflow probability prediction model. When determining the spillover hazard scores of multiple drainage facilities in the current monitoring period, the surrounding population density and the rate of change of harmful substance content can be used to determine the spillover hazard scores of multiple drainage facilities in the current monitoring period. By evaluating the spillover hazards of drainage facilities from two aspects, the potential high-risk drainage facilities can be discovered in time, so as to take preventive measures or carry out necessary repair work, which will help reduce the occurrence of spillover incidents, protect the surrounding environment and the health of residents, and improve the comprehensiveness and accuracy of the spillover hazard score.

[0066] Figure 2 The block diagram of a drainage intelligent monitoring system based on IoT perception according to an embodiment of the present invention is exemplarily shown, and the system includes:

[0067] The facility water level data module is used to obtain the facility water level data of multiple drainage facilities through IoT sensing technology at multiple moments in the current monitoring cycle;

[0068] A facility parameter module, used to obtain facility parameters of a plurality of drainage facilities, wherein the facility parameters include volume data and usage days;

[0069] A facility flow prediction data module, used for inputting the facility water level data and the volume data into a trained facility flow prediction model to obtain predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period;

[0070] The precipitation data module is used to obtain precipitation data of the areas where multiple drainage facilities are located at multiple times in the current monitoring period;

[0071] A predicted overflow probability data module is used to input the precipitation data, the predicted facility flow data, the usage days and the facility water level data into a trained overflow probability prediction model to obtain predicted overflow probability data of multiple drainage facilities at multiple moments in the current monitoring period;

[0072] The harmful substance content and surrounding population density module is used to obtain the harmful substance content of sewage from multiple drainage facilities at multiple moments in the current monitoring period and the surrounding population density of multiple drainage facilities;

[0073] A spillover hazard scoring module, used to determine spillover hazard scores of multiple drainage facilities in the current monitoring period according to the harmful substance content of the sewage and the surrounding population density;

[0074] A grading weight module, used to set grading weights for the spillover hazard score and the predicted spillover probability data;

[0075] The priority control module is used to determine the drainage facilities to be controlled with priority according to the classification weights, the spillover hazard scores and the predicted spillover probability data.

[0076] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0077] It should be understood by those skilled in the art that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be deformed or modified in any way without departing from the principles.

Claims

1. A drainage intelligent monitoring method based on IoT perception, characterized in that: include: At multiple moments in the current monitoring cycle, the water level data of multiple drainage facilities are obtained through IoT sensing technology; Acquiring facility parameters of a plurality of drainage facilities, wherein the facility parameters include volume data and usage days; Inputting the facility water level data and the volume data into a trained facility flow prediction model to obtain predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period; Obtain precipitation data for areas where multiple drainage facilities are located at multiple times during the current monitoring period; Input the precipitation data, the predicted facility flow data, the usage days and the facility water level data into the trained overflow probability prediction model to obtain the predicted overflow probability data of multiple drainage facilities at multiple times in the current monitoring period; Obtain the harmful substance content of sewage from multiple drainage facilities at multiple times during the current monitoring period and the population density around the multiple drainage facilities; Determine the overflow hazard scores of multiple drainage facilities in the current monitoring period according to the harmful substance content of the sewage and the surrounding population density; Setting grading weights for the spillover hazard score and the predicted spillover probability data; Determine the drainage facilities to be controlled with priority according to the classification weights, the spillover hazard scores and the predicted spillover probability data; The training steps of the facility flow prediction model include: Obtain historical facility water level data and historical facility flow data for multiple drainage facilities at multiple times during multiple historical monitoring periods; The historical facility water level data and the volume data are processed by a facility flow prediction model to obtain historical predicted facility flow data of multiple drainage facilities at multiple times in multiple historical monitoring periods; Determining a loss function of the facility flow prediction model based on the historical facility water level data, the volume data, the historical facility flow data, and the historical predicted facility flow data; According to the loss function of the facility flow prediction model, the facility flow prediction model is trained to obtain the trained facility flow prediction model; Determining a loss function of the facility flow prediction model according to the historical facility water level data, the volume data, the historical facility flow data, and the historical predicted facility flow data includes: According to the formula Determine the loss function of the facility flow prediction model ,in, is the volume data of the i-th drainage facility, is the standard volume data of drainage facilities, is the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the historical facility water level data of the ith drainage facility at the first moment of the hth historical monitoring cycle, is the historical facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, The historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, N is the number of drainage facilities, H is the number of historical monitoring periods, M is the number of moments in the monitoring period, i≤N, h≤H, k≤M, and i, h, k, N, H and M are all positive integers.

2. The drainage intelligent monitoring method based on IoT perception according to claim 1 is characterized in that: The training steps of the spillover probability prediction model include: Obtain historical precipitation data for areas where multiple drainage facilities are located at multiple times during multiple historical monitoring periods; Obtain historical facility water level data, historical predicted facility flow data, and historical overflow probability data for multiple drainage facilities at multiple times during multiple historical monitoring periods; Determine the historical usage days of multiple drainage facilities during multiple historical monitoring periods; The historical precipitation data, the historical predicted facility flow data, the historical usage days and the historical facility water level data are processed by the overflow probability prediction model to obtain the historical predicted overflow probability data of multiple drainage facilities at multiple times in multiple historical monitoring periods; Determine the loss function of the spillover probability prediction model based on the historical precipitation data, the historical predicted facility flow data, the historical usage days, the historical facility water level data, the historical spillover probability data, and the historical predicted spillover probability data; The spillover probability prediction model is trained according to the loss function of the spillover probability prediction model to obtain the trained spillover probability prediction model.

3. The drainage intelligent monitoring method based on IoT perception according to claim 2 is characterized in that: Determining the loss function of the spillover probability prediction model according to the historical precipitation data, the historical predicted facility flow data, the historical usage days, the historical facility water level data, the historical spillover probability data and the historical predicted spillover probability data includes: According to the formula Determine the loss function of the spillover probability prediction model ,in, is the historical usage days of the ith drainage facility in the hth historical monitoring period, is the standard use days threshold of the i-th drainage facility, is the historical facility water level data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the standard facility water level data of the i-th drainage facility, is the bottom area of ​​the i-th drainage facility, is the historical precipitation data of the area where the i-th drainage facility is located at the k-th moment in the h-th historical monitoring period, To monitor the adjacent time intervals of the cycle, is the historical predicted facility flow data of the ith drainage facility at the kth moment in the hth historical monitoring period, Increase the water volume for drainage facility standards, is the historical overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period, is the historical predicted overflow probability data of the ith drainage facility at the kth moment in the hth historical monitoring period, N is the number of drainage facilities, H is the number of historical monitoring periods, M is the number of moments in the monitoring period, i≤N, h≤H, k≤M, and i, h, k, N, H and M are all positive integers.

4. The drainage intelligent monitoring method based on IoT perception according to claim 1 is characterized in that: Based on the harmful substance content of the sewage and the surrounding population density, the overflow hazard scores of multiple drainage facilities in the current monitoring period are determined, including: Averaging the harmful substance content of sewage at multiple moments in the current monitoring cycle of each drainage facility to obtain an average value of the content of each harmful substance at multiple moments in the current monitoring cycle of each drainage facility; The overflow hazard scores of multiple drainage facilities in the current monitoring period are determined based on the average value of the harmful substance content and the surrounding population density.

5. The drainage intelligent monitoring method based on IoT perception according to claim 4 is characterized in that: Based on the average value of the harmful substance content and the surrounding population density, the overflow hazard scores of multiple drainage facilities in the current monitoring period are determined, including: According to the formula Determine the overflow hazard score of the i-th drainage facility in the current monitoring period ,in, is the population density around the i-th drainage facility, is the surrounding population density threshold, is the average value of the jth harmful substance content in the ith drainage facility during the current monitoring period, is the j-th hazardous substance content change rate threshold, and is the preset weight, E is the number of types of harmful substances in sewage, j≤E, and j and E are both positive integers.

6. A drainage intelligent monitoring system based on Internet of Things perception, used to execute the drainage intelligent monitoring method based on Internet of Things perception as described in any one of claims 1 to 5, characterized in that: include: The facility water level data module is used to obtain the facility water level data of multiple drainage facilities through IoT sensing technology at multiple moments in the current monitoring cycle; A facility parameter module, used to obtain facility parameters of a plurality of drainage facilities, wherein the facility parameters include volume data and usage days; A facility flow prediction data module, used for inputting the facility water level data and the volume data into a trained facility flow prediction model to obtain predicted facility flow data of multiple drainage facilities at multiple moments in the current monitoring period; The precipitation data module is used to obtain precipitation data of the areas where multiple drainage facilities are located at multiple times in the current monitoring period; A predicted overflow probability data module is used to input the precipitation data, the predicted facility flow data, the usage days and the facility water level data into a trained overflow probability prediction model to obtain predicted overflow probability data of multiple drainage facilities at multiple moments in the current monitoring period; The harmful substance content and surrounding population density module is used to obtain the harmful substance content of sewage from multiple drainage facilities at multiple moments in the current monitoring period and the surrounding population density of multiple drainage facilities; A spillover hazard scoring module, used to determine spillover hazard scores of multiple drainage facilities in the current monitoring period according to the harmful substance content of the sewage and the surrounding population density; A grading weight module, used to set grading weights for the spillover hazard score and the predicted spillover probability data; The priority control module is used to determine the drainage facilities to be controlled with priority according to the classification weights, the spillover hazard scores and the predicted spillover probability data.

Images