Water level signal analyzing and processing method and system

Abstract

The invention discloses a water level signal analyzing and processing method and system, and relates to the technical field of water level signal analyzing and processing. Comprising the steps of placing a plurality of water level gauges and collecting water level signals of the water level gauges; performing primary processing on the collected water level signal; verifying the effectiveness of the sampling point of the water level signal and the effectiveness of the measured value; and carrying out secondary processing on the water level signal, and carrying out automatic early warning according to a secondary processing result. The accuracy and timeliness of water level prediction are improved, flood control and reservoir management are facilitated, abnormal water level conditions can be found and handled in time, and disaster risks are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of water level signal analysis and processing, and in particular to a method and system for analyzing and processing a water level signal. Background Art

[0002] The cascade power stations in the Huoxi River Basin adopt an unmanned (few people on duty) and remote centralized control operation mode. The cascade power stations are connected end to end and distributed in a 40km basin. The water accumulation area is above 1,300m above sea level. It has a subtropical mountain climate and is located in a narrow valley in the mountainous area. It is prone to natural disasters such as earthquakes, mountain torrents, and mudslides. Mudslides caused by floods often cause communication and power line interruptions and road obstructions, making it difficult for dam duty personnel to evacuate and for power station support personnel to arrive in time.

[0003] The reservoir water level is an important basis for opening the flood discharge gates. The accuracy and stability of the water level signal will affect the safe and stable operation of the entire flood discharge gate system. In order to prevent the reservoir water level from exceeding the warning line or even overflowing the dam, and to further improve the dam's ability to withstand extreme natural disasters, it is necessary to carry out intelligent transformation in accordance with higher "unmanned duty and emergency response" requirements. Summary of the invention

[0004] In view of the problem that existing power station support personnel cannot arrive in place in time, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to perform automatic control according to the water level signal.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a method for analyzing and processing water level signals, which includes placing multiple water level meters to collect water level signals from the water level meters; performing a primary processing on the collected water level signals; verifying the validity of the sampling points of the water level signals and the validity of the measured values; performing a secondary processing on the water level signals, and issuing an automatic warning based on the results of the secondary processing.

[0008] As a preferred solution of the method for analyzing and processing water level signals of the present invention, the water level meter includes a pressure water level meter, a radar water level meter and an immersion water level meter.

[0009] As a preferred solution of the water level signal analysis and processing method described in the present invention, the primary processing includes filtering processing, setting threshold values ​​and fault-tolerant processing; the filtering processing includes selecting a sliding window, sorting the data in the window, and selecting the sorted middle value as the output of the window; using a Butterworth low-pass filter to set the cutoff frequency; the threshold value includes a static threshold value, a dynamic threshold value and a change rate threshold value; the fault-tolerant processing includes: using the 3σ principle to detect abnormal values, calculating the difference between the current reading and the reading at the previous moment, if the difference exceeds 3σ, it is marked as abnormal; for data points marked as abnormal, if it is a single abnormal point, it is corrected using linear interpolation; if it is multiple consecutive abnormal points, the previous valid value is used to maintain it.

[0010] As a preferred scheme of the water level signal analysis and processing method described in the present invention, the one-time processing also includes: in the automatic adjustment loop, the analog quantity is selected by the three-average method, and the specific operation is as follows: if the two radar water level gauges and one immersion water level gauge in front of the gate are normal, then the average value of the two radar water level gauges is taken; if one immersion water level gauge in front of the gate is abnormal, then the average value of the two radar water level gauges is taken; if one radar water level gauge in front of the gate is abnormal, then the value of another normal radar water level gauge is taken; if both radar water level gauges are abnormal, then the value of the normal immersion water level gauge is taken.

[0011] As a preferred solution of the water level signal analysis and processing method of the present invention, the secondary processing of the water level signal includes: collecting historical reservoir information data, selecting an additive model for time series decomposition:

[0012] Y(t)=T(t)+S(t)+R(t)

[0013] Among them, Y(t) is time series data, T(t) is the trend component, S(t) is the seasonal component, and R(t) is the residual component. STL is used to decompose the time series and determine the specific value of each component. The long short-term memory network LSTM deep learning model is used to train the water level prediction model based on historical data. The input dimension is the number of features for each time step, and the output dimension is the number of hidden units in the LSTM layer. The output layer formula is:

[0014]

[0015] in, is the predicted water level for the next 6 hours, W y and b y are the weights and biases of the output layer, h t is the hidden state; use mean square error as the loss function:

[0016]

[0017] Where n is the number of samples, is the predicted value, Y i is the true value; the Adam optimization algorithm is used to update the model parameters. The formula of the Adam optimization algorithm is:

[0018]

[0019] Among them, α is the learning rate, and are momentum and RMS estimates, ∈ is a constant, θ t is the updated value of the parameter at time step t.

[0020] As a preferred solution of the water level signal analysis and processing method of the present invention, wherein: automatic warning according to the result of the secondary processing includes: based on historical data, calculating the distribution of water level change rate under different seasons and different initial water level conditions, including the following steps: for each category, that is, the combination of season and initial water level, calculating the mean value μ of the water level change rate 0 and standard deviation σ 0 :Define different seasonal weight coefficients W s ; The effect of initial water level on the change rate is as follows:

[0021] f(h 0 )=1+0.1×(h 0 -h avg ) / h range

[0022] Among them, h 0 is the current water level, h avg is the mean water level, h range is the water level range; based on the rainfall forecast for the next 24 hours:

[0023] R=1+0.05×z 1

[0024] Among them, z 1 To forecast rainfall, the unit is mm; calculate the water level change trend T in the past 7 days g :

[0025] T g =1+0.1×(s z -s q ) / s avg

[0026] Among them, s z is the latest water level, s q is the water level 7 days ago, s avg is the 7-day average water level; dynamic alarm threshold calculation:

[0027] ΔH th =W s ×f(h 0 )×R×T g ×(μ 0 +k.σ 0 )

[0028] Where ΔH th is the dynamic alarm threshold, and k is the sensitivity coefficient.

[0029] As a preferred solution of the water level signal analysis and processing method described in the present invention, the automatic warning based on the result of the secondary processing also includes: real-time collection of water level data, input of a water level prediction model, comparison of the predicted water level with the current dynamic alarm threshold, and triggering of a warning of the corresponding level if the predicted water level exceeds the alarm threshold; setting a three-level water level gradient change alarm, monitoring water level changes according to preset target water levels and water level dead zones, and automatically starting a flood control alarm when the water level rises to the set target water level dead zone; when the water level rises to the upper limit of the target water level dead zone or the water level rise gradient exceeds the set value, an alarm is issued to start the gate linkage control process; after the alarm is issued, if there is no human intervention, an instruction is automatically issued to perform the gate lifting operation.

[0030] In the second aspect, an embodiment of the present invention provides a water level signal analysis and processing system, which includes: a water level signal acquisition module, used to place multiple water level meters and collect water level signals from the water level meters; a primary processing module, used to perform primary processing on the collected water level signals; a validity verification module, used to verify the validity of the sampling points of the water level signals and the validity of the measured values; a secondary processing module, used to perform secondary processing of the water level signals and to issue automatic warnings based on the results of the secondary processing.

[0031] In a third aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the water level signal analysis and processing method as described in the first aspect of the present invention are implemented.

[0032] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the method for analyzing and processing water level signals as described in the first aspect of the present invention are implemented.

[0033] The beneficial effects of the present invention are as follows: the present invention improves the coverage of water level monitoring and the diversity of data by installing multiple water level gauges before and after the gate respectively, and utilizes the advantages of different types of water level gauges, especially radar water level gauges, to effectively measure lower water levels, making up for the shortcomings of pressure water level gauges; through primary processing, validity verification and secondary processing of data, the entire water level signal analysis and processing method has achieved significant improvements in accuracy, stability and intelligence; in summary, the present invention improves the accuracy and timeliness of water level prediction, is helpful for flood control and reservoir management, can promptly detect and handle abnormal water level conditions, reduce disaster risks, and through automatic control, reduces human intervention and improves the intelligence level of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 Flow chart of the analysis and processing method of water level signal.

[0036] Figure 2 This is a schematic cross-sectional view of the gate of the Muzuo Power Station. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0038] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0040] Example 1

[0041] Reference Figure 1-2 , which is the first embodiment of the present invention, and provides a method for analyzing and processing a water level signal, comprising:

[0042] S1: Place multiple water level meters and collect water level signals from the placed water level meters.

[0043] Muzuo Hydropower Station is a high-head diversion hydropower station. The head hub is composed of a river-blocking dam. From left to right along the dam axis are the left bank retaining dam, the water intake dam section, the retaining dam, the 1-hole sand flushing gate, the 3-hole spillway gate, and the right bank retaining dam. The total length of the dam axis is 157.70m, the dam top elevation is 1547.00m, the Muzuo Dam spillway bottom plate elevation is 1528.00m, the normal water storage level of the power station is 1545.00m, the minimum operating water level is 1539.00m, the maximum dam height is 20m, the corresponding storage capacity is 344,000 m3, the regulating storage capacity is 223,000 m3, the design head is 262.7m, the diversion flow is 43.02m3 / s, the diversion tunnel is about 12km long, the tailwater tunnel (channel) is 406m long, and the installed capacity of the power station is 2×50MW. It is a medium-sized hydropower station project. The power station development task is to generate electricity, such as Figure 1 shown.

[0044] At present, there is a pressure water level gauge in front of the trash rack of Muzuo Dam, and two pressure water level gauges behind the rack. Due to the limited installation position, the pressure water level gauge cannot measure the lower water level. In order to increase the accuracy and stability of water level measurement, two radar water level gauges are added in front of the rack and one radar water level gauge is added behind the rack. The water level gauges are connected to the dam LCU to collect water level information.

[0045] After adding radar water level gauges, there are 3 water level gauges before the gate (2 radars and 1 drop-type) and 3 water level gauges after the gate (1 radar and 2 drop-type), and the water level before the gate is used as the reservoir water level control target.

[0046] S2: Process the collected water level signal once.

[0047] Among them, one-time processing includes filtering processing, setting threshold value and fault tolerance processing.

[0048] Specifically, according to the characteristics of the reservoir, a suitable sampling frequency is set, such as sampling once every 2 seconds, and the collected raw water level data is temporarily stored in a buffer.

[0049] S2.1: The filtering process includes: selecting a sliding window, such as 11 sampling points, and sorting the data in the window; selecting the middle value after sorting as the output of the window, sliding the window, and repeating the above steps; using a Butterworth low-pass filter to set the cutoff frequency, such as 0.1 Hz, to remove high-frequency noise, processing the signal through the filter, and outputting the filtered estimate.

[0050] S2.2: Threshold values ​​include static threshold values, dynamic threshold values, and change rate threshold values.

[0051] The process of setting the static threshold value is as follows: according to the characteristics of the reservoir, a reasonable range of water level is set, such as 1530m to 1545m, and any reading outside this range is marked as a potential abnormality.

[0052] The setting of dynamic threshold value includes: calculating the mean μ and standard deviation σ of water level data in the past hour, setting the dynamic threshold value to μ±3σ, and any reading outside this range is marked as potential abnormality.

[0053] The setting of the change rate threshold value includes: calculating the change rate of the water level, such as the change per minute, setting the maximum allowable change rate, such as 0.1m / min, and any reading exceeding this change rate is marked as a potential abnormality.

[0054] S2.3: Fault tolerance processing includes the following steps:

[0055] The 3σ principle is used to detect outliers. The difference between the current reading and the reading at the previous moment is calculated. If the difference exceeds 3σ, it is marked as an abnormality. For data points marked as abnormal, if it is a single abnormal point, linear interpolation is used for correction; if there are multiple consecutive abnormal points, the previous valid value is used to maintain it.

[0056] Furthermore, in the automatic adjustment loop, the selection of analog quantity adopts the method of three averages, which can overcome the system error to a certain extent. The specific operation is as follows:

[0057] If the two radar water level gauges and one immersion water level gauge in front of the gate are both normal, the average value of the two radar water level gauges shall be taken; if one immersion water level gauge in front of the gate is abnormal, the average value of the two radar water level gauges shall be taken; if one radar water level gauge in front of the gate is abnormal, the value of the other normal radar water level gauge shall be taken; if both radar water level gauges are abnormal, the value of the normal immersion water level gauge shall be taken.

[0058] S3: Verify the validity of the water level signal, including checking the validity of the sampling points and the validity of the measured values.

[0059] The validity of the collected water levels includes: verifying the validity of the sampling points and the validity of the measured values.

[0060] Verify the validity of the sampling point: Check the status flag (IV / NT / OV) of the sampling point.

[0061] The process of verifying the validity of the measurements is as follows:

[0062] Whether the measured value is within the valid range: If it exceeds the measurement range, the water level measurement exceeds the limit and an alarm is issued;

[0063] When all water level meter sampling signal channels fail (delay 2s), or the effective water level value has not been updated for more than 10 minutes, a water level meter sampling failure alarm is issued;

[0064] The installation position of the immersion water level is limited, and measurements cannot be made below the installation position, that is, there is a measurement dead zone. To ensure the effectiveness of the measurement, the measurement range is 0.1m above the installation position.

[0065] When the difference deviation between two water level gauges (3) in the same group is too large (the deviation exceeds 10% of the measuring range), the water level of this group will be invalidated and the "water level gauge deviation is too large alarm" will be reported.

[0066] S4: Secondary process the water level signal and issue an automatic warning based on the processing results.

[0067] Preferably, the secondary processing of the water level signal includes:

[0068] Collect relevant data such as reservoir water level, rainfall, seasonal information, etc. in the past few years and standardize the data; use time series decomposition technology to decompose the water level data into trend, seasonal and residual components, analyze the characteristics of each component, and identify the periodic pattern of water level changes.

[0069] Specifically, select the additive model for time series decomposition:

[0070] Y(t)=T(t)+S(t)+R(t)

[0071] Among them, Y(t) is the time series data, T(t) is the trend component, S(t) is the seasonal component, and R(t) is the residual component.

[0072] Use STL to decompose the time series and determine the specific values ​​of each component (trend, seasonality, and residual).

[0073] Machine Learning Model Building:

[0074] The LSTM deep learning model is used to train the water level prediction model based on historical data. The input features include water level data of the past 24 hours, rainfall forecast, seasonal information, etc. The output is the predicted water level for the next 6 hours. The specific operations are as follows:

[0075] The input dimension is the number of features for each time step (such as water level data, rainfall, seasonal information), and the output dimension is the number of hidden units in the LSTM layer. The output layer formula is:

[0076]

[0077] in, is the predicted water level for the next 6 hours, W y and b y are the weights and biases of the output layer, h t In hidden state.

[0078] Use mean square error (MSE) as the loss function:

[0079]

[0080] Where n is the number of samples, is the predicted value, Y i is the true value.

[0081] The Adam optimization algorithm is used to update the model parameters. The formula of the Adam optimization algorithm is:

[0082]

[0083] Among them, α is the learning rate, and are momentum and RMS estimates, ∈ is a constant, θ t is the updated value of the parameter at time step t.

[0084] The dataset is divided into a training set and a validation set, and iterative training is performed. After each round of training, the loss on the validation set is calculated to monitor the model performance.

[0085] Dynamic water level alarm threshold setting:

[0086] Based on historical data, the distribution of water level change rate under different seasons and different initial water level conditions is calculated, and dynamic alarm thresholds are set for different situations. For example, the 90% quantile is set as the first-level alarm threshold, the 95% quantile is set as the second-level alarm threshold, and the 99% quantile is set as the third-level alarm threshold.

[0087] Specifically, water level data from the past few years were collected, including timestamps, water level values, and rainfall, and the data were classified by season (spring, summer, autumn, and winter) and initial water level range (such as low, medium, and high).

[0088] For each category (season and initial water level combination), calculate the mean value μ of the water level change rate 0 and standard deviation σ 0 :

[0089] Define the seasonal weight coefficient W s , reflecting the sensitivity of water level changes in different seasons, for example: spring = 1.2, summer = 1.5, autumn = 1.1, winter = 1.0.

[0090] Define function f(h) to reflect the effect of initial water level on the rate of change:

[0091] f(h 0 )=1+0.1×(h 0 -h avg ) / h range

[0092] Among them, h 0 is the current water level, h avg is the mean water level, h range is the water level range.

[0093] Introduce the rainfall forecast factor R, based on the rainfall forecast for the next 24 hours:

[0094] R=1+0.05×z 1

[0095] Among them, z 1 Forecast rainfall (mm).

[0096] Calculate the water level change trend T for the past 7 days g :

[0097] T g =1+0.1×(s z -s q ) / s avg

[0098] Among them, s z is the latest water level, s q is the water level 7 days ago, s avg It is the 7-day average water level.

[0099] Dynamic alarm threshold calculation:

[0100] ΔH th =W s ×f(h 0 )×R×T g ×(μ 0 +k.σ 0 )

[0101] Where ΔH th is the dynamic alarm threshold, k is the sensitivity coefficient, and different levels of alarms can be set as needed (such as k=2 for level one alarm, k=3 for level two alarm), etc.

[0102] Monthly update 0 and σ 0 , using the exponentially weighted moving average method, adjust W every quarter based on the alarm accuracy of the previous quarter s .

[0103] Real-time monitoring and early warning:

[0104] Collect water level data in real time, input the prediction model, compare the predicted water level with the current dynamic alarm threshold, and trigger an early warning of the corresponding level if the predicted water level exceeds the alarm threshold.

[0105] A three-level water level gradient change alarm is set up to monitor water level changes based on the preset target water level and water level dead zone. When the water level rises to the set target water level dead zone, the flood control alarm is automatically activated; when the water level rises to the upper limit of the target water level dead zone or the water level rise gradient exceeds the set value, an alarm is issued to start the gate joint control process; after the alarm is issued, if there is no human intervention, an instruction is automatically issued to perform the gate lifting operation.

[0106] Furthermore, this embodiment also provides a water level signal analysis and processing system, comprising:

[0107] A water level signal acquisition module is used to place multiple water level meters and collect water level signals from the water level meters; a primary processing module is used to perform primary processing on the collected water level signals; a validity verification module is used to verify the validity of the sampling points of the water level signals and the validity of the measured values; a secondary processing module is used to perform secondary processing of the water level signals and to issue automatic warnings based on the results of the secondary processing.

[0108] This embodiment also provides a computer device suitable for the analysis and processing method of water level signals, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the analysis and processing method of water level signals proposed in the above embodiment.

[0109] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0110] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for analyzing and processing a water level signal as proposed in the above embodiment is implemented.

[0111] In summary, the present invention improves the coverage of water level monitoring and the diversity of data by installing multiple water level meters before and after the gate respectively, and utilizes the advantages of different types of water level meters, especially radar water level meters, which can effectively measure lower water levels and make up for the shortcomings of pressure water level meters; through primary processing, validity verification and secondary processing of data, the entire water level signal analysis and processing method has achieved significant improvements in accuracy, stability and intelligence; in summary, the present invention improves the accuracy and timeliness of water level prediction, which is helpful for flood control and reservoir management. The dynamic early warning system can timely detect and handle abnormal water level conditions and reduce disaster risks. Moreover, through automatic control, human intervention is reduced and the intelligence level of the system is improved.

[0112] Example 2

[0113] Reference Figure 1-2 , which is the second embodiment of the present invention, and this embodiment provides a method for analyzing and processing a water level signal. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0114] In order to verify the effectiveness of the water level signal analysis and processing method, especially its application effect in actual reservoir management, this embodiment takes the Muzuo Hydropower Station as the test object and conducts detailed experimental tests. The Muzuo Hydropower Station is a high-head diversion hydropower station with complex water level management requirements. The test content includes the following steps:

[0115] Test location: Muzuo Hydropower Station; Test object: water level signals before and after the gate; Test equipment: water level meter, data acquisition system (LCU); Test time: June 1, 2024 to June 30, 2024.

[0116] Two radar water level gauges and one pressure water level gauge are placed in front of the gate, and one radar water level gauge and two pressure water level gauges are placed behind the gate. All water level gauges are connected to the dam LCU system to collect and transmit water level data in real time. The sampling frequency is set to once every 2 seconds to ensure high frequency and high accuracy of the data.

[0117] The water level signal is sorted using the sliding window method (window size is 11 sampling points) and the median is taken as the output. Then a Butterworth low-pass filter is used with a cutoff frequency of 0.1 Hz to remove high-frequency noise; the static threshold is set to 1530m to 1545m, the dynamic threshold is μ±3σ (the mean and standard deviation of the water level data in the past hour), and the change rate threshold is 0.1m / min; the 3σ principle is used to detect abnormal values, and if the difference exceeds 3σ, it is marked as abnormal; single abnormal points are corrected by linear interpolation, and multiple consecutive abnormal points are kept by the previous valid value.

[0118] Verify the status flag (IV / NT / OV) of the sampling point and check whether the measured value is within the valid range. If it exceeds the measurement range, an alarm will be issued immediately. When all water level meter sampling signal channels fail or the effective water level value has not been updated for more than 10 minutes, a water level meter sampling failure alarm will be issued. For the measurement dead zone of the immersion water level meter, its measurement range is limited to 0.1m above the installation position.

[0119] The reservoir water level, rainfall and seasonal information of the past five years were collected and the data was standardized. The water level data was decomposed into trend, seasonal and residual components using time series decomposition technology, and the characteristics of each component were analyzed. Based on the long short-term memory network (LSTM) deep learning model, the water level data, rainfall forecast and seasonal information of the past 24 hours were input, and the predicted water level for the next 6 hours was output. If the predicted water level exceeds the alarm threshold, the early warning mechanism is triggered. Some experimental data are as follows:

[0120] Table 1 Water level meter data table

[0121]

[0122] As can be seen from the table above, placing various types of water level gauges before and after the gate ensures the comprehensiveness and diversity of data collection, especially the radar water level gauge shows significant advantages when measuring lower water levels. The data shows that the data consistency of the water level gauges before and after the gate is high, ensuring the accuracy of water level monitoring.

[0123] Moreover, the water level prediction based on historical data and LSTM model significantly improves the accuracy and timeliness of the early warning. The predicted water level in the table is close to the actual water level, indicating the high efficiency of the prediction model. The setting of dynamic alarm threshold enables the system to flexibly respond to changes in different seasons and initial water level conditions, ensuring the accuracy of the early warning.

[0124] In summary, the present invention significantly improves the accuracy, stability and intelligence level of reservoir water level monitoring through a multi-step water level signal analysis and processing method. Compared with the prior art, the present invention has obvious advantages in data processing and early warning mechanism, and can effectively cope with water level changes in complex hydrological environments and ensure the safe operation of reservoirs.

[0125] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for analyzing and processing a water level signal, characterized in that: include, Placing a plurality of water level gauges and collecting water level signals of the water level gauges; Processing the collected water level signal once; Verifying the validity of the sampling points and the validity of the measured values ​​of the water level signal; The water level signal is subjected to secondary processing, and an automatic early warning is issued according to the result of the secondary processing.

2. The method for analyzing and processing water level signals according to claim 1, characterized in that: The water level gauge includes a pressure water level gauge, a radar water level gauge and an immersion water level gauge.

3. The method for analyzing and processing water level signals according to claim 2, characterized in that: The primary processing includes filtering, setting threshold value and fault tolerance processing; The filtering process includes selecting a sliding window, sorting the data in the window, and selecting the sorted middle value as the output of the window; Use a Butterworth low-pass filter to set the cutoff frequency; The threshold values ​​include static threshold values, dynamic threshold values ​​and change rate threshold values; The fault-tolerant processing includes: using the 3σ principle to detect abnormal values, calculating the difference between the current reading and the reading at the previous moment, and marking it as abnormal if the difference exceeds 3σ; For data points marked as abnormal, if it is a single abnormal point, linear interpolation is used for correction; if it is multiple consecutive abnormal points, the previous valid value is used to maintain it.

4. The method for analyzing and processing water level signals according to claim 3, characterized in that: The one-time processing also includes: In the automatic regulation loop, the selection of analog quantity adopts the method of three averages, and the specific operation is as follows: If the two radar water level gauges and one immersion water level gauge in front of the gate are both normal, the average value of the two radar water level gauges shall be taken; if one immersion water level gauge in front of the gate is abnormal, the average value of the two radar water level gauges shall be taken; if one radar water level gauge in front of the gate is abnormal, the value of the other normal radar water level gauge shall be taken; if both radar water level gauges are abnormal, the value of the normal immersion water level gauge shall be taken.

5. The method for analyzing and processing water level signals according to claim 4, characterized in that: The secondary processing of the water level signal includes: Collect historical reservoir information data and select the additive model for time series decomposition: Y(t)=T(t)+S(t)+R(t) Among them, Y(t) is the time series data, T(t) is the trend component, S(t) is the seasonal component, and R(t) is the residual component; Use STL to decompose the time series and determine the specific value of each component; The water level prediction model is trained based on historical data using the long short-term memory network LSTM deep learning model. The input dimension is the number of features for each time step, and the output dimension is the number of hidden units in the LSTM layer. The output layer formula is: in, is the predicted water level for the next 6 hours, W y and b y are the weights and biases of the output layer, h t is a hidden state; Use mean squared error as the loss function: Where n is the number of samples, is the predicted value, Y i is the true value; The Adam optimization algorithm is used to update the model parameters. The formula of the Adam optimization algorithm is: Among them, α is the learning rate, and are momentum and RMS estimates, ∈ is a constant, θ t is the updated value of the parameter at time step t.

6. The method for analyzing and processing water level signals according to claim 5, characterized in that: Automatically warning based on the result of the secondary processing includes: Based on historical data, the distribution of water level change rate under different seasons and different initial water level conditions is calculated, including the following steps: For each category, that is, season and initial water level combination, the mean μ0 and standard deviation σ0 of the water level change rate are calculated: Define different seasonal weight coefficients W s ; The effect of initial water level on the rate of change is as follows: <h2 style=";text-align:left;direction:ltr">f(h0) = 1 + 0.1 × (h0 - h<h2 style=";text-align:left;direction:ltr"> avg <h2 style=";text-align:left;direction:ltr"> ) / h<h2 style=";text-align:left;direction:ltr"> range Among them, h0 is the current water level, h avg is the mean water level, h range is the water level range; Based on the rainfall forecast for the next 24 hours: R=1+0.05×z1 Among them, z1 is the forecast rainfall, in mm; Calculate the water level change trend T for the past 7 days g : T g =1+0.1×(s z -s q ) / s avg Among them, s z is the latest water level, s q is the water level 7 days ago, s avg is the 7-day average water level; Dynamic alarm threshold calculation: ΔH th =W s ×f(h0)×R×T g ×(μ0+k.σ0) Where ΔH th is the dynamic alarm threshold, and k is the sensitivity coefficient.

7. The method for analyzing and processing water level signals according to claim 6, characterized in that: The automatic early warning of the secondary processing result also includes: Collect water level data in real time, input into the water level prediction model, compare the predicted water level with the current dynamic alarm threshold, and trigger the corresponding level of warning if the predicted water level exceeds the alarm threshold; A three-level water level gradient change alarm is set up to monitor water level changes based on the preset target water level and water level dead zone. When the water level rises to the set target water level dead zone, the flood control alarm is automatically activated; when the water level rises to the upper limit of the target water level dead zone or the water level rise gradient exceeds the set value, an alarm is issued to start the gate joint control process; after the alarm is issued, if there is no human intervention, an instruction is automatically issued to perform the gate lifting operation.

8. A water level signal analysis and processing system, based on the water level signal analysis and processing method according to any one of claims 1 to 7, characterized in that: Also includes, A water level signal acquisition module, used to place multiple water level gauges and collect water level signals from the water level gauges; A primary processing module, used for performing primary processing on the collected water level signal; A validity verification module, used to verify the validity of the sampling points of the water level signal and the validity of the measured values; The secondary processing module is used to perform secondary processing of the water level signal and to issue automatic warnings based on the results of the secondary processing.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for analyzing and processing the water level signal according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for analyzing and processing a water level signal according to any one of claims 1 to 7 are implemented.

Images