Drainage facility management system, drainage facility management method, and program for realizing the same

Abstract

To provide a drainage facility management system and a drainage facility management method that can improve an operator's business efficiency.SOLUTION: A drainage facility management system of the present invention comprises a water level measuring instrument and a drainage facility attached by the river, and the drainage facility management system comprises: a past operation result database of the water level measuring instrument and the drainage facility; a water level prediction processing unit that calculates the future water level on the basis of data of the operation result database and current data from the water level measuring instrument; an operator information database for an operator to run to the drainage facility; and a dispatch time calculation unit that calculates the operator's dispatch time on the basis of data in the operator information database and data in the water level prediction processing unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a drainage facility management system, a drainage facility management method, and a program for implementing the same. [Background technology]

[0002] When weather changes suggest that river water levels may rise, pumping station operators must dispatch to the site to operate the drainage facilities. To do so, they check the local weather, anticipate rising river water levels, and then rush to the local drainage facilities. They must make comprehensive decisions based on factors such as when the operating water level will be reached, how many minutes it will take to arrive at the site, and how many minutes it will take to operate the facilities after arriving. Furthermore, determining when to operate the facilities depends on a variety of factors, including simply checking the water level of the upstream river, the operation of nearby drainage facilities, and geographical conditions such as mountainous or flat areas, as well as the amount of rainfall. Therefore, operators often rush to the site before the rain starts, which often results in wasted time. Therefore, there is a need for a simple calculation tool that can be easily adapted by users.

[0003] Patent Document 1 discloses a method for remote monitoring of drainage facilities, in which river water levels are predicted using river information from water level gauges, flow meters, rain gauges, etc. and a calculation simulator, and operating methods are extracted from an operational rules database to provide guidance to operators. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 10-204965 Summary of the Invention [Problem to be solved by the invention]

[0005] In Patent Document 1, water level predictions and operation patterns are guided solely by a calculation simulator that models an actual river on a system. The guidance is provided on the assumption that the operator is in a location where they can operate the drainage equipment's control panel. In this case, considerable CPU power is required for the calculation simulation, and the river model also needs to be tuned periodically.

[0006] An object of the present invention is to provide a drainage facility management system, a drainage facility management method, and a program for implementing the same, which can improve the work efficiency of operators. [Means for solving the problem]

[0007] The drainage equipment management system of the present invention is a drainage equipment management system equipped with water level measuring devices and drainage equipment installed in a river, and is characterized by comprising: a database of past operating records of the water level measuring devices and the drainage equipment, a water level prediction processing unit that calculates future water levels based on data from the operating record database and current data from the water level measuring devices, an operator information database for operators to rush to the drainage equipment, and a dispatch time calculation unit that calculates the dispatch time of the operator based on data from the operator information database and data from the water level prediction processing unit.

[0008] Alternatively, the drainage equipment management method of the present invention is a drainage equipment management method equipped with a water level measuring device and drainage equipment installed in a river, characterized in that it includes a water level prediction step of calculating a future water level based on data in a database of past operating performance of the water level measuring device and the drainage equipment and current data from the water level measuring device, and a dispatch time calculation step of calculating the dispatch time of an operator based on data in an operator information database for an operator to rush to the drainage equipment and the data calculated in the water level prediction step.

[0009] Alternatively, the program of the present invention is a program for a drainage equipment management system equipped with water level measuring devices and drainage equipment installed in a river, which program causes a computer to realize a water level prediction function that calculates future water levels based on data in a database of past operating performance of the water level measuring devices and the drainage equipment and current data from the water level measuring devices, and a dispatch time calculation function that calculates the dispatch time of an operator based on data in an operator information database for an operator to rush to the drainage equipment and data calculated by the water level prediction function. [Effects of the Invention]

[0010] According to the present invention, it is possible to provide a drainage facility management system, a drainage facility management method, and a program for realizing the same, which can improve the work efficiency of operators. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a schematic diagram of a drainage facility management system according to an embodiment of the present invention; [Figure 2] FIG. 10 is a diagram illustrating a predicted water level calculation method according to the present embodiment. [Figure 3] FIG. 10 is a diagram showing the time course of water level and rainfall according to this embodiment. [Figure 4] FIG. 10 is a diagram showing a water level trend according to this embodiment. [Figure 5] FIG. 10 is a diagram showing water levels and past data according to this embodiment. [Figure 6] FIG. 10 is a diagram showing water levels and past data according to this embodiment. [Figure 7] FIG. 10 is a diagram showing water levels and past data according to this embodiment. [Figure 8] FIG. 10 is a diagram illustrating calculation of water level from tide level trend according to the present embodiment. [Figure 9] FIG. 10 is a diagram showing the relationship between pump and gate operation data and water level according to this embodiment. [Figure 10] FIG. 10 is a diagram showing the relationship between pump and gate operation (detailed version) data and water level in this embodiment. [Figure 11]FIG. 10 is a diagram showing past water level trend data according to this embodiment. [Figure 12] FIG. 10 is a diagram showing prediction of operation time using an approximation curve according to the present embodiment. [Figure 13] 1 shows a system flow according to this embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] In an embodiment of the present invention, a water level prediction system for drainage facilities is implemented by extracting data that is similar to past performance data, and by performing a performance comparison type water level prediction, and also by providing guidance on instructions to an operator after the water level prediction. Hereinafter, an embodiment of the present invention will be described with reference to the drawings. [Example]

[0013] In this example, a river drainage pump is used as the target drainage facility. The data acquired as environmental data for calculation includes the water level in the pump well immediately adjacent to the pump, the water level of the inflowing river located upstream of the facility, rain gauge data (rainfall amount) located upstream of the pump facility, the length of the river or the linear distance from the rain gauge measurement point to the pump station, the specifications of nearby pump stations, the current location of the operator, and the location of the pump station.

[0014] The drainage equipment management system of this embodiment mainly comprises an operation record database 1, a water level measuring device 2, a water level prediction processing unit 3, an operator information database 4, and a dispatch time calculation unit 5. It also comprises a dispatch time guidance unit 6.

[0015] The drainage equipment management system is provided with a program for causing a computer to realize the functions of a water level prediction processing unit 3, a dispatch time calculation unit 5, and a dispatch time guidance unit 6. In other words, it is provided with a program for causing a computer to realize a water level prediction function for calculating future water levels based on data from a database 1 of past operation records of the water level measuring device 2 and the drainage equipment and current data from the water level measuring device 2, and a dispatch time calculation function for calculating the dispatch time of an operator based on data from a database 4 of operator information for an operator to rush to the drainage equipment and data calculated by the water level prediction function.

[0016] The water level measuring device is installed in the river and measures the river's water level. The drainage equipment drains water depending on the rise in the river's water level. The operation history database stores past measurements from the water level measuring device and the performance of the drainage equipment. The water level prediction processing unit calculates future water levels based on data from the operation history database and current data from the water level measuring device. The operator information database stores information for operators to rush to the drainage pumping station. The dispatch time calculation unit calculates the dispatch time for operators based on data from the operator information database and data from the water level prediction processing unit. This configuration makes it possible to provide a drainage equipment management system and a drainage equipment management method that can improve the work efficiency of operators. In other words, operators can clearly determine whether or not to dispatch and when to dispatch, thereby improving work efficiency. In addition, the system has a dispatch time guidance unit 6 that can convey appropriate information to operators.

[0017] The main steps include a water level prediction step of calculating future water levels based on data from a database of past operational records of water level measuring devices and drainage equipment and current data from the water level measuring devices, and a dispatch time calculation step of calculating the dispatch time for operators based on data from an operator information database for operators to rush to the drainage equipment and data from the water level prediction processing unit. This drainage equipment management method can improve the work efficiency of operators. Furthermore, since it includes a dispatch time guidance step of transmitting the calculated dispatch time to the operator, appropriate information can be conveyed to the operator.

[0018] The water level prediction processing unit calculates the future river water level. Methods for calculating the future water level include, for example, a method of calculating it from the rate of rise in the pump well water level or a method of calculating it from the tide level.

[0019] We will explain how to calculate the future water level from the rate of rise of the pump well water level. First, the predicted water level calculation method Let the current time and pump well water level be t1 and a1, respectively, the time to reach the site be t2, the pump operation preparation time be t3, the waiting time be t4, the expected operating water level be s1, and the operating time be ut1.

[0020] As shown in Figure 2, for example, if the pump well water level a1 indicates a rise in water level, the rate of water level rise is b1 = (a1 - a0) / (t1 - t0). The expected operating water level s1 can be easily calculated using the general linear function formula y = ax + b, so s1 = b1 × (ut1 - t0) + a0, and the time span between t1 and t0 can be considered the water level data update cycle. The Ministry of Land, Infrastructure, Transport and Tourism's river disaster prevention information water level gauges are updated every 10 minutes, so the basic value is t1 - t0 = 10 minutes.

[0021] A characteristic of water levels is that they rise over time at the beginning of rainfall, but the rate at which the water level rises can be gradual or sudden depending on the intensity of the rainfall and changes in river width. For this reason, it is considered difficult to predict future water levels using the rate of water level rise from the most recent data, as mentioned above. Furthermore, because rain clouds drop rain as they move with the wind, even sudden rainfall will have varying intensity, making it unlikely that the water level will rise quadratically. Therefore, the function to be applied is calculated as a moving average using data from the past few points.

[0022] The selection of the past few points can be freely set by the user, and depending on the river to which it is applied, it is possible to select a large amount of past data for large rivers, which change gradually, and a small amount of past data for small rivers, which change rapidly. Therefore, it is possible to display multiple lines with different setting values ​​for moving average lines, as is often used in foreign exchange trading.

[0023] Although various approximation curves are assumed, such as exponential functions, a linear function is used to make it easier for users to estimate deviations from actual measurements. Also, since calculations using two recent water levels can have a large fluctuation in the values, the calculation results are output as two results: the most recent and a moving average.

[0024] Meanwhile, rainfall patterns can be considered as patterns 1 to 4. Pattern 1 is a prolonged rainfall where the water level rises slowly. Pattern 2 is a sudden rise in the water level in a short period of time due to sudden downpours or typhoons. Pattern 3 is a rainy region where river runoff is slow in mountainous areas and fast in urban areas. Pattern 4 is a sudden rise in the water level from an arbitrary water level at nearby pumps or dams. Correction terms are inserted according to each case.

[0025] In the case of prolonged rainfall (Pattern 1), the water level rises slowly, so it is easiest to approximate it with a linear function. Therefore, no correction is required. In the case of sudden downpours (Pattern 2), the location of occurrence changes depending on the atmospheric pressure pattern and topography, making it difficult to pinpoint the occurrence location and more likely to occur during unstable weather. Since sudden downpours and typhoons are predicted in weather forecasts, we make it possible to input the rainfall index c1 as needed. In the case of direction (Pattern 3), the arrival time at the pumping station varies depending on whether the rain gauge is located in a mountainous or urban area. By comparing past rainfall data with the pump well water level and graphing the delay in the rise in the pump well water level after a sudden increase in rainfall, we consider the delay correction c2. For other equipment in Pattern 4, it may be known from experience how many hours after this equipment operates that the pumping station will begin operating. Alternatively, if it is known that the dam will release the water level several hours after the dam reaches the operating water level, we can ignore the calculation from the water level trend and force the operation time to be output.

[0026] Here, the guerrilla index c1 is generally expressed as rainfall intensity (mm / h) as an index of the intensity of torrential rain and typhoons. Since the higher the rainfall intensity, the faster the rate of water level rise will also increase, it is included as a coefficient for the rate of water level rise. It is often expressed in days. Therefore, for example, a super typhoon would be 800mm / day, a large typhoon would be 500mm / day, and a normal typhoon would be 300mm / day. Torrential rain is expressed as hourly rainfall, with the option of 100mm / h for large typhoons and 50mm / h for normal typhoons.

[0027] The relationship between rainfall and water level is the relationship between the location of the observation rain gauge and the water level gauge at the pumping station, and it is necessary to investigate in advance the time it takes for rain to reach the pumping station.Water level is determined by various factors such as the catchment area into which rainfall flows, the river width, and the slope of the river, so it is calculated simply.

[0028] The target rain gauge is determined from an upstream point near the pumping station, and past rain gauge trend data is compared with the pumping station's water level gauge. When heavy rain falls, rainfall rises suddenly, so the time and the amount of rainfall rise are recorded on the past rain gauge graph, and then the data showing a sudden rise in water level is recorded from past water level gauge data at the pumping station at the same time. This allows the rainfall arrival delay time to1 and the water level rise degree b2 due to the difference in rainfall to be calculated.

[0029] Fig. 3(a) shows the time course of rainfall, and Fig. 3(b) shows the time course of water level. The arrival delay time to1 = t4 - t2, the water level rise rate b2 = s5 - s4, and the guerrilla index c1 = u3 - u2. As Case 1, when c1 < 50 mm / h in the case of normal rain, the guerrilla correction is g1 = 1. As Case 2, when 50 < c1 < 100 mm / h in the case of normal guerrilla, the guerrilla correction (example) is g1 = 1.2. As Case 3, when 100 < c1 in the case of large guerrilla, the guerrilla correction (example) is g1 = 1.4. As Case 4, when c1 < 300 mm / day in the case of normal heavy rain, the guerrilla correction (example) is g1 = 1.2. As Case 5, when 300 < c1 < 500 mm / day in the case of normal typhoon, the guerrilla correction (example) is g1 = 1.4. As Case 6, when 500 < c1 < 800 mm / day in the case of large typhoon, the guerrilla correction (example) is g1 = 1.6. As Case 7, when 800 < c1 in the case of super-large typhoon, the guerrilla correction (example) is g1 = 1.8.

[0030] The guerrilla correction g1 can be calculated from past rainfall data. When the index from super-large to normal rain cannot be obtained from past data, it is obtained by comparison from past values. The arrival delay t01 is determined based on the past rainfall direction (typhoon landing from the east, typhoon landing from the west, etc.) and the above arrival delay is quantified. When it is difficult to quantify from past data, the time delay is input as an empirical value. Summarizing the above as a mathematical formula, a correction term is added to the aforementioned formula, and the future water level becomes s1 = b1 × g1 × (ut1 - t0) + A0. The water level arrival time is calculated by adding the delay corrections c2 and c3. The purpose of this calculation is to construct a simple calculation formula that can be adjusted and used by itself based on past rainfall data and empirical rules.

[0031] When no past information is available, such as when a new water level gauge is installed, the correction term is initially set to a value that has no effect, and adjustments are made by correcting it each time rainfall occurs. To allow for proportional expression, such as c1 = 0.1 when the rainfall intensity is 20 mm / h and c1 = 0.2 when the rainfall intensity is 30 mm / h, c1 = rainfall intensity x coefficient x 0.1, with the coefficient being freely selectable. Taking the above corrections into account, the water level calculation after a certain time period is s1 = (b + c1) x (ut1 - t0) + a0, and the time when water level s1 is reached is ut2 = ut1 + c2, or the forced output ut3 for pattern 4.

[0032] The coefficients are calculated using the above formula and past water level data. The coefficients should be calculated from a large amount of operational data, but since statistical records of past data are often not maintained, the system allows users to freely modify the coefficients.

[0033] Next, we will explain the checkback function. To confirm the accuracy of the created water level prediction formula, past water level data is used. It is difficult to make accurate predictions for all rainfall. This is because behavior changes depending on the situation, such as drainage facilities starting operation depending on the amount of rainfall, flooding occurring, and the time it takes for rainfall from the mountains or land to be reflected in river water levels.

[0034] As check function 1, we will explain the calculation of prediction accuracy using past operating data. When the coefficients are changed, it is assumed that there is a function that allows checkback using past operating data. When using the checkback function, recalculation is performed based on the operating data accumulated in the past, and the result is expressed as a deviation from the true value.

[0035] As check function 2, we explain the calculation of prediction accuracy at singular points in past rainfall. Regarding the selection of accumulated data, if all data is collected, there is a possibility that the prediction will be inaccurate. Considering the characteristic that water level predictions are not used for normal rainfall, but are used for predictions in heavy rainfall that requires pump operation, past data on singular points, such as "trends during a specific typhoon," is called up and displayed from past rainfall data, water level data, and pump operation records. Water level predictions are "baseless assumptions" and have low reliability. Therefore, trend values ​​from past heavy rainfall are one of the criteria for judgment. Furthermore, since the rainfall pattern and path vary depending on the typhoon, we make it possible to select each past typhoon, and at that time, a typhoon path map is also displayed for comparison.

[0036] As shown in Figure 4, in the prediction mode at a singular point, the output is a direct comparison with past rainfall data, so the water level prediction method involves first obtaining water level data a0 at the current time t0. Using past water level trend B, the same water levels K0 and K2 are found. Whether the past water level is on an upward or downward trend is calculated using the average water levels of the past several points (t-3, t-2, t-1). If the water level is on an upward trend, the water levels K0 and K1 from the rainfall record are used to determine the future water level a1 after time (k1-k0). Similarly, if the water level is on a downward trend, the downward water level trend from the rainfall record is used to determine the future water level trend.

[0037] When extracting data showing the same water level from past typhoon data, multiple water levels will be extracted in chronological order. Therefore, some ingenuity is required to make it easy to compare each extracted point with the current water level data in sequence. To achieve this, the comparison method involves horizontally shifting the past water level trend graph and the current water level trend graph, sliding them so that the water levels are aligned at the same position, and moving each extracted point so that it is easy to visually see whether the water level trend changes are similar.

[0038] Figures 5, 6, and 7 show water levels and past data in this embodiment. In Figure 5, three past data points equivalent to water level a0 are extracted. Figure 6 shows past data extraction 1 slid to t0 (extraction 1 pattern). Figure 7 shows past data extraction 3 slid to t0 (extraction 1 pattern). This mode clearly shows past trends, making it easy to make water level predictions, and even if the prediction is wrong, it is easy to estimate the water level rise based on whether it is more rapid than the past water level trend, or vice versa.

[0039] As described above, by calculating the future water level from the rate of rise in the water level in the pump well of the drainage equipment, it is possible to easily and appropriately predict the water level.

[0040] This section explains how to calculate future water levels from tide data. Because tides have a greater impact than rainfall at pumping stations in bay areas, it may be easier to make predictions by estimating the difference from monthly tide tables. Water level predictions are calculated directly from a fixed water level trend. An additional element is the method for obtaining tide tables. Local tide tables are published by the Japan Meteorological Agency, and tide tables can be referenced up to 35 days in the future. Data is based on downloading the current day's tide table, with past tide tables available for comparison. It is known that when a pumping station is located near the sea, the relationship between tides and the internal water level at the pumping station generally follows a synchronized trend. This water level difference occurs when heavy rain falls, causing the pumping station's water level to rise by the height of the hydraulic gradient in the river.

[0041] 8 is a diagram illustrating how water levels are calculated from tide level trends in this embodiment. The pumping station water level c1 (= tide level) is expressed as d1 + α, and the water level at a future time can also be expressed as the value obtained by adding α to the time change in this tide level.

[0042] As described above, by calculating from tide data, water levels can be easily and accurately predicted even in areas such as coastal areas where the influence of tides is greater than the influence of rainfall.

[0043] Next, we will explain the operational status of the drainage equipment. Because the pump well water level is significantly affected by pump operation and gate opening and closing, it is necessary to take into account the status of the pump and gate in addition to water level trend information. In order to simplify the system construction, we will evaluate the water level trend by adding pump operation stoppages and gate opening and closing on the time axis.

[0044] Figure 9 shows the relationship between pump and gate operation data and water levels. When predicting water levels, in addition to checking for matches with past trends, comparison criteria are also added to whether the pump was operating or the gate was open at the same time. In other words, the closest match is a water level trend that meets the following three conditions: (a) similar water level trends, (b) matching pump operation / stop status, and (c) matching gate open / close status. Regarding the results of this match, rather than making a judgment such as "only display if all three conditions are met," the results for each of the three conditions are displayed on the screen, and the user is free to make a judgment based on the results. Pump operation and gate opening / closing have a significant impact on changes in river water levels, so they need to be displayed in as much detail as possible. The following items are displayed to display pump operation and gate operation.

[0045] FIG. 10 is a diagram showing the relationship between pump and gate operation (detailed version) data and water level in this embodiment, and is a graph when all of the following data items can be measured. The pump operation item indicates the time when the pump driver started and the discharge valve was slightly open or more, the pump stop item indicates the time when the discharge valve was fully closed, and the pump displacement amount item displays if it is being measured. The number of operating pumps item displays whether each pump is operating / stopped. For example, it displays the operating status of pump No. 1 and pump No. 2. The pump rotation speed item displays if it is being measured. The discharge valve opening item displays if the flow rate is being controlled by the discharge valve opening. The gate opening item indicates the time when the gate was slightly open or more. The gate closing item indicates the time when the gate was fully closed, and the number of gates item displays whether each gate is open or closed. For example, it displays information on gate No. 1 open and gate No. 2 open.

[0046] The mainstream method for predicting river water levels is to use AI-based machine learning to predict the future water levels of a given river based on past rainfall and water level data at each location. However, the drawback of this method is that it is only accurate for long periods of gentle rain, but when there is sudden heavy rain or rain that is different from past rainfall, for example, when the machine learning is based on the assumption that rain usually falls from the west, but the rain changes to an eastward movement due to the path of a typhoon, the machine learning is not sufficient and the predictions are inaccurate.

[0047] Furthermore, pump operation is determined only by rules as to the water level at which operation begins at a certain point, and the number of pumps to operate at that time depends on the operator's experience. There are also various conditions, such as the fact that after a pump is stopped, it cannot be restarted for a certain period of time, making it difficult to stop the pump, and gates cannot be opened or closed without checking that there are no people nearby. This behavior is unrelated to past water level fluctuations. Therefore, weighting is applied to evaluate the behavior by pattern matching with past data.

[0048] For the water level match rate, for example, the match rate for the water level over the past six hours is calculated. For gate open, the time that the gate has been open in the past 30 minutes is used as the match rate (whether to look at open or closed is determined based on the pump and its behavior. If the predicted river water level drops due to pump operation = opening the gate causes the water level to drop, then comparison is made with the gate open. If opening the gate causes the water level to rise, then comparison is made with the gate closed). For the number of gates, the same calculation is made as for the number of gates open).

[0049] For pump displacement match, the pump displacement over the most recent 30 minutes is used as the match rate, as the most recent pump shutdown behavior has the greatest impact on future water level predictions. If the displacement is not measured for the number of operating pumps, the pump displacement match is switched to the number of operating pumps and calculated accordingly. For pump rotation speed or discharge valve opening, these are also used if the displacement is not measured. Since both the rotation speed and the discharge valve are not controlled, either one can be applied. The pump rotation speed or discharge valve opening over the most recent 30 minutes is used as the match rate. Comparisons are made using the average value of the rotation speed or opening.

[0050] The matching rate is weighted in order of the greatest influence on water level changes. Because floodgates have a greater influence than pump operation, weights are assigned to floodgates and then pumps, and as a result, coefficients are assigned so that the weighting decreases in the aforementioned order.

[0051] A specific example of the calculation is shown below. To find a water level match rate of a%, first extract a graph of the past six hours from the current water level, then compare the match rate of the past six hours of water level trends, and extract the graph that is closest. The match rate can be calculated, for example, by comparing water levels every minute and then accumulating the deviations. Taking calculation time into consideration, for example, select 10 similar data to extract.

[0052] Gate open b% is compared with the gate open data for each minute of the past 30 minutes to see if it matches. The gate open status at that time is compared with the data selected for the water level match rate a% above. For example, if all gates are open and match, the match rate is 100%. Number of gates c% is calculated when there are multiple gates, and the above gate open b% method is calculated for each gate.

[0053] For pump discharge volume d%, similar to gate open b%, the pump operation status at that time is compared with the data selected with the water level match rate a%. As with water level, the discharge volume at each time is compared and the deviation is calculated as the match rate. For the number of operating pumps e%, the calculation is the same as for the number of gates. For pump rotation speed and discharge valve opening f%, the calculation is the same as for pump discharge volume d%. These are weighted.

[0054] Next, we'll explain weighting. For example, A% + b% × 0.8 + c% × 0.8 + d% × 0.7 (or e% × 0.7) + f% × 0.5. The water level trend with the largest total value is the forecast result. In other words, even if the water level matches, if it doesn't match the gate opening / closing or pump operation, or if there are operational delays, the match rate will be low. In fact, there are cases where pumps or gates are not operating due to malfunction or maintenance, so simply comparing past water levels with AI does not capture actual behavior. The above example shows a prediction that combines the water level at a given location with the behavior of the gates and pumps that affect it. However, prediction accuracy can be further improved by using the same calculation to calculate upstream rainfall and water levels at surrounding locations.

[0055] We will explain how to express the data match level. For example, the data match level can be expressed as "80% approximation to the trend graph for the past day." There are ways to express the match rate, but in actual operation, 80% indicates the closeness of the shape, such as "the first half of the trend matches" or "the second half of the trend matches," and is a number that users can use to indicate the expectation that future predictions will show the same trend due to the approximation to the past. Therefore, the match level is displayed as a reference value. The water level forecast result is not displayed on the screen because the total value above would be a number greater than 100%, and only the water level forecast result, a%, is displayed as the forecast value.

[0056] Next, we will explain how to calculate the time from the water level prediction results until the operating water level is reached. Based on the water level prediction results derived above, the future operating time is calculated. There are two methods for calculating the operating time. The first calculation method (operating time prediction 1) is to use the past data of the matching water level trend derived in the calculations described above as is.

[0057] Figure 11 is a diagram showing past water level trend data. The calculation method is to first extract the operating water level at the current time. Then, in the past water level trend, extract the point that is the same as the current water level. Next, read from the graph the time x minutes after the operating water level will be reached (the predicted operating time) in the past water level trend. This is a relatively accurate method when the rainfall is similar to past events, such as long periods of rain, because the water level behavior will also be relatively similar. The operating status of the pump and gate were also compared when the approximate water level trend was extracted as described above, and a trend showing similar behavior has already been selected, so here it is sufficient to use only the water level data without comparing the behavior of the pump and gate.

[0058] As described above, by applying data on past water level trends that match the current trend to the calculation of the time until the operating water level is reached in the dispatch time calculation unit, the time until the operating water level is reached can be easily and appropriately determined.

[0059] FIG. 12 is a diagram showing the prediction of the operation time using an approximation curve according to this embodiment. The second calculation method (operation time prediction 2) predicts the time when the operating water level will be reached from the current water level rise rate. Simply put, it is an extrapolation of a graph with an approximation curve drawn. Predicting the operating water level is based on the worst-case scenario, i.e., the pattern that will reach the operating water level the fastest. This avoids the most important operational delay, and is a safe approach. Therefore, the approximation curve is calculated under different conditions, from 1st to 5th order curves, and the time when the water level is closest to the operating water level is extracted. In FIG. 12, the cubic approximation curve indicates the shortest time, and the predicted operation time calculated using this approximation curve is output as the calculation result.

[0060] As described above, when calculating the time until the operating water level is reached in the dispatch time calculation unit, the time until the operating water level is reached can be easily and appropriately determined by predicting the time from the current water level rise rate.

[0061] Next, we will explain how to exclude abnormal water level values ​​during floods. Factors that affect water level rises during floods include the amount of rain flowing into the river, dam discharges, the opening and closing of nearby gates, and pump operation on the upstream side, and the tide level, the opening and closing of nearby gates, and pump operation on the downstream side. Tsunamis caused by earthquakes also have a downstream impact, but as their unexpected behavior is difficult to predict, the impact of earthquakes is excluded from the calculations. Localized torrential rain is also difficult to predict, so it is also excluded from the calculations.

[0062] Therefore, the order of greatest influence on short-term water level changes is dam discharge, gate opening / closing, pump operation, and rainfall, and the maximum short-term water level change can be said to be a predictable value. Conversely, calculation results that exceed the maximum water level change can be eliminated as abnormal values. This maximum short-term water level change value MAXs is calculated based on past water level data, and when extracting an approximate curve in the calculation method described above (operation time prediction 2), approximate curves that exceed MAXs are eliminated as abnormal values.

[0063] For example, in Figure 12, the cubic curve shows an extreme rise in water level after the current water level, and if the slope value of this slope exceeds the past MAXs value, the cubic curve is treated as an abnormal value, and the next approximation curve, the quadratic curve, is selected.

[0064] Next, we will explain how to eliminate abnormal water level values ​​when the water level is low. Factors that affect the drop in water level when the water level is low on the upstream side include the amount of rain flowing into the river, the suspension of dam discharge, the opening and closing of nearby gates, and the stopping of pumps, while on the downstream side they include the tide level, the opening and closing of nearby gates, and the stopping of pumps. Tsunamis caused by earthquakes are also considered a downstream influence, but as this behavior is unexpected and difficult to predict, the impact of earthquakes is excluded from the calculation. Localized torrential rain is also difficult to predict, so it is excluded from the calculation. Water level drops can also be calculated using the same approach as above, and the minimum value of water level change in the past, MINs, is calculated in the same way and eliminated as an abnormal value.

[0065] This section explains how to combine and display prediction models and calculate dispatch times. Future water levels are calculated and displayed using a water level prediction model. For example, water level F1 and time G1 are calculated from the rate of water level rise, water level F2 and time G2 at the singular point are predicted, and water level F3 and time G3 are predicted based on the tide level. Information specific to each station is organized in advance for these predicted water levels and arrival times.

[0066] Information specific to each station includes, for example, TA1, the travel time from the current location to the pump station; TA2, the travel time from the pump station parking lot to the control room; TA3, the preparation time for operation (checking the surrounding area for safety, contacting various locations in advance); TA4, the advance monitoring time (checking the monitoring screen, etc.); TA5, the advance patrol time (advance patrol of each piece of equipment); and TA6, the time from start-up operation to the start of drainage. TA1 is calculated using the car navigation system or mobile phone map software. TA2 through TA6 can be estimated as approximate times and are entered in advance on the settings screen. Also, because conditions vary depending on the station, for example, TA5 may not be necessary, it is set for each station. TA6 varies depending on the size and model of the pump; for example, it is approximately 10 minutes for a horizontal shaft pump and approximately 1 minute for a vertical shaft motor-driven pump. Since conditions vary depending on the station, with some stations having a fixed pump that will start first and others not, TA6 is set taking this into consideration.

[0067] The following explains how to use the system. First, a notification of rising water levels is received at a location other than the pump (for example, your home). The prediction system is started and a water level prediction calculation is performed. The time TA8 at which the operating water level is reached is calculated. As there are multiple water level predictions, the calculation results for each are displayed. The arrival time TA1 from the current location to the pumping station is calculated using a commercially available navigation system. The pre-set confirmation times TA2 to TA6 are read from the system and the time until drainage starts, TA7 = TA1 + TA2 + TA3 + TA4 + TA5 + TA6, is calculated. The margin time TA9 = time to reach operating water level TA8 - time until drainage starts, TA7, is calculated and the time for operators to dispatch and each calculation result are displayed on the screen. As there are multiple water level predictions, multiple results are also displayed. The user can make a judgment by looking at the multiple calculation results. Water levels change constantly as time passes. The system is designed so that by pressing the recalculation button, calculations can be recalculated using the latest data. [Example]

[0068] Figure 13 shows the system flow according to this embodiment. The system flow taking these factors into consideration is as follows: The drainage equipment management system mainly comprises an operation record database 1, a water level measuring device 2, a water level prediction processing unit 3, an operator information database 4, a dispatch time calculation unit 5, and a dispatch time guidance unit 6.

[0069] The operation record database 1 includes, as water level prediction databases, a past water level trend database, a past pump operation database, a past gate operation database, a water level prediction parameter database, etc. This information is used to predict future water levels and times.

[0070] Current water level data is acquired by the water level measuring device 2. From this water level data and information obtained from the operation record database 1, the water level prediction processing unit 3 calculates the water level prediction result. In addition, the operating water level reaching time is calculated. The dispatch time is calculated in the dispatch time calculation unit 5 from the water level prediction result and the time when the operating water level is reached obtained in the water level prediction processing unit 3, and the time when the operator will arrive from the operator information database 4, and the dispatch time is displayed in the dispatch time guidance unit 6.

[0071] This section explains how to manage multiple stations. Due to the recent labor shortage, there are an increasing number of cases where multiple stations are managed, and the data used for water level prediction also requires the setting values ​​for each location to be stored. For this reason, the parameters for the above calculations are managed in a database (DB), and the system is designed to allow for retrieval and modification as needed. Furthermore, these parameters are values ​​that change, for example, as the operator's proficiency level increases, or as new water level data is added. Furthermore, it must be assumed that this data will be used for purposes other than water level prediction. Therefore, the parameter configuration is a branched DB structure as shown in the figure below. Furthermore, since there are multiple people operating the system, separate DBs are created for parameters that vary in time depending on the person. Parameters also differ depending on whether the operation is performed on-site or remotely.

[0072] The arrival time is calculated using map information software based on the current location and distance. In addition, a database of arrival times is registered to enable validity comparisons (past locations and travel times). However, because arrival times are typically during floods due to heavy rain, there is a possibility of traffic congestion or road closures, which could result in delays from the time calculated by the map information software. For this reason, arrival times should be compared with past results, and adjustments should be made if necessary. In addition, after arrival, the arrival time is automatically calculated backward from the time recorded by the system, such as the time of pump operation, and registered as a record so that it can be accumulated as a record.

[0073] The operator information database 4 includes, for example, a drainage preparation time DB, a pumping station travel time DB (travel time), an operation preparation time DB (each preparation time), a pre-monitoring time DB (each preparation time), a pre-patrol time DB (each preparation time), a drainage start time DB (each operation time), a remote preparation time DB, etc., for operator A. In addition, similar information for another operator B is stored in a DB.

[0074] The dispatch time is displayed, and when the operator arrives at the site and performs the operation, the arrival time is calculated backwards from the time of arrival at the site and registered in the arrival time database.

[0075] It is also possible that, depending on the situation, the operation water level may not be reached in time. In such cases, the calculation is recalculated in the shortened mode. Each preparation time is generally set to the time that is normally considered necessary or to allow for some leeway. If these preparation times are omitted, and if an operation that is not normal is performed, such as operation in linked mode, a confirmation timer is set to take into account the time required to confirm that each device has operated properly. However, when operating in standalone mode, each device is visually confirmed, so the confirmation timer time can be omitted. Therefore, the operation time in standalone mode is registered in advance as the shortened mode. This has the effect of allowing the operator to prepare for emergency operations while rushing to the scene.

[0076] If this still does not make it in time, the system will switch to remote operation mode and operate.If this is still not enough, measures such as contacting neighboring drainage facilities and requesting backup operation will be taken.In this case, by using the alarm mode function to send emails to the operators of each neighboring drainage facility, delays in operation due to communication delays can be avoided.

[0077] After confirming that the operation will be completed in time, the operator will rush to the site and begin operation. Operation procedures are performed within predetermined times, and for example, a countdown is displayed by calculating backwards the pre-monitoring time, providing a backup to prevent delays in operation. After completing each task, the operator can click on the screen to display a record of the actual time taken and a countdown for the next operation. The actual time taken can be referenced later and used for reviewing time settings and for operational training to improve proficiency. In addition, for drainage operation, an operating pattern (number of operating units, rotation speed, operating time, etc.) derived from the results of advance water level predictions is displayed, and the operator simply operates according to that pattern.

[0078] As mentioned above, because the system knows the time of each operation, the operator only needs to manage safety, and the system can automatically perform the operation. As described above, the calculation of the dispatch time is very simple, as it is done by adding each hour from the current time, but in situations where the dispatch time is calculated, the calculation is done while the operator is getting ready, such as getting changed, so by calculating it with the system, and by considering and memorizing the necessary parameters in advance, the effort of thinking about it and remembering the previous setting values ​​can be eliminated.

[0079] Furthermore, if an operator has to deal with, say, ten or so pieces of equipment, the operation time and preparation time will differ, and the arrival time from the current location to each piece of equipment will also differ, so it will be necessary to consider in what order to most efficiently get to the sites from the current location.For this reason, the above-mentioned dispatch time calculation is calculated simultaneously for the multiple pieces of equipment that the operator must head to, and the time to each base is displayed, making it possible to decide the order in which to operate the equipment based on the margin of time.

[0080] Let me explain the patrol plan. Usually, multiple operators manage several airports. In other words, when heavy rain falls and they have to wait at the site, each operator will be stationed in an appropriate location, and they need to consider their placement while coordinating by telephone, etc. To reduce this communication work, after creating a placement plan, operator A presses the distribution button, and the arrival time is distributed to other operators B and C. Operators B and C can then view the information and make adjustments, such as excluding the target airport from the placement plan, and consider their own placement plan.

[0081] We will explain how to improve accuracy. Because there are so many parameters, it is easy to imagine that the water level predictions will be inaccurate. On the other hand, by accumulating past calculation results, the AI ​​will learn the parameter settings and will also be equipped with a function to display parameter guidelines. On the other hand, if the AI ​​is fully applied, the predictions will be corrected in a way that is influenced by normal rainfall values, resulting in predictions that are inaccurate during crucial heavy rainfall events. For this reason, automatic parameter correction will not be incorporated into the function. [Explanation of symbols]

[0082] 1...Operational performance database, 2...Water level measuring equipment, 3...Water level prediction processing unit, 4...Operator information database, 5...Dispatch time calculation unit, 6...Dispatch Time Guidance Section

Claims

1. In a drainage facility management system equipped with water level measuring equipment and drainage facilities installed in a river, A database of past operation records of the water level measuring device and the drainage equipment; a water level prediction processing unit that calculates a future water level based on data from the operational performance database and current data from the water level measuring device; an operator information database for operators to rush to the drainage facility; A drainage equipment management system characterized by comprising a dispatch time calculation unit that calculates the dispatch time of the operator based on data in the operator information database and data from the water level prediction processing unit.

2. The drainage facility management system according to claim 1, A drainage facility management system comprising a dispatch time guidance unit that transmits the dispatch time calculated by the dispatch time calculation unit to the operator.

3. The drainage facility management system according to claim 1, The drainage facility management system is characterized in that the water level prediction processing unit calculates the future water level from the rate of rise in the water level of the pump well of the drainage facility, or calculates it from the tide level.

4. The drainage facility management system according to claim 1, A drainage equipment management system characterized by applying data on past water level trends that match the current trend when calculating the time until the operating water level is reached in the dispatch time calculation unit, or predicting the time when the operating water level will be reached from the current rate of water level rise.

5. A drainage facility management method using a water level measuring device and drainage facilities installed in a river, comprising: a water level prediction step of calculating a future water level based on data in a database of past operation records of the water level measuring device and the drainage equipment and current data from the water level measuring device; A drainage equipment management method characterized by including a dispatch time calculation step of calculating the dispatch time of the operator based on data in an operator information database for the operator to rush to the drainage equipment and the data calculated in the water level prediction step.

6. The drainage facility management method according to claim 5, and transmitting the calculated dispatch time to the operator to provide guidance.

7. The drainage facility management method according to claim 5, A drainage equipment management method characterized in that the water level prediction step includes a step of calculating the future water level from the rate of rise in the water level of the pump well of the drainage equipment, or a step of calculating it from the tide level.

8. The drainage facility management method according to claim 5, A drainage equipment management method characterized by including a step of applying data on past water level trends that match the current trend when calculating the time until the operating water level is reached in the dispatch time calculation step, or a step of predicting the time when the operating water level will be reached from the current water level rise rate.

9. In a program for a drainage facility management system equipped with water level measuring devices and drainage facilities installed in a river, a water level prediction function in the computer that calculates a future water level based on data in a database of past operation records of the water level measuring device and the drainage equipment and current data from the water level measuring device; A program for realizing a dispatch time calculation function that calculates the dispatch time of the operator based on data in the operator information database for the operator to rush to the drainage equipment and data calculated by the water level prediction function.

Images