JP3870407B2 - Inflow prediction device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an inflow amount prediction apparatus capable of accurately predicting an inflow amount of, for example, sewage flowing into a pump well of a sewage treatment plant or rainwater flowing into a rainwater pumping station.
[0002]
[Prior art]
Conventionally, in order to prevent inundation, it is necessary to expand rainwater drainage facilities and operate facilities that make use of existing facilities.In particular, in the operation of existing facilities, the amount of water flowing into the pump well is predicted from the amount of rainfall. Various methods for performing appropriate pump control have been proposed. For example, an inflow prediction apparatus using a modified RRL method for calculating a rainwater outflow hydrograph necessary for designing a rainwater drainage channel based on rainfall is as shown in FIG.
FIG. 3 is a block diagram showing a conventional inflow rate prediction apparatus.
In the figure, 1 is a rain gauge for measuring rainfall, 3 is a rainwater / sewage pump that pumps and discharges rainwater / sewage stored in a pump well 11 according to the storage volume, and 12 is a rainfall measured by the rainmeter 1. A measured value storage device for storing the amount, 9 is an inflow amount prediction device by the modified RRL method, and 8 is a pump operation command device, which outputs an optimal pump operation command from the inflow amount calculated by the inflow amount prediction device 9 .
In particular, the inflow amount predicting device 9 as described above calculates an infiltration region area, a downflow time, a storage amount / outflow amount curve, and calculates an effective rainfall amount, a rainwater inflow water, and an inflow amount. The infiltration area indicates the ratio of the impermeable area to the drainage area and the infiltration area.
Here, the impervious area and infiltrated area are plotted on the aerial photograph of the target drainage area using a combination of two random numbers as the origin, and the number of impervious area plots and infiltrated area plots against the total number of plots. The impervious area and the infiltrated area are calculated at a ratio of numbers.
And the effective rainfall calculates the rainfall which flows into the pipe from the seepage area and the impermeability area calculated by an effective area calculation device (not shown). In addition, the flow-down time is obtained from the time between each manhole using the full flow velocity from the position, shape, pipe diameter, gradient, etc. of the drainage pipe. Furthermore, the storage volume / runoff curve calculates the relationship between the storage volume and runoff volume taking into account the in-pipe storage of rainwater flowing into the pipe. Furthermore, rainwater inflow water is calculated by the unit diagram method using the effective rainfall and flow time, and the inflow is obtained by solving a continuous equation from the function obtained from the storage and outflow curves. It is like that.
[0003]
[Problems to be solved by the invention]
However, in the inflow forecast using the modified RRL method, in order to obtain the effective rainfall, runoff time, storage volume / outflow curve, etc., the relationship between the infiltration zone, non-penetration zone, rundown time, storage volume / inflow volume, etc. There is a problem that it is necessary to calculate a large amount of data, and that the large amount of data must be corrected at any time.
Of these, the infiltration and non-permeability areas must be corrected as needed due to changes in housing conditions and road conditions, and many parameters such as changes in housing conditions and changes in road conditions are changed. Therefore, it takes a lot of time and manpower, and it is difficult to always update the parameters, so that there is a problem that the prediction accuracy is deteriorated.
The present invention has been made to solve the above-described problems, and does not require calculation of enormous data such as the permeation zone, the non-penetration zone, the flow time, the relationship between the storage amount and the inflow amount, and houses and roads. It is an object of the present invention to provide an inflow amount prediction device with high prediction accuracy that does not require a great amount of time and labor without changing many parameters such as changes in conditions.
[0004]
[Means for Solving the Problems]
In order to solve the above problem, the present invention described in claim 1 includes a flow meter for measuring an inflow amount flowing into a pump well of a sewage treatment plant or a rainwater pumping station, a rain meter for measuring a rainfall amount in a target basin, ,
A measured value storage device for storing the measured value of the rainfall and the measured value of the inflow;
From the rainfall and inflow accumulated in the measurement value accumulation device, a latest data creation device that creates a data string of the latest predetermined time;
An autoregressive model creating apparatus for determining three parameters of an autoregressive coefficient of an autoregressive model, a variance of a white noise vector, and an optimal order based on a data string created by the latest data creating apparatus;
A predicted value calculation data creation device that creates a most recent data string for calculating a predicted value of the inflow amount from the data of the measurement value storage device;
A prediction value calculation device that inputs data generated by the prediction value calculation data generation device to the autoregression model whose parameters are determined by the autoregression model generation device and calculates a prediction value of the inflow amount; In the inflow prediction device,
A weather extraction device for checking whether the rainfall is observed based on the data string created by the latest data creation device;
Wherein said rainfall checked in weather extraction device and the inflow of it at the same time, and the measured observed rainfall over certain periods lasted rain period ends after several tens of minutes the flow Iriryou autoregressive model With a data processing device that processes and adds as a data string to be input to the creation device ,
The autoregressive model creation apparatus determines the parameter based on the data string processed by the data processing apparatus.
Further, according to the second aspect of the present invention, the predicted rainfall data predicted several hours ahead based on weather information is used to calculate the amount of rainfall of the predicted value calculation apparatus that has predicted the inflow amount one point ahead from the present . By replacing each term with a point in time and repeatedly predicting the amount of inflow at the next point in time, the predicted value of the amount of inflow in the several hours ahead that predicted the rainfall data was calculated It is.
By the above means, the inflow prediction device of the present invention can perform prediction suitable for the pump well or storm water pumping station of the sewage treatment plant, and the rainfall amount, the inflow amount, and the rainfall amount where the rainfall amount is observed are constant. Since the parameters are determined using the inflow measured for several tens of minutes after the end of the rainy period, the inflow from a model that clearly expresses the relationship between rainfall and inflow. Quantity predictions can be made.
[0005]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of an inflow amount prediction apparatus showing an embodiment of the present invention. In addition, the same code | symbol is attached | subjected about the same component as the past, the description is abbreviate | omitted, and only a different point is demonstrated. The configuration in which the present invention differs from the prior art is as follows. In the figure, 2 is a flow meter for measuring the flow rate flowing into the pump well 11, 4 is a measured value storage device for storing the measured value of the rainfall amount and the measured value of the inflow amount, and 5 is a rainfall amount predicting device, Predicts and accumulates several tens of minutes of rainfall up to several hours ahead based on data such as weather radar. Reference numeral 6 denotes a data creation device that creates data for applying the data input from the measured value storage device 4 to the statistical model. 61 is the latest data creation device, which creates the most recent data of the rainfall amount and the inflow amount accumulated in the measurement value accumulation device 4. Reference numeral 62 denotes a weather extraction device, which checks whether rainfall has been observed in the rain gauge 1 for the last several tens of minutes . Reference numeral 63 denotes a data processing device that creates a data string to be input to the autoregressive model creation device 71. When the result checked by the weather extraction device 62 is clear, the data string is not processed, and when it is raining, only the rainfall amount and the inflow amount measured most recently are added to the data string. However, even if the rainy season, where the observed rainfall lasts for a certain period , is affected by the rain for a while, data is added for several tens of minutes after the rainy season. In this way, the data is processed so that the relationship between rainfall and inflow can be clearly modeled. Reference numeral 7 denotes a prediction device, which inputs a data string created by the data creation device 6 and calculates a predicted value of the inflow amount. Reference numeral 71 denotes an autoregressive model creation device that inputs data created by the data processing device 63 and creates an autoregressive model. This autoregressive model uses measured values for the past several tens of days.
Now, the state of the process at time n is the k-dimensional all variable vector X (n), all variable vectors m times before time n are X (nm), white noise vector is U (n), autoregressive model If the regression coefficient is represented by A (m) and the optimum order of the autoregressive model is represented by M, the autoregressive expression is as follows.
[0006]
[Expression 1]
[0007]
Here, the creation of the autoregressive model results in the determination of three parameters: the autoregressive coefficient, the variance of the white noise vector, and the optimal order of the autoregressive model.
The autoregressive coefficient A (m) is determined by solving the following simultaneous equations for i = 1, 2, 3,..., K with the element Aij (m). However, when the mutual dispersion of Xi and Xj is Rij () and the element of the autoregressive coefficient is Aij (m), the following is obtained.
[0008]
[Expression 2]
[0009]
Aij (m) can be obtained by solving the above simultaneous linear equations for i = 1, 2,..., K.
If the element of the white noise vector U (n) is εi (n), the residual variance value σi ² is as follows.
[0010]
[Equation 3]
[0011]
Note that the optimal order M of the model is a value that minimizes MFPE (M) in the following equation (4) representing the prediction error.
[0012]
[Expression 4]
[0013]
Here, N is the number of data, and ‖d _M ‖ is the variance-covariance matrix estimate of U (n). MFPE is an acronym for Multiple Final Prediction Error.
In this way, the autoregressive coefficient, the variance of white noise, and the optimal model order are obtained, and an autoregressive model is created. Therefore, a relational expression between the inflow and the rainfall necessary for predicting the inflow can be obtained from the autoregressive model. The autoregressive model is updated once a day for the purpose of using the latest data.
Reference numeral 72 denotes a predicted value calculation data creation device that creates a data string to be input to the model created by the autoregressive model creation device 71.
73 is a predicted value calculation device, which predicts several tens of minutes ahead of the inflow that can be statistically estimated from the autoregressive model created by the autoregressive model creating device 71 and the data string created by the predicted value calculation data creating device 72. Calculate the value. A prediction several tens of minutes ahead using the autoregressive model is expressed as follows.
[0014]
[Equation 5]
[0015]
Q ( n ) p: predicted value of inflow at time n Q ( n ): measured value of inflow at time n
Rain ( n ): Rainfall at time n
A ₁₁ (m): Autoregressive coefficient of the residual part with respect to the predicted value of inflow
A ₁₂ (m): Rainfall autoregressive coefficient relative to the inflow prediction value However, forecasting several tens of minutes ahead using the autoregressive model requires more than one point. Therefore, when forecasting more than one point ahead, the predicted value is used for the inflow and the previous rainfall is used for the rainfall. Further, if there is rainfall data predicted to several hours later, by using it to predict the inflow, the predicted rainfall data, the predicted value calculation unit has finished predict the inflow of one time point away from the current rainfall substituting one point in sections by repeating to predict the inflow of the next time one time, it is possible inflow of several hours ahead also predicted. The predicted inflow values Q (0) p, Q (1) p,... Thus obtained are output to the pump operation command device 8.
[0016]
Next, an inflow amount prediction method according to an embodiment of the present invention will be briefly described.
First, the rainfall amount in the target basin and the inflow amount of sewage flowing into the pump well 11 of the sewage treatment plant are measured online and accumulated in the measured value accumulation device 4. Based on the measurement online data stored in the measurement value storage device 4, a data string for the latest several hours is created and processed by the data creation device 6. Thereafter, an autoregressive model is created by the autoregressive model creation device 71 based on the processed data. Subsequently, by using the autoregressive model thus created, the data of the predicted value calculation data creating device 72 and the data of the rainfall amount predicting device 5, the predicted value of the inflow is calculated by the predicted value calculating device 73, and the number after operation Estimate inflow volume suitable for pump wells of sewage treatment plants during the week.
A simulation was performed using actual data by applying the inflow amount prediction apparatus described above. FIG. 2 is an explanatory view for confirming the effect of the inflow amount prediction apparatus according to the embodiment of the present invention, wherein (a) is a graph of the inflow amount prediction value and the actual measurement value obtained by simulation using actual data, and (b). Is the measured value of rainfall corresponding to the predicted inflow and the measured value graph. In the figure, practice indicates actual measurement values, and broken lines indicate simulation prediction values. In carrying out the simulation, only the data for the last few days was used, but it can be seen that the measured value of the inflow and the predicted value are in good agreement as shown in the figure.
Therefore, the inflow prediction apparatus according to the present embodiment is based on the measurement online data of the rainfall in the target basin, the inflow of sewage flowing into the pump well of the sewage treatment plant, and the data sequence of the most recent hours by the data creation device. In the past few weeks, the inflow volume suitable for the pump well of the sewage treatment plant can be predicted using the autoregressive model based on the data after processing. The prediction accuracy at the time of performing the inflow amount prediction based on can be improved. Moreover, unlike the conventional inflow prediction device using the modified RRL method, it is not necessary to calculate enormous data such as the relationship between infiltration area, impermeability area, flow down time, storage volume / inflow volume, Since it is not necessary to change many parameters such as changes in road conditions, a great deal of time and labor can be eliminated.
In addition, although the present Example described the inflow amount prediction apparatus which estimates the inflow amount of the sewage which flows into the pump well of the sewage treatment plant in consideration of the measured value of the rainfall amount in the target basin, the rainwater pump station It is possible to use this inflow amount predicting apparatus for the prediction of rainwater flowing into the water.
[0017]
【The invention's effect】
As described above, according to the present invention, based on the measurement online data of the rainfall of the target basin, the inflow of sewage flowing into the pump well of the sewage treatment plant, the data sequence of the most recent hours is generated by the data creation device. Created and processed, and constructed an inflow prediction device that can predict the inflow suitable for the pumping station of the sewage treatment plant within a few weeks after operation using the autoregressive model based on the data after processing, There is an effect of improving the prediction accuracy when performing the inflow amount prediction based on the past results. In addition, there is no need to calculate a huge amount of data such as the relationship between the permeation zone, non-penetration zone, flow-down time, storage volume / inflow volume, and many parameters such as changes in housing / road conditions are changed. Since it is not necessary to do this, a great deal of time and labor can be eliminated.
[Brief description of the drawings]
FIG. 1 is a block diagram of an inflow amount prediction apparatus showing an embodiment of the present invention.
2A and 2B are explanatory diagrams for confirming the effect of the inflow amount prediction device according to the present invention, in which FIG. 2A is a graph of an inflow amount predicted value and a measured value obtained by simulation using actual data, and FIG. It is an actual measurement value of rainfall corresponding to the predicted amount and actual value graph.
FIG. 3 is a block diagram of an inflow amount prediction apparatus showing a conventional technique.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Rain gauge 2 Flow meter 3 Rainwater / sewage pump 4 Measurement value storage device 5 Rainfall prediction device 6 Data creation device 61 Recent data creation device 62 Weather extraction device 63 Data processing device 7 Prediction device 71 Autoregressive model creation device 72 Predicted value Calculation data creation device 73 Predicted value calculation device 8 Pump operation command device

Claims (2)

A flow meter for measuring the amount of inflow flowing into the pump well of the sewage treatment plant or rainwater pumping station, a rain meter for measuring the amount of rainfall in the target basin,
A measured value storage device for storing the measured value of the rainfall and the measured value of the inflow;
From the rainfall and inflow accumulated in the measurement value accumulation device, a latest data creation device that creates a data string of the latest predetermined time;
An autoregressive model creating apparatus for determining three parameters of an autoregressive coefficient of an autoregressive model, a variance of a white noise vector, and an optimal order based on a data string created by the latest data creating apparatus;
A predicted value calculation data creation device that creates a most recent data string for calculating a predicted value of the inflow amount from the data of the measurement value storage device;
From the predicted value calculation device that calculates the predicted value of the inflow amount by inputting the data created by the predicted value calculation data creation device to the autoregressive model in which the parameter is determined by the autoregressive model creation device In the inflow amount prediction device
A weather extraction device for checking whether the rainfall is observed based on the data string created by the latest data creation device;
The autoregressive model is created based on the amount of rainfall checked by the weather extraction device and the amount of inflow at the same time, and the amount of inflow measured in the tens of minutes after the end of the rainy period in which the observed rainfall has continued for a certain period of time. A data processing device that processes and adds as a data string to be input to the device,
The inflow prediction apparatus, wherein the autoregressive model creation apparatus determines the parameter based on a data string processed by the data processing apparatus.   The forecasted rainfall data predicted several hours ahead based on weather information is replaced one by one with the rainfall term of the forecast value calculation device that has finished predicting the inflow volume one point ahead from the present, and the next The inflow amount prediction apparatus according to claim 1, wherein a predicted value of the inflow amount in the several hours ahead in which the rainfall data is predicted is calculated by repeatedly predicting the inflow amount at each time point.