JP3379008B2 - Flow forecasting system - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a flow predicting system for predicting a flow condition of a fluid, traffic, etc., and particularly, for monitoring and predicting a flow condition of a fluid in a river network, a pipe network, etc., and a road. The present invention relates to a flow prediction system that monitors and predicts the flow state of vehicles in a transportation network. Regarding flow prediction systems such as river networks and pipe networks, more specifically, drainage pump stations that determine appropriate operating conditions based on the flow conditions and rainfall information in the pipeline networks of cities and rivers and their surrounding watersheds. Regarding driving support system.

[0002]

2. Description of the Related Art Conventional flow predicting systems for predicting the flow condition of fluids and traffic include, for example, predicting the flow condition of traffic based on a system for predicting the state of traffic congestion on a road network and a prediction result. There is a system that appropriately controls signals at intersections, and a system that predicts the state of fluid flow is one that predicts the state of inflow into the pumping station for the purpose of supporting driving of the pumping station.

As a system for predicting the state of traffic congestion on a road network, there is a traffic flow predicting device described in Japanese Patent Laid-Open No. 8-249586, which uses a neural network to detect data from a vehicle detector in each section in advance. The relationship with the traffic jam section was learned, and during operation, the traffic jam section was predicted based on the detection data from the vehicle detector of each section.

As an example of using a flow prediction system to appropriately control a signal at an intersection, there is a traffic signal control method described in JP-A-7-014093, which includes information such as a speed sensor for detecting traffic information and traffic congestion. The neural network was trained so that the traffic signal could be controlled optimally from the relationship with the time.

Regarding the operation support of the drainage pumping station, the impervious area of rainwater increases with the progress of urbanization of the drainage area, and during heavy rain, the rainwater tends to rapidly flow into the drainage pumping station. Driving support system is required. That is, in conventional drainage pump station operation based on operator's judgment, the operator flexibly made judgments by associating temporal changes such as water level and rainfall at the pump station with past discharge amount, It is difficult to improve the driving skill because the burden on the operator is heavy and the driving frequency is low during heavy rain, and an appropriate driving support system is required.

In the rainwater pump operation support system described in JP-A-6-026093, guidance on the amount of inflow to the drainage facility is given from the pump well water level, the predicted rainfall data, the pipe water level data, and the like. Specifically, it predicts the rainfall 30 minutes ahead based on the measured rainfall sent from the rain gauge, predicts the inflow to the rainwater drainage facility, and displays the rainwater pump guidance online. There is.

In the rainwater pump operation controller described in Japanese Examined Patent Publication No. 8-033157, the future rainfall amount is predicted based on the state of the rain cloud and the amount of rainfall on the ground, and the drainage network of the urban area is drained. Is analyzed to control the operation of the drainage pump. In the runoff analysis, for example, a nonlinear hyperbolic partial differential equation is solved under the assumption of uniform flow to obtain the rainwater transfer time.

In the rainwater pump operation control device described in Japanese Patent Laid-Open No. 5-311727, from the rainfall amount and the water level in the pipe,
It predicts the margin of discharge amount to the predicted rainwater inflow to the pump well, and controls pump selection, start / stop, and rotation speed. From the rainfall and manhole water level acquired from appropriate places in the jurisdiction, fuzzy reasoning is used to determine the margin of discharge amount to the predicted rainwater inflow to the pump well.

Also, from page 785 to page 787 of the "28th Annual Sewerage Research Conference Lectures" (June 1991), the flexible judgment of skilled operators is entrusted to the neural network and reflected in driving. The method is described. That is, the time-series data of the inflow amount to the drainage pump station, the inflow water level, the rainfall amount, and the discharge amount from the drainage pump station are used as the input data, and the discharge amount from the drainage pump station at the next time is used as the output data, and a neural net is created. Constitute. Prior to the actual driving, the relationship between the input data and the output data is learned in advance, and during the actual driving, the amount of discharge at the next time is predicted from the input data obtained up to the current time to support the driving. is doing.

[0010]

[Problems to be Solved by the Invention] Since the tendency of traffic congestion on the road traffic network varies depending on the day of the week, it is important to grasp the tendency for each day of the week. Therefore, if the relationship between the traffic jam situation for each day of the week and the actual measurement data from the vehicle detector or the like is learned by a neural network in advance, the usual traffic jam situation can be predicted. Japanese Patent Laid-Open No. 8 above
-249586 and traffic flow prediction device described in JP-A-7-01
The traffic signal control method described in No. 4093, after learning the traffic state in advance using a neural network,
The system was applied to predict or control traffic conditions.
Particularly, in the traffic signal control method described in JP-A-7-014093, one week is learned as a trial period, and the trial period is repeated until the prediction accuracy is improved to some extent.

However, the above two conventional techniques cannot predict a sudden traffic condition that occurs after learning. Sudden traffic conditions mean conditions that are difficult to learn in advance, such as traffic jams during homecoming rush hours such as Obon and Golden Week, and temporary traffic jams due to road construction. In order to cope with such a sudden traffic condition, it is necessary to flexibly change the result learned in advance according to the condition.

Next, a drainage pump operation support system will be described. As shown in FIG. 16, with the spread of paved roads, most of the rainwater flowing into the city does not permeate the ground and flows into the underground pipeline network. Rainwater in the pipeline network further flows into the drainage pump station through the inflow trunk line and is discharged to the river from the drainage pump station.

In the rainwater pump operation support system described in the above-mentioned JP-A-6-026093, the measured rainfall amount sent from the rain gauge for about 15 minutes or 30 minutes from the past to the present is averaged to obtain 30 It predicts the amount of rainfall ahead. Therefore, when a significantly different operation is predicted based on the operator's judgment, such as during heavy rain in which a large error occurs in the rainfall prediction due to the averaging process, the operator manually operates the vehicle, which imposes a heavy burden on the operator.

The inflow of rainwater into the drainage pump station largely depends not only on the amount of rainfall but also on the flow state in the pipeline network.
The pipe network into which rainwater flows in merges with the inflow trunk line at a plurality of confluence points, and the inflow amount of rainwater from each individual pipe network to the inflow trunk line determines the inflow amount to the drainage pump station. Moreover,
This complicates the phenomenon because rainwater flows into the drainage station with a time delay depending on the distance from each confluence to the drainage station.

In the rainwater pump operation control device described in Japanese Patent Publication No. 8-033157, the outflow analysis in which the operation is completed in a short time is taken into consideration in terms of time and space in consideration of the operation time required for pump operation. A uniform flow analysis that ignores changes is used to determine the rainwater transfer time in the sewer network.

However, in a complicated network of pipes in a city, the flow of rainwater changes temporally and spatially, and especially during a sudden heavy rain when the operation load on the drainage pump station is large, the flow of rainwater is not uniform. Since the characteristics are remarkable, the accurate transfer time cannot be calculated by the uniform flow analysis. On the other hand, in order to solve the temporally and spatially varying flow state during heavy rain for a complicated pipeline network, it is necessary to perform a large-scale flow analysis, and it is not possible to handle drainage operation in an emergency.

In the rainwater pump operation control device described in JP-A-5-311727, the amount of rainfall is "small",
"Normal", "many", "very many", manhole water level,
"Low", "Medium", "High", "Very high" functions are assigned, and fuzzy inference is used to calculate the discharge capacity margin function for the predicted rainwater inflow to the pump well "No need",
"Necessary" and "very necessary" are determined, and the selection of pumps, start / stop, and rotation speed are controlled.

However, in the event of a sudden heavy rain in a complicated pipeline network, the correspondence between the measured data such as the rainfall amount and the manhole water level and the margin of the discharge amount with respect to the predicted rainwater inflow into the pump well is preliminarily accurate. It is difficult to describe in.

Although it is possible to predict and control the drainage pump station by learning the correspondence relationship of the margins in advance, it is difficult to learn all the flow states for a complicated pipeline network. In the operation support system described in the "28th Sewer Research Conference Lecture Collection", the time series data of the inflow amount to the drainage pump station, the inflow water level, the rainfall amount, and the discharge amount from the drainage pump station are used as input data. Learn the relationship with the discharge amount from the drainage pump station at the next time in advance, and at the time of actual operation,
From the input data obtained up to the current time, the discharge amount at the next time is predicted to support driving.

However, all input / output relationships must be learned in advance, and if a phenomenon outside the learning range occurs, an error will occur.

Further, when the pipeline network is changed due to the expansion of the city, it is necessary to newly learn. Above "Second
The operation support system described in "8th Sewer Research Conference Lecture Collection" does not understand the effect of each pipeline network, so if there is any change in the pipeline network, the learning results obtained so far will be used. If used, a large error may occur.

Furthermore, it is necessary to pay attention to the flow state of the river where the rainwater is discharged from the drainage pump station. That is, while predicting the water level of the river downstream from the drainage pumping station, it is necessary to control the discharge amount from the pumping station to the river and select the discharge river so that the water level does not exceed the limit value.

The object of the present invention is to improve the predicting means for predicting the downstream flow state from the upstream flow state of fluid, traffic, etc., by comparing the predicted value and the measured value to the optimum state at all times, and sudden flow It is to provide a flow prediction system with a flexible configuration that can appropriately predict the flow state of the downstream side even for a phenomenon.

[0024]

In order to achieve the above object, the present invention provides a flow predicting system for monitoring a flow condition of an upstream such as fluid or traffic and predicting a flow condition of a downstream such as fluid or traffic. The flow prediction system predicts the target flow state of the downstream of the fluid or traffic based on the flow state of the upstream of the fluid or traffic and a predetermined prediction parameter, and the flow state of the upstream or downstream of the fluid or traffic. And a calculation means for calculating a prediction parameter to be used by the prediction means based on the actually measured flow state, and the prediction of the downstream target flow state in the prediction means and the prediction parameter in the calculation means are created by parallel processing, and the prediction means is used. Predicted value and
Compared with the measured flow state in the downstream, the error is larger than the specified value.
If not, use the measured flow state value as a prediction parameter,
If the error is within the specified value, the calculated prediction parameters
We propose a flow prediction system that includes a determination unit that uses a computer.

The flow predicting system is a predicting means for predicting a target flow condition of a downstream such as a fluid or a traffic based on a flow condition of the upstream such as a fluid or a traffic and a predetermined prediction parameter, and an upstream flow condition such as a fluid or a traffic. And a calculation means for calculating a prediction parameter used by the prediction means on the basis of the state, the actually measured flow state of the downstream, and the target flow state of the downstream by the prediction means, and the prediction of the target flow state of the downstream in the prediction means and the prediction by the calculation means. Create parameters and in parallel processing
The predicted value by the prediction means and the measured flow state downstream.
If the error is larger than the specified value, the measured flow state value is
When used as a prediction parameter and the error is within the specified value
If there is a decision means that uses the calculated prediction parameters
May be that.

The flow predicting system also includes a predicting means for predicting a target target flow condition of the downstream of the fluid or the traffic based on the upstream flow condition of the fluid or the traffic and a predetermined predictive parameter, and an upstream of the fluid or the traffic. Whether the calculation means for calculating the prediction parameter used in the prediction means based on the flow state of the above and the measured flow state in the downstream and the prediction value by the prediction means and the measured flow state in the downstream are calculated to calculate the prediction parameter in the calculation means Equipped with determination means to determine whether or not
If the error due to
When used as a prediction parameter and the error is within the specified value
If so, the calculated prediction parameter can be used .

The flow predicting system further comprises a predicting means for predicting a target flow condition of the downstream of the fluid or the traffic based on the flow condition of the upstream of the fluid or the traffic and a predetermined prediction parameter, and an upstream of the fluid or the traffic. The calculation means for calculating the prediction parameter used by the prediction means based on the flow state of the, the measured flow state of the downstream, and the target flow state of the downstream by the prediction means, and the predicted value by the prediction means and the measured flow state of the downstream are compared. And a determination unit that determines whether or not the prediction parameter is calculated by the calculation unit, and the error due to the comparison is a predetermined value.
If it is larger than the measured flow state value,
If the error is within the specified value, the calculated prediction
Parameters can be used .

In any of the above cases, the predicting means includes a plurality of sub predicting means for predicting the flow state of each confluence section or each branching section, and a main predicting means for predicting the downstream target flow state. The configuration may be adopted.

The flow predicting system predicts a target flow state of the downstream of the fluid or traffic based on the upstream flow state of the fluid or traffic and a predetermined prediction parameter, and an upstream flow of the fluid or traffic. And a plurality of sub-prediction means for predicting the flow state of each merging portion or each branching portion, and a calculation means for calculating a prediction parameter used in the prediction means based on the state and the downstream measured flow state. a main prediction means for predicting downstream target flow state
Consists of the predicted value by the prediction means and the measured flow state of the downstream
If the error is larger than the specified value, the measured flow state is compared.
If the market price is used as a prediction parameter and the error is within the specified value,
Determining means that uses the calculated prediction parameter if there is
It is also possible to have a configuration including.

The predicting means is a means for predicting the target flow state of the merging portion or the branch portion on the downstream side based on the predicted value of the flow state of the merging portion or the branch portion on the upstream side.

The predicting means can also include fixed sub-predicting means having a fixed parameter for predicting the target flow state at each of the merging portions or each of the branch portions or downstream.

In the present invention, the predicted downstream target flow state and the downstream measured flow state are sequentially compared, and the prediction parameter for prediction is updated based on the comparison result.
Since the prediction is made within the predetermined error, it is possible to obtain a flow prediction system capable of appropriately predicting the downstream flow state even for a sudden flow state.

For example, the predicting means for predicting the operating condition of the drainage pump station based on the signal from the rainwater collecting area and a predetermined predictive parameter is a sub-device for predicting the flow condition at the confluence of each rainwater conduit and the inflow trunk line. The prediction means and the main prediction means for predicting the driving support information are provided, and the calculation means for calculating the prediction parameter is used for the prediction of the sub-parameter calculation means and the driving support information for calculating the sub-parameter used for the prediction of the merging section flow state. And a main parameter calculating means for calculating a main parameter. While sequentially comparing the predicted results of the flow condition at the merging section with the measured flow conditions, the prediction parameters themselves are modified to predict the operation of the drainage pump station, so the operation of the drainage pump station can be performed even for sudden events such as heavy rain. I can help you properly.

Therefore, in addition to the operation of the drainage pump station, for example, in the traffic flow monitoring device, not only in normal times,
It is possible to obtain a flow prediction system that can appropriately predict sudden changes in traffic conditions that are difficult to learn in advance, that is, traffic congestion such as during consecutive holidays.

[0035]

BEST MODE FOR CARRYING OUT THE INVENTION << Embodiment 1 >> Next, an embodiment in which the flow prediction system according to the present invention is applied to a drainage pump operation support system will be described with reference to FIGS.

FIG. 1 is a block diagram showing the configuration of the drainage pump station driving support system according to the first embodiment of the present invention. This drainage pump station driving support system includes a data input unit 102,
Prediction means 101 for predicting the flow state based on the data from the data input part 102 and predetermined prediction parameters Am, Asi, a driving support part 103 for supporting the operation of the drainage pump station based on the prediction data from the prediction means 101, and data input The calculation unit 201 calculates a prediction parameter based on the input data from the unit 102 and transmits the new prediction parameter to the prediction unit.

FIG. 2 is a flow chart showing an example of a support procedure of the drainage pumping station operation support system of FIG.

In step 10, the prediction parameters used by the prediction means 101 are initialized. These parameters are updated to new parameters one after another by the arithmetic means 201 during the operation of this system. As the initial value, the latest parameter updated at the previous operation may be used. Alternatively, for each rain feature, for example, the rainy season parameters and the typhoon parameters may be separately stored and selected. When operating for the first time, it is necessary to obtain the parameters in advance, but it is also possible to use the pre-learning, the non-linear conduit analysis, etc. as described in the prior art.

In step 20, data is input. This data is actually measured in the procedure 70. The data input unit 102 sequentially reads four types of actual measurement data. The four types of actual measurement data are data indicating the flow state of the i-th tributary in a pipe network or a river having a plurality of tributary rivers, or the tributary data Isi that affects the flow state. And the main stream data Im that affects the flow state of the inflow trunk line or the river or the flow state
And the operation support information Cm indicating the flow condition of the drainage pumping station and its downstream, the flow condition at the merging part of the i-th pipe and the inflow trunk line or the merging part of the i-th tributary river and the river. It is a joining part fluidized state Csi which shows a fluidized state.

FIG. 3 is a diagram showing an input / output relationship of data of the driving support system. For example, as tributary data,
The flow rate Isij at the point j in the i-th pipe, the precipitation Isik at the point k around the i-th pipe, the water level Isjk at the point k of the j-th tributary river, etc. are input. As the mainstream data, the precipitation amount Imi at the point i around the inflow trunk line, the water level Imj at the point j in the river, the meteorological data Imk at the point k, and the like are input. As the operation support information, the water level Cmi at the drainage pumping station, the discharge amount Cmj from the drainage pumping station to the river, the water level Cmk at the point k downstream of the drainage pumping station in the river, and the like are input. The flow condition of the confluence is as follows: flow rate Csi flowing into the inflow trunk from the conduit at the confluence of the i-th conduit and the inflow trunk, and from the tributary river to the river at the confluence of the j-th tributary river and the river. Input the inflow rate Csj.

In step 30, the predicting means 101 predicts the driving support information and the merging section flow state. Prediction means 10
1 is a sub-prediction unit 121 that predicts the flow state of the merge section.
And a main prediction unit 111 that predicts the driving support information, and each prediction unit includes a sub-parameter that predicts the flow state of the merging section and a main parameter that predicts the driving support information. The prediction means 101 is
At time T = T1, mainstream data Im (T1) and driving support information Cm
(T1), tributary data Isi (T1), confluence flow state Csi (T
Read 1). The prediction unit 101 stores past data up to time T = T1 and predicts a future state from these time series data.

First, in the sub-prediction means 121, the tributary data Isi (T1) from the i-th conduit or tributary river and the confluence of the conduit and the inflow trunk line or the tributary river / river confluence are collected. Flow state Csi (T1) is read and time T =
Using the past data Isi (T) and Csi (T) up to T1 and the subparameter Asi (T1) at time T = T1, the predicted value Csi of the confluent flow state at time T = T1 + ΔTsi after ΔTsi is used. Predict '(T1 + ΔTsi).

FIG. 4 is a diagram showing an example of the relationship between the tributary data and the predicted flow rate. As tributary data, the flow rate Isij (T1) at the upstream point j of the i-th pipe and the position k around it.
When the rainfall amount Isik (T1) at time T = T1 is input to the predicting means 101 at time T = T1, the past data from time t = T1 + ΔTsi to the inflow trunk line at time T = T1 + ΔTsi. Flow rate Csi ′ (T1 + ΔT
si) is predicted.

Simultaneously with the prediction of the flow state Csi '(T1 + ΔTsi) at the junction, the main prediction means 111 causes the mainstream data Im.
(T1) and driving support information Cm (T1) are read, and time T =
From the time T = T1 of the past mainstream data Im (T) up to T1, the driving support information Cm (T), the confluence flow state Csi (T), and the predicted value Csi ′ of the confluence flow state just predicted. T1
Using the prediction result Cm '(T) up to + ΔTsi and using the main parameter Am (T1) at time T = T1, the driving support information C at time T = T1 + ΔTm after ΔTm
Predict m '(T1 + ΔTm).

In step 40, the driving of the drainage pump station is supported based on the predicted value of the driving support information predicted by the prediction means 101. For example, as the driving support information, the current time T = T1
From ΔTm after time T = T1 + ΔTm, the water level Cmi ′ (T1 + ΔTm) at the drainage pump station, the future discharge amount Cmj ′ (T1 + ΔTm) from the drainage pumping station to the future water level Cmk ′ (T1 + ΔTm) at the river downstream of the drainage pumping station ) And the like, and the driving support unit 103 outputs information such as the number of operating drainage pumps and the number of rotations.

In step 50, the calculation means 201
Calculate prediction parameters. The calculation means 201 is provided with a sub-parameter calculation means 221 for calculating a sub-parameter used for predicting the merged part flow state, and a main parameter calculation means 211 for calculating a main parameter used for prediction of the driving support information. The sub-parameter calculation means 221 uses the tributary data from the i-th pipe or tributary river to determine the flow state of the confluence at a predetermined reference value Si.
Sub-parameter Asi 'so that it can be calculated with an error within
Calculate (T1 + ΔTsi + ΔT1si). In the main parameter calculating means 211, the main parameter Am ′ (T1 + ΔTm + ΔT) is calculated so that the driving support information can be calculated from the mainstream data and the flow state of the merging portion with an error within a predetermined reference value M.
1m) is calculated. For example, the subparameter Asi '(T1 +
ΔTsi + ΔT1si) and main parameter Am '(T1 +
For the calculation of ΔTm + ΔT1m), a neural network having the tributary data and the mainstream data as input signals and the junction flow state and driving support information as teacher signals may be configured.

The calculation of the prediction parameter in step 50 and the prediction of the prediction data in step 30 are processed in parallel.

In step 60, the prediction parameter calculated by the calculation means 201 is replaced with the prediction parameter of the prediction means 101. At this time, the main parameter A after calculation
m '(T1 + ΔTm + ΔT1m) is the main parameter Am (T1)
And each subparameter Asi '(T1 + Δ
Replace Tsi + ΔT1si) with each subparameter Asi (T1). By the time the phenomenon at the confluence is transmitted to the drainage pump station or its downstream river, each subparameter Asi '(T1 + ΔTsi +
If the calculation of ΔT1si) is completed, before the phenomenon at the confluence part is transmitted to the drainage pumping station or its downstream river, it is possible to predict the driving assistance information using the new subparameter after the computation and appropriately perform the driving assistance.

In the first embodiment, each prediction parameter is corrected based on the actual measurement data in parallel with the prediction of the confluence state and the driving support information so that the prediction value is within a predetermined error. In particular, in order to correct the prediction parameter for each tributary, correct the prediction parameter according to the phenomenon before the phenomenon is transmitted to the drainage pump station or its downstream river,
It can be reflected in driving support information. As a result, even if a sudden phenomenon such as heavy rain occurs, it becomes possible to provide driving assistance in accordance with the phenomenon.

The flow state prediction procedure 30 and the prediction parameter calculation procedure 50 of the first embodiment are always processed in parallel. However, the data used by the prediction means 101 and the calculation means 201 are not the data at the same time. Time T = T
In 1, the latest data, that is, tributary data Isi
(T1), main stream data Im (T1), flow state Csi (T
1) When the driving support information Cm (T1) is input to the data input unit 102, the prediction unit 101 reads the latest data and predicts the flow state at the junction and the driving support information. On the other hand, the sub-parameter calculating means 221 has time T = T1.
When the latest confluence flow state Csi (T1) at is read, the time T is calculated by using the tributary data and the confluence flow state until ΔTsi past time T = T1−ΔTsi already read.
The sub-parameters are calculated so that the junction flow state Csi (T1) at = T1 can be calculated with an error within a predetermined reference value. Therefore, the actual measured values from the current time to the past time ΔTsi are not used for the calculation of the subparameter.

Therefore, the predicted value predicted by the sub predictor 121 is used for the calculation of the sub parameter. As shown in FIG. 1, the sub-prediction unit 121 uses the tributary data up to time T = T1 and the measured value of the flow state of the merging portion to calculate the time T.
= Confluence state Csi '(T1 + ΔTsi) at T1 + ΔTsi
Is predicted and transmitted to the sub-parameter calculation means 221.
The sub-parameter calculation means 221 measures the latest tributary data up to time T = T1 and the actual measurement value of the confluence flow state, and the predicted value Cs of the confluence flow state up to time T = T1 + ΔTsi.
Based on i ′ (T1 + ΔTsi), the subparameter Asi ′
Calculate (T1 + ΔTsi + ΔT1si).

The same applies to the calculation of the main parameter in the main parameter calculation means 211. That is, the main prediction means 111, based on the mainstream data until time T = T1, the driving support information, and the actual measurement value of the flow state of the merging portion, the driving support information Cm ′ (T1 at time T = T1 + ΔTm.
+ ΔTm) is predicted and transmitted to the main parameter calculating means 211. The main parameter calculation means 211 determines the time T
= Main parameter Am '(T1 + ΔTm + ΔT1m) based on the latest mainstream data up to T1, actual measurement value of driving support information, flow state of the merging section, and predicted value Cm' (T1 + ΔTm) of driving support information up to time T = T1 + ΔTm. Is calculated.

In the first embodiment, the latest measured value can be utilized by using the predicted value of the prediction means for the calculation of the predicted parameter. Therefore, even for a phenomenon that changes more drastically, the prediction parameter can be corrected according to the phenomenon and reflected in the driving support information.

In the first embodiment, in order to predict the driving support information, the mainstream data actually measured on the inflow trunk line and its surroundings is used. However, directly reflecting the upstream data far from the drainage pump station in the driving support information leads to a decrease in the prediction accuracy and an increase in the calculation time of the prediction parameters. Therefore, in the same way as predicting the flow state of the confluence with the inflow trunk for each conduit, the upstream main trunk of the confluence is regarded as a tributary to the confluence, and is predicted in the same way as the conduit. Predict the flow condition at the confluence of the inflow trunk line.

As shown in FIG. 4, when a plurality of conduits merge with the inflow trunk line, it is possible to predict the flow state of the downstream merging portion by using the predicted value of the flow state of the merging portion on the upstream side. Good. FIG. 5 is a block diagram showing a configuration in which the sub-prediction unit 121 shown in FIG. 1 is divided for each conduit, FIG. 6 is a flowchart showing the support procedure, and the procedure 30 shown in FIG. 2 is divided. Shows the procedure.

Downstream of the confluence section h between the pipe h and the inflow trunk line,
When there is a confluence part i between the pipe i and the inflow trunk, the flow rate Ishj (T1) at the upstream point j of the pipe h, the rainfall Ishk (T1) at the position k in the vicinity, and the tributary of the inflow trunk. The tributary data such as the flow rate Isgj (T1) at the upstream point g when it is regarded as is read from the data input unit 102, the flow state Csh ′ (T1 + ΔTsh) at the merging portion h is predicted, and the upstream point j of the pipe i Based on the tributary data such as the flow rate Isij (T1) and the rainfall amount Isik (T1) at the position k in the vicinity thereof, and the predicted value Csh ′ (T1 + ΔTsh) of the flow state at the confluence h. Predict the flow state Csi '(T1 + ΔTsi). Similarly, the predicted value Csi ′ (T1 + ΔTs) of the flow state at the merging portion i
i) is used to predict the predicted value Csj ′ (T1 + ΔTsj) of the flow state at the downstream confluence j. In this way, the flow state on the downstream side is predicted from the flow state on the merge section on the upstream side, and the driving support information is predicted from the flow state on the merge section immediately before the drainage pumping station.

According to this embodiment, since the inflow trunk line between the merging portions is also regarded as a tributary, the flow state of the merging portion is predicted, and the flow state of the downstream side is predicted from the flow state of the upstream side. It is not necessary to predict the driving support information from the measured data at the upstream point, and the driving support information can be predicted with high accuracy,
The calculation of the prediction parameter can also be executed in a short time.

Second Embodiment A second embodiment in which the present invention is applied to a drainage pump station operation support system will be described with reference to FIGS. 7 to 11.

FIG. 7 is a block diagram showing the configuration of the drainage pump station driving support system according to the second embodiment of the present invention. This flow prediction system includes a data input unit 102, data from the data input unit 102 and a predetermined prediction parameter A.
Prediction means 101 for predicting the flow state based on m and Asi
And a driving support unit 103 that supports driving of a drainage pump station based on prediction data from the prediction unit 101, and a data input unit 10
A determination unit 104 that determines whether to correct the driving support information based on the data from 2 and the prediction data from the prediction unit 101 and determines whether to correct the prediction parameter used by the prediction unit 101, and a data input unit. The calculation means 201 calculates a prediction parameter based on the input data from 102 and transmits a new prediction parameter to the prediction means. That is, it differs from the first embodiment shown in FIG. 1 only in the configuration of the part relating to the determination means 104.

FIG. 8 is a flow chart showing a support procedure of the drainage pumping station operation support system shown in FIG. Step 10
2 to 40 are almost the same as steps 10 to 40 in FIG.

In procedure 10, the prediction parameters used in the prediction means 101 are initialized, and in procedure 20, the procedure 70
Enter the data actually measured in. In the data input unit 102, similar to the first embodiment of FIG. 1, data or flow indicating the flow state of the i-th tributary river of the i-th divert or a plurality of tributary rivers in the pipe network. Tributary data Isi that influences the state, data indicating the flow state of an inflow trunk line or river, or main stream data Im that affects the flow state,
The operation support information Cm indicating the flow condition of the drainage pumping station and its downstream, the flow condition at the confluence of the i-th pipe and the inflow trunk or the flow condition at the confluence of the i-th tributary river and the river. Four types of flow conditions Csi at the junction are shown.

In step 30, the predicting means 101 predicts the driving support information and the merging section flow state. Similar to the first embodiment of FIG. 1, the prediction unit 101 includes a sub prediction unit 121 and a main prediction unit 111, and each prediction unit is
Sub-parameters for predicting the flow condition at the junction respectively,
It has main parameters for predicting driving support information. The sub-prediction unit 121 reads the tributary data Isi (T1) and the junction flow state Csi (T1) at time T = T1, and the past data Isi (T) until time T = T1 is read.
Csi (T) and subparameter As at time T = T1
Based on i (T1), the predicted value Csi '(T1 + ΔTsi) of the flow state at the merging portion at the time T = T1 + ΔTsi after ΔTsi is predicted. The main prediction means 111 reads the mainstream data Im (T1) and the driving support information Cm (T1) at time T = T1, and the past mainstream data Im (T) and driving support information Cm until the time T = T1. (T), the merging portion flow state Csi (T), and the time T = T1 of the predicted value Csi 'of the merging portion flow state just predicted.
Based on the prediction result Cm '(T) from T1 to ΔTsi, the driving assist information Cm' (T1 + ΔTm) at time T = T1 + ΔTm after ΔTm is predicted using the main parameter Am (T1) at time T = T1. .

In step 40, the drainage pump station is drivingly supported based on the predicted value of the driving support information predicted by the prediction means 101.

At step 80, the judging means 104
The error of the prediction data predicted by the prediction means 101 is calculated. The predicted data is the flow state Csi '(T1) of the merging portion at time T = T1 + ΔTsi after ΔTsi from the current time T = T1.
+ ΔTsi) and the driving support information Cm ′ (T1 + ΔTm) at time T = T1 + ΔTm after ΔTm. In order to calculate the error of these prediction data, the actual measurement value of the confluence state Csi (T1 + ΔTsi) at time T = T1 + ΔTsi and the time T
= The measured value of the driving support information Cm (T1 + ΔTm) at T1 + ΔTm is read. Next, the measured value Cm (T1
+ ΔTm) and its predicted value Cm ′ (T1 + ΔTm) are compared,
Error of driving support information Em (T1 + ΔTm) = | Cm (T1 + ΔTm) −Cm ′ (T1 + ΔTm) | (1) and the measured value Csi of the flow state at the merging portion at each merging portion
(T1 + ΔTsi) and its predicted value Csi ′ (T1 + ΔTsi) are compared, and the error Esi (T1 + ΔTsi) = | Csi (T1 + ΔTsi) −Csi ′ (T1 + ΔTsi) |

Actually, the error of the equation (2) is calculated at the time T = T1 + ΔTsi, then the error of the equation (1) at the time T = T1 + ΔTm is calculated, and the magnitude of the error is determined each time. Move to the next step. Although the procedure of this determination is shown as procedure 80 in FIG. 8, the details are shown in FIG.
The procedure is as shown in.

In step 81, the predicted value Csi '(T1 + ΔT) of the confluence flow state from the sub prediction means 121 at time T = T1.
si) is read, and at time T = T1 + ΔTsi, data input section 1
The measured value Csi (T1 + ΔTsi) of the flow state of the merging portion is read from No. 02, and the error Esi (T1 + ΔTsi) is calculated by the mathematical expression (2).

The error Esi (T1 + ΔTsi) is a predetermined reference value Si
If it is within the range, it is not necessary to correct both the driving support information and the prediction parameter, so the procedure returns to step 20 and new data is input.

The error Esi (T1 + ΔTsi) is a predetermined reference value Si
If it is larger, it is necessary to improve the prediction accuracy of the flow state of the merging portion, and a correction command of the subparameter Asi (T1) is transmitted to the subparameter calculation means 221 of the calculation means 201. Since the predicted value of the merging section flow state is used to predict the driving support information, it is desirable to correct the main parameter Am (T1) by using the actual measurement value of the merging section flow state. Therefore, the main parameter Am (T1) is corrected. The command is also transmitted to the main parameter calculating means 211 of the calculating means 101. Further, the main predicting means 111 predicts the driving support information Cm '(T1 + ΔTm) using the predicted value Csi ′ (T1 + ΔTsi) of the flow state at the merging portion with a large error, and performs the driving support based on the result. , You may not be driving properly. Therefore, in order to bring the vehicle closer to the appropriate driving, a command for correcting the driving support information is transmitted to the main parameter calculating means 211.

In step 82, the predicted value Cm '(T1 + ΔT of the driving support information from the main prediction means 111 at time T = T1.
m) is read, and at time T = T1 + ΔTm, the data input unit 10
From 2 the measured value Cm (T1 + ΔTm) of the driving support information
The error Em (T1 + ΔTm) is calculated in (1).

If the error Em (T1 + ΔTm) is within the predetermined reference value M, it is not necessary to correct both the driving support information and the prediction parameter, and the process returns to step 20 to input new data.

If the error Em (T1 + ΔTm) is larger than the predetermined reference value M, it is necessary to improve the prediction accuracy of the driving support information, and a correction command for the main parameter Am (T1) is sent to the main parameter calculating means 211 of the calculating means 101. Send.

In step 90, the main parameter calculation means 211 determines at time T = T1 + ΔTsi the determination means 10
When a command to correct the driving support information is received from 4, time T =
The correction value Cm ″ (T1 + ΔTm) of the driving support information is predicted based on the measured values of the mainstream data up to T1 + ΔTsi, the driving support information, and the flow state of the merge section, and the corrected value Cm ″ (T1 + ΔTm) is transmitted to the driving support unit 103. Based on the corrected driving support information, driving support is provided again.

Therefore, in step 90, when predicting the correction value Cm ″ (T1 + ΔTm) of the driving support information, the actual measurement value Csi (T1 + ΔTsi) is used instead of the predicted value Csi ′ (T1 + ΔTsi) of the flow state at the merging portion with a large error. ), Time T =
It is possible to increase the accuracy of the predicted value of the driving support information by adding the measured data of T1 to T = T1 + ΔTsi.

In the second embodiment, the predicting means is divided into the sub predicting means 121 for predicting the flow state of each confluence and the main predicting means 111 for predicting the flow state around the machine area and downstream thereof. In step 90, the predicted value of the flow state at each merging portion is compared with the actual measured value to check the prediction accuracy. The flow state at the confluence then reaches the drainage pump station and its downstream after a certain time passes through the inflow trunk line and the river. Therefore, even if the error in the predicted value of the flow state at the junction is large, if the driving support information is corrected based on the measured value, the driving support can be carried out again before the phenomenon reaches the drainage pumping station or its downstream river. That is, according to the second embodiment, not only the driving support information of the drainage pump station is predicted from the information of each tributary and its surroundings, but also the error of the predicted value is confirmed at the merging portion,
When the error is large, the driving support information is corrected using the actual measurement value at the merging portion, so that the driving support can be performed with higher accuracy than before when necessary.

In step 50, when the prediction parameter correction command is received from the determination means 104, the calculation means 201 calculates the prediction parameter. The configuration of the calculation means and the calculation procedure are the same as in the first embodiment of FIG. That is, the calculation means 201 calculates the sub-parameter used for the prediction of the flow state of the merging portion, and the sub-parameter calculation means 221.
And a main parameter calculating means 211 for calculating a main parameter used for prediction of driving support information. The sub-parameter calculator 221 uses the i-th tributary data to determine the flow state of the merging portion to a predetermined reference value Si.
Sub-parameter Asi 'so that it can be calculated with an error within
Calculate (T1 + ΔTsi + ΔT1si). In the main parameter calculating means 211, the main parameter Am ′ (T1 + ΔTm + ΔT) is calculated so that the driving support information can be calculated with an error within a predetermined reference value M from the mainstream data and the merging section flow state.
1m) is calculated.

In step 60, the prediction parameter calculated by the calculation means 201 is replaced with the prediction parameter of the prediction means 101. Main parameter A after calculation at this time
m '(T1 + ΔTm + ΔT1m) is the main parameter Am (T1)
And each subparameter Asi '(T1 + Δ
Replace Tsi + ΔT1si) with each subparameter Asi (T1). By the time the phenomenon at the confluence is transmitted to the drainage pump station or its downstream river, each subparameter Asi '(T1 + ΔTsi +
If the calculation of ΔT1si) is completed, before the phenomenon at the confluence part is transmitted to the drainage pumping station or its downstream river, the new sub-parameter after the calculation can be used to predict the driving support information and appropriately support the operation.

In the second embodiment, each predicted value, that is, the predicted value of the confluence state and the driving support information is sequentially compared with the measured value, and the predicted parameter may be calculated only when the predicted parameter needs to be corrected. Therefore, the calculation of unnecessary prediction parameters can be eliminated, and the load on the driving support system can be reduced.

Similar to the first embodiment, the prediction data may be used for the calculation of the prediction parameter. That is, as shown in FIG. 7, the sub-prediction unit 121, based on the tributary data up to time T = T1 and the actual measurement value of the confluence flow state, merges flow state Csi ′ (T1 + ΔTsi at time T = T1 + ΔTsi.
Tsi) is predicted and transmitted to the sub-parameter calculation means 221. The sub-parameter calculation means 221 has a time T = T1.
Based on the latest tributary data up to and the actual measurement value of the confluent flow state and the predicted value Csi ′ (T1 + ΔTsi) of the confluent flow state until time T = T1 + ΔTsi.
i '(T1 + ΔTsi + ΔT1si) is calculated. Further, the main prediction means 111 predicts the driving support information Cm ′ (T1 + ΔTm) at the time T = T1 + ΔTm based on the mainstream data until the time T = T1, the driving support information, and the actual measurement value of the flow state of the merging portion, It is transmitted to the main parameter calculation means 211.
The main parameter calculation means 211 is based on the latest mainstream data up to time T = T1, the driving assistance information, the actual measurement value of the merging section flow state, and the predicted value Cm ′ (T1 + ΔTm) of the driving assistance information until time T = T1 + ΔTm. Then, the main parameter Am ′ (T1 + ΔTm + ΔT1m) is calculated.

In the second embodiment, the latest measured value can be utilized by using the predicted value of the prediction means for the calculation of the predicted parameter. Therefore, even for a phenomenon that changes more drastically, the prediction parameter can be corrected according to the phenomenon and reflected in the driving support information.

As in the first embodiment, the inflow trunk upstream of the confluence is regarded as a tributary to the confluence, and it is predicted in the same manner as the conduit, and the flow state at the confluence by the inflow trunk is predicted. May be. That is, as shown in FIG. 4, when a plurality of conduits merge with the inflow trunk, the downstream merged flow state may be predicted by using the predicted value of the upstream merged flow state. .

FIG. 10 is a sub prediction means 121 shown in FIG.
FIG. 11 is a diagram showing a configuration in which is divided into pipes, and FIG.
9 is a flowchart showing the support procedure, which is a division of the procedure 30 shown in FIG. 8. The configuration and the procedure are almost the same as the configuration shown in FIG. 5 and the procedure shown in FIG. 6 of the first embodiment, respectively. Only the procedure of transmitting the prediction result of the flow state of the merging portion to the determining unit 104 and determining the error of the prediction result for each merging portion is different.

According to the second embodiment, since the inflow trunk line between the merging portions is also regarded as a tributary, the flow state of the merging portion is predicted, and the flow state of the downstream side is predicted from the flow state of the upstream side. Further, it is not necessary to predict the driving support information from the actual measurement data at the upstream point, the driving support information can be predicted with high accuracy, and the calculation of the prediction parameter can be executed in a short time.

Example 3 Referring to FIGS. 12 and 13,
Example 3 in which the present invention is applied to a drainage pump operation support system
Will be explained. FIG. 12 is a block diagram showing the configuration of the third embodiment in which the present invention is applied to the drainage pumping station operation support system.

If a complicated conduit network or river network is divided for each tributary, it can be composed of a single conduit and a junction and an inflow trunk or river. When predicting the outflow state from such a single tributary to the inflow trunk or river, it is divided into the influence of the flow state at the upstream point of the tributary and the influence of the inflow state of the tributary from that point to the junction. You can think about it.

For example, in the case of FIG. 4, the influence (for example, outflow amount) of the flow state (for example, flow rate) Isij (T1) at the upstream point j of the i-th tributary on the confluence part i separated by the distance Δx (for example, outflow amount) and The time can be expressed by a simple empirical formula Fs, ignoring the influence of the state of inflow into the tributary from the point j to the junction i. That is, the predicted value Csij ″ (T1 + ΔTsi) considering only the flow state at the upstream point j of the flow state at the junction i
Can be expressed by Csij ″ (T1 + ΔTsi) = Fs (Isij (T1), Δx) (3). Generally, the empirical formula Fs is a function of the flow state at the upstream point and the distance to the confluence.

The third embodiment shown in FIG. 12 corresponds to FIG.
Sub predicting means 121 and sub parameter calculating means 2
21 is divided into two kinds of prediction means and calculation means, respectively, and the influence of the flow state at the upstream point on the merging portion is calculated by a mathematical formula.
The prediction is performed in (3), and similarly to the first embodiment of FIG. 1 or the second embodiment of FIG. 7, the influence of the inflow state into the tributary is predicted with the flow state and the predictable parameter Asis.

In FIG. 12, the sub prediction means 121 is
Fixed sub-prediction means 122 for predicting the flow state of the merging part considering only the flow state at the upstream point by using the function Fs of the mathematical expression (3), and the tributary from the upstream point to the merging point using the variable prediction parameter. The variable sub-prediction unit 123 for predicting the flow state of the merged portion is also provided in consideration of the inflow state of the.

The fixed sub-prediction means 122 reads the upstream data Isij (T1) giving the flow state at the upstream point of the tributary from the data input unit 102, and only the flow state at the upstream point is read by the function Fs of the equation (3). The flow state Csij ″ (T1 + ΔTsi) of the merging portion in consideration of the above is predicted.

The variable sub-prediction unit 123 considers only the flow state at the upstream point predicted by the fixed sub-prediction unit 122, and the confluence state Csij ″ (T1 + ΔTsi) and the tributary from the upstream point to the confluence. From the basin data Isik (T1) giving the inflow state, the confluence state Csij '(T1 + ΔTsi) is predicted using the variable prediction parameter Asis (T1).

Prediction value Csij '(T1 + Δ
Tsi) is transmitted to the main prediction means 111, and is used when predicting the driving support information and executing the driving support in the driving support unit 103, as in the first embodiment of FIG. 1 or the second embodiment of FIG. 7. To be

On the other hand, the sub-parameter calculation means 221
Using the function Fs of Equation (3), the variable sub-prediction means 122 for predicting the confluence flow state that considers only the flow state at the upstream point, and the variable sub-prediction means 123 for calculating the variable prediction parameters used Sub-parameter calculation means 2
And 23.

In the sub-parameter calculation means 221,
The fixed sub-prediction means 122 reads the upstream data Isij (T1) and predicts the merged part flow state Csij ″ (T1 + ΔTsi) considering only the flow state at the upstream point. The merging section flow state Csij ″ (T1 + ΔTsi) and the basin data Isik (T1) considering only the flow state at the upstream point predicted by the means 122
From the above, the variable prediction parameter Asis' (T is set so that the flow state Csij (T1 + ΔTsi) of the merging portion can be predicted within the error Si.
1 + ΔTsi + ΔT1si), and the variable sub-prediction unit 12
Replace with the variable sub-parameter Asis (T1) of 3.

Alternatively, as in the second embodiment shown in FIG. 7, the determination means 10 determines the predicted value Csij ′ (T1 + ΔTsi) of the flow state at the junction.
4 to calculate the error of the driving support information and the error of the merging section flow state from the driving support information and the predicted value of the merging section flow state and each measured value, based on Equations (1) and (2). May be. If the error in the flow condition at the junction is larger than a predetermined reference value Si, it is necessary to improve the prediction accuracy of the flow condition at the junction, so the variable prediction parameter Asis (T
The correction command of 1) is transmitted to the sub-parameter calculation means 221. When the sub-parameter calculation means 221 reads this correction instruction, the variable prediction parameter Asis ′ (T1 + ΔTsi + ΔT1si) is calculated as in the case where the determination means 104 is not provided, and the variable sub-parameter Asi of the variable sub-prediction means 123 is calculated.
Replace with s (T1).

Further, similar to the embodiment of FIG. 5, the downstream merged portion flow state may be predicted by using the predicted value of the upstream merged fluid state. At this time, as shown in FIG. 4, if the upstream end of the inflow trunk line from the merging portion h to the merging portion i is the merging portion h and the downstream end is the merging portion i, the flow state (eg, flow rate) at the merging portion h The influence of Csh (T1) on the confluence part i separated by the distance Δy (for example, the outflow amount) and the time are as follows.
If ignoring the effect of the inflow state on the inflow trunk line from the to the confluence part i, it can be expressed by a simple empirical formula Fm. That is, the predicted value Csih ″ (T1 + ΔTsh) considering only the flow state of the flow state at the merge portion i is expressed by Csih ″ (T1 + ΔTsh) = Fm (Csh (T1), Δy) (4) it can. In general, the empirical formula Fm is a function of the flow state at the upstream point and the distance to the junction.

The influence of the inflow trunk line between the merging portions on the merging portion at the downstream end should be considered in consideration of the flow state at the merging portion at the upstream end and the peripheral state that affects the flow state of the inflow trunk line between the merging portions. It can be treated like a tributary, since it is enough. That is, in FIG. 12, the upstream merging portion flow state is used as the upstream data, and the downstream merging portion flow state is used as the merging portion flow state, and the fixed sub-prediction means uses the function Fs.
You can use Fm instead of.

In FIG. 13, the inflow trunk line of each junction is treated in the same manner as a tributary, and the predicted value of the flow state of the upstream junction is used to
It is a block diagram which shows an example of a structure of the means which estimates the downstream junction flow state. In FIG. 13, the prediction means 10
1 includes a plurality of downstream end flow state prediction means 124. For example, when predicting the flow state of the downstream end i, the downstream end flow state predicting means 124 reads the predicted value Csh ″ (T1 + ΔTsh) of the flow state at the merging portion on the upstream side, and the function Fm.
Is used to predict the confluent flow state Csi ″ (T1 + ΔTsh) on the downstream side due to the confluent flow state on the upstream side. Next, the predicted value Csi ″ (T1 + ΔTsh) is predicted by the sub-prediction means 121. From the predicted value of the confluence flow state due to the effect of the tributary, the confluence state Csi ′ (T1 +
Predict ΔTsi). This predicted value Csi ′ (T1 + ΔTsi)
Is the flow state Csj ′ (T1 + ΔTs
Used to predict j). The last downstream end is just before the drainage pump station, and the flow state here corresponds to the driving support information.

In the third embodiment, the sub-prediction means is a fixed sub-prediction means considering only the flow state at the upstream point, and a variable sub-prediction means considering the inflow state into the tributary from the upstream point to the confluence. It is divided. The fixed sub-prediction means
The flow state of the merging portion is predicted using a predetermined physical model, and the variable sub-prediction means uses the variable sub-parameter calculated from the past flow state to make the prediction. When the prediction is performed using only the variable subparameters that depend on the past flow state, the prediction error may be large for a flow state that is completely different from the past flow state. When the fixed sub-prediction method is used, the influence of the flow state at the upstream point, which is dominant on the flow state at the confluence, is predicted by a universal physical model, so there is no prediction error even for sudden flow states. It is small and can provide appropriate driving assistance even during typhoons. However, since it is difficult to predict all flow states with high accuracy using only the physical model, the variable sub-prediction is used in order to correct the results of the physical model and to consider the inflow state into the tributary from the upstream point to the confluence. It is desirable to use the means together.

Fourth Embodiment A fourth embodiment in which the flow prediction system of the present invention is applied to a traffic flow prediction system will be described with reference to FIGS. 14 and 15.

FIG. 14 is a diagram for explaining a flow state around an intersection which is a prediction target of the fourth embodiment in which the flow prediction system of the present invention is applied to the traffic flow prediction system. When there are a plurality of intersections, the merging and branching of vehicles from the branch road to the main road at each intersection is predicted, and the mainstream condition, that is, the traffic condition on the downstream main road is predicted.

FIG. 15 is a block diagram showing an example of the configuration of the fourth embodiment of the traffic flow prediction system according to the present invention.
The traffic flow prediction system of FIG.
Predicting means 101 for predicting a flow state based on the data from the data input section 102 and predetermined prediction parameters Am, Asi, a data output section 105 for outputting the flow state predicted by the predicting means 101, and a data input section 102. Determination means 104 for determining whether to correct the prediction parameter used in the prediction means 101 based on the data from the data and the prediction data from the prediction means 101, and the prediction parameter based on the input data from the data input unit 102. And a calculation means 201 for calculating and transmitting a new prediction parameter to the prediction means.

The predicting means 101 comprises a sub predicting means 121 and a main predicting means 111, and each predicting means is
Each has subparameters that predict the inflow and outflow conditions at the intersection and main parameters that predict the mainstream condition. In the sub-prediction means 121, at time T = T1, the intersection upstream state Ui (T1), inflow state IDi (T1), IUi (T
1), the outflow state ODi (T1), OUi (T1) is read, and based on the past upstream state until the time T = T1, the inflow state, the outflow state, and the subparameter Asi (T1) at the time T = T1. , Predicted value IDi ′ (T1 + ΔTi) of inflow state and outflow state at time T = T1 + ΔTi after ΔTi, IUi ′
(T1 + ΔTi), ODi ′ (T1 + ΔTi), OUi ′ (T1 + ΔTi
Predict Ti). In the main prediction means 111, time T =
Inflow / outflow status IDi (T1), IUi (T1), ODi
(T1), OUi (T1) and mainstream state Ri (T1) are read,
Time T = T of the predicted value of the past inflow / outflow state and the mainstream state until the time T = T1 and the mainstream state just predicted.
Using the prediction results from 1 to T1 + ΔTi, using the main parameter Am (T1) at time T = T1,
The mainstream state Ri ′ (T1 at time T = T1 + ΔTri after ΔTri
Predict + ΔTri). The predicted mainstream state is transmitted to the data output unit 105 and output to the outside.

In the judging means 104, the predicting means 10
The error of the prediction data predicted in 1 is calculated. The error determination is performed in the same manner as in the second embodiment, and the measured value and the predicted value of each inflow and outflow state are compared, or the predicted value and the measured value of the mainstream state are compared, and if the error is larger than a predetermined reference value. , A correction command for executing the calculation of the sub parameter or the main parameter is sent to the calculation means 201.

The calculating means 201 comprises a sub-parameter calculating means 221 for calculating a sub-parameter and a main parameter calculating means 211 for calculating a main parameter. The sub-parameter calculation means 221 calculates the sub-parameter so that the inflow and outflow states can be calculated with an error within a predetermined reference value from the upstream state. The main parameter calculation means 211 calculates the main parameter from the inflow / outflow state and the upstream mainstream state so that the downstream mainstream state can be calculated with an error within a predetermined reference value. The calculated sub-parameters or main parameters are replaced with the sub-parameters and main parameters of the prediction means 101.

According to the above embodiment, the prediction parameters can be corrected in parallel with the prediction of the inflow / outflow state and the mainstream state based on the actually measured data so that the prediction is within a predetermined error. Therefore, it is possible to appropriately predict sudden changes in traffic conditions that are difficult to learn in advance, such as traffic jams such as Golden Week and homecoming rush hours at Obon.

[0105]

According to the present invention, the predicted downstream target flow state and the downstream actually measured flow state are sequentially compared, and the prediction parameter for prediction is updated based on the comparison result. It is possible to obtain a flow prediction system that can appropriately predict the downstream flow condition even with respect to the flow condition. For example,
In parallel with the prediction of the inflow / outflow state and the mainstream state, each prediction parameter is corrected based on the actual measurement data so that the prediction falls within a predetermined error. for that reason,
It is possible to appropriately predict sudden changes in traffic conditions that are difficult to learn in advance, that is, traffic congestion such as during consecutive holidays.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a drainage pump station driving support system according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an example of a support procedure of the drainage pumping station operation support system of FIG.

FIG. 3 is a diagram showing a data input / output relationship of a driving support system.

FIG. 4 is a diagram showing an example of a relationship between tributary data and a predicted flow rate.

5 is a block diagram showing a configuration in which the sub-prediction unit shown in FIG. 1 is divided into pipes.

FIG. 6 is a flowchart showing a support procedure of the system of FIG.

FIG. 7 is a block diagram showing a configuration of a drainage pump station driving support system according to a second embodiment of the present invention.

8 is a flowchart showing a support procedure of the drainage pumping station operation support system of FIG.

FIG. 9 is a diagram showing details of a determination procedure.

10 is a diagram showing a configuration in which the sub-prediction unit shown in FIG. 7 is divided into pipes.

11 is a flowchart showing a support procedure of the system of FIG.

FIG. 12 is a block diagram showing a configuration of a drainage pump station driving support system according to a third embodiment of the present invention.

FIG. 13 is a block diagram showing an example of a configuration of means for predicting a downstream junction flow state by using an upstream main junction flow state as a tributary and using an upstream junction flow state prediction value; Is.

FIG. 14 is a diagram illustrating a flow state around an intersection, which is a prediction target of Example 4 in which the flow prediction system of the present invention is applied to a traffic flow prediction system.

FIG. 15 is a block diagram showing an example of the configuration of a fourth embodiment of the traffic flow prediction system according to the present invention.

FIG. 16 is a diagram showing a relationship between a pipe network and a river network and a drainage pump station.

[Explanation of symbols]

101 Prediction means 102 data input section 103 Driving Support Department 104 determination means 111 Main prediction means 121 Sub prediction means 122 Fixed Confluence Part Flow Condition Prediction Means 123 Variable Confluence Part Flow State Prediction Means 124 Downstream end flow state prediction means 201 computing means 211 Main parameter calculation means 221 Sub-parameter calculation means 223 Variable prediction parameter calculation means

─────────────────────────────────────────────────── --Continued front page (56) Reference JP-A-3-134703 (JP, A) JP-A-9-95917 (JP, A) JP-A-5-250007 (JP, A) JP-A-4- 34602 (JP, A) JP-A-7-98794 (JP, A) JP-A-8-123538 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G05B 13/04 E02B 1 / 00 G08G 1/00 G06F 17/00

Claims (8)

(57) [Claims] 1. The upstream flow condition of fluid, traffic, etc. is monitored.
Flow prediction for predicting downstream flow conditions such as fluid and traffic
In the system, The flow prediction system is used for upstream flows such as fluid and traffic.
Fluids and traffic based on conditions and predetermined predictive parameters
Prediction means for predicting downstream target flow conditions such as
Flow condition of upstream and traffic and measured flow condition of downstream
The prediction parameter used in the prediction means based on
And a calculation means for calculating Prediction and calculation of downstream target flow state in the prediction means
Parallel processing with prediction parametersmake, The predicted value by the predicting means and the measured flow state of the downstream
If the error is larger than the specified value, the measured flow
Using the state value as the prediction parameter, the error
If it is within the fixed value, the calculated prediction parameter is
Equipped with determination means to use Flow prediction system characterized by
Tem. 2. The upstream flow condition of fluid, traffic, etc. is monitored.
Flow prediction for predicting downstream flow conditions such as fluid and traffic
In the system, The flow prediction system is used for upstream flows such as fluid and traffic.
Fluids and traffic based on conditions and predetermined predictive parameters
Prediction means for predicting downstream target flow conditions such as
Flow condition of upstream such as traffic and traffic, measured flow condition of downstream, before
Based on the target flow state of the downstream by the prediction means,
A calculator for calculating the prediction parameters used in the measuring means
With steps and Prediction and calculation of downstream target flow state in the prediction means
Parallel processing with prediction parametersmake, The predicted value by the predicting means and the measured flow state of the downstream
If the error is larger than the specified value, the measured flow
Using the state value as the prediction parameter, the error
If it is within the fixed value, the calculated prediction parameter is
Equipped with determination means to use Flow prediction system characterized by
Tem. 3. An upstream flow condition of fluid, traffic, etc. is monitored.
Flow prediction for predicting downstream flow conditions such as fluid and traffic
In the system, The flow prediction system is used for upstream flows such as fluid and traffic.
Fluids and traffic based on conditions and predetermined predictive parameters
Prediction means for predicting downstream target flow conditions such as
Flow condition of upstream and traffic and measured flow condition of downstream
The prediction parameter used in the prediction means based on
Calculating means for calculating
The flow rate is measured by comparing the measured flow state in the downstream with the calculation means.
And a determination means for determining whether or not to calculate the parameterEquipment
e, If the error due to the comparison is larger than a predetermined value, the actual measurement is performed.
Using the flow state value as the prediction parameter, the error
If is within a predetermined value, the calculated prediction parameter is calculated.
Use A flow prediction system characterized in that 4. The upstream flow condition of fluid, traffic, etc. is monitored.
Flow prediction for predicting downstream flow conditions such as fluid and traffic
In the system, The flow prediction system is used for upstream flows such as fluid and traffic.
Fluids and traffic based on conditions and predetermined predictive parameters
Prediction means for predicting downstream target flow conditions such as
Flow condition of upstream such as traffic and traffic, measured flow condition of downstream, before
Based on the target flow state of the downstream by the prediction means,
A calculator for calculating the prediction parameters used in the measuring means
Stage, the predicted value by the predicting means and the measured flow downstream
Comparing with the state and calculating the prediction parameter by the calculating means
And the determination means to determine whether or notPrepare, If the error due to the comparison is larger than a predetermined value, the actual measurement is performed.
Using the flow state value as the prediction parameter, the error
If is within a predetermined value, the calculated prediction parameter is calculated.
Use A flow prediction system characterized in that 5. The flow prediction system according to any one of claims 1 to 4, wherein the prediction means includes a plurality of sub-prediction means for predicting a flow state of each confluence part or each branch part. And a main prediction means for predicting a downstream target flow state. 6. The upstream flow condition of fluid, traffic, etc. is monitored.
Flow prediction for predicting downstream flow conditions such as fluid and traffic
In the system, The flow prediction system is used for upstream flows such as fluid and traffic.
Fluids and traffic based on conditions and predetermined predictive parameters
Prediction means for predicting downstream target flow conditions such as
Flow condition of upstream and traffic and measured flow condition of downstream
The prediction parameter used in the prediction means based on
And a calculation means for calculating The predicting means is a flow state of each merging portion or each branching portion.
Of multiple sub-prediction means for predicting the
Main prediction means to predictConsists of The predicted value by the predicting means and the measured flow state of the downstream
If the error is larger than the specified value, the measured flow
Using the state value as the prediction parameter, the error
If it is within the fixed value, the calculated prediction parameter is
Equipped with determination means to use Flow prediction system characterized by
Tem. 7. The flow predicting system according to claim 6, wherein the predicting unit is a target flow state of the downstream merging portion or the branch portion based on a predicted value of a flow state of the upstream merging portion or the branch portion. A flow forecasting system characterized by being a means for forecasting. 8. The flow predicting system according to claim 5, wherein the predicting unit includes a fixed parameter for predicting a target flow state at each of the merging portions, each of the branch portions, or the downstream. A flow forecasting system comprising fixed sub forecasting means.