JP5032445B2 - Unknown water amount prediction device and unknown water prediction system - Google Patents

Description

  The present invention relates to an unknown water amount prediction apparatus and an unknown water prediction system for predicting an unknown water amount flowing into a sewage treatment plant.

  Currently, combined sewers and split sewers are generally known as sewer systems.

  In the combined sewer, sewage and rainwater are caused to flow through the same pipe to perform sewage treatment and rainwater drainage treatment.

  In recent years, the development of shunt sewerage has progressed. In the sewerage sewer system, sewage and rainwater are treated separately by flowing down separate pipes for sewage and rainwater.

  However, in the sewerage sewerage system, there are problems such as damage to the sewage pipes and inflow of rainwater into the sewage pipes due to misconnection between the sewage pipes and the rainwater pipes. Has occurred. Rainwater flowing into the sewage pipe is called unknown water. As a result, problems such as an influence on operation and an increase in treatment costs occur due to the sewage exceeding the sewage treatment capacity of the sewage treatment plant flowing into the sewage treatment plant.

Therefore, in order to reduce the influence of unknown water, it is necessary to grasp the point where the unknown water is generated and repair the point. However, since the pipe is buried underground, it is very difficult to grasp the point. In the investigation of the location of unknown water, collect and analyze pipe burial conditions, sewage treatment plant operation data, survey flow measurement data, etc. As a result, the location where the unknown water is generated is narrowed down, but this narrowing down requires a lot of labor, time and cost. Therefore, there is a demand for a technique for narrowing down the location where unknown water is generated in a short time and in a simple manner. For example, Patent Document 1 discloses an unknown water generation distribution estimation device as a technique for narrowing down the generation location of such unknown water.
Japanese Patent No. 3857670

  However, even if the unknown water location can be narrowed down by the technique of Patent Document 1, it takes a considerable amount of time to identify and repair the location where the unknown water is generated. Therefore, a large amount of sewage containing unknown water flows into the sewage treatment plant for a long time until the location where unknown water is generated is repaired. For this reason, in a sewage treatment plant, there exists a subject that it is difficult to respond | correspond to the amount of inflowing sewage.

  The present invention was devised to solve the above-described problems, and provides an unknown water amount prediction device and an unknown water prediction system that enable a sewage treatment plant to easily cope with sewage containing unknown water. It is aimed.

  In order to achieve the above object, the invention according to claim 1 of the present invention substitutes rainfall amounts at a plurality of points in a processing target region into an unknown water amount prediction formula calculated by principal component analysis, and performs sewage treatment. An unknown water amount prediction device that calculates an unknown water amount flowing into a field as a predicted value.

  The invention according to claim 2 is characterized in that the unknown water amount prediction formula is calculated by performing a principal component analysis using an unknown amount of water based on a past amount of rainfall and past measured values as elements. The unknown water amount prediction apparatus according to Item 1.

  In the invention according to claim 3, the unknown water amount prediction formula is set based on a first principal component formula having the highest contribution ratio among a plurality of principal component formulas calculated by principal component analysis. The unknown water amount prediction apparatus according to claim 1 or 2, wherein the apparatus is an unknown water amount prediction apparatus.

  The invention according to claim 4 is characterized in that the unknown water amount prediction formula is calculated by performing a principal component analysis on a higher-order element of the rainfall amount. The unknown water amount prediction apparatus according to item 1.

  Further, the invention according to claim 5 calculates the delay time having a high correlation between the rainfall amount and the unknown water amount, and performs principal component analysis on the rainfall amount and the unknown water amount in consideration of the delay time. A water amount prediction formula is calculated, It is an unknown water amount prediction apparatus of any one of Claims 1-4 characterized by the above-mentioned.

  In the invention according to claim 6, the rainfall is clustered into a plurality of rainfall areas, a principal component analysis is performed for each of the clustered rainfall areas, and an unknown water amount prediction formula is calculated for each of the rainfall areas. It is calculated, It is an unknown water quantity prediction apparatus of any one of Claims 1-5 characterized by the above-mentioned.

  Moreover, the invention which concerns on Claim 7 multiplies each unknown water quantity calculated from the several principal component formula calculated by the principal component analysis by the contribution rate of a corresponding principal component formula, and a plurality of said unknown water quantity and contribution rate The unknown water amount as claimed in any one of claims 1 to 6, wherein the unknown water amount is calculated as a predicted value by dividing the sum of the products of the two and the sum of the contribution ratios. It is a prediction device.

  Moreover, the invention which concerns on Claim 8 is based on the unknown water quantity prediction apparatus of any one of Claims 1-7, and the unknown water quantity as a predicted value transmitted from the said unknown water quantity prediction apparatus, An unknown water prediction system including a water quality calculation device that calculates the quality of sewage flowing into a sewage treatment plant.

  The invention according to claim 9 is characterized in that the water quality calculation device calculates and predicts water quality by a prediction model constructed by learning using a neural network technique. It is an unknown water prediction system described.

  According to the present invention, since the unknown water amount can be calculated and predicted by the unknown water amount prediction formula calculated by the principal component analysis, the sewage treatment plant can easily cope with sewage containing unknown water. .

(First embodiment)
Hereinafter, an unknown water prediction system including an unknown water amount prediction device according to a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is an overall configuration diagram of an unknown water amount prediction system according to the first embodiment.

  The unknown water prediction system 1 according to the first embodiment is for predicting by calculating an unknown water amount Qu at the time of rainfall, and transmitting the unknown water amount Qu to the sewage treatment plant 52 as a predicted value.

  First, the sewage treatment system 51 that is a treatment target of the unknown water prediction system 1 will be described. The sewage treatment system 51 is a diversion sewer system. As shown in FIG. 1, the sewage treatment system 51 includes a sewage pipe 53 and a sewage treatment plant 52 buried in the ground. The pipe 53 is drained with sewage discharged from a household or business establishment in the region 61 to be treated. In addition to the pipe 53, a rainwater pipe is also buried in the ground, but its illustration and description are omitted. The entire ellipse shown in FIG. 1 is a processing target area 61 of the sewage treatment system 51.

  One end of the pipe 53 is connected to the sewage treatment plant 52. That is, the sewage flows in the pipe 53 in the direction of the arrow 55 and is sent to the sewage treatment plant 52. Here, the pipe 53 is installed so that only sewage is drained. However, unknown water due to rain or the like may flow into the pipe 53 due to damage of the pipe 53 or incorrect connection of the rainwater pipe.

  As shown in FIG. 1, the unknown water prediction system 1 according to the first embodiment includes a rain radar 2, a rainfall amount information processing device 3, a sewage amount measurement device 4, an unknown water amount calculation device 5, and an unknown water amount prediction device. 6 is provided. Each configuration 2 to 6 includes hardware such as an arithmetic processing unit such as a CPU, a storage unit capable of storing various information and programs, a connection unit for connecting them, and an input / output unit for inputting / outputting each information to / from the outside. Wear is provided (not shown).

  The rain radar 2 is for measuring the distribution and intensity of rainfall in the region 61 to be processed. The rainfall radar 2 observes the intensity and direction of radio waves scattered and returned by hitting raindrops, and the transmission / reception time of radio waves, and transmits the information to the rainfall information processing apparatus 3 online.

The rainfall amount information processing device 3 processes the above-described observation information transmitted from the rainfall radar 2. Specifically, rainfall information processing apparatus 3 converts the radio wave strength and direction, rainfall _r 1 ~r _n transceiver time based on the zone ₆₂ 1 through 62 _n (multiple locations). The areas 62 _{1 to} 62 _n referred to here are obtained by dividing the region 61 to be processed into n pieces. Rainfall information processing apparatus 3 transmits the converted rainfall r ₁ ~r _n into unknown water predictor 6 online.

  The sewage amount measuring device 4 measures the sewage amount Qs that flows through the pipe 53 and is sent to the sewage treatment plant 52. The amount of sewage Qs referred to here includes not only the amount of sewage but also an unknown amount of water Qr during rainfall. The sewage amount measuring device 4 transmits the measured sewage amount Qs to the unknown water amount calculating device 5 online.

  The unknown water amount calculation device 5 subtracts the sewage amount Qsf in fine weather stored in the storage means (not shown) from the sewage amount Qs measured by the sewage amount measurement device 4 to obtain the unknown water amount Qr (= Qs) as a measurement value. -Qsf) is calculated. The unknown water amount calculation device 5 transmits the unknown water amount Qr as a measurement value to the unknown water amount prediction device 6 online.

The unknown water amount prediction device 6 sets the following unknown water amount prediction formula (1) by principal component analysis, and substitutes _n rainfalls r _{1 to} rn into the unknown water amount prediction formula (1), The unknown water amount Qu as a predicted value is calculated. Note that w _1m (m = 1, 2,... N) is a weighting coefficient.
Qu = w ₁₁ · r ₁ + w ₁₂ · r ₂ +... + W _1n · r _n (1)

  First, the setting method of the unknown water amount prediction formula (1) described above by principal component analysis will be specifically described.

First, different times of rainfall _r 1 ~r _n and at least the data of unknown water Qr measurements (n + 1) pieces measurement and calculation. Then, by principal component analysis of these (n + 1) or more past rainfall _r 1 ~r _n and data for the last unknown water Qr, eigenvectors _a i1 _~a in the i-th principal _{component, b} i ( i = 1, 2,..., n + 1) is calculated. As a result, the following (n + 1) principal component formulas are calculated. These are defined as the first principal component formula (1 ₁ ) to (n + 1) principal component formula (1 _{n + 1} ). Z _i is the i-th principal component score.

Here, the contribution ratio of the first principal component formula (1 ₁ ) is very high as compared to the contribution ratio of the second principal component formula (1 ₂ ) or less. In addition, the contribution ratio is further reduced below the third principal component formula (1 ₃ ). Therefore, by omitting the second principal component equation _{(1 2)} below, employs the first principal component equation that best represents the correlation of rainfall _r 1 ~r _n and unknown water Qr only _{(1 1)} .

Wherein each principal component axis passes through the center of gravity of the data of rainfall _r 1 ~r _n and unknown water Qr. Further, the principal component axis passes through the principal component axis having a large dispersion in the order of the first principal component axis and the second principal component axis. The first principal component axis 71 and the second principal component axis 72 are shown in FIG. In FIG. 2, the vertical axis indicates the amount of unknown water, and the horizontal axis indicates the amount of rainfall in each area (note that although it is actually (n + 1) th space, it cannot be shown, the unknown water amount and the amount of rainfall are not shown here. Schematically shown as a relationship). Each point represents the data of each rainfall _r 1 ~r _n and unknown water Qr. As shown in FIG. 2, the dispersion of the first principal component axis 71 is the largest among the principal component axes, and the dispersion below the second principal component axis 72 is very small. Therefore, the data of rainfall r ₁ ~r _n and unknown amount of water Qr is an element of the principal component analysis, abundant in the vicinity of the first principal component axis 71. Furthermore, the data of rainfall _r 1 ~r _n and unknown water Qr are abundant in the vicinity of the center of gravity.

Here, the center of gravity is a point where the principal component score Z ₁ = 0. Therefore, the principal component score Z ₁ = 0 is substituted into the first principal component formula (1 ₁ ). Furthermore, the weight coefficient w _1m = − (a _1m / b ₁ ) and the unknown water amount Qr ₁ = Qu. As a result, a plurality of past zone 62 _1-62 precipitation of _n r ₁ ~r _n and unknown water prediction expression by principal component analysis of unknown water Qr calculated based on the volume of sewage Qs is past measurements as elements ( 1) is derived.

  The unknown water amount prediction device 6 transmits the calculated unknown water amount Qu, which is a predicted value, to the sewage treatment plant 52 online.

  Next, operation | movement of the unknown water prediction system 1 by 1st Embodiment mentioned above is demonstrated.

  When the rain is observed, the rain radar 2 measures the distribution and intensity of the rain in the entire region 61 to be processed. Thereafter, the rainfall radar 2 transmits the measured observation information to the rainfall amount information processing device 3.

Next, the rainfall information processing apparatus 3, based on the observation information, calculates the rainfall _r 1 ~r _n of each zone ₆₂ 1 through 62 _n. Thereafter, the rainfall information processing device 3 transmits the information of rainfall r ₁ ~r _n of a plurality of points to the unknown water predicting device 6.

  On the other hand, the sewage amount measuring device 4 measures the sewage amount Qs flowing through the pipe 53 and being sent to the sewage treatment plant 52. Thereafter, the sewage amount measuring device 4 transmits information on the measured sewage amount Qs to the unknown water amount calculating device 5.

  The unknown water amount calculation device 5 calculates the unknown water amount Qr as a measured value by subtracting the sewage amount Qsf in the fine weather stored in the storage unit from the transmitted sewage amount Qs in the rain. Thereafter, the unknown water amount calculation device 5 transmits information of the calculated unknown water amount Qr to the unknown water amount prediction device 6.

Perform these processes described above, at least (n + 1) times and sends an unknown amount of water Qr rainfall r ₁ ~r _n and measurements into unknown water predictor 6 (n + 1) or more times.

Then, unknown water predicting device 6, unknown water Qr rainfall r ₁ ~r _n and measurements, after being transmitted (n + 1) or more, performs the principal component analysis. Thereby, the unknown water amount prediction apparatus 6 calculates the weighting coefficient w _1m and calculates the unknown water amount prediction formula (1).

Then, unknown water predicting device 6, the unknown amount of water prediction equation has been calculated by the principal component analysis as described above (1), by substituting rainfall r ₁ ~r _n newly transmitted from rainfall information processing apparatus 3 Then, the unknown water amount Qu as a predicted value is calculated.

  Next, the unknown water amount prediction device 6 transmits the calculated unknown water amount Qu to the sewage treatment plant 52 as a predicted value. In the sewage treatment plant 52, based on the unknown amount of water Qu, the operation state of equipment for treating sewage is monitored and controlled.

As described above, unknown water predicting device 6 according to the first embodiment, the unknown amount of water prediction equation set by principal component analysis (1), by substituting the newly transmitted rainfall r ₁ ~r _n The unknown water amount Qu as a predicted value can be calculated. Then, the unknown water amount prediction device 6 transmits the unknown water amount Qu to the sewage treatment plant 52, so that the sewage treatment plant 52 can treat the sewage based on the predicted unknown water amount Qu. As a result, the sewage treatment plant 52 can easily cope with sewage containing unknown water.

In addition, since the unknown water amount prediction device 6 can predict the unknown water amount Qu by the unknown water amount prediction formula (1), the above-described effects can be easily obtained without performing complicated processing by an arithmetic processing unit or the like. Further, unknown water predicting device 6, it is possible to obtain information of unknown water Qr rainfall r ₁ ~r _n and measurements online, it can be predicted more easily calculate the unknown water Qu.

(Second Embodiment)
Next, a second embodiment in which the rain radar according to the above-described embodiment is changed will be described. FIG. 3 is an overall configuration diagram of an unknown water amount prediction system according to the second embodiment. In addition, the same code | symbol is attached | subjected to the structure similar to embodiment mentioned above, and description is abbreviate | omitted.

As shown in FIG. 3, the unknown water prediction system _{1 </} b _{> A according} to the second embodiment includes a plurality of rain gauges 63 _{1 to} 63 _n installed at different points. Rain gauge ₆₃ 1 to 63 _n measures the amount of rainfall _r 1 ~r _n of the installed location, and transmits to the unknown water predicting device 6. The rain gauge 63 _m (m = 1, 2,..., N) includes a reed that rotates when a predetermined amount of rain enters. The configuration of the rain gauge 63 _m is not particularly limited.

In the unknown water prediction system 1A according to the second embodiment, the rainfall amount information processing device 3 in the first embodiment can be omitted by installing the rain gauge 63 _m at a plurality of points.

(Third embodiment)
Next, 3rd Embodiment which changed the unknown water amount prediction formula of embodiment mentioned above is described. In addition, the same code | symbol is attached | subjected to the structure similar to embodiment mentioned above, and description is abbreviate | omitted.

The In third embodiment according Unknown water predicting device 6, the correlation between the unknown water Qu predicted value and the rainfall r ₁ ~r _n is higher so as to delay time t ₁ ~t _n rainfall r Considering setting the unknown water prediction equation (1) by principal component analysis ₁ ~r _n. May also constitute unknown water predicting device 6 so as to substitute the rainfall r ₁ ~r _n considering the delay time t ₁ ~t _n unknown water prediction equation (1).

Here, the delay time t _m, reaching time from zone 62 _m of unknown water corresponding to rainfall r _m to reach the sewage treatment plant 52, the time rain reaches the surface, the rain from the surface to Kanmizo 53 It can be determined in consideration of the flow rate of sewage flowing through the pipe tub 53 and the like.

For example, if the unknown water corresponding to the rainfall r ₁ reaches the sewage treatment plant 52 with a delay time t later than the unknown water corresponding to the rainfall r ₂ , a delay time t before the rainfall r ₁ is assumed. setting the unknown water prediction equation (1) by principal component analysis rainfall r _2. Then, its unknown water prediction equation (1), substitutes the rainfall _r 1 ~r _n considering the delay time _{t m.} Thereby, the prediction precision of unknown water quantity Qu can be improved more.

(Fourth embodiment)
Next, 4th Embodiment which changed the unknown water amount prediction formula of embodiment mentioned above is described. In addition, the same code | symbol is attached | subjected to the structure similar to embodiment mentioned above, and description is abbreviate | omitted.

In Unknown water predicting device 6 according to the fourth embodiment, set Unknown water prediction expression which will be described later by performing a principal component analysis in consideration of the squared components of rainfall r ₁ ~r _n a (2). Specifically, by principal component analysis of unknown water Qr rainfall _r 1 ~r _n and measurements, as shown below (2n + 1) pieces of main component formula _{(2 1) ~ (2 2n} + 1) Generate.

Next, the contribution ratios of the principal component formulas (2 ₁ ) to (2 _{2n + 1} ) are considered. Again, considering that the contribution ratio of the first principal component formula (2 ₁ ) is very high, only the first principal component formula (2 ₁ ) is adopted. Further, when the principal component scores Z ′ ₁ = 0 and Qu ₁ = Qu are substituted and transformed, the following equation (2) is obtained. This is defined as the unknown water amount prediction formula (2). However, it is assumed that weighting factor w _1m = − (a _1m / b ₁ ) and weighting factor w ′ _1m = − (a ′ _1m / b ₁ ).

In Unknown water predicting device 6 according to the fourth embodiment, because it uses unknown water prediction equation in consideration of the square of rainfall r ₁ ~r _n (2), high accuracy than in consideration of the non-linear element unknown water Qu prediction can be realized.

In Unknown water predicting device 6 according to the fourth embodiment described above, in consideration of rainfall r ₁ ² ~r _n ² of the square, was calculated unknown water predictive equation (2), the cube higher order rainfall The unknown water amount prediction formula may be calculated in consideration of the amounts r ₁ ^{3 to} r _n ^{3 and the} like.

(Fifth embodiment)
Next, a fifth embodiment in which the unknown water amount prediction formula of the above-described embodiment is changed will be described. In addition, the same code | symbol is attached | subjected to the structure similar to embodiment mentioned above, and description is abbreviate | omitted.

  In the unknown water amount prediction apparatus 6 according to the fifth embodiment, the unknown water amount prediction formula is set in consideration of the contribution ratio of each principal component formula.

Hereinafter, the setting method of the unknown water amount prediction formula (3) will be specifically described. First, as in the first embodiment, (n + 1) principal component equations (1 ₁ ) to (1 _{n + 1} ) are calculated by principal component analysis. Then, after substituting _Z m = 0 to m-th principal component equation _{(1 m),} substituting rainfall _r 1 ~r _m. Note that m = 1, 2,... N. Thereby, the unknown water amount Qu _m is calculated from each principal component equation (1 _m ).

Here, the contribution ratio of the m-th principal component formula (1 _m ) is C _m . Then, multiplying the contribution rate C _m and the corresponding respective unknown water Qu _m. The sum obtained by adding all the products (= Qu _m · C _m ) calculated in this way is divided by the sum obtained by adding all the contribution ratios C _m . As a result, the unknown water amount prediction formula becomes the following formula (3). The unknown water amount prediction device 6 calculates the unknown water amount Qu as a predicted value by the unknown water amount prediction formula (3).

By calculating such a contribution rate C _m unknown water Qu into consideration, we can achieve a more accurate prediction of unknown water Qu.

In consideration of the contribution ratio C _m , for example, the third principal component formula (1 ₃ ) and the following may be omitted. Moreover, you may apply this embodiment to 4th Embodiment.

(Sixth embodiment)
Next, a sixth embodiment in which the unknown water amount prediction formula of the above-described embodiment is changed will be described. In addition, the same code | symbol is attached | subjected to the structure similar to embodiment mentioned above, and description is abbreviate | omitted.

In Unknown water predicting device 6 according to the sixth embodiment, as shown in FIG. 4, and clustering rainfall r ₁ ~r _n into two rainfall regions 74 ₁ and rainfall region 74 _2, the first main each component shaft ₇₁ 1, 71 ₂ and the second principal component axis ₇₂ 1, 72 ₂ may be set. Then, from the first principal component equation corresponding to the regions ₇₄ 1, 74 _2, sets the regions ₇₄ 1, 74 ₂ unknown water prediction equation. Then, by substituting the rainfall r ₁ ~r _n each unknown water prediction equations, unknown water Qu predicted value is calculated.

Here, a method of determining the dividing line 73 of clustering rainfall r ₁ ~r _n is not particularly limited. For example, as a method for determining the dividing line 73, a k-means method or the like can be given. In the k-means method, clusters are appropriately formed and elements are allocated, and then the center of gravity of the elements is calculated for each cluster. Each element is then assigned to the cluster with the nearest centroid. After that, the k-means method is a method of repeating these processes until the elements assigned to the cluster are not changed.

As another determination method, a circle around the center of gravity of rainfall _r 1 ~r _n, predetermined proportion (e.g., 80%) rainfall _r 1 ~r _n of, in the circle Determine the diameter of the circle to enter. Of the tangent lines of the circle, a tangent line parallel to the vertical axis (Q axis) may be used as the dividing line 73.

Thus, the unknown amount of water predicting device 6, corresponding to the rainfall r ₁ ~r _n, it is possible to predict the unknown amount of water Qu, it is possible to further improve the accuracy of prediction.

In the above-described embodiment, the unknown water amount prediction formula is calculated by approximating the principal component score Z ₁ = 0, but the target component of the principal component analysis is classified for each rainfall amount by a plurality of principal component scores. (Clustering) An unknown water amount prediction formula may be calculated for the rainfall region corresponding to each principal component score. Thus, for the remote elements from Z 1 _{= 0,} it is possible to improve the prediction accuracy of unknown water.

(Seventh embodiment)
Next, 7th Embodiment which changed the unknown water prediction system of embodiment mentioned above is described. FIG. 5 is an overall configuration diagram of an unknown water amount prediction system according to the seventh embodiment. FIG. 6 is a configuration diagram of a water quality calculation device according to the seventh embodiment. In addition, the same code | symbol is attached | subjected to the structure similar to embodiment mentioned above, and description is abbreviate | omitted.

  The unknown water prediction system 1B according to the seventh embodiment further includes a water quality calculation device 7.

  The water quality calculation device 7 calculates and predicts the water quality SSu of the sewage that changes due to unknown water, using a prediction model constructed by learning using neural network technology. In addition, the water quality here is SS concentration (suspended substance concentration).

  The water quality calculation device 7 includes a sewage flow rate Qf in fine weather, a sewage water quality SSf in fine weather, a clear weather period Tf that is an elapsed period from the previous rain, and a measured sewage water quality SSr in rainy weather. Is transmitted from the storage unit 52a of the monitoring control system of the sewage treatment plant 52. Further, the unknown water amount Qu predicted from the unknown water amount prediction device 6 is transmitted to the water quality calculation device 7.

  The water quality computing device 7 includes an input layer 21, an intermediate layer 22, and an output layer 23. The input layer 21, the intermediate layer 22, and the output layer 23 are connected by a predetermined weight coefficient.

  The input layer 21 receives the flow rate Qf of sewage in fine weather, the quality of sewage SSf in fine weather, the clear weather period Tf, and the unknown water amount Qu predicted by the unknown water amount prediction device 6.

  In the intermediate layer 22, the weighting coefficient is changed so that an error between the information input to the input layer 21 and the information input to the output layer 23 becomes small.

  In the output layer 23, the actually measured quality of sewage SSr during rainfall is input as teaching information.

  In addition, the water quality calculation device 7 transmits the predicted water quality SSu to the sewage treatment plant 52. In the sewage treatment plant 52, sewage treatment is performed based on the transmitted water quality SSu.

  As described above, the unknown water prediction system 1B can predict the water quality SSu by calculation and transmit the predicted water quality SSu to the sewage treatment plant 52. Thereby, in the sewage treatment plant 52, it is possible to perform processing corresponding to the water quality SSu that is predicted to be thinner than usual due to unknown water. As a result, the sewage treatment plant 52 can efficiently operate equipment for sewage treatment.

  In addition, you may comprise so that the water quality calculating apparatus 7 may calculate water quality simply by the weighted average using a weighting coefficient. Moreover, although SS density | concentration (suspended-substance density | concentration) was employ | adopted as an example of water quality, BOD (biochemical oxygen demand), COD (chemical oxygen demand), TN (total nitrogen amount), T- You may employ | adopt the water quality by P (total phosphorus amount) etc.

  As mentioned above, although this invention was demonstrated in detail using embodiment, this invention is not limited to embodiment described in this specification. The scope of the present invention is determined by the description of the claims and the scope equivalent to the description of the claims. Hereinafter, modified embodiments in which the above-described embodiment is partially modified will be described.

  Numerical values and the like shown in the configuration of the above-described embodiment can be changed as appropriate.

  Further, in the above-described embodiment, the unknown water amount prediction device, the unknown water amount calculation device, the water quality calculation device, and the like are illustrated outside the sewage treatment plant, but part or all of these configurations are included in the monitoring control system of the sewage treatment plant. It may be configured as a part.

  In addition, rainfall information may be obtained from meteorological information instead of the rainfall radar or rain gauge.

It is a whole block diagram of the unknown water quantity prediction system by a 1st embodiment. It is a figure explaining a principal component axis in principal component analysis. It is a whole block diagram of the unknown amount prediction system by a 2nd embodiment. It is a figure explaining the clustering in 6th Embodiment. It is a whole block diagram of the unknown water quantity prediction system by 7th Embodiment. It is a block diagram of the water quality calculating apparatus in 7th Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1, 1A, 1B Unknown water prediction system 2 Rain radar 3 Rainfall information processing device 4 Sewage amount measurement device 5 Unknown water amount calculation device 6 Unknown water amount prediction device 7 Water quality calculation device 21 Input layer 22 Intermediate layer 23 Output layer 51 Sewage treatment system 52 Sewage Treatment Plant 52a Storage Unit 53 of Monitoring Control System Pipe 61 Treatment Target Areas 62 _{1 to} 62 _n Areas 63 _{1 to} 63 _n Rain Gauge 71 First Main Component Axis 72 Second Main Component Axis 73 Dividing Lines 71 ₁ and 71 ₂ First principal component axis 72 ₁ , 72 ₂ Second principal component axis 74 ₁ Rainfall region 74 ₂ Rainfall region C _{1 to} C _n Contribution rate Qf Flow rate Qr Unknown water amount (measured value)
Qu Unknown water volume (predicted value)
Qu ₁ ~Qu _n unknown amount of water Qs sewage amount Qsf sewage amount _r 1 ~r _n rainfall SSf water quality SSr water quality SSu water quality t delay time Tf sunny period _w 11 _{~w 1n} weighting factor
Z ₁ ~Z _{n + 1} principal component scores

Claims (9)

  An unknown amount of water that is calculated by substituting rainfall at multiple points in the treatment target area into the unknown amount prediction formula calculated by principal component analysis, and calculating the unknown amount of water flowing into the sewage treatment plant as a predicted value Prediction device.   The unknown water amount prediction apparatus according to claim 1, wherein the unknown water amount prediction formula is calculated by performing principal component analysis using an unknown amount of water based on a past amount of rainfall and past measurement values as elements.   The unknown water amount prediction formula is set based on a first principal component formula having the highest contribution ratio among a plurality of principal component formulas calculated by performing principal component analysis. Item 3. An unknown water amount prediction apparatus according to Item 2.   The unknown water amount prediction apparatus according to any one of claims 1 to 3, wherein the unknown water amount prediction formula is calculated by performing a principal component analysis on a higher-order element of the rainfall amount.   Calculating an unknown water amount prediction formula by calculating a delay time having a high correlation between the rainfall amount and the unknown water amount, and analyzing principally the rainfall amount and the unknown water amount in consideration of the delay time. The unknown water amount prediction apparatus according to any one of claims 1 to 4. Clustering rainfall into multiple rainfall areas,
6. The unknown according to claim 1, wherein a principal component analysis is performed for each clustered rainfall region, and an unknown water amount prediction formula is calculated for each of the rainfall regions. Water volume prediction device.   Multiply each unknown water amount calculated from a plurality of principal component formulas calculated by principal component analysis by the contribution rate of the corresponding principal component formula, and add up the product of the plurality of unknown water amounts and contribution rate, The unknown water amount prediction apparatus according to claim 1, wherein an unknown water amount as a predicted value is calculated by dividing by the sum of the contribution ratios. The unknown water amount prediction apparatus according to any one of claims 1 to 7,
An unknown water prediction system comprising: a water quality calculation device that calculates the quality of sewage flowing into a sewage treatment plant based on an unknown water amount as a predicted value transmitted from the unknown water amount prediction device.   The unknown water prediction system according to claim 8, wherein the water quality calculation device calculates and predicts water quality using a prediction model constructed by learning using neural network technology.