Aeration tank controlling apparatus, wastewater treatment apparatus, and aeration tank controlling method

Abstract

To provide an aeration tank management device, a wastewater treatment device, and an aeration tank management method for easily determining the quality of an operation state without depending on an individual technique of a skilled person.SOLUTION: The detection unit 20 of the aeration tank management device 10 includes the odor sensor 22 and the water quality meter 24, and the processing unit 50 includes a data storage unit 56 in which detection results and treatment contents are stored and a data calculation unit 58 that performs machine learning on the detection results and the treatment contents and generates a clustering output model in which a result of clustering is output data using the odor detected by the odor sensor 22 and / or the water quality measured by the water quality meter 24 as input data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an aeration tank management device, a wastewater treatment device, and an aeration tank management method. [Background technology]

[0002] There are a wide variety of factors that need to be managed in the operation and management of wastewater treatment facilities, and in many cases wastewater treatment facilities centered on aeration tanks are managed through water quality management and microbial management. However, most of these are large-scale operations, and comprehensive judgments rely heavily on experience, so they often rely on the personal skills of experienced personnel. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 9-94595 [Patent Document 2] Japanese Patent Application Laid-Open No. 2014-83524 Summary of the Invention [Problem to be solved by the invention]

[0004] Patent Document 1 discloses a technique for controlling the aerobic and anaerobic states in a wastewater treatment reaction tank using an oxidation-reduction potential electrode (ORP). However, this technique has the drawback that it is not possible to accurately grasp the nitrification and denitrification states when the wastewater contains substances that change the pH or redox substances.

[0005] Patent Document 2 proposes a method for directly monitoring the microbial activity of wastewater undergoing biological treatment, in which the intensity of the fluorescent light emitted by a compound that represents the metabolic activity of the microorganisms is measured when the wastewater is irradiated with excitation light of a predetermined wavelength, and the microbial activity is measured from the relationship between the activated sludge concentration and the fluorescence intensity when the activated sludge concentration is changed, thereby controlling the wastewater treatment device.However, it is estimated that this method would be difficult to apply when the biota that makes up the microbial community itself has changed significantly.

[0006] The present invention has been made in view of the above, and aims to provide an aeration tank management device, a wastewater treatment device, and an aeration tank management method that enable easy determination of the operating condition of wastewater treatment facilities centered around an aeration tank, without relying on large-scale water quality management, microbial management, or the personal skills of an expert. [Means for solving the problem]

[0007] The aeration tank management device of the present invention is an aeration tank management device provided in an aeration tank, and comprises a detection unit and a processing unit. The detection unit comprises an odor sensor that detects odors from the aeration tank and a water quality meter that measures the water quality of the treatment liquid in the aeration tank. The processing unit comprises an input unit into which treatment details to be performed on the aeration tank in response to the results detected by the odor sensor and / or the results measured by the water quality meter are input, a data storage unit that stores the results and treatment details, and a data calculation unit that reads the results and treatment details from the data storage unit, performs machine learning on the read results and treatment details, and generates a clustering output model using the odor detected by the odor sensor and / or the water quality measured by the water quality meter as input data and the results of clustering them as output data. [Effects of the Invention]

[0008] The present invention is characterized by the use of detection of odor components in addition to water quality, and by its adaptability to changes in the quality of the water flowing into the wastewater treatment facility. This makes it possible to provide an aeration tank management device, wastewater treatment device, and aeration tank management method that can easily determine whether the operation is good or bad without relying on the personal skills of an experienced person. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a block diagram showing the configuration of a wastewater treatment device according to the present invention. [Figure 2] FIG. 2 is a flow chart showing the diagnostic procedure in the aeration tank management method of the present invention. [Figure 3]FIG. 3 is a diagram showing an example of a cluster in the aeration tank management method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0010] (Wastewater treatment equipment) A wastewater treatment device 1 according to an embodiment of the present invention will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the wastewater treatment device 1. As shown in FIG. 1, the wastewater treatment device 1 includes an aeration tank 100 and an aeration tank management device 10. The aeration tank 100 is a tank that treats wastewater through the action of microorganisms. The aeration tank management device 10 is a device that performs processes related to the management of the aeration tank 100. Management of the aeration tank 100 broadly includes measures to ensure that wastewater is appropriately treated in the aeration tank 100, such as control of a blower provided in the aeration tank 100, control of the amount of chemicals added, and addition of food for the microorganisms. Management of the aeration tank 100 also includes daily inspections of the aeration tank 100.

[0011] (Aeration tank management device) The aeration tank management device 10 includes a detection unit 20 and a processing unit 50. The detection unit 20 is a unit that detects values ​​indicating the state of the aeration tank 100, such as odors from the aeration tank 100, and / or acquires information indicating the state of the aeration tank 100. The processing unit 50 is a unit that processes the values ​​detected by the detection unit 20 and generates information useful for managing the aeration tank 100.

[0012] (Detection unit) The detection unit 20 includes an odor sensor 22, a water quality meter 24, and an operating equipment information acquisition unit 26. The odor sensor 22, the water quality meter 24, and the operating equipment information acquisition unit 26 may be provided in the aeration tank 100.

[0013] The odor sensor 22 is a sensor that detects odors from the aeration tank 100. The odor sensor 22 used may be a plurality of odor sensors corresponding to individual gases that may emerge from the aeration tank 100. Alternatively, the odor sensor 22 may be an odor sensor that measures the overall odor, for example, the intensity (relative concentration) of a reducing gas.

[0014] The water quality meter 24 is a device that measures the water quality of the treatment liquid in the aeration tank 100. The water quality may include, for example, pH, DO (dissolved oxygen concentration), MLSS (activated sludge concentration), and ORP (oxidation-reduction potential).

[0015] The operating equipment information acquisition unit 26 is a unit that acquires operating information and / or equipment information of the aeration tank 100. The operating information may include, for example, the date and time when the aeration tank 100 was operated, the operating status of the blower, and the status of chemical injection. The equipment information may include, for example, the flow rate, water temperature, and air temperature. Furthermore, the operating equipment information acquisition unit 26 may acquire information on production equipment and pretreatment equipment such as pressurized flotation. This information can be used in the data preprocessing section, which will be described later.

[0016] The detection unit 20 inputs the results detected by the odor sensor 22 and / or the results measured by the water quality meter 24 and / or the operating information and / or equipment information acquired by the operating equipment information acquisition unit 26 to the processing unit 50.

[0017] (Processing section) The processing unit 50 includes an input unit 54, a data storage unit 56, a data calculation unit 58, and a memory unit 60. The input unit 54 is a unit into which the details of the treatment to be performed on the aeration tank 100 in response to the results detected by the odor sensor 22 and / or the results measured by the water quality meter 24 are input. The data storage unit 56 is a unit into which the data input from the detection unit 20 and the data input from the input unit 54 are stored. In detail, the data storage unit 56 stores at least one of the results detected by the odor sensor 22, the results measured by the water quality meter 24, the operating information and / or equipment information acquired by the operating equipment information acquisition unit 26, and the details of the treatment to be performed on the aeration tank 100 that are input in response to the results detected by the odor sensor 22 and / or the results measured by the water quality meter 24.

[0018] The data calculation unit 58 is a part that generates a clustering output model. The data calculation unit 58 reads out the results detected by the odor sensor 22 and / or the results measured by the water quality meter 24 from the data storage unit 56. The data calculation unit 58 performs machine learning on the read out results and generates a clustering output model. The clustering output model is a model that takes the odor detected by the odor sensor 22 and / or the water quality measured by the water quality meter 24 as input data and outputs the results of clustering them. By using the clustering output model, it is possible to provide an aeration tank management device 10 and a wastewater treatment device 1 that reduce the human burden in managing the aeration tank 100.

[0019] The clustering output model generated by the data calculation unit 58 is not limited to the above example. Another example of the clustering output model is a clustering output model generated by including in the data to be machine-learned at least one of the operation information and the equipment information acquired by the operating equipment information acquisition unit 26, in addition to at least one of the odor detected by the odor sensor 22 and the water quality measured by the water quality meter 24. This clustering output model enables more accurate clustering.

[0020] Furthermore, when generating a clustering output model using the above-described machine learning, the data for machine learning can be preprocessed before the machine learning. For example, as part of the preprocessing, the time of wastewater inflow into the aeration tank 100, the measurement season, and other information can be added to the data for machine learning. Furthermore, to further improve the accuracy of the clustering output model, the data for machine learning can be created taking into account the quality of the raw wastewater, the flow rate of the raw wastewater, the time it takes for the raw wastewater to enter the aeration tank, temperature fluctuations, and the operating environment of the wastewater treatment device 1. For example, if it takes half a day for the raw wastewater to enter the aeration tank, the water quality and flow rate data of the raw wastewater half a day ago can be associated with the data in the aeration tank at the current time. Furthermore, the operating environment of the wastewater treatment device 1 can include, for example, the retention time in the water receiving tank.

[0021] A more specific explanation will be given. For example, consider a case where wastewater from a manufacturing facility is used as raw wastewater, and this raw wastewater flows into an aeration tank via a flow control tank or raw water layer. The raw wastewater data, such as water quality data and flow rate data, used to form a normal cluster are acquired at the time the raw wastewater is discharged from the manufacturing facility. The aeration tank data, such as odor data and water quality data, are acquired in the aeration tank. Typically, there is a time lag before wastewater from the manufacturing facility flows into the aeration tank. Therefore, the raw wastewater data and aeration tank data are associated. For example, assume that it takes 12 hours for wastewater from the manufacturing facility to flow into the aeration tank. In this case, the raw wastewater data from 12 hours ago is used as the aeration tank data to form a normal cluster. This makes it easier to form more accurate clusters.

[0022] The storage unit 60 is a part that stores the clustering output model generated by the data calculation unit 58. The clustering output model is read out from the storage unit 60 and used for clustering.

[0023] (Aeration tank management method) The flow of the aeration tank management method using the aeration tank management device 10 will be described. In the aeration tank management method, the aeration tank 100 is diagnosed. The diagnosis flow is roughly divided into four steps, step 1 to step 4. This will be described with reference to FIG. 2. FIG. 2 is a flow chart showing the diagnosis flow of this embodiment. In FIG. 2, S1 indicates step 1. The same applies to the other steps. Each of the following steps can be executed by a processing unit 50, such as a data calculation unit 58.

[0024] (S1) S1 is the normal state cluster formation step. S1 is also called the cluster identification step. In S1, the clusters formed during normal operation are identified. A cluster is a group formed by items with similar states. Examples of states include water quality, generated gas composition, and sludge properties. Clustering is the process of dividing into clusters. Clustering is also called cluster analysis. Methods that can be used for clustering include, for example, k-means clustering, hierarchical clustering, and DBSCAN.

[0025] The cluster formation for the normal state in S1 can be performed using, for example, a clustering output model stored in the memory unit 60 of the processing unit 50. A normal cluster can be formed by inputting data on the odor from the aeration tank 100 and / or the water quality of the treated liquid in the aeration tank 100, which is output during normal operation, into the clustering output model.

[0026] Furthermore, when forming a normal cluster, it is preferable to properly manage the aeration tank 100 on a daily basis and accumulate data while keeping the aeration tank 100 in a normal state.

[0027] (S2) S2 is the step for detecting abnormal conditions. In S2, clusters of abnormal conditions are also formed. Like S1, S2 is also called the cluster understanding step. In S2, abnormal conditions are detected based on data from normal conditions. In other words, abnormal conditions are detected based on the degree of deviation from the normal state.

[0028] An alarm is set to be issued when an abnormality is detected. When an abnormal condition occurs, initially, the person goes to the site and takes action. Once the action is taken and it is confirmed that the condition has returned to normal, the action taken is assigned to the abnormal data as a label. The process of assigning a label indicating the situation to each abnormal condition cluster formed is called labeling. Examples of labels used in labeling include (1) nutrient deficiency, (2) excess excess sludge, and (3) oxygen deficiency.

[0029] When forming an abnormal cluster, a simulated state in which the treated liquid in the aeration tank 100 is abnormal may be created, and the abnormal cluster may be formed based on data from that state. Also, multiple types of states in which the treated liquid in the aeration tank 100 is abnormal may be created. This makes it possible to form abnormal clusters corresponding to a variety of abnormal states in a short period of time. It is also possible to understand how the position of the abnormal cluster changes depending on the abnormal state. This makes it possible to improve the accuracy of diagnosis when the abnormal state changes while the wastewater treatment device 1 is in operation.

[0030] (S3) S3 is the step of clustering normal / abnormal states. By going through the process of transition from S1 to S1 and S2, that is, the process of transition between normal and abnormal, it becomes possible to grasp whether the state is normal or abnormal through clustering. Furthermore, by performing clustering, abnormal states may also be classified into several types of clusters.

[0031] S3 is also referred to as a diagnosis step. In S3, the state of the aeration tank 100 may be diagnosed based on the degree of dissociation between pre-prepared clusters in a normal state and clusters in an abnormal state and the current state of the aeration tank 100. The degree of dissociation of the current state of the aeration tank 100 from the normal clusters is called the degree of cluster dissociation. In S3, a diagnosis of whether the aeration tank 100 is normal or abnormal may be made based on the degree of cluster dissociation.

[0032] The degree of cluster dissociation will now be explained. When the clusters are visualized on a graph, the degree of cluster dissociation corresponds to the visual distance between a point on the graph showing the current state of the aeration tank 100 and a normal cluster. When there are multiple points on the graph showing the current state of the aeration tank 100, the degree of dissociation can also be determined from the perspective of the difference between the shape of the cluster formed by these points and the visual cluster shape from a normal cluster.

[0033] As a method for visualizing the clusters described above, it is also possible to reduce the dimension of a large amount of state data, that is, perform dimensionality compression from feature quantities. The method of dimensionality compression is not particularly limited, and for example, PCA (principal component analysis) or t-SNE can be used. Visualizing the clusters makes it easier to grasp the number of clusters and / or the cluster shapes. While visualizing the clusters and / or reducing the dimension of the data are effective methods, they are not essential for creating a diagnostic model. In S3, after diagnosing whether the aeration tank 100 is normal or abnormal, a status diagnosis is performed in the event of an abnormality.

[0034] (S4) S4 is the step of control based on the clustering results. In S4, action is taken based on the cluster to which the abnormality belongs. By labeling the status of each abnormal state for each abnormal state cluster in S3, it becomes possible to take action according to the abnormal state based on the labeling results.

[0035] In S4, for example, the abnormal cluster closest to the current state of the aeration tank 100 is determined based on the newly output data on the odor from the aeration tank 100 and / or the water quality of the treated liquid in the aeration tank 100 and the proximity to each previously identified abnormal cluster, and the classification of the abnormal state labeled in the abnormal cluster can be identified. Then, measures to resolve the abnormal state can be determined based on the abnormal state classification of the labeled abnormal cluster, and the aeration tank can be controlled accordingly.

[0036] Examples of control of the aeration tank 100 included in measures to resolve abnormal conditions include the following controls. If the measure is "sludge removal," the sludge removal valve is automatically opened and the sludge removal pump is turned on to perform fixed-quantity sludge removal control. If the measure is "microbial food addition," fixed-quantity chemical injection control for microbial food is automatically performed. If the measure is "antifoam agent addition," fixed-quantity chemical injection control for antifoam agent is automatically performed.

[0037] (Variation) In the above description, normal and abnormal clusters are used as examples of clusters. An example of diagnosing the aeration tank 100 by referring to the normal and / or abnormal clusters has been described. The clusters used for diagnosis are not limited to normal and / or abnormal clusters. This description will be made with reference to FIG. 3. FIG. 3 is a diagram illustrating examples of clusters according to an embodiment of the present invention. FIG. 3 illustrates three clusters: a normal cluster 301, a feed addition cluster 302, and a sludge extraction cluster 303. The normal cluster 301 is an example of the normal cluster described in the above description of the embodiment. The feed addition cluster 302 is a cluster formed when microbial feed is added to the aeration tank 100. The sludge extraction cluster 303 is a cluster formed when an increased amount of excess sludge is extracted from the aeration tank 100. In this way, by identifying the clusters formed when certain measures are taken in the aeration tank 100, it is possible to more accurately determine the measures to be taken when the aeration tank 100 is diagnosed as abnormal. This is because it is possible to determine how the clusters will move when a specific measure is taken in the aeration tank 100. Depending on the positional relationship between the plot showing the aeration tank 100 in an abnormal state and the normal cluster 301, it becomes easier to determine the measures to be taken.

[0038] Furthermore, normal clusters and abnormal clusters may change with the seasons. This is because the temperature and manufacturing processes in factories change with the seasons, and the quality of wastewater may also change. Therefore, the accuracy of anomaly detection can be improved by changing the clusters used depending on the season. For example, summer clusters and winter clusters corresponding to the seasonal fluctuations between summer and winter may be formed. Then, for example, a summer model using the summer clusters and a winter model using the winter clusters may be created as diagnostic models that can respond to the seasonal fluctuations between summer and winter. The number of abnormal clusters and the accompanying number of labels may be changed by season.

[0039] Although the present invention has been described above as an embodiment, it is not limited to the above-described embodiment, and various changes, modifications, and combinations are possible.

[0040] The aeration tank management method of this embodiment allows the blower control, chemical injection amount control, and microbial food addition, which have conventionally been performed manually to manage the aeration tank 100, to be performed without the knowledge of a skilled engineer. Furthermore, it also eliminates the need for the labor of daily inspections performed to detect abnormal conditions, such as microscopic observation and measurement of SV30 (activated sludge settling rate).

[0041] Although the present invention has been described above as an embodiment, it is not limited to the above-described embodiment, and various changes, modifications, and combinations are possible.

[0042] <1> An aeration tank management device provided in an aeration tank, a detection unit and a processing unit, The detection unit an odor sensor that detects odors from the aeration tank; a water quality meter that measures the water quality of the treated liquid in the aeration tank, The processing unit an input unit into which a treatment to be performed on the aeration tank in response to the result detected by the odor sensor and / or the result measured by the water quality meter is input; a data storage unit in which the results and the treatment details are stored; an aeration tank management device comprising: a data calculation unit that reads out each of the results and the treatment details from the data storage unit, performs machine learning on each of the read out results and treatment details, and generates a clustering output model in which the odor detected by the odor sensor and / or the water quality measured by the water quality meter are used as input data and the results of clustering them are used as output data. <2> In the above-described aeration tank management device, the detection unit further includes an operating equipment information acquisition unit that is provided in the aeration tank and acquires operating information and / or equipment information of the aeration tank, The data storage unit further stores the operation information and / or facility information acquired by the operation facility information acquisition unit, The data calculation unit performs machine learning on at least one of the odor detected by the odor sensor and the water quality measured by the water quality meter, and at least one of the operating information and equipment information acquired by the operating equipment information acquisition unit, and generates a clustering output model. <3> A wastewater treatment device comprising the above-mentioned aeration tank management device and an aeration tank. <4> a cluster grasping step of forming a normal cluster consisting of data output during normal operation regarding data on the odor from the aeration tank and / or the water quality of the treated liquid in the aeration tank; and a diagnostic step of detecting an abnormality based on the degree of cluster dissociation, which is the degree of dissociation from the normal cluster, of the newly output data. <5> The above-described aeration tank management method, wherein the aeration tank management method proceeds to the diagnosis step after a predetermined period of time has elapsed since the cluster identification step. <6> In the above-described aeration tank management method, in the cluster identification step, an abnormal cluster is formed consisting of data output during abnormal operation regarding data on the odor from the aeration tank and / or the water quality of the treated liquid in the aeration tank. <7> In the above-described aeration tank management method, in the cluster grasping step, each abnormal cluster is previously labeled with a label corresponding to a known abnormal state classification. <8> In the above-described aeration tank management method, measures are taken to resolve the abnormal state based on the classification of the abnormal state of the labeled abnormal cluster. <9> In the above-mentioned aeration tank management method, the content of the treatment is determined using a model in which data on the odor from the aeration tank and / or the water quality of the treated liquid in the aeration tank is input data and the content of the treatment is output data. <10> In the above-described aeration tank management method, the abnormal cluster is formed based on the data output in a plurality of simulated abnormal states of the treated liquid in the aeration tank. <11> In the above-described aeration tank management method, the data output during the normal operation in the cluster identification step includes at least one of the time it takes for wastewater to flow from its discharge source into the aeration tank, and the season in which the data on the odor and / or the water quality of the treated liquid in the aeration tank was acquired.

[0043] [Contribution to the United Nations-led Sustainable Development Goals (SDGs)] This disclosure includes matters that contribute to achieving Goal 6 of the SDGs (Sustainable Development Goals), "Clean water and sanitation," and Goal 9, "Industry, innovation and infrastructure." [Explanation of symbols]

[0044] 1 Wastewater treatment equipment 10 Aeration tank management device 20 Detector 22 Odor Sensor 24 Water quality meter 26 Operational equipment information acquisition unit 50 Processing section 54 Input section 56 Data storage unit 58 Data Calculation Unit 60 Storage section 100 aeration tank 301 Normal Cluster 302 Feed Addition Cluster 303 Sludge Extraction Cluster

Claims

1. An aeration tank management device provided in an aeration tank, A detection unit and a processing unit are provided, The detection unit an odor sensor that detects odors from the aeration tank; a water quality meter that measures the water quality of the treated liquid in the aeration tank, The processing unit an input unit into which a treatment to be performed on the aeration tank in response to the result detected by the odor sensor and / or the result measured by the water quality meter is input; a data storage unit in which the results and the treatment details are stored; an aeration tank management device comprising: a data calculation unit that reads out each of the results and the treatment details from the data storage unit, performs machine learning on each of the read out results and treatment details, and generates a clustering output model in which the odor detected by the odor sensor and / or the water quality measured by the water quality meter are used as input data and the results of clustering them are used as output data.

2. The detection unit further includes an operating equipment information acquisition unit that is provided in the aeration tank and acquires operating information and / or equipment information of the aeration tank, The data storage unit further stores the operation information and / or facility information acquired by the operation facility information acquisition unit, The aeration tank management device of claim 1, wherein the data calculation unit performs machine learning on at least one of the odor detected by the odor sensor and the water quality measured by the water quality meter, and at least one of the operating information and equipment information acquired by the operating equipment information acquisition unit, to generate a clustering output model.

3. A wastewater treatment device comprising the aeration tank management device according to claim 1 or 2 and an aeration tank.

4. a cluster grasping step of forming a normal cluster consisting of data output during normal operation regarding data on the odor from the aeration tank and / or the water quality of the treated liquid in the aeration tank; and a diagnostic step of detecting an abnormality based on the degree of cluster dissociation, which is the degree of dissociation from the normal cluster, of the newly output data.

5. 5. The aeration tank management method according to claim 4, wherein the step of diagnosing is carried out after a predetermined period of time has elapsed since the step of determining clusters.

6. 6. The aeration tank management method according to claim 4 or 5, wherein in the cluster identification step, an abnormal cluster is formed consisting of data output during abnormal operation regarding data on the odor from the aeration tank and / or the water quality of the treated liquid in the aeration tank.

7. 7. The aeration tank management method according to claim 6, wherein in the cluster grasping step, each abnormal cluster is previously labeled with a label corresponding to a classification of a known abnormal state.

8. The aeration tank management method according to claim 7, wherein measures are taken to resolve the abnormal state based on the classification of the abnormal state of the labeled abnormal cluster.

9. The aeration tank management method according to claim 8, wherein the content of the treatment is determined using a model in which data on the odor from the aeration tank and / or the water quality of the treated liquid in the aeration tank is input data and the content of the treatment is output data.

10. The aeration tank management method according to claim 6, wherein the abnormal cluster is formed based on the data output in a plurality of simulated abnormal states of the treated liquid in the aeration tank.

11. 5. The aeration tank management method according to claim 4, wherein the data output during normal operation in the cluster identification step includes at least one of the time it takes for wastewater to flow from its source into the aeration tank, and the season in which the data on the odor and / or the water quality of the treated liquid in the aeration tank was acquired.

Images