WO2014157750A1

WO2014157750A1 - Apparatus and method for providing causative factors for state of quality of effluent water from sewage treatment plant

Info

Publication number: WO2014157750A1
Application number: PCT/KR2013/002573
Authority: WO
Inventors: 김창원; 김예진; 김효수; 김민수; 이승훈; 박문화
Original assignee: 부산대학교 산학협력단
Priority date: 2013-03-27
Filing date: 2013-03-28
Publication date: 2014-10-02
Also published as: KR20140117959A

Abstract

With the quality of effluent water from a sewage treatment plant being the complex product of external causative factors, including changes in the flow rate and property of influent water and changes in environmental factors such as temperature and rain, and internal causative factors including an operator's operational measures to stably maintain the quality of effluent water in response to the external causative factors, the present invention relates to an apparatus and a method for providing causative factors for the state of the quality of effluent water from a sewage treatment plant, the apparatus and method being capable of: extracting types and patterns for a series of undesirable states of effluent water quality, in case of the occurrence thereof, from cumulative data which is a record of the causal relationship between previously measured external and internal causative factors in processes and the effluent water quality; identifying the causes of the present effluent water quality; and providing same to an operator. Provided according to the present invention is an apparatus for providing causative factors for the state of the quality of effluent water from a sewage treatment plant, the apparatus comprising: a data collection unit for collecting influent/effluent water quality data and process operating data which are required for the provision of causative factors for the state of the quality of effluent water from a sewage treatment plant; a data processing unit for grouping the data collected by the data collection unit into sets at predetermined time intervals and processing same; a processing state determination unit for determining the state of processing the effluent water quality data among the processed sets of data by using a discriminant function prepared in advance; a decision tree application unit for applying a decision tree algorithm prepared in advance to the determined processing state and searching for a relevant rule among the rules which constitute a decision tree; and a process state causing factor provision unit for determining the causative factors for the state of the quality of the effluent water from the sewage treatment plant on the basis of the process operating data and influent/effluent water quality data comprising the rule searched for by the decision tree application unit and for providing same to an operator, wherein the process operating data includes at least one among the amount of aeration, amount of sludge disposal, amount of return sludge, amount of chemical feed, precipitability, and concentration of suspended matter in a reactor of the sewage treatment plant.

Description

Apparatus and method for providing causal factors for the condition of outflow water quality of sewage treatment plant

The present invention relates to an apparatus and method for providing causal factors for the condition of effluent quality. In more detail, the effluent quality of the sewage treatment plant is an external cause including fluctuations in the flow rate and properties of the influent and changes in environmental factors such as temperature and rainfall, and in response to the driver's ability to maintain the effluent quality in a stable manner. A cumulative result of internal causal factors, including operational measures, that records the causal relationship between external and internal causal factors and runoff quality of a previously measured process in the event of any series of undesirable runoff conditions. The present invention relates to an apparatus and method for providing a causal factor for the condition of the sewage treatment plant effluent quality which can extract the type and pattern from the data to find out the cause of the effluent quality present and provide it to the driver.

Biological sewage treatment facility is a process to remove organic matter, nitrogen and phosphorus, which are included in influent, by activated sludge. The performance of the process is changed every hour. Corresponds to the concentration, it is determined by the operation for changing and maintaining the driving conditions efficiently taken by the driver. The process operator decides and executes these actions based on the know-how existing in the driver's empirical knowledge. In this case, the knowledge of which run variables are modified to keep the process stable in the case of any water quality situation is very subjective. Subsequently, even if it is intended to be extracted through interviews or literature studies, it is not easily objectified, and thus it is difficult to extract into visualized documents or intangible rules. The process of determining what effluent quality is in the process and the process of searching for the cause of the effluent are called sewage treatment process diagnosis.

In addition, the runoff quality of the process is the result of a combination of external factors such as fluctuations in the sewage treatment plant, external factors such as temperature and rainfall, and operational variables of the process. Uneasy. If it is possible to find out what specific non-process and internal causes are the causes of the runoff quality, the driver can modify the operating parameters and keep the process more stable. will be.

Diagnosis or operation of such sewage treatment plant processes has often required complex mathematical modeling. The purpose was to provide the process operator with model simulation how much of the operating variables to change. However, the data collection and coefficient optimization steps are necessary to optimize the coefficient values of the model until these models are able to better simulate the behavior of the target process. In providing change to the driver, the accuracy has changed, making it difficult to achieve reliability for field applications.

When a process operator discovers poor effluent quality, the operating effluent quality and process performance may be caused by extra-process factors that are difficult to control directly, or even when there is a situation where the influence of disturbance can be sufficiently reduced. You may want to know if it has been a major contributor to current runoff.

In this regard, there is an urgent need to develop a process operation support system and method that provides internal and external factors that contribute to the current state of the process.

The present invention has been made to solve such a problem, the present invention is the outflow water quality of the sewage treatment plant is to respond to external causes including variations in the flow rate and characteristics of the influent and changes in environmental factors such as temperature, rainfall This is a complex result of the internal cause factors including the driver's operation plan to maintain the effluent quality stably.In the event of any series of undesirable effluent conditions, the external and internal causes of the previously measured process Apparatus for providing a causal factor for the condition of the sewage treatment plant to extract the type and pattern from the cumulative data, which is a record of the causal relationship between the factor and the effluent quality, to identify the cause of the present effluent and provide it to the driver. The purpose is to provide a method.

In addition, the present invention allows the search process of the cause factor for the state of the runoff quality is automatically derived by a series of predefined algorithms to enable the objective cause driver to be derived and provided without depending on human resources. The purpose of the present invention is to provide an apparatus and method for providing a causal factor on the condition of effluent quality which can provide decision support.

According to the present invention, a data collection unit for collecting the inflow / outflow water quality data and process operation data required for the provision of the cause factor for the state of the outflow water quality of the sewage treatment plant; A data processing unit for processing the set data collected by the data collection unit at predetermined time intervals; A process state discriminating unit which determines a process state of the outflow water quality data among the processed data by using a discrimination function prepared in advance; A decision tree applying unit searching for a corresponding rule among rules constituting the decision tree by applying a decision tree algorithm prepared in advance to the determined process state; And determining the cause factor for the condition of the outflow water quality of the sewage treatment plant based on the process operation data and the inflow / outflow water quality data constituting the rule searched by the decision tree applying unit. Providing unit; Wherein, the process operation data is a sewage treatment plant outflow, characterized in that it comprises at least one of the aeration amount, sludge waste amount, sludge transport amount, chemical injection amount, sedimentation capacity and concentration of suspended solids in the reactor Provided is a causal factor providing device for the condition of water quality.

On the other hand, the process state discriminating unit is a data preparation unit for collecting the BOD, COD, SS, TN, TP data of the effluent quality of a certain section from the data collection unit or a separate database and set and process at a predetermined time interval; An effluent quality grouping unit for grouping the types of effluent quality by clustering all data hierarchically by clustering from the nearest data set using hierarchical clustering method for the data processed by the data preparation unit; And a discriminant function deriving unit for deriving a discriminant function to be used as a means for determining which type of new effluent quality data belongs to the type of the grouped effluent quality using Fisher's linear discriminant analysis. The process state discriminating unit may determine a process state of the outflow water quality data among the processed data using the discrimination function derived by the discriminating function extracting unit.

On the other hand, the decision tree is generated by the decision tree algorithm separately for the grouped effluent quality, the decision tree algorithm is characterized by the following equation.

[Revision 29.05.2013 under Rule 91]

Where Pi is the fraction of S belonging to class i, A is a variable, and Sv is a subset of S when variable A has the value v.

Meanwhile, the process state causal factor providing unit classifies the process operation data among the process operation data and the inflow / outflow water quality data constituting the rule searched by the decision tree applying unit as an internal process cause factor of the sewage treatment plant. The inflow / outflow water quality data is categorized as an external process cause factor of the sewage treatment plant to determine whether the cause factor for the condition of the outflow water quality of the sewage treatment plant is an internal process factor or an external process cause factor and is provided to the driver. It is characterized by giving.

In addition, according to the present invention, the data input step of receiving the inflow / outflow water quality data and process operation data required for the provision of the cause factor for the state of the outflow water quality from the sewage treatment plant; A data processing step of setting and processing the input data at predetermined time intervals; A process state determination step of determining a process state of the outflow water quality data among the processed data by using a determination function prepared in advance; A decision tree applying step of searching for a corresponding rule among rules constituting the decision tree by applying a decision tree algorithm prepared in advance to the determined process state; And a process state causation factor providing step for determining a causal factor for the state of the outflow water quality of the sewage treatment plant based on the process operation data and the inflow / outflow water quality data constituting the searched rule. The process operation data is characterized in that it comprises at least one or more of aeration amount, sludge waste amount, sludge conveyance amount, chemical injection amount, sedimentation capacity and the concentration of suspended solids in the reactor.

On the other hand, the process state determination step is to collect the data of the BOD, COD, SS, TN, TP of the effluent quality of a certain section from the data collection unit or a separate database, and set the processing at a predetermined time interval to process the processed data Using hierarchical clustering, we clustered from the nearest dataset to cluster all the data hierarchically to group the type of runoff quality. After deriving a discrimination function to be used as a means for determining whether to belong, using Fisher's linear discriminant analysis, and using the derived discriminant function to determine the process state of the effluent quality data of the processed data.

[Revision 29.05.2013 under Rule 91]

On the other hand, the process factor providing step is to classify the process operation data among the process operation data and the inflow / outflow water quality data constituting the rules found in the decision tree application step as an internal process factor of the sewage treatment plant The inflow / outflow water quality data is categorized as an external process cause factor of the sewage treatment plant to determine whether the cause factor for the condition of the outflow water quality of the sewage treatment plant is an internal process factor or an external process cause factor and is provided to the driver. It is characterized by giving.

The present invention can perform the operation support of the process by searching for the cause of the current effluent quality in the operation of the sewage treatment plant, and can provide a simple and key process cause factor in the process operation of the sewage treatment plant. The search process of causal factors is automatically derived by a series of predefined algorithms, which makes it possible to derive and provide objective causal factors without depending on human resources.

In addition, the present invention can distinguish whether it is greatly influenced by external factors of the process (changes in inflow quality and flow rate or changes in weather conditions) or internal factors of the process. Integrating sewage collection, sewage treatment plant operation, and discharge water quality management through sewage pipes can clarify whether to improve the operation of sewage treatment plants.

1 is a block diagram showing a causal factor providing apparatus for the state of the sewage treatment plant effluent quality according to an embodiment of the present invention.

Figure 2 is a flow chart showing a method for providing a causal factor for the state of the sewage treatment plant effluent quality in accordance with an embodiment of the present invention.

3 is a diagram illustrating an example of a decision tree composed of rules regarding cause factors derived in accordance with an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, in adding reference numerals to the components of each drawing, it should be noted that the same reference numerals are used as much as possible even if displayed on different drawings. In describing the present invention, when it is determined that the detailed description of the related well-known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

1 is a block diagram showing a causal factor providing apparatus for the state of the sewage treatment plant effluent quality according to an embodiment of the present invention, Figure 2 is a causal factor providing method for the state of effluent water quality according to an embodiment of the present invention. 3 is a diagram illustrating an example of a decision tree composed of rules regarding cause factors derived according to an embodiment of the present invention.

Referring to Figure 1, the causative factor providing apparatus 10 for the state of the sewage treatment plant effluent quality according to the present invention is the data collection unit 100, the data processing unit 200, the process state determination unit 300, decision making It includes a tree applying unit 400 and the process factor providing unit 500.

The data collection unit 100 collects inflow / outflow water quality data and process operation data required to provide a cause factor for the condition of the outflow water quality of the sewage treatment plant. The inflow / outflow water quality data includes inflow / outflow flow rates and inflow / outflow component concentrations (BOD ₅ , COD _Mn , SS, TN, TP, etc.). In more detail, the inflow / outflow water quality data is usually measured once a day and recorded, or measured several times a day, and as an average value of the result, all of the inflow and outflow water of the sewage treatment plant existed as one value per day. Recorded data of water quality items and inflows. The inflow / outflow water quality data may include weather factors such as temperature, water temperature, rainfall, sunshine, and humidity that may affect the treatment performance of the sewage treatment plant.

The process operation data includes at least one or more of aeration amount, sludge waste amount, sludge conveyance amount, chemical injection amount, sedimentation capacity, and suspended solids concentration in the reactor. In more detail, process operation data are typically modified once a day by the operator and applied to the process, and the results are typically the operating factors of the sewage treatment plant where the results are present as records, aeration volume, DO concentration and sludge in the reactor. Recorded results of operator factors including one or more of conveyed volume, sludge waste volume and sedimentation capacity (SVI, SV30).

The data processing unit 200 serves to process and set the data collected by the data collection unit 100 at predetermined time intervals. The data processing unit 200 averages the data collected at different time intervals to exist as one representative value per day or week, or processes the data by various statistical methods for selecting other representative values and collected at predetermined time intervals. This means that data is present in a set.

The process state discriminating unit 300 serves to determine the process state of the outflow water quality data among the processed data by using a determination function prepared in advance. That is, the process state discriminating unit 300 determines which state the effluent quality currently in use belongs to by utilizing the state type of the effluent quality derived by hierarchical cluster analysis and the discrimination function derived for each type. will be.

In more detail, the process state discrimination unit 300 includes a data preparation unit 310, a effluent quality grouping unit 320, and a determination function extracting unit 330. The data preparation unit 310 collects BOD, COD, SS, TN, and TP data of effluent quality from a data collection unit 100 or a separate database, and sets and processes the data at predetermined time intervals. do. The effluent quality grouping unit 320 clusters all the data hierarchically by clustering the data processed by the data preparing unit 310 from the closest data set using a hierarchical clustering method. It serves to group. The discrimination function derivation unit 330 serves to derive a discrimination function to be used as a means for determining what type of new effluent quality data belongs to the type of the grouped effluent quality using Fisher's linear discrimination analysis. do. Therefore, the process state discriminating unit 300 determines the process state of the outflow water quality data among the processed data by using the discrimination function derived by the discriminating function extracting unit 330.

The hierarchical clustering method mentioned above clusters the datasets having the closest distance or excludes the farthest datasets by performing the distance calculation between each dataset using the distance calculation method proposed by Ward. How to do it. By analyzing the dendrograms that group and group the datasets, the person who wants to cluster the data determines the number of clusters by identifying the dendrograms. Derived clusters are assigned to each type of water quality by cluster by identifying the characteristics of the datasets included in the cluster (concentration range, mean and standard deviation of each water quality item, etc.).

The decision tree application unit 400 serves to search for a corresponding rule among rules constituting the decision tree by applying a decision tree algorithm prepared in advance to the determined process state. That is, the decision tree application unit 400 is prepared in advance for each of the current outflow water quality currently found from the process state discriminating unit 300, the current inflow water quality and environmental factors, processes Apply operational data to determine which of the various rules make up a decision tree.

In more detail, the decision tree is generated by the decision tree algorithm separately for the grouped effluent quality, and the decision tree algorithm is derived by the following equation.

[Revision 29.05.2013 under Rule 91]

The decision tree rearranges the data again according to the type of the grouped water quality, and sets the input variables as influent water quality factors, influent flow rates, and operation history variables to act as classification criteria for rules to be generated in the future. It exists as a collection of rules that make the target variable one or more types of water quality. The rule generated from the decision tree is composed of some causes for one effluent quality, unlike the existing decision trees to find out whether the target is classified as good / bad / normal. There is a distinction as a rule to derive the number of cases, and it is desirable to have one decision tree for one or two combinations of the target runoff quality.

The process state causal factor providing unit 500 may determine the state of the outflow water quality of the sewage treatment plant based on the process operation data and the inflow / outflow water quality data constituting the rule searched by the decision tree application unit 400. The cause factor is judged and provided to the driver. Therefore, the process state cause factor providing unit 500 of the process operation data and the inflow / outflow water quality data constituting the rule searched by the decision tree applying unit 400 is the process operation data in the process of the sewage treatment plant. The inflow / outflow water quality data is classified as an out-of-process cause factor of the sewage treatment plant, and whether the cause factor for the outflow water quality of the sewage treatment plant is an in-process or out-of-process cause factor. Determine and provide to the driver.

Referring to Figure 2 describes the method for providing a causal factor for the condition of the effluent quality of the sewage treatment plant according to the present invention.

The first step is a data input step of receiving inflow / outflow water quality data and process operation data required for providing a causative factor for the condition of outflow water quality from the sewage treatment plant (S110).

The second step is a data processing step of processing by setting the input data at predetermined time intervals (S120).

The third step is a process state determination step of determining a process state of the outflow water quality data among the processed data by using a determination function prepared in advance (S130). The process state determination step (S130) is to collect the data of the BOD, COD, SS, TN, TP of the effluent quality of a certain section from the data collection unit 100 or a separate database to set and process at a predetermined time interval Using hierarchical clustering, the processed data is clustered from the nearest dataset to cluster all the data hierarchically to group the type of runoff quality, and the new runoff quality data for the type of grouped runoff quality. After deriving a discriminant function to be used as a means for determining which type belongs to using Fisher's linear discriminant analysis, the derived discriminant function is used to determine the process state of the effluent quality data among the processed data.

The fourth step is a decision tree application step of searching for a corresponding rule among rules constituting the decision tree by applying a decision tree algorithm prepared in advance to the determined process state (S140). The decision tree is generated by a decision tree algorithm separately for the grouped effluent quality, and the decision tree algorithm is derived by the following equation.

[Revision 29.05.2013 under Rule 91]

The fifth step is a process state causal factor providing step for determining the causal factors for the condition of the outflow water quality of the sewage treatment plant on the basis of the process operation data and the inflow / outflow water quality data constituting the searched rules ( S150). In the process state providing factor (S150), the process operation data among the process operation data and the inflow / outflow water quality data constituting the rule searched in the decision tree application step (S140) are internal causes of the sewage treatment plant. The inflow / outflow water quality data is classified as an external process cause factor of the sewage treatment plant to determine whether the cause factor for the effluent quality of the sewage treatment plant is an internal process factor or an external process cause factor. Give it to the driver.

Hereinafter, a method of providing a causal factor for the condition of the outflow water quality of the sewage treatment plant mentioned in the present invention will be described with reference to Examples.

In this embodiment, the inflow water quality of the measured value per day of the sewage treatment plant and the value of the effluent quality measured once per hour in order to derive the type of effluent quality of the sewage treatment plant and to search for and provide the causative factor, The data was processed by collecting the values of the operating variables that were measured and recorded once a day. This means that the collected effluent quality values exist as a daily average so that the collected data exist as one data set per day. 1 data set.

In order to extract the type of runoff quality that has existed in the sewage treatment plant, data from a certain period of the accumulated historical data were selected and prepared to exist as one data set per day, and types were grouped through hierarchical cluster analysis. . In principle, the data analyst judges the result of hierarchical cluster analysis to subjectively specify the number of groups. However, in the present invention, the number of groups between 5 and 7 is set as desirable so that the invention technology can be applied without subjective intervention of the data analyst. It could be done. [Table 1] below shows the average value of five water quality items in the sewage treatment plant selected because five clusters existed when the number of small clusters was determined when five to seven clusters were designated.

Table 1

Item cluster		BOD	COD	SS	TN	TP
5	One	1.7877	7.8802	3.0387	11.1056	1.1957
	2	2.9679	8.3069	4.0626	11.7548	1.4858
	3	2.5552	7.7299	4.2388	9.7709	1.2188
	4	1.6517	6.8542	3.2975	7.2424	0.9602
	5	1.3296	6.8516	3.3157	8.4893	1.5432

In addition, a series of characteristics can be assigned to each type of runoff based on the above average values, as shown in Table 2 below.

[Revision 29.05.2013 under Rule 91]
TABLE 2

For each of the types 1 to 5 derived as described above, it is possible to derive a discrimination function that can easily determine which type the new dataset belongs to. That is, in the present invention, a group classified by hierarchical cluster analysis is taken as an external standard, and a functional formula for each classified group, that is, a discriminating functional formula, is derived through Fisher's linear discriminant analysis.

Given a new data set, the derived discriminant function gives a conclusion that it is the group representing the largest value by comparing the magnitudes of the values computed by the discriminant function of each group, so that any of the existing classified groups Enable evaluation of belonging to groups. In addition, a cross-validation technique was used to evaluate the accuracy of group classification by the discriminant function.

The cross-validity technique is generally performed by the following method. Given that the total number of observations is N, N-1 observations except for the I (I = 1, 2, ..., N) observations are classified. In other words, it is used as an analysis sample to derive the discrimination function, and the I observation is used as the evaluation sample. By applying a classification rule derived using N-1 observations to the I observation, the group to which the I observation belongs belongs is determined, and this procedure is performed from the first observation to the N observation (that is, from I = 1 Up to N) is applied sequentially. The accuracy of group classification by the derived classification rule is then evaluated by comparing the group classified by these rules with the actual group given by the external criteria.

Therefore, using the discriminant function derived through Fisher's linear discriminant analysis, the process state of effluent quality data is determined from the processed data, and the decision tree algorithm is applied to the determined process state by applying the prepared decision tree algorithm. The results of constructing the decision tree by searching for the corresponding rule among the rules are shown well in FIG. 3. The decision tree of FIG. 3 constitutes a decision tree with respect to two or more types of water quality in order to enhance the classification performance, that is, as a result of generating the decision tree by tying together three types and five types having the most different group characteristics. In addition, it can be seen that each rule consists of variables that refer to the combined cause of the internal and external cause factors of the process. As a result, the factor of operation status of the process is "The SS removal rate and the TP removal rate of the treatment process are not good because of the SRT (the cause of the effluent water quality) which was kept short due to the shortening of the residence time in the reactor due to the rainfall. Process status and cause can be provided to the operator.

The apparatus and method for providing the causative agent for the condition of the effluent quality of the sewage treatment plant according to the present invention are compared to the process operation support system based on the existing complicated mathematical model, and to the operator, "The current effluent quality is good and / or normal. / Bad, good removal of nitrogen and phosphorus / normal / bad, the main cause is improperly maintained aeration tank DO concentration despite routine external disturbance factors / MLSS concentration in the reaction tank is too low / waste sludge amount is small Because it's too short / etc. " One sentence, which contains complex or clear information, is provided as a causal factor for the condition of the effluent quality so that the driver can support objective decision-making.

The above description is merely illustrative of the present invention, and those skilled in the art to which the present invention pertains may various modifications without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed herein are not intended to limit the present invention but to describe the present invention, and the spirit and scope of the present invention are not limited by these embodiments. It is intended that the scope of the invention be interpreted by the following claims, and that all descriptions within the scope equivalent thereto shall be construed as being included in the scope of the present invention.

The present invention extracts the type and pattern from the cumulative data, which is a record of the causal relationship between the external and internal causative factors of the existing process and the effluent quality, if any series of undesirable effluent conditions occur. It can be widely used in sewage treatment plants because it grasps the condition of outflow water in sewage treatment plants by providing the driver where the cause is.

Claims

A data collection unit for collecting inflow / outflow water quality data and process operation data necessary for providing a causative factor on the condition of outflow water quality of the sewage treatment plant;

A data processing unit for processing the set data collected by the data collection unit at predetermined time intervals;

A process state discriminating unit which determines a process state of the outflow water quality data among the processed data by using a discrimination function prepared in advance;

A decision tree applying unit searching for a corresponding rule among rules constituting the decision tree by applying a decision tree algorithm prepared in advance to the determined process state; And

Process condition cause factor which provides the driver by determining the cause factor for the condition of outflow water quality of sewage treatment plant based on process operation data and inflow / outflow water quality data constituting the rule searched by the decision tree application unit. Including;

The process operation data is a cause factor for the state of the sewage treatment plant effluent quality, characterized in that it comprises at least one or more of the aeration amount, sludge waste amount, sludge transport amount, chemical injection amount, sedimentation capacity and the concentration of suspended solids in the reactor. Providing device.
The method of claim 1,

The process state discriminating unit

A data preparation unit for collecting BOD, COD, SS, TN, and TP data of effluent quality from a data collection unit or a separate database, and processing the set data at predetermined time intervals;

An effluent quality grouping unit for grouping the types of effluent quality by clustering all data hierarchically by clustering from the nearest data set using hierarchical clustering method for the data processed by the data preparation unit; And

And a discriminant function deriving unit for deriving a discriminant function to be used as a means for determining which type of data of the new effluent quality belongs to the type of the grouped effluent quality using Fisher's linear discriminant analysis.

And said process state discriminating unit determines a process state of effluent quality data among said processed data using said discrimination function derived by said discriminating function extracting unit.
[Revision 29.05.2013 under Rule 91]

The method of claim 2,

The decision tree is generated by the decision tree algorithm separately for the grouped effluent quality, and the decision tree algorithm is a causal factor for the state of the sewage treatment plant effluent quality, which is derived by the following equation. Providing device.

Where Pi is the fraction of S belonging to class i, A is a variable, and Sv is a subset of S when variable A has the value v.
The method of claim 3, wherein

The process state cause factor providing unit classifies the process operation data among the process operation data and the inflow / outflow water quality data constituting the rule searched by the decision tree applying unit as an internal process cause factor of the sewage treatment plant, The inflow / outflow water quality data is classified as an external process cause factor of the sewage treatment plant to determine whether the cause factor for the condition of the outflow water quality of the sewage treatment plant is the intrinsic or non-process causal factor and provides it to the driver Apparatus for providing a causal factor for the condition of the effluent quality of the sewage treatment plant.
A data input step of receiving inflow / outflow water quality data and process operation data necessary for providing a cause factor on the condition of outflow water quality from the sewage treatment plant;

A data processing step of setting and processing the input data at predetermined time intervals;

A process state determination step of determining a process state of the outflow water quality data among the processed data by using a determination function prepared in advance;

A decision tree applying step of searching for a corresponding rule among rules constituting the decision tree by applying a decision tree algorithm prepared in advance to the determined process state; And

And a process state causation factor providing step for determining a causal factor on the condition of the outflow water quality of the sewage treatment plant based on the process operation data and the inflow / outflow water quality data constituting the searched rule.

The process operation data is a cause factor for the state of the sewage treatment plant effluent quality, characterized in that it comprises at least one of the aeration amount, sludge waste amount, sludge conveyance amount, chemical injection amount, sedimentation capacity and the concentration of suspended solids in the reactor. How to Provide.
The method of claim 5,

The process state determining step is to collect the data of the BOD, COD, SS, TN, TP of the effluent quality from a data collection unit or a separate database and set and process at a predetermined time interval for the processed data Using hierarchical clustering, cluster from the nearest dataset to cluster all the data hierarchically to group the type of runoff quality, and to determine what type of new runoff quality data belongs to the type of the grouped runoff quality. After deriving a discriminant function to be used as a means for discriminant using Fisher's linear discriminant analysis, using the derived discriminant function, the process state of the runoff quality data among the processed data is determined. How to provide causal factors for the condition of water quality.
[Revision 29.05.2013 under Rule 91]

The method of claim 6,

The decision tree is generated by a decision tree algorithm separately for the grouped effluent quality, and the decision tree algorithm is a causal factor for the condition of the effluent quality of the sewage treatment plant, which is derived by the following equation. How to Provide.

Where Pi is the fraction of S belonging to class i, A is a variable, and Sv is a subset of S when variable A has the value v.
The method of claim 7, wherein

The process state causal factor providing step is to classify the process operation data among the process operation data and the inflow / outflow water quality data constituting the rules found in the decision tree application step as the in-process cause factors of the sewage treatment plant. The inflow / outflow water quality data is classified as an external process cause factor of the sewage treatment plant to determine whether the cause factor for the condition of the outflow water quality of the sewage treatment plant is the intrinsic or non-process causal factor and provides it to the driver. Method for providing a causal factor for the condition of the effluent quality of the sewage treatment plant.