Situation diagnosis support system for water treatment base

Abstract

To reduce a time required for computation in situation diagnosis of a water treatment base.SOLUTION: A state diagnosis support system for a water treatment base having a plurality of water treatment units comprises: a storage unit which stores a hierarchical model, with respect to each of the plurality of water treatment units, including unit models representing relations between a plurality of situation nodes indicating situations of detection objects and observation information nodes by a Bayesian network and describing a relation between models of the next lower layer in each of a plurality of layers into which the water treatment base is divided; an arithmetic unit which uses probability distributions set to the respective nodes and inputted observation information to correct the probability distributions and selects a situation node where there is a high possibility that a situation occurs. The arithmetic unit includes a plurality of computers, and each of the computers is assigned to each model in each layer to perform computations in parallel.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a status diagnosis support system for a water treatment plant. [Background technology]

[0002] Conventionally, diagnostic systems have been known that detect abnormalities or signs of abnormalities in equipment in power plants, water treatment plants, etc., and instruct maintenance work required for the detected abnormalities. For example, Patent Document 1 describes a plant operation system that monitors signs of abnormalities in a plant based on the plant's operation history, inputs measurement parameters that are determined to have abnormality signs into an abnormality diagnostic system as abnormality sign data, and the abnormality diagnostic system compares the abnormality sign data with abnormality model patterns and uses a Bayesian network as a statistical model to identify the cause of the abnormality.

[0003] When applying Bayesian network technology to the diagnosis of water treatment facilities, unit models that model the causal relationships for each water treatment unit within the water treatment facility, such as coagulation tanks, reverse osmosis membrane separation devices, and electrodeionization devices, are created, and the unit models are then directly connected to each other to create a facility model.

[0004] However, when targeting large-scale water treatment facilities such as water supply sites where hundreds to thousands of water treatment units are installed, a huge facility model is created in which many unit models are connected in series, and calculations for many unit models are performed sequentially, which takes a long time and makes it difficult to perform regular diagnosis of the site situation (for example, once per hour). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-62729 Summary of the Invention [Problem to be solved by the invention]

[0006] An object of the present invention is to provide a water treatment plant status diagnosis support system that can shorten the time required for calculations to diagnose the status of a water treatment plant. [Means for solving the problem]

[0007] [1] A status diagnosis support system for a water treatment facility having multiple water treatment units, a storage unit that includes, for each of the plurality of water treatment units, a unit model that expresses the relationship between a plurality of situation nodes that represent the status of a detection target and an observation information node using a Bayesian network, the water treatment base is divided into a plurality of layers, and a hierarchical model is stored in which the relationship between models in the layer one layer below is described in each layer; a calculation unit that modifies the probability distribution set for each node using the input observation information and selects a situation node where a situation is likely to occur; Equipped with The calculation unit has a plurality of computers, each assigned to a model at each level, and performs calculations in parallel.

[0008] [2] The hierarchical model is Multiple unit models and A process model that describes the relationships between unit models, and A system model that describes the relationships between process models, and A plant model that describes the relationships between system models, and The situation diagnosis support system according to [1], [Effects of the Invention]

[0009] According to the present invention, the time required for calculations to diagnose the status of a water treatment plant can be reduced. [Brief explanation of the drawings]

[0010] [Figure 1]1 is a schematic configuration diagram of a water treatment plant status diagnosis support system according to an embodiment of the present invention. [Figure 2] FIG. 1 is a diagram illustrating an example of the configuration of a water treatment facility. [Figure 3] FIG. 1 is a diagram illustrating an example of a hierarchical model of a water treatment facility. DETAILED DESCRIPTION OF THE INVENTION

[0011] The present invention will be described in detail below with reference to the drawings.

[0012] As shown in FIG. 1, a situation diagnosis support system 2 according to this embodiment provides situation diagnosis support for a water treatment facility 1 in which a plurality of water treatment units are installed.

[0013] The water treatment facility 1 includes one or more water treatment plants. The water treatment plants include multiple water treatment systems. Each water treatment system includes multiple water treatment processes. Each water treatment process includes multiple water treatment systems. Each water treatment system includes multiple water treatment units.

[0014] The water treatment unit is equipped with various sensors that measure the amount of chemicals added, the amount of water supplied, the amount of treated water, the quality of treated water, etc. The water treatment unit is not particularly limited, but examples thereof include a coagulation tank, a turbidity membrane device, an MF membrane (microfiltration membrane) separation device, an RO membrane (reverse osmosis membrane) separation device, and an electrodeionization device.

[0015] For example, as shown in FIG. 2, a water treatment plant has a water treatment system A and a water treatment system B to which treated water from the water treatment system A is supplied.

[0016] Water treatment system A has water treatment processes A1, A2, and A3 connected in series. Water treatment system B has water treatment processes B1 and B2 connected in series.

[0017] Water treatment process A1 has water treatment units A1a and A1b connected in series, water treatment process A2 has water treatment units A2a and A2b connected in series, and water treatment process A3 has water treatment units A3a and A3b connected in series.

[0018] The water treatment process B1 has water treatment units B1a and B1b connected in series, and the water treatment process B2 has water treatment units B2a and B2b connected in series.

[0019] The numerous water treatment units installed at the water treatment facility 1 are classified (grouped) into multiple water treatment processes according to their functions. The multiple water treatment processes are classified into multiple water treatment systems. The multiple water treatment systems are classified into one or multiple water treatment plants.

[0020] The situation diagnosis support system 2 includes a calculation system 20, a storage unit 21, and a display unit 22. The calculation system 20 has a plurality of probability distribution calculation devices 201. The probability distribution calculation device 201 is a computer having a CPU, RAM, ROM, and the like.

[0021] The situation diagnosis support system 2 may include a communication unit that communicates with the sensors at the water treatment plant 1. The situation diagnosis support system 2 can acquire sensor measurement values ​​in real time using the communication unit.

[0022] The display unit 22 is a liquid crystal display or the like, and displays the results of calculations by the calculation system 20 .

[0023] The memory unit 21 stores a model (causal relationship model) that expresses, in a graph structure, the qualitative causal relationship between the situation at the water treatment plant 1 and the observation information. For example, using a Bayesian network, the relationship between the situation (condition) to be detected and the observation information is modeled for each water treatment unit. The observation information includes values ​​measured by sensors installed in the water treatment equipment, the results of analyzing sampled water with an analyzer, etc.

[0024] A Bayesian network expresses probability distributions using frequency distributions based on discretization of numerical values. For example, discretization is expressed using five state variables: LL, L, D, H, and HH. When expressing numerical information, LL has the engineering meaning of very low, L is low, D is design assumption, H is high, and HH is very high.

[0025] When expressing the occurrence of an event, LL has the engineering meaning of being extremely unlikely to have occurred compared to the design assumptions, L being unlikely to have occurred, D being an occurrence at the level of the design assumptions, H being highly likely to have occurred compared to the design assumptions, and HH being extremely likely to have occurred.

[0026] The above discretization is based on an engineering meaning, so there is an order, and it is possible to implement discretization of normalized numerical intervals. For example, the numerical range of 0.0 to 2.0 is taken as the normalized numerical range, and this numerical range is divided into five intervals, with LL corresponding to 0.0 to 0.4, L to 0.4 to 0.8, D to 0.8 to 1.2, H to 1.2 to 1.6, and HH to 1.6 to 2.0.

[0027] Input nodes, situation nodes, observation nodes, etc. are used as logical node components based on the discrete representation of the Bayesian network.

[0028] An input node represents the input information itself, is a node that does not receive input from other nodes, and defines the input boundary of the model.

[0029] A situation node is a node that represents the situation to be detected, and represents a situation that can occur in on-site equipment (a trouble event that should not occur in principle). When the values ​​of HH and H in the probability distribution are large, it indicates that the situation is likely to actually occur on-site.

[0030] An observation node is a node that represents observed information. Observation information includes measurements from sensors installed in the water treatment unit and the results of analyzing sampled water using an analytical device. Sensor measurement values ​​include those transmitted online and those visually confirmed on-site.

[0031] For example, in the observation node for treated water flow rate, when the values ​​of HH and H in the probability distribution are large, it means that there is a high probability that the treated water flow meter will indicate a high value. Furthermore, when actual observation information is obtained, a reasonable five-level state variable is determined in accordance with the threshold information set for each node, and a probability distribution is given in which the probability of that state variable is 100%.

[0032] In model construction, a probability distribution (called a "prior probability distribution") for the input node is set. In addition, a conditional probability distribution that expresses the probabilistic relationship with the input node is set for the situation node, observation node, etc.

[0033] The prior probability distribution and the conditional probability distribution may be values ​​based on expert experience and knowledge including subjective opinions, or may be values ​​based on frequency information from a large amount of actual data.

[0034] The probability distribution calculation device 201 retrieves the causal relationship model from the storage unit 21, and each time observation information is obtained, it modifies the probability distribution based on the prior probability distribution and the new observation information to obtain the posterior probability distribution. The probability distribution calculation device 201 also calculates the probability distribution of each node using Bayes' theorem (called the "posterior probability distribution" by inference) based on the probability distribution of the input information and the conditional probability distribution of each node. The probability distribution calculation device 201 also performs a process (called "Bayesian update") of reflecting the posterior probability distribution of the input information in the prior probability distribution.

[0035] If the causal relationship models (unit models) of the many water treatment units at the water treatment site 1 are connected in series (in sequence), one huge model will be calculated. Each model is the smallest calculation unit, and it is not possible to reduce the total calculation time by increasing the number of servers and performing parallel calculations, so as a result, many unit models are calculated in sequence, which takes a very long time for calculations.

[0036] Therefore, in this embodiment, the model of the water treatment facility 1 is expressed in multiple hierarchical layers. Specifically, it is divided into hierarchical layers of plant, system, process, and unit. For example, the water treatment plant shown in FIG. 2 is described in a hierarchical model as shown in FIG. 3. In FIG. 3, the A1a unit model represents a causal relationship model of the water treatment unit A1a. The A1 process model represents a causal relationship model of the water treatment process A1. The A system model represents a causal relationship model of the water treatment system A. The plant model represents a causal relationship model of the plant.

[0037] The relationships between unit models are described in the process model at the next higher level. Similarly, the relationships between processes are described in the system model at the next higher level. The relationships between systems are described in the plant model at the next higher level. Although not shown in the figure, if there are multiple plants, the relationships between the plants are described in the factory model at the next higher level. This ensures that the upstream and downstream relationships between unit models are not lost.

[0038] In the process model, nodes (events) of the unit model are placed and connected to each other. Similarly, in the system model, nodes (events) of the process model are placed and connected to each other. In the plant model, nodes (events) of the system model are placed and connected to each other.

[0039] The storage unit 21 stores data of such a hierarchical model.

[0040] In this hierarchical model, models (modules) are not connected to each other, and the probability calculation results of each model are not propagated. Therefore, it is preferable to provide the calculation system 20 with probability distribution calculation devices 201 (computers) equal to or greater than the number of models in the hierarchical model, and perform the probability calculations for each model in parallel.

[0041] For example, in the configuration shown in FIG. 10 unit models: A1a unit model, A1b unit model, ..., B2b unit model Five process models: A1 process model, A2 process model, A3 process model, B1 process model, and B2 process model. Two process models: A system model and B system model 1 plant model Since there are a total of 18 models, 18 probability distribution calculation devices 201 (computers) are provided in the calculation system 20, and one probability distribution calculation device 201 is assigned to each model, and calculations are performed in parallel.

[0042] As a result, each model becomes the smallest unit for probability calculation, regardless of the number of models in water treatment plant 1. Therefore, even if the scale of water treatment plant 1 is enormous and the number of water treatment units to be calculated increases, by adding the same number of computers in probability distribution calculation device 201 as the number of models, it becomes possible to execute calculations for site-wide situation diagnosis in the order of the time required for diagnosing a single model.

[0043] An example of derivation of causal relationships in the hierarchical model of Figure 3 is explained below. When a situation occurs in which the plant suffers a shortage of treated water, a probability calculation is performed for the plant model, which describes the situation nodes of the A system model and the B system model. This determines the situation of a shortage of treated water in the B system, which is likely to be the cause of the shortage in the plant's treated water. It also determines the situation of a shortage of treated water in the A system, which is likely to be the cause of the shortage in treated water in the B system. The relationship between the A system model and the B system model, which are independent of each other, can be shown in the higher-level plant model.

[0044] By performing probability calculations for the A system model, we can determine the situation of a shortage of treated water volume in the A3 process, which is likely to be the cause of a shortage of treated water volume in the A system.We can also determine the situation of a shortage of treated water volume in the A2 process, which is likely to be the cause of a shortage of treated water volume in the A3 process.We can also determine the situation of a shortage of treated water volume in the A1 process, which is likely to be the cause of a shortage of treated water volume in the A2 process.

[0045] By performing probability calculations for the A1 process model, we can determine situations where the A1b unit treated water volume is insufficient, which is likely to cause a shortage of treated water in the A1 process. We can also determine situations where the A1a unit treated water quality deviation is likely to cause a shortage of treated water in the A1b unit.

[0046] By performing probability calculations for the A1a unit model, we can determine situations in which the A1a unit chemical feed rate is insufficient, which is likely to cause deviations in the quality of the A1a unit treated water.We can also determine situations in which the A1a unit equipment malfunctions, which is likely to cause insufficient chemical feed rates in the A1a unit.

[0047] By performing similar calculations for all models, it becomes possible to derive causal relationships across hierarchical levels.

[0048] As described above, according to this embodiment, models are described in a hierarchical manner, and the relationships between models at the same level are described in the model at the level one level higher. Since models are not directly connected to each other, the probability calculations for each model can be executed in parallel on multiple computers, reducing the time required to calculate situation diagnosis and making it practical to execute scheduled situation diagnosis for large-scale bases.

[0049] In the above embodiment, an example has been described in which the elements constituting the water treatment facility are divided into a hierarchy of factory > plant > system > process > unit, but the granularity of the hierarchy is not limited to this.

[0050] The present invention is not limited to the above-described embodiments, and the components can be modified and embodied in practice without departing from the spirit of the invention. Furthermore, various inventions can be created by appropriately combining multiple components disclosed in the above-described embodiments. For example, some components may be omitted from all the components shown in the embodiments. Furthermore, components from different embodiments may be appropriately combined. [Explanation of symbols]

[0051] 1. Water treatment facility 2. Situation diagnosis support system 201 Probability distribution calculation device

Claims

1. A status diagnosis support system for a water treatment facility having a plurality of water treatment units, a storage unit that includes a unit model that expresses, for each of the plurality of water treatment units, a relationship between a plurality of situation nodes that represent the status of a detection target and an observation information node using a Bayesian network, the water treatment base being divided into a plurality of layers, and that stores a hierarchical model in which the relationship between models in the layer one level below is described in each layer; a calculation unit that modifies the probability distribution set for each node using the input observation information and selects a situation node where a situation is likely to occur; Equipped with The calculation unit has a plurality of computers, each assigned to a model at each level, and performs calculations in parallel.

2. The hierarchical model is Multiple unit models and A process model that describes the relationships between unit models, and A system model that describes the relationships between process models, and A plant model that describes the relationships between system models, and The situation diagnosis support system according to claim 1 , comprising:

Images