Control system, control method, and control program

Abstract

Provided is a process of acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant (1), classifying the state of the intermediate material or the product material by analyzing the image information via an AI analysis model (102a) that has been trained by performing machine learning by using an image group for learning, deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material, and instructing an operation manipulation of the field instrument based on the decided manipulation condition.

Description

CONTROL SYSTEM, CONTROL METHOD, AND CONTROL PROGRAM

[0001] The present invention relates to a control system, a control method, and a control program.

[0002] Conventionally, there is a known mineral processing process of selectively extracting, in a plant, valuable minerals, such as gold, silver, copper, nickel, from a mined ore. In the mineral processing process, for example, a method referred to as a froth flotation (hereinafter, appropriately referred to as "flotation process") of collecting a target mineral by putting fine ores that are obtained by performing size reduction and particle size separation into a water tank together with a chemical reagent referred to as a foaming reagent, and then, generating bubbles by agitating the pulverized and sized particles, and causing the mineral to clump with the bubbles and floating the clumped mineral is used.

[0003] In addition, a lot of impure substances are included in the ore, there is a demand to efficiently collect a high-purity object as much as possible. Accordingly, in the flotation process, there is a need to appropriately grasp the state of each of the bubbles including the size of the respective bubbles and the like, and maintain the bubbles in an appropriately state. In order to support this, there is a proposed technology of accurately measuring the size and the quantity of each of the bubbles by having a structure in which a camera that captures images of the states of the bubbles, and, even when the size of the bubbles becomes small, the bubbles hardly stay at the imaging positions of the camera (for example, see NPL 1).

[0004] [NPL 1] International Publication Pamphlet No. WO 2019 / 189117

[0005] However, in the conventional technology, there is a room to improve efficiency of the control of the mineral processing process in the plant.

[0006] In the mineral processing process, there is a need to appropriately maintain not only the size and the quantity of the bubbles but also an amount of discharge of the bubbles, and, in order to control this, there is a need to appropriately manipulate various elements, such as an injection rate of a chemical reagent, an inside tank level, and a supply rate of air.

[0007] In the present circumstances, image information on an image of a captured bubble is treated as important data that is used for control of an amount of discharge of the bubbles in real time. However, a process of deciding the manipulation condition related to the above described various elements or the like is performed by an operator on the basis of the image information, and so that this process is personalized work.

[0008] An object of the present invention is to provide a control system, a control method, and a control program capable of making mineral processing process control performed in a plant more efficient.

[0009] According to an aspect, a control system includes: an acquisition unit that acquires image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; a classification unit that classifies the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; a decision unit that decides a manipulation condition for a field instrument according to the state of the intermediate material or the product material that has been classified by the classification unit; and an instruction unit that instructs an operation manipulation of the field instrument based on the manipulation condition that has been decided by the decision unit.

[0010] According to an aspect, a control method that causes a computer to execute a process includes: acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; classifying the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material; and instructing an operation manipulation of the field instrument based on the decided manipulation condition.

[0011] According to an aspect, a control program that causes a computer to execute a process includes: acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; classifying the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material; and instructing an operation manipulation of the field instrument based on the decided manipulation condition.

[0012] According to an aspect of one embodiment, it is possible to provide a control system, a control method, and a control program capable of making mineral processing process control performed in a plant more efficient.

[0013] Fig. 1 is an explanation diagram of a mineral processing process.Fig. 2 is an explanation diagram illustrating, in outline, a control method according to a conventional technology.Fig. 3 is a magnified diagram illustrating an M1 portion illustrated in Fig. 2.Fig. 4 is a diagram illustrating one example of a manipulation condition for appropriately maintaining an amount of discharge of bubbles.Fig. 5 is an explanation diagram illustrating, in outline, a control method according to an embodiment.Fig. 6 is a diagram illustrating a classification example of a state of the bubbles.Fig. 7 is a diagram illustrating another manipulation example of each of the states of the bubbles.Fig. 8 is a diagram illustrating an example of the overall configuration of a plant control system according to the embodiment.Fig. 9 is a block diagram illustrating an example of a configuration of a concentration device according to the embodiment.Fig. 10 is a block diagram illustrating an example of a configuration of an AI analysis device according to the embodiment.Fig. 11 is an explanation diagram (No. 1) illustrating an AI analysis model according to the embodiment.Fig. 12 is an explanation diagram (No. 2) illustrating an AI analysis model according to the embodiment.Fig. 13 is a block diagram illustrating an example of a configuration of an advanced control device according to the embodiment.Fig. 14 is a diagram illustrating one example of preset information according to the embodiment.Fig. 15 is a flowchart illustrating the flow of the process performed by a control system according to the embodiment.Fig. 16 is a diagram of a hardware configuration illustrating one example of a computer that implements a function of the AI analysis device according to the embodiment.

[0014] Preferred embodiments of a control system, a control method, and a control program disclosed in the present invention will be described in detail below with reference to the accompanying drawings. Furthermore, the present invention is not limited to the embodiments. In addition, the same components are denoted by the same reference numerals and an overlapping description will be omitted. Each of the embodiments can be used in any appropriate combination as long as they do not conflict with each other.

[0015] In addition, in the following, it is assumed that the control system according to the embodiment is a plant control system 500 (see Fig. 5, etc.). The plant control system 500 includes an advanced control device 200 (see Fig. 5, etc.). The advanced control device 200 is one example of a control device that performs advanced control (advanced process control (APC)) of a plant.

[0016] Furthermore, in the following, a case in which the advanced control device 200 decides, in a concentration process included in a mineral processing process performed in the plant, on the basis of image information on a liquid surface outermost layer in a flotation cell 61 (see Fig. 2, etc.), each of manipulated variables MV, such as a chemical reagent injection rate, an air supply rate, an inside cell level, with respect to the flotation cell 61 will be used as a main example.

[0017] <Outline of control method according to the present embodiment> First, an outline of the control method according to the embodiment will be described with reference to Fig. 1 to Fig. 7. Fig. 1 is an explanation diagram illustrating a mineral processing process. Fig. 2 is an explanation diagram illustrating, in outline, a control method according to a conventional technology. Fig. 3 is a magnified diagram of an M1 portion illustrated in Fig. 2. Fig. 4 is a diagram illustrating one example of a manipulation condition for appropriately maintaining an amount of discharge of bubbles. Fig. 5 is an explanation diagram illustrating, in outline, the control method according to the embodiment. Fig. 6 is a diagram illustrating a classification example of the state of the bubbles. Fig. 7 is a diagram illustrating another manipulation example of each of the states of the bubbles.

[0018] As illustrated in Fig. 1, the mineral processing process includes a size reduction process (Step S1), a particle size separation process (Step S2), a concentration process (Step S3), and a dewatering process (Step S4).

[0019] The size reduction process is a process of finely pulverizing a mined ore so as to obtain a powder and granular material. The particle size separation process is a process of further finely screening the pulverized ore. The concentration process is a process of separating minerals from the ore by using the property of the minerals (a relative density, an electrical characteristic, a chemical property, etc.). The concentration process greatly affects the degree of purity and an amount of collection of the object.

[0020] The control method according to the present embodiment is applied at Step S3 that is enclosed by the broken lines in a rectangular shape. Moreover, in the present embodiment, it is assumed that a flotation process is used in the concentration process. In a case of the flotation process, the ore that has been subjected to the particle size separation process is converted to a pulp in the form of a liquid by adding a liquid in the concentration process and is then put into the flotation cell 61 after having injected a chemical reagent referred to as a foaming reagent that is mixed with a surfactant, fats and oils, or the like.

[0021] The dewatering process is a process of collecting froth layers that are formed on the outermost layer of the liquid surface (hereinafter, referred to as a "liquid surface outermost layer") in the flotation cell 61 in the concentration process, and extracting the target minerals by separating moisture and the chemical reagent from the froth layer.

[0022] The process performed at Step S3 will be more specifically described. As illustrated in Fig. 2, the concentration process is performed by a flotation machine 60. The flotation machine 60 includes the flotation cell 61, an agitation mechanism 62, a chemical reagent injection mechanism 63, an air supply mechanism 64, an inside cell level adjustment mechanism 65, a bubble discharge mechanism 66, and a camera 67.

[0023] The flotation cell 61 is a vessel to which the pulp injected with the chemical reagent by way of the particle size separation process is input. The flotation cell 61 is formed in, for example, a cylindrical shape with bottom. The agitation mechanism 62 is a mechanism for agitating an inside of the flotation cell 61. The agitation mechanism 62 agitates the inside of the flotation cell 61 by rotating about the central axis of, for example, the flotation cell 61.

[0024] The chemical reagent injection mechanism 63 is a mechanism for injecting the chemical reagent into the inside of the flotation cell 61. The air supply mechanism 64 is a mechanism for supplying air into the inside of the flotation cell 61. The inside cell level adjustment mechanism 65 is a mechanism for adjusting the inside cell level of the flotation cell 61. The inside cell level is the height of the liquid surface of the inside of the flotation cell 61. The chemical reagent injection mechanism 63, the air supply mechanism 64, and the inside cell level adjustment mechanism 65 are implemented by, for example, a valve or the like.

[0025] The bubble discharge mechanism 66 is a mechanism for discharging the froth layer that is formed on the liquid surface outermost layer in the flotation cell 61. If the agitation mechanism 62 agitates the inside of the flotation cell 61 while the air supply mechanism 64 supplying air, the pulp bubbles by the injected chemical reagent, and the bubbles rise to the liquid surface outermost layer in the flotation cell 61 and forms the froth layer.

[0026] At this time, as illustrated in Fig. 3, the hydrophobic property particles included in the pulp float together with the bubbles by clinging on to the bubbles, and the hydrophilic property particles are precipitated. In general, country rock of an ore usually has a hydrophilic property, and the mineral included in the ore usually has a hydrophobic property. The flotation method is a method of separating the country rock and the mineral from the ore by using this principle.

[0027] A description will be given here by referring back to Fig. 2. The camera 67 captures an image of the liquid surface outermost layer in the flotation cell 61. Furthermore, the camera 67 transmits the image information including the captured image data to an operator O in real time or at a constant period. Moreover, the operator O checks the image information by, for example, a Distributed Control Systems (DCS), the advanced control device 200, a monitor for viewing the image information, or the like that are included in the plant control system 500.

[0028] By the way, in the mineral processing process, there is a need to appropriately maintain not only the size and the quantity of the bubbles but also the amount of discharge of the bubbles, and, in order to control this, there is a need to appropriately manipulate various elements that are related to a movement of the flotation machine 60.

[0029] In also the control method according to the conventional technology, the image information on an image having captured therein bubbles is treated as important data that is used to control the state of the bubbles including the amount of discharge of the bubbles in real time. However, in the conventional technology, as illustrated in Fig. 2, a process of, for example, decision of the manipulation condition related to the above described various elements based on the image information is performed by the operator O, so that the decision process is performed as personalized work.

[0030] Furthermore, in a case of the flotation machine 60 illustrated in Fig. 2, examples of the manipulation condition decided by the operator O include, as illustrated in Fig. 4, a chemical reagent injection rate of a chemical reagent injected by the chemical reagent injection mechanism 63, an air supply rate of air supplied by the air supply mechanism 64, an inside cell level obtained by the inside cell level adjustment mechanism 65, and the like.

[0031] Accordingly, the control method according to the embodiment is adopted such that the plant control system 500 acquires image information on an image having captured therein a state of bubbles that are produced in the mineral processing process performed in the plant, classifies the state of the bubbles by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning, decides a manipulation condition for the flotation machine 60 according to the classified state of the bubbles, and instructs an operation manipulation of the flotation machine 60 on the basis of the decided manipulation condition.

[0032] Specifically, as illustrated in Fig. 5, in the control method according to the embodiment, an artificial intelligence (AI) analysis device 100 included in the plant control system 500 acquires the image information on the image having captured therein, by the camera 67 included in the flotation machine 60, the state of the bubbles contained in the flotation cell 61 in real time or at a constant period. Moreover, it can be said that the state of the bubbles mentioned here is one example of a state of an intermediate material or a product material that is produced in the mineral processing process.

[0033] Then, the AI analysis device 100 conducts AI analysis on the acquired image information. The AI analysis device 100 conducts the AI analysis by using an AI analysis model 102a (see Fig. 10) obtained by training an image group for learning having captured therein, in advance, the state of various bubbles that are present on the liquid surface outermost layer (froth layer) in, for example, the flotation cell 61.

[0034] The AI analysis model 102a is a deep neural network (DNN) model trained by, for example, an algorithm for deep learning. In the case where an image for classification having captured therein the state of the bubbles contained in the actual flotation cell 61 is input, the AI analysis model 102a outputs a class value that indicates the state of the bubbles. The class value is a classification value for classifying the state of the bubbles.

[0035] Then, the AI analysis device 100 transmits the state of the bubbles that has been classified by the AI analysis to the advanced control device 200. The advanced control device 200 decides the manipulation condition in accordance with the state of the bubbles on the basis of the state of the bubbles that has been classified by the AI analysis device 100. Then, the advanced control device 200 instructs the operation control of the flotation machine 60 based on the decided manipulation condition.

[0036] As illustrated in Fig. 6, the state of the bubbles are roughly classified into, for example, four states indicated by states #1 to #4. The state #1 indicates the state in which the size of each of the bubbles is maintained in an appropriate size, and the gap between the bubbles is moderately tight. In the case of the state #1, the froth layer is slowly discharged at an appropriately speed in also the bubble discharge mechanism 66.

[0037] The state #2 indicates some bubbles are disappearing in some part. In the case of the state #2, the state enters a state in which a part of the flotation cell 61 corresponding to, for example, an edge of the flotation cell 61 or the like becomes exposed.

[0038] The state #3 indicates the state in which some bubbles are converted into a slurry, and a gap between the bubbles is becoming filled with the slurry. In the case of the state #3, the slurry leaks out with respect to, for example, the bubble discharge mechanism 66 described above.

[0039] The state #4 indicates a state in which almost no bubble is found. In the case of the state #4, almost entire of the bottom surface of the bubble discharge mechanism 66 becomes exposed in, for example, the bubble discharge mechanism 66 described above.

[0040] Then, the advanced control device 200 according to the embodiment decides the manipulation condition with respect to each of the states #1 to #4 as illustrated in Fig. 7. If it is classified, by the AI analysis device 100, that the state of the bubbles is, for example, the state #1, the advanced control device 200 decides the manipulation condition such that all of the chemical reagent injection rate, the air supply rate, and the inside cell level are maintained.

[0041] Furthermore, if it is classified, by the AI analysis device 100, that the state of the bubbles is, for example, the state #2, the advanced control device 200 decides the manipulation condition such that the chemical reagent injection rate and the air supply rate are increased and the inside cell level is increased.

[0042] Furthermore, if it is classified, by the AI analysis device 100, that the state of the bubbles is, for example, the state #3, the advanced control device 200 decides the manipulation condition such that the chemical reagent injection rate is increased, the air supply rate is maintained or decreased, and the inside cell level is decreased.

[0043] Furthermore, if it is classified, by the AI analysis device 100, that the state of the bubbles is, for example, the state #4, the advanced control device 200 decides the manipulation condition such that the chemical reagent injection rate and the air supply rate are increased and the inside cell level is maintained.

[0044] Moreover, the states of the bubbles illustrated in Fig. 6 and Fig. 7 are provided as one example, and are not limited to the state of the bubbles classified by the advanced control device 200. Therefore, the state of the bubbles may be classified into three or less, or may be classified into five or more. Furthermore, the manipulation condition may be appropriately decided in accordance with the classification of these states of the bubbles.

[0045] Furthermore, in Fig. 7, an example in which three elements corresponding to the chemical reagent injection rate, the air supply rate, and the inside cell level are manipulated has been given, but these elements may be suitably used in combination as needed. For example, air agitated by the agitation mechanism 62 flows into the flotation machines 60, so that, there may be a case in which some of the flotation machines 60 does not need air supplied by the air supply mechanism 64. In this case, the advanced control device 200 decides the two elements corresponding to the chemical reagent injection rate and the inside cell level as the manipulation condition. Furthermore, the advanced control device 200 may also decide a manipulation condition that is different from the manipulation condition illustrated in, for example, Fig. 7 in terms of a balance with the other process data obtained in the plant.

[0046] In this way, in the control method according to the embodiment, the structure has been constituted such that the plant control system 500 acquires the image information on the image having captured therein the state of the bubbles that are produced in the mineral processing process performed in the plant, classifies the state of the bubbles by analyzing the image information via the AI analysis model 102a that has been trained by machine learning by using the image group for learning, decides the manipulation condition for the flotation machine 60 according to the classified state of the bubbles, and instructs the operation manipulation of the flotation machine 60 on the basis of the decided manipulation condition.

[0047] Therefore, with the control method according to the embodiment, it is possible to automatically decide the manipulation condition for the flotation machine 60 on the basis of the image information on the image having captured therein the state of the bubbles in real time without depending on personalized work. In other words, with the control method according to the embodiment, it is possible to improve efficiency of mineral processing process control that is performed in the plant. In the following, an example of a configuration of the plant control system 500 that includes the AI analysis device 100 and the advanced control device 200 according to the embodiment will be more specifically described.

[0048] <Example of overall configuration of plant control system 500> Fig. 8 is a diagram illustrating an example of the overall configuration of the plant control system 500 according to the embodiment. The plant control system 500 is a plant system that performs operation control and monitoring of a plant 1.

[0049] The plant control system 500 includes the plant 1, a DCS 9, the AI analysis device 100, the advanced control device 200, a protocol server 300, and an integrated server device 400.

[0050] The plant 1 is one example of various kinds of plants using minerals, petroleum, petrochemicals, chemical substances, gas, or the like. The plant 1 includes a factory or the like provided with various facilities for obtaining product materials. Examples of the product materials include mining raw materials, mining products, liquefied natural gas (LNG), a resin (plastic, nylon, etc.), and a chemical substance product, and the like. Examples of the facilities include a factory facility, a machine facility, a production facility, an electric-generating facility, a storage facility, and a facility in a wellhead for mining minerals, petroleum, natural gas or the like, and the like.

[0051] In the plant 1, various kinds of devices for producing the intermediate material and the product materials, various kinds of field instruments for acquiring information related to the state of the inside of the plant 1, and the like are included.

[0052] Each of the field instruments is an instrument that corresponds to, for example, a sensor or the like and that is set in various locations in the plant 1. The field instrument is broadly divided (classified) into, for example, a sensor instrument and a manipulation instrument. The sensor instrument is an instrument for acquiring (detecting, measuring, etc.), for example, a physical quantity. Examples of the sensor instrument includes a pressure sensor, a temperature sensor, a pH sensor, a speed sensor, an acceleration sensor, and the like. The manipulation instrument is an instrument for manipulating, for example, a physical quantity. Examples of the manipulation instrument include a valve, a pump, a fan, and the like that are included in various kinds of devices and that are used to produce an intermediate material and a product material, and is driven by a motor, an actuator, and the like.

[0053] In the present embodiment, the plant 1 is provided such that at least the mineral processing process is able to be performed. The plant 1 according to the embodiment includes a size reduction device 2, a particle size separation device 4, a concentration device 6, and a dewatering device 8.

[0054] The size reduction device 2 is a device that carries out a function of the above described size reduction process. The particle size separation device 4 is a device that carries out a function of the above described particle size separation process. The concentration device 6 is a device that carries out a function of the above described concentration process. The dewatering device 8 is a device that carries out a function of the above described dewatering process.

[0055] The concentration device 6 is provided with the plurality of flotation machines 60. Fig. 8 illustrates an example of a case in which the concentration device 6 includes the four flotation machines 60-1, 60-2, 60-3, and 60-4; however, the number of flotation machines 60 may be one, or may be five or more.

[0056] As illustrated in Fig. 8, the plurality of flotation machines 60 are connected in, for example, multiple stages. In the case where the multi-stage flotation machine 60 is used, the concentration device 6 is able to set individual object for each of the manipulations performed in the respective stages. For example, the concentration device 6 roughly separates and collects a group of mixtures each having a significantly different flotation property by using the flotation machine 60 that is located in the first stage. Furthermore, the concentration device 6 performs separation and collection, in the flotation machine 60 that is located in the second stage and the subsequent stages, by adding a chemical reagent that has an effect of increasing the difference of flotation property to a mixture similar to the flotation property. In this way, the concentration device 6 allows the flotation machine 60 that is located in each of the stages to perform froth flotation, so that it is possible to enhance the processing efficiency of the mineral processing process.

[0057] The DCS 9 includes a plurality of process controller and a plurality of manipulation monitoring devices. Each of the process controllers and each of the field instruments included in the plant 1 are connected so as to be able to communicate with each other via a field bus or the like. Operation control devices 10, 20, 40, and 80 are one example of the process controllers. Monitoring devices 30, 50, 70, and 90 are one example of the manipulation monitoring devices.

[0058] Each of the operation control devices 10, 20, 40, and 80 collects data from each of the field instruments, and notifies the advanced control device 200 of the collected data as process data. Furthermore, each of the operation control devices 10, 20, 40, and 80 performs process control on the basis of the instruction received from the advanced control device 200 based on the process data. Moreover, Fig. 8 illustrates an example of a configuration that enables, for convenience of description, the single advanced control device 200 to give an instruction to each of the operation control devices 10, 20, 40, and 80, but it may also be possible for each of the operation control devices 10, 20, 40, and 80 to perform the process control, on the basis of the instruction received from the plurality of advanced control devices 200, corresponding to the size reduction process, the particle size separation process, the concentration process, and the dewatering process.

[0059] The operation control device 10 performs the process control on the concentration device 6. The operation control device 20 performs the process control on the size reduction device 2. The operation control device 40 performs the process control on the particle size separation device 4. The operation control device 80 performs the process control on the dewatering device 8.

[0060] Each of the monitoring devices 30, 50, 70, and 90 is a device that is used to monitor or arbitrarily manipulate the plant 1. Each of the monitoring devices 30, 50, 70, and 90 causes the operator O to monitor a movement of each of the field instruments, a situation of various kinds of processes, and the like by displaying the various kinds of information that are related to the plant 1.

[0061] Furthermore, each of the monitoring devices 30, 50, 70, and 90 receives a desired instruction manipulation with respect to each of the field instruments, and notifies the operation control devices 10, 20, 40, and 80 of the received instruction manipulation. Each of the operation control devices 10, 20, 40, and 80 controls the respective field instruments on the basis of the received instruction manipulation.

[0062] The monitoring device 30 is a device that is used to monitor or arbitrarily manipulate the concentration device 6. The monitoring device 50 is a device that is used to monitor or arbitrarily manipulate the size reduction device 2. The monitoring device 70 is a device that is used to monitor or arbitrarily manipulate the particle size separation device 4. The monitoring device 90 is a device that is used to monitor or arbitrarily manipulate the dewatering device 8.

[0063] The operation control devices 10, 20, 40, and 80, the monitoring devices 30, 50, 70, and 90, and the protocol server 300 are connected so as to be able to communicate with each other via a network N1. The AI analysis device 100, the advanced control device 200, the protocol server 300, and the integrated server device 400 are connected so as to be able to communicate with each other via a network N2.

[0064] For the network N1, for example, a control bus that is constituted for a special purpose is used. The data transmitted by the network N1 is data that is related to the plant 1 and includes data that is used to perform control or the like of the plant 1. Real time control is included in this control. Accordingly, the network N1 is constituted, for example, in a dual manner as a duplex network in terms of ensuring reliability of data transmission.

[0065] In a case of the duplex network, the network N1 may transmits the same data in parallel by using two communication paths. In this case, even if a failure occurs in one of the communication paths constituted in a dual manner, the network N1 is able to maintain the transmission (sending and receiving) of the data by using the other communication path. As an example of this type of the network N1, it is possible to use a Vnet / IP (registered trademark) or the like. Furthermore, for the network N2, it is possible to use a local area network (LAN) or the like. Furthermore, regarding the network N2, it is also possible to constitute the network N2 in a dual manner as a duplex network in terms of ensuring reliability of data transmission or the like.

[0066] Moreover, Fig. 8 does not always illustrate a physical configuration. Therefore, the network topologies in the plant control system 500 are not limited to the bus type illustrated in Fig. 8.

[0067] The AI analysis device 100 acquires image information received from each of the cameras 67 that are included in the concentration device 6 by way of the operation control device 10 and the protocol server 300. Furthermore, the AI analysis device 100 conducts AI analysis on the acquired image information, and classifies the state of the bubbles that are present in each of the flotation machines 60. Furthermore, the AI analysis device 100 transmits the classified state of the bubbles to the advanced control device 200.

[0068] The advanced control device 200 acquires the process data that has been sent from each of the operation control devices 10, 20, 40, and 80 by way of the protocol server 300. The advanced control device 200 acquires each of the flotation speeds or the like that has been sent from the respective flotation machines 60 included in the concentration device 6, as the process data that has been sent from the operation control device 10.

[0069] Furthermore, the advanced control device 200 decides the manipulation condition according to the state of the bubbles that has been classified by the AI analysis device 100. Furthermore, the advanced control device 200 instructs the operation control of the concentration device 6 on the basis of the decided manipulation condition. In other words, the advanced control device 200 notifies, by way of the protocol server 300, the operation control device 10 of the above described manipulated variable MV of the various elements indicating the decided manipulation condition.

[0070] The protocol server 300 is a device that carries out a function of, for example, conversion as well as reception and delivery of a format of the data between the network N1 and the network N2.

[0071] For example, a protocol (the data format, the communication standard, etc.) used in the network N1 is different from a protocol used in the network N2, and the protocol server 300 performs conversion or the like of the protocols for the data between these networks. The data format in the network N1 may conform to the original standard of each of the process controllers that perform operation control of the plant 1. The data format in the network N2 may conform to the protocol used in the AI analysis device 100, the advanced control device 200, and the integrated server device 400, that is, for example, a protocol used in Open Platform Communications (OPC), or the like. The protocol server 300 may be an OPC server.

[0072] The integrated server device 400 is a device that integrally manages the plant control system 500. The integrated server device 400 collects, by way of the network N1, the protocol server 300, and the network N2, an alarm notified from, for example, each of the process controllers, and then, displays an alarm message and a state of the alarm. The operator O is able to check the entire situation of the plant control system 500 including the alarm message and the current alarm state by way of the display content that is displayed on the integrated server device 400.

[0073] <Example of configuration of concentration device 6> Next, Fig. 9 is a block diagram illustrating an example of a configuration of the concentration device 6 according to the embodiment. Moreover, in Fig. 9, and also in Fig. 10 and Fig. 13 that will be described later, only the components that are needed for description of the present embodiment are illustrated, and illustrations of general components are omitted.

[0074] Furthermore, in descriptions using Fig. 9, Fig. 10, and Fig. 13, the descriptions of each of the already described components will be appropriately simplified or omitted.

[0075] As illustrated in Fig. 9, the concentration device 6 includes the flotation machine 60. The concentration device 6 includes, for example, the four flotation machines 60-1, 60-2, 60-3, and 60-4.

[0076] The flotation machine 60 includes the flotation cell 61, the agitation mechanism 62, the chemical reagent injection mechanism 63, the air supply mechanism 64, the inside cell level adjustment mechanism 65, the bubble discharge mechanism 66, the camera 67, a communication unit 68, and various kinds of sensors 69.

[0077] The flotation cell 61, the agitation mechanism 62, the chemical reagent injection mechanism 63, the air supply mechanism 64, the inside cell level adjustment mechanism 65, the bubble discharge mechanism 66, and the camera 67 have already been described; therefore, the descriptions thereof will be omitted.

[0078] The communication unit 68 connects the concentration device 6 and the operation control device 10 so as to be able to communicate with each other. The communication unit 68 is implemented by a network interface card (NIC), or the like. The communication unit 68 transmits, for example, the image information on the image that has been captured by the camera 67 and the process data that has been acquired by the various kinds of sensors 69 toward the operation control device 10. Furthermore, the communication unit 68 receives a manipulation signal that indicates the manipulated variable MV that is transmitted from, for example, the advanced control device 200 by way of the operation control device 10. The concentration device 6 causes the chemical reagent injection mechanism 63, the air supply mechanism 64, and the inside cell level adjustment mechanism 65 to operate on the basis of the received manipulation signal.

[0079] <Example of configuration of AI analysis device 100> Next, Fig. 10 is a block diagram illustrating an example of a configuration of the AI analysis device 100 according to the embodiment. As illustrated in Fig. 10, the AI analysis device 100 includes a communication unit 101, a storage unit 102, and a processing unit 103.

[0080] The communication unit 101 connects the AI analysis device 100, the advanced control device 200, and the DCS 9 so as to be able to communicate with each other by way of the network N2, the protocol server 300, and the network N1. The communication unit 101 is implemented by a NIC or the like.

[0081] The storage unit 102 is implemented by a storage device, such as a random access memory (RAM), a flash memory, or a hard disk drive (HDD). The storage unit 102 stores therein a control program according to the embodiment executed by the processing unit 103. Furthermore, the storage unit 102 stores therein various kinds of information that are used in the information processing that is performed by the processing unit 103.

[0082] In the example illustrated in Fig. 10, the storage unit 102 stores therein the AI analysis model 102a. The AI analysis model 102a is the analysis model that has been subjected to AI learning such that a class value indicating the state of the bubbles is output when an image for classification, as described above, that has been obtained by capturing the state of the bubbles that are present in the froth layer in the actual flotation cell 61 is input. Moreover, in a description below, AI learning may also be reread as "machine learning".

[0083] Fig. 11 is an explanation diagram (No. 1) of the AI analysis model 102a according to the embodiment. Furthermore, Fig. 12 is an explanation diagram (No. 2) of the AI analysis model 102a according to the embodiment. As illustrated in Fig. 11, the AI analysis model 102a is generated by performing AI learning by a learning unit 103a, which will be described later, by using the image group for learning of the image for learning having captured therein various states of the bubbles that are present in the froth layer in the flotation cell 61.

[0084] The learning unit 103a causes the AI analysis model 102a to train various kinds of feature amounts each of which indicates the state of the bubbles, such as the size of the bubbles, the shape of the bubbles, a distribution of the bubbles, and a distribution of the slurries, that are included in the image group for learning.

[0085] Furthermore, as illustrated in Fig. 12, the AI analysis model 102a outputs a class value that indicates the state of the bubbles to a classification unit 103c when the image for classification having captured therein the state of the bubbles that are present in the froth layer contained in the actual flotation cell 61 is input by the classification unit 103c that will be described later.

[0086] Moreover, the various kinds of feature amounts each of which indicates the state of the bubbles trained by the AI analysis model 102a may include, for example, a time series variation in the state of the bubbles, an appearance of a part of the flotation cell 61 that is a portion exposed described above, and the like. Accordingly, each of the cameras 67 that captures the image group for learning and the image for classification may be installed any place as long as each of the cameras 67 is able to capture the state of the bubbles that are obtained after the bubbles are formed as the froth layer.

[0087] A description will be given here by referring back to Fig. 10. The processing unit 103 corresponds to what is called a processor. The processing unit 103 is implemented by a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), or the like.

[0088] The processing unit 103 reads the control program according to the embodiment that is stored in the storage unit 102 and then executes the read control program by using the RAM as a work area. The processing unit 103 may also be implemented by an integrated circuit, such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

[0089] The processing unit 103 includes the learning unit 103a, an acquisition unit 103b, the classification unit 103c, and a transmission unit 103d, and implements or executes the function and the operation of the information processing that will be described below. Moreover, the internal configuration of the processing unit 103 is not limited to the configuration illustrated in Fig. 10, but another configuration may be used as long as the configuration in which the information processing that will be described below is able to be performed is used. Furthermore, the connection relation among each of the processing units included in the processing unit 103 is not limited to the connection relation illustrated in Fig. 10, but another connection relation may also be used.

[0090] The learning unit 103a generates the AI analysis model 102a by performing AI learning on the basis of the above described image group for learning. The image group for learning may also be an image group of images that have been captured by the camera 67 and that have been acquired by the acquisition unit 103b by way of the communication unit 101, or an image group of images that have been collected or generated by another device and that have been acquired by the acquisition unit 103b by way of the communication unit 101, a recording medium, or the like.

[0091] The acquisition unit 103b acquires, by way of the communication unit 101 as needed, image information sent from the camera 67.

[0092] The classification unit 103c classifies the state of the bubbles contained in the image for classification by inputting the image information that functions as the image for classification and that has been acquired by the acquisition unit 103b to the AI analysis model 102a and receiving an output that indicates the state of the bubbles from the AI analysis model 102a. Furthermore, the classification unit 103c notifies the transmission unit 103d of the classified state of the bubbles.

[0093] The transmission unit 103d transmits the classification result that has been classified by the classification unit 103c to the advanced control device 200 via the communication unit 101.

[0094] <Example of configuration of advanced control device 200> Next, Fig. 13 is a block diagram illustrating an example of a configuration of the advanced control device 200 according to the embodiment. As illustrated in Fig. 13, the advanced control device 200 includes a communication unit 201, a storage unit 202, and a processing unit 203.

[0095] The communication unit 201 connects the AI analysis device 100, the advanced control device 200, and the DCS 9 so as to be able to communicate with each other by way of the network N2, the protocol server 300, and the network N1. The communication unit 201 is implemented by a NIC, or the like.

[0096] The storage unit 202 is implemented by a storage device, such as a RAM, a flash memory, or an HDD. The storage unit 202 stores therein the control program according to the embodiment executed by the processing unit 203. Furthermore, the storage unit 202 stores therein various kinds of information that are used in the information processing executed by the processing unit 103.

[0097] In the example illustrated in Fig. 13, the storage unit 202 stores therein preset information 202a. The preset information 202a is information in which the manipulation condition according to the state of the bubbles that is classified by the AI analysis device 100 has been set in advance. Fig. 14 is a diagram illustrating one example of the preset information 202a.

[0098] As illustrated in Fig. 14, in the preset information 202a, each of the manipulated variables MV of, for example, the chemical reagent injection rate, the air supply rate, the inside cell level, and the like according to each of the states #1 to #4, … of the bubbles is set in advance. In the example illustrated in Fig. 14, if the image for classification is classified to be in the state #1 by the AI analysis device 100, the manipulated variables MV of the chemical reagent injection rate, the air supply rate, and the inside cell level are decided to be manipulated variables MV11, MV12, and MV13, respectively. Similarly, if the image for classification is classified to be in the state #2 by the AI analysis device 100, the manipulated variables MV of the chemical reagent injection rate, the air supply rate, and the inside cell level are decided to be manipulated variables MV21, MV22, and MV23, respectively.

[0099] Similarly, if the image for classification is classified to be in the state #3 by the AI analysis device 100, the manipulated variables MV of the chemical reagent injection rate, the air supply rate, and the inside cell level are decided to be manipulated variables MV31, MV32, and MV33, respectively. Similarly, if the image for classification is classified to be in the state #4 by the AI analysis device 100, the manipulated variables MV of the chemical reagent injection rate, the air supply rate, and the inside cell level are decided to be manipulated variables MV41, MV42, and MV43, respectively.

[0100] A description will be given here by referring back to Fig. 13. The processing unit 203 corresponds to what is called a processor. The processing unit 203 is implemented by a CPU, an MPU, a GPU, or the like.

[0101] The processing unit 203 reads the control program according to the embodiment that is stored in the storage unit 202 and then executes the read control program by using the RAM as a work area. The processing unit 203 may also be implemented by an integrated circuit, such as an ASIC, an FPGA, or the like.

[0102] The processing unit 203 includes an acquisition unit 203a, a decision unit 203b, and an instruction unit 203c, and implements or executes the function and the operation of the information processing that will be described below. Moreover, the internal configuration of the processing unit 203 is not limited to the configuration illustrated in Fig. 13, but another configuration may be used as long as the configuration in which the information processing that will be described below is able to be performed is used. Furthermore, the connection relation among each of the processing units included in the processing unit 203 is not limited to the connection relation illustrated in Fig. 13, but another connection relation may also be used.

[0103] The acquisition unit 203a acquires the process data sent from each of the process controllers, for example, from the operation control device 10 by way of the communication unit 201. Furthermore, the acquisition unit 203a acquires, via the communication unit 201, the classification result sent from the AI analysis device 100.

[0104] The decision unit 203b extracts the target manipulation condition from the preset information 202a on the basis of the state of the bubbles that has been classified by the AI analysis device 100, and decides the manipulation condition that is to be instructed to the operation control device 10. Moreover, the decision unit 203b may also decide not only the state of the bubbles but also the manipulation condition in accordance with various kinds of process data sent from the operation control device 10 acquired by the acquisition unit 103b.

[0105] In this case, the manipulation condition related to each of the states of the bubbles for each flotation speed is set to, for example, the preset information 202a in advance. Then, the decision unit 203b decides the manipulation condition from the preset information 202a on the basis of a combination of the flotation speed that has been acquired as the process data and the state of the bubbles that has been classified by the AI analysis device 100.

[0106] The instruction unit 203c transmits, via the communication unit 201, each of the manipulated variables MV of the manipulation condition decided by the decision unit 203b toward the operation control device 10, and instructs the operation control device 10 to perform the operation control of the concentration device 6 on the basis of each of the manipulated variables MV.

[0107] <Flow of process performed by plant control system 500> In the following, the flow of the process performed by the plant control system 500 will be described with reference to Fig. 15. Fig. 15 is a flowchart illustrating the flow of the process performed by the plant control system 500 according to the embodiment.

[0108] Moreover, Fig. 15 illustrates the flow of the process performed in the case where the AI analysis model 102a has been trained and is stored in advance in the storage unit 102 included in the AI analysis device 100. Furthermore, the flow of the process illustrated in Fig. 15 is repeated during a period of time in which the mineral processing process is performed.

[0109] As illustrated in Fig. 15, first, the acquisition unit 103b included in the AI analysis device 100 acquires the image information on the image having captured therein the state of the bubbles contained in the flotation cell 61 (Step S101). Then, the classification unit 103c inputs the image information that has been acquired by the acquisition unit 103b to the AI analysis model 102a, and conducts the AI analysis on the image information (Step S102).

[0110] Then, the classification unit 103c classifies the state of the bubbles on the basis of the output from the AI analysis model 102a (Step S103).

[0111] Subsequently, the decision unit 203b included in the advanced control device 200 decides, on the basis of the preset information 202a, the manipulation condition according to the state that has been classified by the classification unit 103c included in the AI analysis device 100 (Step S104). Then, the instruction unit 203c included in the advanced control device 200 instructs the operation control device 10 on the operation control of the concentration device 6 based on the decided manipulation condition (Step S105). After the processes described above, a single cycle of the flow of the process illustrated in Fig. 15 has been completed.

[0112] <Effects> As described above, the plant control system 500 (corresponding to one example of a "control system") according to the embodiment includes the acquisition unit 103b that acquires image information on an image having captured therein a state of an intermediate material or a product material produced in the mineral processing process performed in the plant 1, the classification unit 103c that classifies the state of the intermediate material or the product material by analyzing the image information via the AI analysis model 102a (corresponding to one example of the "analysis model") trained by performing machine learning by using the image group for learning, the decision unit 203b that decides the manipulation condition for a field instrument according to the state of the intermediate material or the product material that has been classified by the classification unit 103c, and the instruction unit 203c that instructs an operation manipulation of the field instrument based on the manipulation condition that has been decided by the decision unit 203b. Therefore, with the plant control system 500 according to the embodiment, it is possible to conduct analysis of the image information via the AI analysis model 102a that has been trained by performing machine learning and it is possible to decide the manipulation condition for the field instrument on the basis of the analysis result, thereby leading to automation of the mineral processing process control. In other words, with the plant control system 500 according to the embodiment, it is possible to improve efficiency of mineral processing process control that is performed in the plant 1. Furthermore, with the plant control system 500 according to the embodiment, it is possible to improve a yield and the degree of purity of the object in the mineral processing process, and it is thus possible to provide a clear benefit to a customer.

[0113] Furthermore, the acquisition unit 103b acquires the image information on the image having captured therein the state of the bubbles that are produced in the froth flotation performed in the mineral processing process. The instruction unit 203c instructs, on the basis of the manipulation condition that has been decided from the image information on the image having captured therein the state of the bubbles, the operation manipulation of the field instrument that is included in the concentration device 6 that performs the froth flotation. Therefore, with the plant control system 500 according to the embodiment, it is possible to perform automatic control of the field instrument, such as the chemical reagent injection mechanism 63, the air supply mechanism 64, and the inside cell level adjustment mechanism 65, that is included in the concentration device 6 that performs the froth flotation in the mineral processing process.

[0114] Furthermore, the concentration device 6 includes the flotation cell 61 to which a pulp of an ore that has been subjected to particle size separation is input. The acquisition unit 103b acquires, in the froth flotation, the image information on the image having captured therein the state of the bubbles that are present on the froth layer that is formed in the liquid surface outermost layer in the flotation cell 61. Therefore, with the plant control system 500 according to the embodiment, it is possible to simply and easily improve automation of the mineral processing process control on the basis of the state of the bubbles that are present on the froth layer that is able to be easily captured as an image.

[0115] Furthermore, the AI analysis model 102a is trained such that, when the image information that is used for classification has been input, a classification value that indicates the state of the bubbles is output. The decision unit 203b decides the manipulation condition on the basis of the preset information 202a in which the manipulation condition has been set to each of the classification values in advance. Therefore, with the plant control system 500 according to the embodiment, it is possible to perform the automatic mineral processing process control in the froth flotation without depending on the conventional personalized method.

[0116] Furthermore, the decision unit 203b decides the manipulation condition about the manipulated variable MV corresponding to at least one of the injection rate of the chemical reagent to be injected into the flotation cell 61 in the froth flotation, the supply rate of sir to be supplied to the flotation cell 61 in the froth flotation, and the inside cell level that is the height of the liquid surface in the flotation cell 61 indicated in the froth flotation. Therefore, with the plant control system 500 according to the embodiment, it is possible to automatically decide the manipulation condition such that the state of the bubbles including the amount of discharge of the bubbles is able to be appropriately maintained by the manipulated variable MV corresponding to at least one of the injection rate of the chemical reagent, the supply rate of the air, and the inside cell level.

[0117] Furthermore, when it is classified that, regarding the state of the bubbles, the size of the bubbles is smaller than the size of the bubbles that are in an appropriate state, the decision unit 203b decides the manipulation condition such that at least the injection rate of the chemical reagent is increased. Therefore, with the plant control system 500 according to the embodiment, in the case of the state #2 described above, it is possible to automatically decide the manipulation condition such that the state of the bubbles including the amount of discharge of the bubbles is able to be appropriately maintained.

[0118] Furthermore, when, regarding the state of the bubbles, some of the bubbles are converted into a slurry, the decision unit 203b decides the manipulation condition such that at least the supply rate of air is maintained or increased. Therefore, with the plant control system 500 according to the embodiment, in the case of the state #3 described above, it is possible to automatically decide the manipulation condition such that the state of the bubbles including the amount of discharge of the bubbles is able to be appropriately maintained.

[0119] Furthermore, when it is classified that, regarding the state of the bubbles, almost no bubble is found, the decision unit 203b decides the manipulation condition such that the injection rate of the chemical reagent and the supply rate of air are increased, and the inside cell level is maintained. Therefore, with the plant control system 500 according to the embodiment, in the case of the state #4 described above, it is possible to automatically decide the manipulation condition such that the state of the bubbles including the amount of discharge of the bubbles is able to be appropriately maintained.

[0120] Furthermore, the control method according to the embodiment causes a computer that is the AI analysis device 100 and / or the advanced control device 200 to execute a process of acquiring the image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in the plant 1, classifying the state of the intermediate material or the product material by analyzing the image information via the AI analysis model 102a that has been trained by performing machine learning by using the image group for learning, deciding a manipulation condition for a field instrument according to the state of the classified intermediate material or the classified product material, and instructing an operation manipulation of the field instrument based on the decided manipulation condition. Therefore, with the control method according to the embodiment, it is possible to conduct the image information via the AI analysis model 102a that has been trained by performing the machine learning, and it is possible to decide the manipulation condition for the field instrument on the basis of the obtained analysis result, thereby leading to automation of the mineral processing process control. In other words, with the control method according to the embodiment, it is possible to improve efficiency of mineral processing process control that is performed in the plant 1. Furthermore, with the control method according to the embodiment, it is possible to improve a yield and the degree of purity of the object in the mineral processing process, and it is thus possible to provide a clear benefit to a customer.

[0121] Furthermore, the control program according to the embodiment causes a computer that is the AI analysis device 100 and / or the advanced control device 200 to execute a process of acquiring the image information on an image having captured therein a state of an intermediate material or a product material that is produced in the mineral processing process performed in the plant 1, classifying the state of the intermediate material or the product material by analyzing the image information via the AI analysis model 102a that has been trained by performing machine learning by using the image group for learning, deciding a manipulation condition for a field instrument according to the state of the classified intermediate material or the classified product material, and instructing an operation manipulation of the field instrument based on the decided manipulation condition. Therefore, with the control program according to the embodiment, it is possible to conduct the image information via the AI analysis model 102a that has been trained by performing the machine learning, and decide the manipulation condition for the field instrument on the basis of the obtained analysis result, thereby leading to automation of the mineral processing process control. In other words, with the control program according to the embodiment, it is possible to improve efficiency of mineral processing process control that is performed in the plant 1. Furthermore, with the control program according to the embodiment, it is possible to improve a yield and the degree of purity of the object in the mineral processing process, and it is thus possible to provide a clear benefit to a customer.

[0122] <Other embodiments> In the above explanation, a description has been given of the embodiments according to the present invention; however, the present invention may also be implemented with various kinds of embodiments other than the embodiments described above.

[0123] <Camera 67> In the embodiment described above, a case has been as an example in which a single piece of the camera 67 is provided in the flotation machine 60, but two or more of the cameras 67 may be provided in the flotation machine 60. Furthermore, in this case, the plurality of AI analysis models 102a corresponding to the installation locations of the cameras 67 may be provided, and the decision unit 203b may decide the manipulation condition for the operation control device 10 by using the analysis results in combination obtained from the plurality of AI analysis model 102a.

[0124] <System> 　The flow of the processes, the control procedures, the specific names, and the information containing various kinds of data or parameters indicated in the above specification and drawings can be arbitrarily changed unless otherwise stated.

[0125] Furthermore, the components of each unit illustrated in the drawings are only for conceptually illustrating the functions thereof and are not always physically configured as illustrated in the drawings. In other words, the specific shape of a separate or integrated device is not limited to the drawings. Specifically, all or part of the device can be configured by functionally or physically separating or integrating any of the units depending on various loads or use conditions.

[0126] For example, in the embodiment described above, the AI analysis device 100 and the advanced control device 200 are constituted as different individual computers, but the AI analysis device 100 and the advanced control device 200 may also be constituted as an integrated computer. Furthermore, the AI analysis device 100 may also be implemented as, for example, a public cloud. In this case, the AI analysis device 100 conducts the AI analysis in response to a supply of the image information on the image obtained by the camera 67 from, for example, the advanced control device 200, and provides a cloud service to return the classification result of the state of the bubbles to the advanced control device 200.

[0127] Furthermore, all or any part of each of the processing functions performed by the each of the devices can be implemented by a CPU and by programs analyzed and executed by the CPU or implemented as hardware by wired logic.

[0128] <Hardware> The operation control devices 10, 20, 40, and 80, the monitoring devices 30, 50, 70, and 90, the AI analysis device 100, the advanced control device 200, the protocol server 300, and the integrated server device 400 according to the embodiment described above are implemented by a computer 1000 having the configuration illustrated in, for example, Fig. 16. In the following, an explanation will be given by using the AI analysis device 100 as an example. Fig. 16 is a diagram of a hardware configuration illustrating one example of the computer 1000 that implements the function of the AI analysis device 100 according to the embodiment.

[0129] As illustrated in Fig. 16, the computer 1000 includes a communication device 1000a, a secondary storage device 1000b, a memory 1000c, and a processor 1000d. Furthermore, each of the units illustrated in Fig. 16 is connected by a bus or the like with each other.

[0130] The communication device 1000a is a NIC or the like, and communicates with another device. The secondary storage device 1000b is implemented by a HDD, or the like, and stores therein the programs and the databases that operate the functions illustrated in Fig. 10.

[0131] The processor 1000d operates the thread that executes each of the function described above in Fig. 10 and the like by reading the programs that execute the same process as that performed by each of the processing units illustrated in Fig. 10 and the like from the secondary storage device 1000b or the like and loading the read programs in the memory 1000c. For example, the thread executes the same functions as those performed by each of the processing units included in the AI analysis device 100. Specifically, the processor 1000d reads, from the secondary storage device 1000b or the like, the programs having the same functions as those performed by the learning unit 103a, the acquisition unit 103b, the classification unit 103c, the transmission unit 103d, and the like. Then, the processor 1000d executes the thread for executing the same processes as those performed by the learning unit 103a, the acquisition unit 103b, the classification unit 103c, the transmission unit 103d, and the like.

[0132] In this way, the computer 1000 is operated as an information processing apparatus that performs various kinds of processing methods by reading and executing the programs. Furthermore, the computer 1000 is also able to implement the same functions as those described above in the embodiment by reading the above described programs from a recording medium by a medium reading device and executing the read programs. Moreover, the programs described here are not limited to be executed by only the computer 1000. For example, For example, the present invention may also be similarly used in a case in which a computer or a server each having another hardware configuration executes a program or in a case in which another computer and a server cooperatively execute the program with each other.

[0133] The programs may be distributed via a network, such as the Internet. Furthermore, the programs may be executed by storing the programs in a recording medium that can be read by a computer readable recording medium, such as a HDD, a flexible disk (FD), a CD-ROM, a magneto-optical disk (MO), a digital versatile disk (DVD), or the like, and read the programs from the recording medium by the computer. The recording medium in which these programs are recorded is also one mode of the present disclosure.

[0134] <Others> Some examples of combinations of the disclosed technical features are described in the following.

[0135] (1) 　A control system comprising: 　an acquisition unit that acquires image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; 　a classification unit that classifies the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; 　a decision unit that decides a manipulation condition for a field instrument according to the state of the intermediate material or the product material that has been classified by the classification unit; and 　an instruction unit that instructs an operation manipulation of the field instrument based on the manipulation condition that has been decided by the decision unit. (2) 　The control system according to (1), wherein 　the acquisition unit acquires image information on an image having captured therein a state of bubbles that are produced in froth flotation performed in the mineral processing process, and 　the instruction unit instructs the operation manipulation of the field instrument included in a concentration device that performs the froth flotation based on the manipulation condition that has been decided from the image information on the image having captured therein the state of the bubbles. (3) 　The control system according to (2), wherein 　the concentration device includes a flotation cell to which a pulp of an ore that has been subjected to size separation is input, and 　the acquisition unit acquires the image information on the image having captured therein the state of the bubbles that are present on the froth layer that is formed in a liquid surface outermost layer in the flotation cell in the froth flotation. (4) 　The control system according to (3), wherein 　the analysis model is trained such that, when the image information that is used for classification is input, a classification value that indicates the state of the bubbles is output, and 　the decision unit decides the manipulation condition based on preset information in which the manipulation condition has been set to each of the classification values in advance. (5) 　The control system according to (3) or (4), wherein the decision unit decides the manipulation condition about a manipulated variable corresponding to at least one of an injection rate of a chemical reagent to be injected into the flotation cell in the froth flotation, a supply rate of air to be supplied to the flotation cell in the froth flotation, and an inside cell level that is a height of a surface of a liquid in the flotation cell indicated in the froth flotation. (6) 　The control system according to (5), wherein, when it is classified that, regarding the state of the bubbles, a size of the bubbles is smaller than a size of the bubbles that are in an appropriate state, the decision unit decides the manipulation condition such that at least the injection rate of the chemical reagent is increased. (7) 　The control system according to (5) or (6), wherein, when, regarding the state of the bubbles, some of the bubbles are converted into a slurry, the decision unit decides the manipulation condition such that at least the supply rate of the air is maintained or increased. (8) 　The control system according to (5), (6), or (7), wherein, when it is classified that, regarding the state of the bubbles, almost no bubble is found, the decision unit decides the manipulation condition such that the injection rate of the chemical reagent and the supply rate of the air are increased and the inside cell level is maintained. (9) 　A control method that causes a computer to execute a process comprising: 　acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; 　classifying the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; 　deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material; and 　instructing an operation manipulation of the field instrument based on the decided manipulation condition. (10) 　A control program that causes a computer to execute a process comprising: 　acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; 　classifying the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; 　deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material; and 　instructing an operation manipulation of the field instrument based on the decided manipulation condition.

[0136] 1 plant 2 size reduction device 4 particle size separation device 6 concentration device 8 dewatering device 10, 20, 40, 80 operation control device 30, 50, 70, 90 monitoring device 60 flotation machine 61 flotation cell 62 agitation mechanism 63 chemical reagent injection mechanism 64 air supply mechanism 65 inside cell level adjustment mechanism 66 bubble discharge mechanism 67 camera 68 communication unit 69 various kinds of sensors 100 AI analysis device 101 communication unit 102 storage unit 102a AI analysis model 103 processing unit 103a learning unit 103b acquisition unit 103c classification unit 103d transmission unit 200 advanced control device 201 communication unit 202 storage unit 202a preset information 203 processing unit 203a acquisition unit 203b decision unit 203c instruction unit 300 protocol server 400 integrated server device 500 plant control system

Claims

1. A control system comprising: 　　an acquisition unit that acquires image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; 　　a classification unit that classifies the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; 　　a decision unit that decides a manipulation condition for a field instrument according to the state of the intermediate material or the product material that has been classified by the classification unit; and 　　an instruction unit that instructs an operation manipulation of the field instrument based on the manipulation condition that has been decided by the decision unit.

2. The control system according to claim 1, wherein 　　the acquisition unit acquires image information on an image having captured therein a state of bubbles that are produced in froth flotation performed in the mineral processing process, and 　　the instruction unit instructs the operation manipulation of the field instrument included in a concentration device that performs the froth flotation based on the manipulation condition that has been decided from the image information on the image having captured therein the state of the bubbles.

3. The control system according to claim 2, wherein 　　the concentration device includes a flotation cell to which a pulp of an ore that has been subjected to particle size separation is input, and 　　the acquisition unit acquires the image information on the image having captured therein the state of the bubbles that are present on the froth layer that is formed in a liquid surface outermost layer in the flotation cell in the froth flotation.

4. The control system according to claim 3, wherein 　　the analysis model is trained such that, when the image information that is used for classification is input, a classification value that indicates the state of the bubbles is output, and 　　the decision unit decides the manipulation condition based on preset information in which the manipulation condition has been set to each of the classification values in advance.

5. The control system according to claim 3, wherein the decision unit decides the manipulation condition about a manipulated variable corresponding to at least one of an injection rate of a chemical reagent to be injected into the flotation cell in the froth flotation, a supply rate of air to be supplied to the flotation cell in the froth flotation, and an inside cell level that is a height of a surface of a liquid in the flotation cell indicated in the froth flotation.

6. The control system according to claim 5, wherein, when it is classified that, regarding the state of the bubbles, a size of the bubbles is smaller than a size of the bubbles that are in an appropriate state, the decision unit decides the manipulation condition such that at least the injection rate of the chemical reagent is increased.

7. The control system according to claim 5, wherein, when, regarding the state of the bubbles, some of the bubbles are converted into a slurry, the decision unit decides the manipulation condition such that at least the supply rate of the air is maintained or increased.

8. The control system according to claim 5, wherein, when it is classified that, regarding the state of the bubbles, almost no bubble is found, the decision unit decides the manipulation condition such that the injection rate of the chemical reagent and the supply rate of the air are increased and the inside cell level is maintained.

9. A control method that causes a computer to execute a process comprising: 　　acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; 　　classifying the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; 　　deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material; and 　　instructing an operation manipulation of the field instrument based on the decided manipulation condition.

10. A control program that causes a computer to execute a process comprising: 　　acquiring image information on an image having captured therein a state of an intermediate material or a product material that is produced in a mineral processing process performed in a plant; 　　classifying the state of the intermediate material or the product material by analyzing the image information via an analysis model that has been trained by performing machine learning by using an image group for learning; 　　deciding a manipulation condition for a field instrument according to the classified state of the intermediate material or the product material; and 　　instructing an operation manipulation of the field instrument based on the decided manipulation condition.

Images