Prediction system and chemical solution

Abstract

[Problem] To provide a prediction system and a chemical solution that make it possible to predict the occurrence of troubles on paper, thereby preventing the occurrence of troubles on paper. [Solution] The present invention provides a prediction system and a chemical solution. The prediction system is provided with: a measurement unit 30 for performing water quality measurement of a pulp suspension; an imaging unit 40 for imaging paper that has passed through a dry part; and a control unit 20 connected to the measurement unit and the imaging unit. The water quality measurement is at least one selected from the group consisting of suspended solids measurement, turbidity measurement, ash content measurement, oxidation-reduction potential measurement, total raw material concentration measurement, raw material yield measurement, ash content yield measurement, conductivity measurement, water temperature measurement, and pH measurement. The control unit 20 includes: an acquisition means 21 for acquiring image data and a measurement result; a detection means 22 for detecting an incipient defect; an extraction means 23 for extracting a feature amount of the incipient defect; a classification means 24 for classifying the incipient defect by type; a model acquisition means 25 for acquiring a prediction model indicating a relationship between measurement information and defect count information; a setting means 26 for setting a threshold to a calculated measurement result obtained from the prediction model; and a transmission means 27 for transmitting an alert when an actually measured measurement result exceeds the threshold.

Description

Prediction system and chemical solution

[0001] The present invention relates to a prediction system and chemical solution for predicting the occurrence of defects in paper when paper is produced by papermaking a pulp suspension.

[0002] The manufacture of paper involves a raw material process in which a dried pulp sheet is defibrated, a preparation process in which additives such as fillers and sizing agents are added to this and stirred and mixed to form a pulp suspension, a wet part in which this pulp suspension is separated using a papermaking machine to remove the water contained in the pulp suspension and form wet paper, or a wet part in which the water contained in the pulp suspension is removed and compressed with a press roll to form wet paper, a dry part in which the wet paper is dried to form paper, and a reel part in which the paper is wound up.

[0003] In the papermaking process, from the viewpoint of productivity, paper (including the "wet paper") is transported at an extremely high speed. Therefore, if defects such as paper breaks (where the paper is cut) or pitch contamination (where pitch adheres to the paper) occur, the yield will decrease significantly. Therefore, various technologies have been developed in the papermaking process with the aim of minimizing the occurrence of defects such as paper breaks and pitch contamination.

[0004] For example, a monitoring system is known that includes a paper machine for producing paper, an application device for applying a chemical solution to a portion of the paper machine that comes into direct or indirect contact with the paper while the paper machine is operating, a control panel for setting application conditions for the application device, a surveillance camera for monitoring the portion to be monitored, and a control device connected to the surveillance camera via a network (see, for example, Patent Document 1). Such a monitoring system targets a dry part and monitors using video captured by the surveillance camera.

[0005] Also known is a defect classification system for classifying defect information based on defects in paper that has undergone a dry part in the papermaking process after the stock preparation process into corresponding defect cause items from among a plurality of defect cause items based on preset defect causes, the defect classification system comprising: an imaging means for imaging the paper that has undergone the dry part with an imaging device and acquiring the image data of the image; a detection means for detecting paper defects in the image data; an extraction means for extracting feature quantities of the defects; a calculation means for calculating certainty factors of the defect cause items based on the defect feature quantities with respect to a classification model with preset reference feature quantities and calculating the certainty factors for each defect cause item; a display means for displaying the certainty factors; and a classification means for classifying the defect information into the defect cause item with the greatest certainty factor among the plurality of certainty factors (see, for example, Patent Document 2). This defect classification system extracts paper defects from the image data that has been captured and classifies them into predetermined defect cause items.

[0006] Patent No. 6697132 Patent No. 7390085

[0007] However, the monitoring system described in Patent Document 1 is limited to monitoring the dry part, and is not sufficient in terms of predicting the occurrence of defects in paper. The defect classification system described in Patent Document 2 can take measures by classifying defects, but is not sufficient in terms of predicting the occurrence of defects in paper.

[0008] The present invention has been made in consideration of the above circumstances, and aims to provide a prediction system and chemical solution that can predict the occurrence of defects in paper, thereby preventing the occurrence of defects in paper.

[0009] The inventors of the present invention conducted extensive research to solve the above-mentioned problems and came to the conclusion that there may be a correlation between measurement information from water quality measurements of pulp suspensions and defect count information related to predictive defects. They then created a prediction model from this information and found that the above-mentioned problems could be solved by setting a threshold value for the calculation measurement results obtained from the prediction model, which led to the completion of the present invention.

[0010] The present invention provides a prediction system for predicting the occurrence of defects in paper when paper is manufactured by making a pulp suspension using a papermaking machine having at least a wet part and a dry part, the prediction system comprising a measurement unit for measuring the water quality of the pulp suspension, an imaging unit for imaging the paper after it has passed through the dry part, and a control unit connected to the measurement unit and the imaging unit via a network, wherein the water quality measurement is at least one selected from the group consisting of suspended solids measurement, turbidity measurement, ash content measurement, oxidation-reduction potential measurement, raw material total concentration measurement, raw material yield measurement, ash content measurement, electrical conductivity measurement, water temperature measurement, and pH measurement, and the control unit receives the image data from the imaging unit. a detection means for detecting at least a precursor defect that is a sign of a malfunction in the image data; an extraction means for extracting feature quantities of the precursor defect; a classification means for classifying the precursor defect by type from the feature quantities; a model acquisition means for acquiring a prediction model that shows the relationship between measurement information, which is time-series data of the water quality measurement results, and defect number information, which is time-series data of precursor defects for each type; a setting means for setting a threshold value for the calculated measurement results obtained from the prediction model; and a transmission means for transmitting an alert when the actual measurement results exceed the threshold value.

[0011] Preferably, the control unit further includes an addition command means for issuing a command to add or increase a chemical solution to the pulp suspension when the actual measurement result exceeds a threshold value. More preferably, the chemical solution includes at least one selected from the group consisting of a pitch control agent, a slime control agent, a paper strength agent, a sizing agent, a retention agent, a coagulant, and aluminum sulfate.

[0012] In the prediction system of the present invention, it is preferable that the defect is a defect caused by pitch contamination, the water quality measurement is at least one selected from the group consisting of suspended solids measurement, turbidity measurement, ash content measurement, raw material retention measurement, ash retention measurement, and electrical conductivity measurement, and the predictive defect is a micro-pitch defect caused by micro-pitch. In this case, it is preferable that the control unit further has addition command means for issuing a command to add or increase the addition of a chemical solution containing at least one selected from the group consisting of a pitch control agent, a paper strength agent, a retention agent, a coagulant, and aluminum sulfate to the pulp suspension when the actual measurement result exceeds a threshold value.

[0013] In the prediction system of the present invention, it is preferable that the defect is a defect due to paper breakage, the water quality measurement is at least one selected from the group consisting of oxidation-reduction potential measurement, water temperature measurement, and pH measurement, and the predictive defect is a peeling defect caused by peeling, a blending defect caused by blending, or an edge split defect caused by edge splitting. In this case, it is preferable that the control unit further has addition command means for issuing a command to add or increase a chemical solution containing a slime control agent to the pulp suspension when the actual measurement result exceeds a threshold value.

[0014] The present invention relates to a chemical solution that is added or increased to a pulp suspension when the actual measurement result exceeds a threshold value in the above-mentioned prediction system, and that includes at least one selected from the group consisting of a pitch control agent, a slime control agent, a paper strength agent, a sizing agent, a retention agent, a coagulant, and aluminum sulfate.

[0015] In the prediction system of the present invention, the control unit's acquisition means, detection means, extraction means, and classification means can detect predictive defects in image data and classify the predictive defects by type. Incidentally, even predictive defects have different trends in their progression until defects occur, making classification extremely important. Furthermore, in the prediction system, the control unit's model acquisition means acquires a prediction model, the setting means sets a threshold for the calculated measurement results obtained from the prediction model, and the transmission means issues an alert when the actual measurement results exceed the threshold. Therefore, by measuring water quality over time, it is possible to recognize an increase in the number of specific predictive defects. In other words, an increase in the number of specific predictive defects can be predicted as a possibility that they may become paper defects in the future.

[0016] In the prediction system of the present invention, a chemical solution can be added or increased to the pulp suspension in response to a command from the addition command means of the control unit. This improves the results of water quality measurements and prevents paper defects. In this case, if the chemical solution contains at least one selected from the group consisting of a pitch control agent, a slime control agent, a paper strength agent, a sizing agent, a retention agent, a coagulant, and aluminum sulfate, paper defects can be sufficiently prevented.

[0017] In the prediction system of the present invention, for example, to predict defects due to pitch contamination, a prediction model can be obtained from measurement information obtained from at least one water quality measurement selected from the group consisting of suspended solids measurement, turbidity measurement, ash content measurement, raw material yield measurement, ash retention measurement, and conductivity measurement, along with information on the number of minute pitch defects, and this prediction model can be used. By actually measuring the corresponding water quality, it is possible to predict the possibility that the corresponding water quality measurement will result in a paper defect due to pitch contamination in the future. In this case, by issuing a command from the addition command means to add or increase the amount of a chemical solution containing at least one selected from the group consisting of a pitch control agent, paper strength agent, retention agent, coagulant, and aluminum sulfate to the pulp suspension, the measurement results of the corresponding water quality measurement can be improved, thereby preventing the occurrence of paper defects due to pitch contamination.

[0018] In the prediction system of the present invention, for example, to predict defects due to paper breaks, a prediction model can be obtained and used from measurement information obtained from at least one water quality measurement selected from the group consisting of oxidation-reduction potential measurement, water temperature measurement, and pH measurement, along with defect count information for peeling defects, incorporation defects, or edge split defects. By actually measuring the corresponding water quality, it is possible to predict the possibility of future paper breaks. In this case, by issuing a command from the addition command means to add or increase the amount of a chemical solution containing a slime control agent to the pulp suspension, the measurement results of the corresponding water quality measurement can be improved, preventing paper breaks from occurring.

[0019] The chemical solution of the present invention contains at least one selected from the group consisting of pitch control agents, slime control agents, paper strength agents, sizing agents, retention agents, coagulants, and aluminum sulfate, and therefore improves the results of water quality measurements and prevents problems from occurring in the paper.

[0020] FIG. 1 is a schematic diagram illustrating a paper machine using a prediction system according to the present embodiment. FIG. 2 is a schematic diagram illustrating a measurement unit in the prediction system according to the present embodiment. FIG. 3 is a schematic diagram illustrating an example of an installation location for the measurement unit of the prediction system according to the present embodiment. FIG. 4 is a block diagram illustrating a control unit in the prediction system according to the present embodiment. FIG. 5 is a graph illustrating the relationship between measurement information, which is data over time on the measurement results of the amount of suspended solids, and defect number information, which is data over time on the number of fine pitch defects, in the prediction system according to the present embodiment. FIG. 6 is a graph illustrating the relationship between measurement information, which is data over time on the measurement results of the amount of suspended solids, and defect number information, which is data over time on the total number of fine pitch defects and non-fine normal-sized pitch defects, in the prediction system according to the present embodiment. FIG. 7 is a graph illustrating the relationship between measurement information, which is data over time on the measurement results of the oxidation-reduction potential, and paper break occurrence information, which is data over time on the occurrence or non-occurrence of a paper break, in the prediction system according to the present embodiment.

[0021] Preferred embodiments of the present invention will be described in detail below, with reference to the drawings as necessary. In the drawings, identical elements are designated by the same reference numerals, and duplicate explanations will be omitted. Furthermore, unless otherwise specified, positional relationships such as up, down, left, and right are based on the positional relationships shown in the drawings. Furthermore, the dimensional ratios of the drawings are not limited to those shown.

[0022] In this specification, "pulp suspension" refers to a suspension of dispersed pulp (hereinafter referred to as "pulp dispersion"), a raw material for paper, with or without additives. Pulp suspensions include not only virgin raw materials but also white water recovered in the wet part and a mixture of virgin raw materials and the white water. "Defects" in paper refer to a state in which poor quality occurs in paper, such as defects due to pitch contamination and defects due to paper breaks. "Precursor defects" are defects that are predictive of paper defects. "Pitch" refers to impurities contained in pulp, such as sticky substances derived from adhesive tape or glue or sticky substances derived from wood. "Fine pitch defects" are defects caused by minute pitches with a maximum diameter of less than N mm. The value of N can be set arbitrarily. It is generally set to less than 10 mm or less than 8 mm. The setting may also be changed depending on the desired quality of the paper. For example, the value should be less than 2 mm for white cardboard and less than 15 mm for corrugated board. Even if minute pitch defects are too small to cause defects in themselves, they can gradually grow and cause defects due to pitch contamination. Layer separation defects are defects that occur when foreign matter gets mixed in between layers of paper, causing poor adhesion between the layers and resulting in separation. Even if peeling defects do not cause defects in themselves, they can gradually grow and cause defects due to sheet breaks. Defects from stock preparation are defects in which foreign matter mixed in the pulp suspension is mixed into the paper, causing the foreign matter to appear on the surface of the paper or holes to form around the foreign matter. Even if stock preparation defects do not cause defects in themselves, they can gradually grow and cause defects due to sheet breaks. Edge cracks are defects in which the edges of the paper tear as drying progresses due to foreign matter adhering to both edges of the paper or poor formation. Even if edge separation defects do not cause defects in themselves, they can gradually grow and cause defects due to sheet breaks.

[0023] Fig. 1 is a schematic diagram illustrating a paper machine in which the prediction system according to this embodiment is used. As shown in Fig. 1, the paper machine 10 has at least a head box 1 into which a pulp suspension prepared in a preparation process is introduced, a wet part 2a consisting of a wire part 2 that pours the pulp suspension onto a wire and drops the water to form a wet paper, and a press part 3 that presses the wet paper via a felt to drop excess water, a dry part 4 that heats and dries the wet paper, a reel part 5 that winds the dried wet paper on a spool or the like, and a white water pit 6 installed below the wire part 2.

[0024] In the papermaking machine 10, the pulp suspension is made into paper by passing through these parts. The paper to be made is not particularly limited as long as it can be made by a papermaking process, and examples of the paper that can be used include so-called western paper such as printing paper, newsprint, coated paper, packaging paper, thin paper, household paper such as toilet paper and tissue paper, miscellaneous paper, and so-called layered paperboard such as cardboard base paper, white paperboard, colored paperboard, paper tube base paper, building material base paper, and various types of backing paper.

[0025] In the wire part 2, the water (white water) that falls is stored in the white water pit 6 installed below. In the press part 3, the water (white water) discharged from the wet paper is absorbed by a felt and collected by a felt suction box (not shown) installed on the felt. The collected water is then separated into gas and liquid in a separator, and only the white water is sent to the white water pit and stored there. The white water stored in the white water pit 6 contains pulp dispersion and is discharged from the white water pit 6 at appropriate times to be either disposed of as waste or recycled.

[0026] The prediction system according to this embodiment is a system for predicting the occurrence of defects in paper caused by pitch, and is performed using the papermaking machine 10 described above. The prediction system includes a measurement unit 30 for measuring the water quality of the pulp suspension fed from the head box 1 to the wire part 2, an imaging unit 40 for imaging the paper after it has passed through the dry part 4, and a control unit 20 connected to the measurement unit 30 and the imaging unit 40 via a network. The network may be wired or wireless.

[0027] The measurement unit 30 has a measuring device for measuring the water quality of the pulp suspension, which is at least one selected from the group consisting of suspended solids (SS), turbidity, ash content, oxidation-reduction potential (ORP), total organic carbon (TOC), dissolved oxygen, biochemical oxygen demand, chemical oxygen demand, total feedstock concentration, feedstock yield, ash retention, conductivity, water temperature, and pH. Among these, it is preferable that the measurement unit 30 has a measuring device for measuring the water quality of the pulp suspension, which is at least one selected from the group consisting of suspended solids (SS), turbidity, ash content, oxidation-reduction potential (ORP), total feedstock concentration, feedstock yield, ash retention, conductivity, water temperature, and pH. Note that these measuring devices are all known, and commercially available devices can be used as appropriate.

[0028] Here, "suspended solids" refers to the amount of suspended solids (SS) with a diameter of 2 mm or less suspended in a sample (pulp suspension). Specifically, it measures the amount of material that passes through a 2 mm sieve and remains on a 1 μm filter. "Turbidity" measures the degree of cloudiness of a sample (pulp suspension). "Ash content" measures the amount of inorganic non-combustible material in a sample (pulp suspension). "Oxidation-reduction potential" is the potential difference between the oxidizing and reducing power of a sample (pulp suspension). "Total organic carbon" is the total amount of organic matter present in a sample (pulp suspension) expressed as the amount of carbon contained in the organic matter. "Dissolved oxygen" measures the amount of oxygen dissolved in a sample (pulp suspension). "Biochemical oxygen demand" measures the amount of oxygen consumed when organic matter in a sample (pulp suspension) is decomposed by microorganisms. "Chemical oxygen demand measurement" refers to the amount of oxygen required to chemically oxidize organic matter in a sample (pulp suspension). "Total raw material concentration measurement" refers to the concentration of all solids in the raw material (pulp suspension). "Raw material retention measurement" refers to the ratio of the solids concentration in the pulp suspension fed into the headbox 1 to the value obtained by subtracting the solids concentration in the white water recovered by the wire part 2 from that solids concentration. "Ash retention measurement" refers to the ratio of the inorganic non-combustible matter concentration in the pulp suspension fed into the headbox 1 to the value obtained by subtracting the inorganic non-combustible matter concentration in the white water recovered by the wire part 2 from that inorganic non-combustible matter concentration. "Conductivity measurement" refers to measuring the conductivity of a sample (pulp suspension). "Water temperature measurement" refers to measuring the water temperature of a sample (pulp suspension). "pH measurement" refers to measuring the pH of a sample (pulp suspension).

[0029] In the prediction system according to this embodiment, when it is desired to predict defects due to pitch contamination, it is preferable to perform at least one water quality measurement selected from the group consisting of suspended solids measurement, turbidity measurement, ash content measurement, raw material yield measurement, ash content retention measurement, and conductivity measurement. This makes it possible to predict the possibility of future defects due to pitch contamination in paper. Furthermore, when it is desired to predict defects due to pitch contamination, it is more preferable to perform suspended solids measurement as a water quality measurement. In this case, the prediction becomes easier and more accurate.

[0030] For example, when measuring the amount of suspended solids, the higher the value of the amount of suspended solids, the more impurities such as micro-pitch and inorganic dispersions are contained in the pulp suspension. Therefore, when the value of the amount of suspended solids in a pulp suspension is higher than the normal value, the impurities in the pulp suspension also increase, and it is thought that the increase in impurities brought into the paper machine will increase the number of micro-pitch defects, leading to pitch contamination.

[0031] When measuring turbidity, the higher the turbidity value, the more impurities the pulp suspension contains, such as micro-pitch and inorganic dispersions. Therefore, when the suspended solids content of a pulp suspension is higher than normal, the impurities in the pulp suspension also increase, and the increased impurities carried into the paper machine increase the number of micro-pitch defects, which is thought to lead to pitch contamination.

[0032] When measuring ash content, the higher the ash content, the more impurities such as micro-pitch and inorganic dispersions are contained. Therefore, when the suspended solids content of a pulp suspension is higher than the normal value, the impurities in the pulp suspension also increase, and the increased impurities carried into the paper machine increase the number of micro-pitch defects, which is thought to lead to pitch contamination.

[0033] When measuring raw material yield, the lower the raw material yield value, the more impurities such as micro-pitch and inorganic dispersions are contained in the pulp suspension. Therefore, when the raw material yield value of the pulp suspension is lower than the normal value, the impurities in the pulp suspension are also increased, and the increase in impurities brought into the paper machine is thought to increase the number of micro-pitch defects, leading to pitch contamination.

[0034] When measuring ash retention, the lower the ash retention value, the more impurities such as micro-pitch and inorganic dispersions are contained in the pulp suspension. Therefore, when the ash retention value of the pulp suspension is lower than the normal value, the impurities in the pulp suspension also increase, and it is thought that the increase in impurities brought into the paper machine increases the number of micro-pitch defects, leading to pitch contamination.

[0035] When measuring electrical conductivity, the higher the conductivity value, the more impurities such as micro-pitch and inorganic dispersions are contained in the pulp suspension. Therefore, when the conductivity value of the pulp suspension is higher than the normal value, the impurities in the pulp suspension also increase, and it is thought that the increase in impurities brought into the paper machine increases the number of micro-pitch defects, leading to pitch contamination.

[0036] In the prediction system according to this embodiment, when it is desired to predict defects due to paper breaks, it is preferable to perform at least one water quality measurement selected from the group consisting of oxidation-reduction potential measurement, dissolved oxygen measurement, biochemical oxygen demand measurement, chemical oxygen demand measurement, water temperature measurement, and pH measurement, and it is more preferable to perform at least one water quality measurement selected from the group consisting of oxidation-reduction potential measurement, water temperature measurement, and pH measurement. This makes it possible to predict the possibility of future defects due to paper breaks. Furthermore, when it is desired to predict defects due to paper breaks, it is more preferable to perform oxidation-reduction potential measurement as the water quality measurement. In this case, the prediction becomes easier and more accurate.

[0037] For example, when measuring oxidation-reduction potential, the greater the drop in potential, the more likely the environment is for anaerobic bacteria, such as sulfate-reducing bacteria, to grow. Microorganisms, such as bacteria, mold, and algae, form slime. Therefore, when the oxidation-reduction potential of a pulp suspension is lower than normal, it means that a large amount of slime has been generated in the pulp suspension. This leads to an increase in the number and size of peeling defects, incorporation defects, edge split defects, and other defects, which can lead to paper breaks.

[0038] When measuring water temperature, the lower the water temperature, the more optimal the temperature range for bacteria to grow, promoting slime generation. It also promotes the aggregation / precipitation of pitch and inorganic components contained in the pulp suspension, causing incorporation defects. Therefore, when the temperature of the pulp suspension is lower than normal, a lot of slime is generated in the pulp suspension, or a lot of aggregated dirt is generated. This increases the amount of slime or pitch carried into the papermaking machine, resulting in an increase in the number of peeling defects, incorporation defects, edge split defects, and minute black spot defects, leading to paper breaks and other defects.

[0039] When measuring pH, the higher the pH, the more the aluminum ions in the aluminum sulfate in the pulp suspension become hydroxide complex ions, reducing the charge number, and the less effective they are at flocculating with the anion trash hydrated in the pulp suspension, leading to increased aggregation of the pitch particles themselves. This increases the amount of minute pitch in the pulp suspension, and when it is carried into the paper machine, the number of peeling defects, incorporation defects, edge split defects, minute black spot defects, etc. increases, leading to paper breaks and defects.

[0040] The measurement unit 30 periodically measures the water quality of the pulp suspension. That is, changes in the water quality measurement results over time are monitored. The measurements may be performed every few minutes, every few hours, or even once a day. The measurement unit 30 is connected to the control unit 20 via a network. Therefore, measurement commands to the measurement unit 30 are issued by the control unit 20, and the measurement results obtained by the measurement unit 30 are sent to the control unit 20. Details of this will be described later.

[0041] 2 is a schematic diagram illustrating a measurement unit in the prediction system according to this embodiment. As shown in FIG. 2, measurement unit 30 includes branch pipe 15 attached to any pipe X through which the pulp suspension flows, pipe Y branched by branch pipe 15, measurement device 11 attached to branched pipe Y, converter 12 capable of receiving a measurement result signal from measurement device 11, and gateway 13 connected to converter 12. In measurement unit 30, measurement device 11 measures the water quality of the pulp suspension that has flowed into branched pipe Y.

[0042] Next, the measuring device 11 transmits a measurement result signal representing the measurement result to the converter 12. Upon receiving the measurement result signal, the converter 12 converts the measurement result signal into a numerical value and transmits it to the gateway 13. The gateway 13 then transmits the numerical value (measurement result) to the control unit 20. The pulp suspension that flows into the branched pipe Y is measured by the measuring device 11 and then returned to a pipe at an arbitrary position further upstream.

[0043] FIG. 3 is a schematic diagram illustrating an example of the installation location of the measurement unit of the prediction system according to this embodiment. In FIG. 3, "P" denotes a pump, and a pulp suspension is sent in the direction of the arrow. As shown in FIG. 3, in the raw material process, pulp is sent to a refiner 31, where the pulp is continuously defibrated, beaten, refined, and other processes to produce a pulp dispersion. Next, in the preparation process, water containing the pulp dispersion from the refiner 31 is introduced into a raw material chest 32 and a machine chest 33, where predetermined additives are added, and a pulp suspension is prepared in a seed box 34. At this time, a portion of the pulp suspension is returned from the seed box 34 to the machine chest 33.

[0044] The pulp suspension in the seed box 34 is then sent to the head box 1 via a screen 35 that removes foreign matter. Meanwhile, the white water stored in the white water pit 6 is sent to the white water silo 36, where it is again stored. Because the white water stored in the white water silo 36 contains pulp dispersion, a portion of it is recycled. In other words, in this case, the pulp suspension used as the raw material is prepared by mixing a portion of the white water with the pulp suspension (virgin raw material) sent from the seed box 34, and is introduced into the head box 1 via the screen 35, as described above.

[0045] In the prediction system, the measurement unit 30 has the branch pipe 15 as described above, and therefore can be connected to any position as long as there is a pipe X through which the pulp suspension flows. The measurement unit 30 can be installed, for example, in the pipe on the outlet side of the raw material chest 32, the pipe through which the pulp suspension returns from the seed box 34 to the machine chest 33, the pipe between the seed box 34 and the junction with the white water, the pipe on the outlet side of the white water silo 36, or the pipe on the outlet side of the screen 35 (the inlet side of the bed box 1). The measurement unit 30 may be installed in one or more locations. Incidentally, by installing the measurement unit 30 in the pipe between the seed box 34 and the junction with the white water, it is possible to measure the pulp suspension as virgin raw material; by installing the measurement unit 30 in the pipe on the outlet side of the white water silo 36, it is possible to measure the pulp suspension as white water; and by installing the measurement unit 30 in the pipe on the outlet side of the screen 35, it is possible to measure the pulp suspension actually introduced from the head box 1. The specific installation location may be determined in light of actual measurement results and customer issues.

[0046] Returning to FIG. 1 , the imaging unit 40 has an imaging device for imaging the state of the paper between the dry part 4 and the reel part 5. Incidentally, lighting and other features may also be included as necessary. Examples of such imaging devices include a video camera, a line sensor camera, and an area sensor camera. Because the imaging device is installed downstream of the dry part in the papermaking process in the paper transport direction, there is no need to enter the wet part 2a or the dry part 4 to install the imaging device, making preparation and maintenance extremely safe and easy. Furthermore, defects occurring in the stock material process, preparation process, and papermaking process can be reliably detected.

[0047] The imaging unit 40 continuously captures images of the paper using an imaging device to acquire image data. The resolution of the image data is preferably 10 to 500 MHz, from the perspective of the size of minute defects to be detected, as described below. The imaging unit 40 is connected to the control unit 20 via a network. Therefore, imaging commands to the imaging unit 40 are issued by the control unit 20, and the image data acquired by the imaging unit 40 is sent to the control unit 20. This allows the control unit 20 to monitor the state of the paper over time as it goes through the processes up to the dry part 4.

[0048] Fig. 4 is a block diagram showing a control unit in the prediction system according to this embodiment. As shown in Fig. 4, the control unit 20 includes an acquisition unit 21, a detection unit 22, an extraction unit 23, a classification unit 24, a model acquisition unit 25, a setting unit 26, a transmission unit 27, and an addition command unit 28. The control unit 20 may be an ordinary computer having a calculation unit, a storage unit, an input unit, an output unit (display unit), etc.

[0049] As described above, the acquisition means 21 acquires image data from the imaging unit 40 and acquires measurement results from the measurement unit 30. The image data and measurement results acquired by the acquisition means 21 are then stored in a storage unit (not shown). Note that, in order to associate the image data and measurement results with each other, it is preferable to acquire data obtained at the same time as much as possible.

[0050] The detection means 22 is a means for detecting, in the captured image data, at least a defect that has occurred in the paper and is a sign of a problem. As described above, since the imaging device is installed downstream of the dry part 4 in the paper transport direction, the defect that is detected in the image data will be one that has occurred in the stock process, the preparation process, or a part before the dry part 4 in the papermaking process.

[0051] Examples of predictive defects include the adhesion of minute pitch (minute pitch defect), incorporation of foreign matter (incorporation defect), peeling due to the incorporation of foreign matter (peeling defect), and tearing of the paper edge (edge ​​split defect). Incidentally, an example of a defect that does not cause paper defects is the adhesion of insects (insect incorporation defect). Note that minute pitch defects are particularly predictive defects of pitch contamination, as are incorporation defects and peeling defects. Edge split defects are particularly predictive defects of paper breakage.

[0052] The detection of predictive defects is carried out by digitizing the data using intensity measurement, RGB measurement, shading processing, pattern search, edge detection, etc. Note that, from the viewpoint of accuracy, the detection of predictive defects is preferably carried out by intensity measurement or RGB measurement, and more preferably by both intensity measurement and RGB measurement. For example, when digitization is carried out by intensity measurement, the image is converted into a black and white binary image using a grayscale, and the shading is digitized by dividing it into 256 levels, for example, from 0 to 255. When digitization is carried out by RGB measurement, a specific color (e.g., blue) can be digitized.

[0053] The extraction means 23 is a means for extracting feature quantities of the predictive defects using the data of the predictive defects digitized by the detection means 12. The extraction means 23 can appropriately use a filter method, a wrapper method, an embedding method, or the like, to extract feature quantities. The extracted feature quantities are preferably at least one or more selected from the group consisting of the digitized size of the predictive defect, the shape of the predictive defect, the shading of the predictive defect, and the position of the predictive defect. The position of the predictive defect refers to at least the position where the predictive defect occurs in the width direction of the paper. The extracted feature quantities are stored in a memory unit as predictive defect information.

[0054] The classification means 24 classifies the predictive defects into types based on the feature amounts of the predictive defects extracted by the extraction means 12. That is, a classification model with preset reference feature amounts is acquired, and the classification model is made to calculate the confidence level for each type of predictive defect. The higher the confidence level, the more likely the defect falls into that type. Note that defects with a low confidence level are temporarily classified as "other" and then visually classified.

[0055] Here, the classification model is one in which reference features are learned by machine learning from an accumulation of actual predictive defects and their feature values. Furthermore, when new data on predictive defects and their feature values ​​is obtained, the classification model can be further trained with this data. In other words, the reference features can be changed. This further improves the accuracy of the confidence level calculated by the classification model.

[0056] As described above, in the prediction system, the control unit 20 has the acquisition means 21, detection means 22, extraction means 23, and classification means 24, so it is possible to detect predictive defects in image data and classify the predictive defects by type. In this way, in the prediction system, even predictive defects have different increasing trends until the occurrence of a defect depending on the type, so by deliberately classifying them, it is possible to more accurately predict the occurrence of defects in paper.

[0057] The model acquisition means 25 acquires a prediction model that shows the relationship between measurement information, which is data on the results of water quality measurements over time, and defect count information, which is data on each type of predictive defect over time. The predictive defect data over time includes data on the number of predictive defects over time, data on the size of predictive defects over time, and data on the location of predictive defects (e.g., their position in the width direction of the paper) over time. In other words, the prediction model learns the relationship between the accumulated measurement information and defect count information through machine learning. Furthermore, when new measurement information and defect count information are obtained, the prediction model can further learn them. In other words, the relationship between the two can be changed. This further improves the accuracy of paper defect predictions.

[0058] The setting means 26 is a means for setting a threshold value for the calculated measurement results obtained from the prediction model. In the setting means 26, the threshold value can be set arbitrarily. For example, the value of the calculated measurement results at a stage before the number of micro-pitch defects increases to become a defect can be set as the threshold value. Note that the threshold value may be set in multiple stages depending on the timing and amount of chemical agent injection.

[0059] The transmitting means 27 transmits an alert when the actual measurement result exceeds a threshold value. The alert may be transmitted by voice or a warning sound, or the warning may be displayed on a display unit (not shown).

[0060] As described above, the prediction system has the model acquisition means 25, setting means 26, and transmission means 27, and by measuring water quality over time and obtaining actual measurement results, it is possible to recognize an increase in the number of specific predictive defects. In other words, an increase in the number of specific predictive defects makes it possible to predict the possibility that they will become paper defects in the future.

[0061] The addition command means 28 is a means for issuing a command to add or increase the amount of chemical solution to the pulp suspension when the actual measurement result exceeds a threshold value. In the prediction system, the addition command means 28 can issue a command to add or increase the amount of chemical solution to the pulp suspension. This improves the measurement results from the water quality measurement and prevents problems from occurring in the paper. The location at which the chemical solution is added can be selected appropriately, for example, by adding it to the seed box 34.

[0062] The chemicals preferably contain at least one selected from the group consisting of pitch control agents, slime control agents, paper strength agents, sizing agents, retention agents, coagulants, and aluminum sulfate. The chemicals may also contain other known additives. In this case, paper defects can be sufficiently prevented.

[0063] Here, when the prediction system predicts that the actual measurement results exceed the threshold and that defects due to pitch contamination will occur, it is more preferable that the chemical solution added to the pulp suspension contain at least one selected from the group consisting of pitch control agents, paper strength agents, retention aids, coagulants, and aluminum sulfate, among the above. In this case, the measurement results of the corresponding water quality measurement (e.g., suspended solids) are improved, and defects due to pitch contamination in the paper can be prevented. Furthermore, when the prediction system predicts that the actual measurement results exceed the threshold and that defects due to paper breaks will occur, it is more preferable that the chemical solution added to the pulp suspension contain a slime control agent, among the above. In this case, the measurement results of the corresponding water quality measurement (e.g., oxidation-reduction potential) are improved, and defects due to paper breaks can be prevented.

[0064] The chemical solution according to the present invention is added or increased to a pulp suspension when the actual measurement results in the above-described prediction system exceed a threshold. The chemical solution contains at least one selected from the group consisting of pitch control agents, slime control agents, paper strength agents, sizing agents, retention aids, coagulants, and aluminum sulfate. The chemical solution may also contain additives such as fillers, sizing agents, dispersants, emulsifiers, paper strength agents, chelating agents, pH adjusters, preservatives, viscosity adjusters, solid lubricants, wetting agents, anti-dusting agents, release agents, adhesives, surface modifiers, detergents, paper strength agents, retention aids, anti-slip agents, and softeners. This improves the results of water quality measurements and makes it possible to prevent paper defects.

[0065] Although the preferred embodiments of the present invention have been described above, the present invention is not limited to the above embodiments.

[0066] In the prediction system according to this embodiment, the papermaking machine 10 has a wet part 2a (wire part 2 and press part 3), a dry part 4, and a reel part 5, but is not limited to these. For example, the wet part 2a may have only the wire part 2 without the press part 3. Incidentally, the "wet part" refers to a part preceding the dry part. Specifically, if the paper is cardboard, it refers to the wire part 2 and press part 3, and if the paper is household paper or the like that does not require the press part 3, it refers to the wire part. In the above-described embodiment, the wet part is described as a wire part and press part. The papermaking machine 10 may also have a calender part, etc. Furthermore, the papermaking machine 10 may be equipped with a processing machine that cuts and recovers paper instead of the reel part 5. Similarly, the preparation process includes, but is not limited to, a refiner 31, a raw material chest 32, a machine chest 33, a seed box 34, a screen 35, a white water silo 36, etc.

[0067] In the prediction system according to this embodiment, white water and virgin raw material are mixed to form the raw material pulp suspension, but only virgin raw material may be used as the raw material pulp suspension, or only white water may be used as the raw material pulp suspension.

[0068] In the prediction system according to this embodiment, the measurement unit 30 includes a branch pipe 15, a measurement device 11, a converter 12, and a gateway 13, but is not limited to this configuration as long as it is capable of measurement.

[0069] In the prediction system according to this embodiment, the detection means 22 detects predictive defects, but may also detect defects that are not predictive defects (hereinafter referred to as "normal defects") together with the predictive defects. In this case, the extraction means may extract feature quantities of the predictive defects and normal defects, and the classification means may classify the predictive defects and normal defects based on the feature quantities.

[0070] In the prediction system according to this embodiment, the addition command means 28 issues a command to add or increase the amount of chemical solution to the pulp suspension, but the chemical addition may be performed by a device or manually by a person. Reference example

[0071] (Regarding the correlation between the amount of suspended solids and fine pitch) Figure 5 is a graph showing the relationship between measurement information (shown as "surface white water SS" in Figure 5), which is data over time on the measurement results of the amount of suspended solids, and defect number information (shown as "pitch defects (all sizes)" in Figure 5), which is data over time on the number of fine pitch defects, in the prediction system of this embodiment. Note that the measurement results of the amount of suspended solids in Figure 5 were measured using a measuring device attached to the pipe on the outlet side of the white water silo 36. As shown in Figure 5, the occurrence trend of defect number information for fine pitches shows a correlation with the measurement information. For this reason, the prediction model is created based on the relationship between measurement information, which is data over time on the measurement results of the amount of suspended solids, and defect number information, which is data over time on the number of fine pitch defects.

[0072] (Correlation between Suspended Solids and Total Pitch) Figure 6 is a graph showing the relationship between measurement information (shown as "Surface White Water SS" in Figure 6), which is data on the measurement results of suspended solids over time, and defect count information (shown as "Total Defects" in Figure 6), which is data on the total number of minute pitch defects and non-minute, normal-sized pitch defects (hereinafter referred to as "Normal Pitch Defects") over time, in the prediction system of this embodiment. The suspended solids measurement results in Figure 6 were measured using a measuring device attached to the outlet pipe of the white water silo 36. As shown in Figure 6, the occurrence trend of defect count information for minute pitch and normal pitch defects is overshadowed by the occurrence number of normal pitch defects other than minute pitch, and therefore no sufficient correlation with the measurement information can be found. Incidentally, minute pitches grow into normal pitch defects, and normal pitch defects correspond to the pitch contamination described above.

[0073] (Regarding the correlation between oxidation-reduction potential and paper breaks) FIG. 7 is a graph showing the relationship between measurement information, which is data on the results of oxidation-reduction potential measurements over time, and paper break occurrence information, which is data on the occurrence of paper breaks over time, in the prediction system according to this embodiment. The measurement results in FIG. 7 were obtained by measuring oxidation-reduction potential using measuring devices attached to the pipes on the outlet side of the white silo 36 and the pipes on the outlet side of the raw material chest 32. Paper breaks occur when peeling defects grow or when the locations of precursor defects are concentrated in the selvage area, and are confirmed by a paper break sensor attached to the paper machine or by visual inspection. As shown in FIG. 7, the occurrence trend of paper break occurrence information shows a correlation with the measurement information.

[0074] The prediction system of the present invention can be used as a system for predicting the occurrence of defects in paper when a pulp suspension is made into paper using a papermaking machine.The prediction system of the present invention makes it possible to predict the occurrence of defects in paper, thereby making it possible to prevent the occurrence of defects in paper.The chemical solution of the present invention makes it possible to prevent the occurrence of defects in paper.

[0075] DESCRIPTION OF SYMBOLS 1...Headbox 10...Paper machine 11...Measuring device 12...Converter 13...Gateway 15...Branch pipe 2...Wire part 20...Control unit 21...Acquisition means 22...Detection means 23...Extraction means 24...Classification means 25...Model acquisition means 26...Setting means 27...Transmission means 28...Addition command means 2a...Wet part 3...Press part 30...Measuring unit 31...Refiner 32...Raw material chest 33...Machine chest 34...Seed box 35...Screen 36...White water silo 4...Dry part 40...Imaging unit 5...Reel part 6...White water pit

Claims

1. A prediction system for predicting defects in paper when paper is manufactured by making a pulp suspension using a papermaking machine having at least a wet part and a dry part, comprising: a measurement unit for measuring the water quality of the pulp suspension; an imaging unit for imaging the paper after it has passed through the dry part; and a control unit connected to the measurement unit and the imaging unit via a network, wherein the water quality measurement is at least one selected from the group consisting of suspended solids measurement, turbidity measurement, ash measurement, oxidation-reduction potential measurement, raw material total concentration measurement, raw material yield measurement, ash yield measurement, conductivity measurement, water temperature measurement, and pH measurement, and the control unit comprises: acquisition means for acquiring image data from the imaging unit and acquiring measurement results from the measurement unit; detection means for detecting at least precursor defects in the image data that are precursors of the defects; extraction means for extracting feature quantities of the precursor defects; and classification means for classifying the precursor defects by type based on the feature quantities. A prediction system having a model acquisition means for acquiring a prediction model showing the relationship between measurement information, which is data over time on the measurement results of the water quality measurement, and defect number information, which is data over time on the precursor defects for each type; a setting means for setting a threshold for the calculated measurement results obtained from the prediction model; and a transmission means for transmitting an alert when the actual measurement results exceed the threshold.

2. The prediction system according to claim 1, wherein the defect is a defect caused by pitch contamination, the water quality measurement is at least one selected from the group consisting of suspended solids measurement, turbidity measurement, ash measurement, raw material yield measurement, ash yield measurement, and conductivity measurement, and the precursor defect is a micro-pitch defect caused by micro-pitch.

3. A prediction system as described in claim 1, wherein the defect is a defect caused by paper breakage, the water quality measurement is at least one selected from the group consisting of oxidation-reduction potential measurement, water temperature measurement, and pH measurement, and the precursor defect is a peeling defect caused by peeling, a weaving defect caused by weaving, or an edge split defect caused by edge splitting.

4. A prediction system as described in claim 1, wherein the control unit further comprises an addition command means for issuing a command to add or increase the amount of chemical solution to the pulp suspension when the actual measurement result exceeds the threshold value.

5. The prediction system according to claim 4, wherein the chemical solution includes at least one selected from the group consisting of pitch control agents, slime control agents, paper strength agents, sizing agents, retention aids, coagulants, and aluminum sulfate.

6. A prediction system as described in claim 2, wherein the control unit further has an addition command means for issuing a command to add or increase the amount of a chemical solution containing at least one selected from the group consisting of a pitch control agent, a paper strength agent, a retention agent, a coagulant, and aluminum sulfate to the pulp suspension when the actual measurement result exceeds the threshold value.

7. A prediction system as described in claim 3, wherein the control unit further has an addition command means for issuing a command to add or increase the amount of a chemical solution containing a slime control agent to the pulp suspension when the actual measurement result exceeds the threshold value.

8. A chemical solution that is added or increased to the pulp suspension when the actual measurement result exceeds the threshold value in the prediction system described in claim 4, the chemical solution containing at least one selected from the group consisting of a pitch control agent, a slime control agent, a paper strength agent, a sizing agent, a retention agent, a coagulant, and aluminum sulfate.

Images