Method and apparatus for detecting abnormality of filter cloth of horizontal filter press

Abstract

The present invention provides an apparatus for detecting an abnormality of a filter cloth of a horizontal filter press, the apparatus comprising: a sensing unit including a plurality of sensors for acquiring sensing information at a slurry inlet, a filtrate outlet, and an air inlet of the horizontal filter press; and a control unit for selecting one abnormality detection algorithm among a plurality of abnormality detection algorithms on the basis of training completeness and determining whether the filter cloth of the horizontal filter press is abnormal on the basis of output values of the abnormality detection algorithm, wherein the output values are derived using the sensing information as input values.

Description

Method and device for detecting filter media abnormalities in a horizontal filter press

[0001] The present disclosure relates to a technique for detecting filter cloth abnormalities in a horizontal filter press.

[0002] Horizontal filter presses for collecting brine lithium are used to remove impurities from brine slurry and collect clean filtrate. The filter press process must ensure the quality of the filtrate for subsequent processes. If the filter press filter cloth becomes defective, subsequent processes will be impeded by impurities in the filtrate, necessitating immediate replacement.

[0003] Examples of abnormalities in filter press filter media include filter media hardening and filter media tearing. Filter media hardening refers to the phenomenon in which filter media, which is composed of cloth, hardens after repeated long-term wetting and drying. Filter media hardening can cause a decrease in the filtrate collection rate as slurry becomes trapped between the hardened fibers. Filter media tearing refers to the phenomenon in which the filter media bursts due to the pressure increase caused by the inability to discharge the filtrate due to hardening. Filter media tearing can cause a loss of filter function, as unfiltered slurry is discharged through the outlet in the torn space.

[0004] Therefore, a specific plan is required to minimize process downtime due to filter media abnormalities in the filter press and to recognize the status of the filter media installed in the filter press and suggest measures to the manager.

[0005] These embodiments are intended to provide a technique for detecting whether an abnormality may occur in a filter cloth during the operation of a horizontal filter press.

[0006] In addition, the present embodiments aim to provide a method and device for detecting filter cloth abnormalities in a horizontal filter press, which can recognize the status of filter cloths installed in the filter press and suggest measures to a manager in order to minimize process downtime due to filter cloth abnormalities.

[0007] Additionally, the present embodiments aim to provide a step-by-step learning program that can advance the operation time of an anomaly detection algorithm by resolving the data bias problem.

[0008] In order to solve the above-described problem, one embodiment of the present disclosure is provided for a device for detecting an abnormality in a filter cloth of a horizontal filter press, the device including a sensing unit including a plurality of sensors that obtain sensing information from a slurry inlet, a filtrate discharge outlet, and an air inlet of the horizontal filter press, and a control unit that determines one abnormality detection algorithm based on a learning completion degree among a plurality of abnormality detection algorithms, and determines whether the filter cloth of the horizontal filter press is abnormal based on an output value of the abnormality detection algorithm derived by applying the sensing information as an input value.

[0009] In addition, one embodiment can provide a method for detecting an abnormality in a filter cloth of a horizontal filter press, the method including the steps of obtaining sensing information from a slurry inlet, a filtrate discharge outlet, and an air inlet of the horizontal filter press, the step of determining one abnormality detection algorithm based on a learning completion degree among a plurality of abnormality detection algorithms, and the step of determining whether the filter cloth of the horizontal filter press is abnormal based on an output value of the determined abnormality detection algorithm derived by applying the sensing information as an input value.

[0010] According to the present embodiment, a method and device for detecting filter cloth abnormalities in a horizontal filter press can be provided, which can recognize the status of filter cloths installed in the filter press and suggest measures to a manager in order to minimize process downtime due to filter cloth abnormalities.

[0011] Additionally, by providing a step-by-step learning program that can address data bias issues and accelerate the operation of anomaly detection algorithms, the cost of losses due to process delays can be reduced.

[0012] FIG. 1 is a drawing for explaining the configuration of a filter cloth abnormality detection device of a horizontal filter press according to one embodiment.

[0013] FIG. 2 is a drawing for explaining the operation of a horizontal filter press according to one embodiment.

[0014] FIG. 3 is a drawing for explaining a filter plate provided in a horizontal filter press according to one embodiment.

[0015] FIG. 4 is a drawing for explaining the operation of a filter anomaly detection device according to one embodiment.

[0016] FIG. 5 is a diagram for explaining an anomaly detection algorithm based on a fuzzy inference device according to one embodiment.

[0017] FIG. 6 is a diagram for explaining an ANFIS-based anomaly detection algorithm according to one embodiment.

[0018] FIG. 7 is a diagram illustrating a 1D CNN-based anomaly detection algorithm according to one embodiment.

[0019] FIG. 8 is a drawing for explaining a procedure for detecting an abnormality in a filter cloth of a horizontal filter press according to one embodiment.

[0020] FIG. 9 is a drawing for explaining the procedure of a step-by-step abnormality detection method for a filter cloth of a horizontal filter press according to one embodiment.

[0021] Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. When adding reference numerals to components in each drawing, identical components may have the same numerals as much as possible even if they are shown in different drawings. In addition, when describing the present embodiments, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the technical idea of ​​the present invention, the detailed description may be omitted. When "includes," "has," "consists of," etc. are used in this specification, other parts may be added unless "only" is used. When a component is expressed in the singular, it may include a case in which the plural is included unless specifically stated otherwise.

[0022] Additionally, terms such as first, second, A, B, (a), (b), etc. may be used to describe components of the present disclosure. These terms are only intended to distinguish the components from other components, and the nature, order, sequence, or number of the components are not limited by the terms.

[0023] In a description of the positional relationship of components, when it is described that two or more components are "connected," "combined," or "connected," it should be understood that the two or more components may be directly "connected," "combined," or "connected," but that the two or more components may also be further "interposed" with another component to be "connected," "combined," or "connected." Here, the other component may be included in one or more of the two or more components that are "connected," "combined," or "connected" to each other.

[0024] In the description of the temporal flow relationship related to components, operation methods, or manufacturing methods, for example, when the temporal or flow relationship is described as “after”, “following”, “next to”, “before”, etc., it may also include cases where it is not continuous, unless “immediately” or “directly” is used.

[0025] Meanwhile, when numerical values ​​or corresponding information (e.g., levels, etc.) for components are mentioned, even without separate explicit description, the numerical values ​​or corresponding information may be interpreted as including an error range that may occur due to various factors (e.g., process factors, internal or external impact, noise, etc.).

[0026] Hereinafter, a device for detecting filter cloth abnormalities of a horizontal filter press according to embodiments of the present disclosure will be described in detail with reference to related drawings.

[0027] FIG. 1 is a drawing for explaining the configuration of a filter cloth abnormality detection device of a horizontal filter press according to one embodiment.

[0028] Referring to FIG. 1, a filter cloth abnormality detection device (100) for detecting a filter cloth abnormality of a horizontal filter press may include a sensing unit (110) including a plurality of sensors that obtain sensing information from a slurry inlet, a filtrate outlet, and an air inlet of the horizontal filter press, and a control unit (120) that determines one abnormality detection algorithm among a plurality of abnormality detection algorithms based on a learning completion degree, and determines whether a filter cloth of the horizontal filter press is abnormal based on an output value of the abnormality detection algorithm derived by applying the sensing information as an input value. The configuration of the filter cloth abnormality detection device (100) illustrated in FIG. 1 is an example and is not limited thereto, and other components may be further included as needed.

[0029] Filtration using a horizontal filter press is a process that forces slurry into a sealed filtration chamber and separates solids and liquids through multiple, layered filter plates. For example, a horizontal filter press for collecting lithium from brine can be used to remove impurities in brine slurry and collect a clean filtrate. The filter press process is divided into a filling process (injecting slurry), a filtration process (discharging the filtrate), a compression process (injecting air to compress the cake), a suction process (discharging the compressed air), and a discharge process (discharging the dried cake). If a filter press filter media malfunctions, subsequent processes can be impeded by impurities in the filtrate, so it is important to promptly identify any abnormalities in the filter media.

[0030] As mentioned above, abnormalities in filter press filter media include filter media hardening and filter media tearing. When filter media hardening occurs, the flow rate at the slurry inlet decreases and the pressure increases during the filling and filtration processes. Furthermore, the flow rate at the filtrate outlet decreases and the pressure drops significantly. Furthermore, the pressure at the air inlet increases during the compression process. When filter media tearing occurs, the flow rate at the slurry inlet remains similar to normal during the filling and filtration processes, while the pressure decreases. Furthermore, the flow rate and pressure at the filtrate outlet decrease, and turbidity increases. Furthermore, the pressure at the air inlet decreases during the compression process as air leaks through the torn gaps. Therefore, filter media abnormalities can be diagnosed by examining changes in the flow rate and pressure of the slurry injection during the filling and filtration processes, changes in the flow rate, pressure, and turbidity at the filtrate outlet, and changes in the pressure at the air inlet during the compression process.

[0031] To this end, the sensing unit (110) may include a flow rate sensor and a pressure sensor installed at the slurry inlet of the horizontal filter press, a flow rate sensor, a pressure sensor, and a turbidity sensor installed at the filtrate outlet, and a pressure sensor installed at the air inlet. The sensing unit (110) may obtain sensing information through each sensor. The sensing information may include flow rate and pressure information detected at the slurry inlet, flow rate, pressure, and turbidity information detected at the filtrate outlet, and pressure information detected at the air inlet.

[0032] The sensing unit (110) can transmit the acquired sensing information to the control unit (120). For example, the sensing information may be configured to be continuously transmitted in real time during the filter press process. Alternatively, the sensing information may be configured to be transmitted at the request of the control unit (120). Alternatively, the sensing information may be configured to be transmitted according to predetermined conditions, such as a preset time interval or cycle.

[0033] The control unit (120) may be implemented in software, hardware, or a combination thereof in various devices capable of executing the method according to the technical idea of ​​the present disclosure, such as a processor, a computer, or other processing device.

[0034] The control unit (120) is connected to the sensing unit (110) in a communication manner, and can control the operation of the sensing unit (110) and monitor whether a malfunction occurs in the sensing unit (110). The control unit (120) can receive sensing information from the sensing unit (110). In one example, the control unit (120) can request the sensing unit (110) to transmit sensing information.

[0035] The control unit (120) can store the sensing information received from the sensing unit (110) in a database. The sensing information can be stored according to predetermined criteria, such as the sensed time information or the sensed location.

[0036] For example, a 1D CNN (Convolutional Neural Network) algorithm can be applied to detect system anomalies using such time-series sensor data. However, anomaly detection using 1D CNN requires a sufficient, unbiased data set. This is because, typically, the number of abnormal data is small compared to normal data, and the reliability of prediction results using the 1D CNN algorithm is low until sufficient abnormal data is secured. Therefore, it may be difficult to apply a 1D CNN-based anomaly detection algorithm in the field until sufficient abnormalities in the filter press pores have occurred.

[0037] Accordingly, the control unit (120) may determine one anomaly detection algorithm to derive a conclusion value based on the learning completion level of each anomaly detection algorithm that performs learning among the plurality of anomaly detection algorithms. For example, the plurality of anomaly detection algorithms may include a fuzzy inference engine-based anomaly detection algorithm, an adaptive network-based fuzzy inference system (ANFIS)-based anomaly detection algorithm, and a 1D CNN (convolutional neural network)-based anomaly detection algorithm. In this case, the fuzzy inference engine, ANFIS, and 1D CNN algorithms may include fuzzy inference engine, ANFIS, and 1D CNN algorithms that are known before or after the present disclosure, as long as they do not contradict the technical spirit of the present disclosure.

[0038] The control unit (120) can determine the learning completion level based on the accuracy of the ANFIS-based anomaly detection algorithm and the accuracy of the 1D CNN-based anomaly detection algorithm. The accuracy of the ANFIS-based anomaly detection algorithm and the accuracy of the 1D CNN-based anomaly detection algorithm can be measured by comparing the model's prediction results with actual data.

[0039] In this case, the control unit (120) can receive an answer value regarding whether the filter cloth of the horizontal filter press is abnormal through the user interface unit. That is, the control unit (120) can input an answer value regarding the actual state of the filter cloth (hardened filter cloth, torn filter cloth, or normal state) through the user interface unit. In one example, the control unit (120) can compare the output values ​​derived based on the sensing information with the answer values ​​to obtain a precise ratio of cases that match in all cases.

[0040] Additionally, the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm can each have their target precision preset as a first target value and a second target value. Each target value can be individually set as needed and is not limited to a specific value.

[0041] The control unit (120) can determine whether the current precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value. If the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value, the control unit (120) can determine the 1D CNN-based anomaly detection algorithm as the anomaly detection algorithm for deriving a conclusion value. That is, if the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value, the control unit (120) can determine whether the filter cloth of the horizontal filter press is abnormal based on the output value of the 1D CNN-based anomaly detection algorithm.

[0042] The control unit (120) can determine whether the precision of the ANFIS-based anomaly detection algorithm is higher than the second target value when the precision of the 1D CNN-based anomaly detection algorithm is lower than the first target value. When the precision of the ANFIS-based anomaly detection algorithm is higher than the second target value, the control unit (120) can determine the ANFIS-based anomaly detection algorithm as the anomaly detection algorithm for deriving a conclusion value. That is, when the precision of the 1D CNN-based anomaly detection algorithm is lower than the first target value and the precision of the ANFIS-based anomaly detection algorithm is higher than the second target value, the control unit (120) can determine whether the filter cloth of the horizontal filter press is abnormal based on the output value of the ANFIS-based anomaly detection algorithm.

[0043] If the precision of the 1D CNN-based anomaly detection algorithm is lower than or equal to the first target value and the precision of the ANFIS-based anomaly detection algorithm is also lower than or equal to the second target value, the control unit (120) may determine the fuzzy inference-based anomaly detection algorithm as the anomaly detection algorithm for deriving a conclusion. That is, if the precision of the 1D CNN-based anomaly detection algorithm is lower than or equal to the first target value and the precision of the ANFIS-based anomaly detection algorithm is lower than or equal to the second target value, the control unit (120) may determine whether the filter cloth of the horizontal filter press is abnormal based on the output value of the fuzzy inference-based anomaly detection algorithm.

[0044] The control unit (120) can output information to the user via the user interface regarding the presence or absence of a filter abnormality. If the control unit (120) determines that a filter abnormality has occurred, it can output the type of filter abnormality and an alarm for filter plate replacement. For example, the user interface unit is not limited to specific hardware, software, or a combination thereof, as long as it can input and output information according to operations authorized by the user.

[0045] Since filter anomaly data is acquired based on the assumption that actual filter anomalies occur, it can take considerable time to balance the anomaly and normal data. Therefore, if a 1D CNN-based anomaly detection algorithm is directly applied to the less biased anomaly data, the accuracy may fall below the initial target value. Therefore, as described above, the 1D CNN-based anomaly detection algorithm, the ANFIS-based anomaly detection algorithm, and the fuzzy inference-based anomaly detection algorithm can be applied in stages to prioritize the output value prediction.

[0046] As described above, the control unit (120) can receive an answer value regarding whether or not the filter pad of the horizontal filter press is abnormal through the user interface unit. That is, information regarding the actual state of the filter pad corresponding to the output value derived through the abnormality detection algorithm can be received. The control unit (120) can determine whether the sensing information used to derive the result value corresponds to abnormal data of the filter pad based on the answer value corresponding to the actual state of the filter pad. The control unit (120) can store data identified as filter pad abnormality data and use it for future learning.

[0047] The control unit (120) can perform learning of the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm by reflecting the correct answer value. That is, in a state of less bias in the anomaly data, the learning that reflects the prediction result and the output value prediction for the filter anomaly data can be repeated every time an anomaly occurs in the filter. Accordingly, as the bias state of the anomaly data is resolved, the accuracy of the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm can be increased.

[0048] The control unit (120) can learn linguistic variables in layer 1 and fuzzy rules in layer 4 for the ANFIS-based anomaly detection algorithm. Since ANFIS combines fuzzy logic and neural networks, learning of linguistic variables reflects the characteristics of fuzzy logic, and learning of fuzzy rules can be performed in a manner similar to learning of neural networks. Through such learning, the ANFIS-based anomaly detection algorithm can learn complex fuzzy rules for input data and derive an output value for whether or not a filter is abnormal using the learning results.

[0049] In Layer 1 of the ANFIS-based anomaly detection algorithm, linguistic variables of the input data are defined and learned, and these variables can be used to transform the input data into fuzzy sets. For example, the ANFIS-based anomaly detection algorithm defines linguistic variables using basis functions, such as the Gaussian function, and the parameters of these functions can be learned. In this case, learning can be performed using gradient descent or a similar optimization algorithm.

[0050] Additionally, in Layer 4 of the ANFIS-based anomaly detection algorithm, the weights of fuzzy rules can be learned. For example, learning can be performed to adjust the rule weights to minimize the error between the model output and the actual output. In this case, learning can be performed using the recursive least squares method or a similar optimization algorithm.

[0051] Additionally, the control unit (120) can perform learning for a 1D CNN-based anomaly detection algorithm through backpropagation. In this case, backpropagation can be used to learn in the direction of minimizing a loss function by adjusting the weights and bias of the neural network model.

[0052] For example, the control unit (120) may derive an output value by applying the weights and activation functions of each layer through a forward propagation process in a 1D CNN-based anomaly detection algorithm. In this case, a loss function representing the difference between the predicted output value and the actual value may be derived. In the backpropagation process, the loss function may be differentiated with respect to the weights and biases to derive the gradient of each parameter. Using the gradient, the weights and biases may be adjusted through gradient descent or a similar optimization algorithm. The updating of the weights and biases may be performed in a direction that minimizes the loss function. If this process is repeated, the number of filter anomaly data used for learning increases, the loss function is minimized, and the performance of the 1D CNN-based anomaly detection algorithm may be improved.

[0053] The control unit (120) can distribute the weights of the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value. That is, when the state of the biased filter anomaly data is resolved and the performance of the 1D CNN-based anomaly detection algorithm improves to a target precision or higher, the 1D CNN-based anomaly detection algorithm model can be distributed and the presence or absence of anomalies in the filter can be determined through the algorithm. Accordingly, the 1D CNN-based anomaly detection algorithm can be applied immediately even when the filter anomaly data is biased.

[0054] For example, the control unit (120) may store additional information related to the filter media within the horizontal filter press in the database. For example, information about the filter media may be stored based on the installation location of the filter media within the horizontal filter press. If the filter media is determined to be defective and the filter media is replaced, the control unit (120) may store or update information regarding the replacement timing of the filter media. The control unit (120) may learn a pattern regarding the replacement timing of the filter media, and may output a replacement warning alarm in advance if there is a risk of a filter media defect occurring.

[0055] This approach can minimize process downtime due to filter media abnormalities by recognizing the status of filter media installed in a filter press and suggesting corrective action to managers. Furthermore, by providing a step-by-step learning program that addresses data bias and accelerates the implementation of anomaly detection algorithms, the cost of losses resulting from process delays can be reduced.

[0056]

[0057] *Below, with reference to the relevant drawings, each embodiment related to a device for detecting filter abnormalities in a horizontal filter press will be described in detail.

[0058] FIG. 2 is a diagram for explaining the operation of a horizontal filter press according to one embodiment. FIG. 3 is a diagram for explaining a filter plate provided in a horizontal filter press according to one embodiment. FIG. 4 is a diagram for explaining the operation of a filter cloth anomaly detection device according to one embodiment. FIG. 5 is a diagram for explaining a fuzzy inference-based anomaly detection algorithm according to one embodiment. FIG. 6 is a diagram for explaining an ANFIS-based anomaly detection algorithm according to one embodiment. FIG. 7 is a diagram for explaining a 1D CNN-based anomaly detection algorithm according to one embodiment.

[0059] Referring to FIG. 2, a schematic internal structure diagram of a horizontal filter press (10) viewed from one side is illustrated. Filtration using a horizontal filter press is a process of forcing slurry into a sealed filter chamber between a head (16) and a follower (17) and separating solids and liquids through a plurality of layered filter plates (14). The filter press process is divided into a filling process in which slurry is injected through a slurry inlet (11), a filtration process in which filtrate is discharged through a filtrate discharge port (12), a compression process in which air is injected through an air inlet (13) to compress a cake (20), a suction process in which compressed air is discharged, and a discharge process in which the dried cake is discharged.

[0060] Referring to Fig. 3, a drawing of filter plates (14) stacked on a horizontal filter press is shown as viewed from the direction in which slurry is injected. A filtrate discharge port (12) may be provided at each of the four corners of the filter plate (14). A slurry inlet (11) is provided at the center of the filter plate (14), and a filter cloth (15) may be provided next to it. If the filter cloth (15) is worn or torn, it will affect the injection of slurry, the discharge of filtrate, and the injection of air.

[0061] Referring again to FIG. 2, a slurry inlet sensing unit (111) including a flow rate sensor and a pressure sensor may be installed at the slurry inlet (11) of the horizontal filter press. In addition, a filtrate outlet sensing unit (112) including a flow rate sensor, a pressure sensor, and a turbidity sensor may be installed at the filtrate outlet (12). In addition, an air inlet sensing unit (113) including a pressure sensor may be installed at the air inlet (13). Each sensing unit may obtain sensing information through each sensor. The sensing information may include flow rate and pressure information detected at the slurry inlet, flow rate, pressure, and turbidity information detected at the filtrate outlet, and pressure information detected at the air inlet.

[0062] Referring to FIG. 4, a filter abnormality detection device (100) for detecting filter abnormalities in a horizontal filter press may include a sensing unit (110) including a slurry inlet sensing unit, a filtrate discharge sensing unit, and an air inlet sensing unit, and a control unit (120) including a sensor module management unit (121), a data management unit (122), an algorithm unit (123), and a user interface unit (124). According to an example, the control unit (120) may be implemented as software, hardware, or a combination thereof operating within a computer.

[0063] The sensor module management unit (121) may include a sensor data collection unit and may be connected to the sensing unit (110) to control the operation of the sensor module and monitor whether a malfunction occurs in the sensor module. The sensor module management unit (121) may receive sensing information from the sensing unit (110). In one example, the sensor module management unit (121) may request the sensing unit (110) to transmit sensing information.

[0064] The data management unit (122) may include a database and store sensing information received from the sensing unit (110). The sensing information may be stored according to predetermined criteria, such as sensing time information or sensing location.

[0065] The user interface unit (124) may include monitoring, status input / output, and data query functions. As an example, the user interface unit (124) may include various input interfaces for receiving user-authorized operations and various output interfaces for outputting information according to input information or control of the control unit (120).

[0066] The algorithm section (123) may include a step-by-step anomaly detection algorithm for overcoming data bias, where abnormal data is less than normal data. The step-by-step anomaly detection learning algorithm for overcoming data bias may include a fuzzy inference-based anomaly detection algorithm, an ANFIS-based anomaly detection algorithm, and a 1D CNN-based anomaly detection algorithm.

[0067] The sensing unit (110) can obtain the flow rate, pressure of the slurry inlet, flow rate, pressure, and turbidity of the filtrate outlet during the filter press filling and filtration process. In addition, the sensing unit (110) can obtain sensor data on air pressure during the compression process.

[0068] The algorithm unit (123) may include a step-by-step anomaly detection learning algorithm for overcoming data bias in order to diagnose the status of the horizontal filter press filter cloth. The algorithm unit (123) may apply the step-by-step anomaly detection learning algorithm for overcoming data bias to evaluate the learning completion rate, detect anomalies in the filter press filter cloth, record the correct answer value, perform learning, and perform distribution.

[0069] The algorithm unit (123) can evaluate the accuracy of the 1D CNN-based anomaly detection algorithm and the accuracy of the ANFIS-based anomaly detection algorithm in order to evaluate the learning completion level.

[0070] In order to detect an abnormality in a filter press filter cloth, the algorithm unit (123) can output the state of the filter press filter cloth through a fuzzy inference-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm and the precision of the ANFIS-based anomaly detection algorithm are lower than their respective target precisions. The algorithm unit (123) can output the state of the filter press filter cloth through the ANFIS-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is lower than the target precision and the precision of the ANFIS-based anomaly detection algorithm is higher than the target precision. The algorithm unit (123) can output the state of the filter press through the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the target precision.

[0071] Referring to FIG. 5, the fuzzy inference-based anomaly detection algorithm may include an input, a fuzzifier, a fuzzy inference engine, a fuzzy rule, a defuzzifier, a state classifier, and an output.

[0072] The input of the fuzzy inference-based anomaly detection algorithm can be the average value of sensor data measured during the filling, filtering, and compression processes.

[0073] The fuzzifier of the fuzzy inference-based anomaly detection algorithm can convert the input sensor values ​​into each linguistic variable and membership function.

[0074] The fuzzy inference engine of the fuzzy inference-based anomaly detection algorithm can derive the influence of each language variable according to the input value.

[0075] The fuzzy rules of the fuzzy inference-based anomaly detection algorithm can be written as truth tables based on expert knowledge-based If / Then Rules.

[0076] The defuzzifier of the fuzzy inference-based anomaly detection algorithm can convert the filter status into a quantitative value and output it.

[0077] The state classifier of the fuzzy inference-based anomaly detection algorithm can classify the state of the nearest filter press filter based on the output value of the defuzzifier.

[0078] The output of the fuzzy inference-based anomaly detection algorithm can output the status of the classified filter.

[0079] Referring to FIG. 6, the ANFIS-based anomaly detection algorithm may include input, layer 1, layer 2, layer 3, layer 4, layer 5, a state classifier, and output.

[0080] The input of the ANFIS-based anomaly detection algorithm can be the average value of sensor data measured during the filling, filtering, and compression processes.

[0081] Layer 1 of the ANFIS-based anomaly detection algorithm can calculate the membership value of each language node from the input sensor data.

[0082] Layer 2 of the ANFIS-based anomaly detection algorithm can be output by multiplying the values ​​of each node.

[0083] Layer 3 of the ANFIS-based anomaly detection algorithm can be normalized by the strength of influence of each node.

[0084] Layer 4 of the ANFIS-based anomaly detection algorithm can be calculated as the contribution of each node based on fuzzy rules.

[0085] Layer 5 of the ANFIS-based anomaly detection algorithm can be output as the sum of each node.

[0086] The state classifier of the ANFIS-based anomaly detection algorithm can classify the state of the nearest filter press filter based on the output value of layer 5.

[0087] The output of the ANFIS-based anomaly detection algorithm can output the status of the classified filter.

[0088] Referring to FIG. 7, the 1D CNN-based anomaly detection algorithm may include an input, a convolutional layer, a pooling layer, a flattening layer, a fully connected layer 1, and a fully connected layer 2.

[0089] The input of the 1D CNN-based anomaly detection algorithm can be time series data from sensors measured during the filling, filtering, and compression processes.

[0090] The convolutional layer of the 1D CNN-based anomaly detection algorithm can be used to extract potential features by applying convolutional filtering.

[0091] The pooling layer of a 1D CNN-based anomaly detection algorithm can be used to reduce the size of an image or series while retaining important features identified in the convolutional layer.

[0092] The flattening layer of the 1D CNN-based anomaly detection algorithm can be used to convert the final convolutional layer created through selective pooling into an array rather than a matrix for use as input to an artificial neural network.

[0093] The fully connected layer 1 and fully connected layer 2 of the 1D CNN-based anomaly detection algorithm can be used to classify the state of the filter layer through a flattened matrix in the form of a one-dimensional array.

[0094] The algorithm unit (123) can receive the status of the filter press filter cloth input by the user through the status input / output function of the user interface until sufficient filter cloth abnormality data is secured to record the correct answer value. The algorithm unit (123) can record the received correct answer value.

[0095] The algorithm unit (123) can perform learning of an ANFIS-based anomaly detection algorithm and a 1D CNN-based anomaly detection algorithm.

[0096] The training of the ANFIS-based anomaly detection algorithm can be performed by training the linguistic variables of Layer 1 of the ANFIS-based anomaly detection algorithm and training the fuzzy rules of Layer 4 of the ANFIS-based anomaly detection algorithm. The training of the linguistic variables of Layer 1 of the ANFIS-based anomaly detection algorithm can be trained through propagation using the gradient descent algorithm. The training of the linguistic fuzzy rules of Layer 4 of the ANFIS-based anomaly detection algorithm can be optimized through the recursive least squares method.

[0097] Training of 1D CNN-based anomaly detection algorithms can be performed through backpropagation.

[0098] The algorithm unit (123) can distribute the weights of the 1D CNN when the precision of the 1D CNN-based anomaly detection algorithm is higher than the target precision.

[0099] This approach can minimize process downtime due to filter media abnormalities by recognizing the status of filter media installed in a filter press and suggesting corrective action to managers. Furthermore, by providing a step-by-step learning program that addresses data bias and accelerates the implementation of anomaly detection algorithms, the cost of losses resulting from process delays can be reduced.

[0100]

[0101] Hereinafter, a method for detecting filter cloth abnormalities in a horizontal filter press capable of performing some or all of the embodiments described with reference to FIGS. 1 to 7 will be described with reference to the drawings. The above description may be omitted to avoid redundant description, and in this case, the omitted content may be substantially equally applied to the following description, as long as it does not contradict the technical spirit of the invention.

[0102] Fig. 8 is a diagram for explaining a procedure (800) of a method for detecting anomalies in a filter cloth of a horizontal filter press according to one embodiment. Fig. 9 is a diagram for explaining a procedure (900) of a method for detecting anomalies in a filter cloth of a horizontal filter press according to one embodiment.

[0103] Referring to FIG. 8, the filter abnormality detection device can obtain sensing information from the slurry inlet, filtrate outlet, and air inlet of the horizontal filter press (S810).

[0104] A filter cloth abnormality detection device may include a flow sensor and a pressure sensor installed at a slurry inlet of a horizontal filter press, a flow sensor, a pressure sensor, and a turbidity sensor installed at a filtrate outlet, and a pressure sensor installed at an air inlet. The filter cloth abnormality detection device may obtain sensing information through each sensor. The sensing information may include flow rate and pressure information detected at the slurry inlet, flow rate, pressure, and turbidity information detected at the filtrate outlet, and pressure information detected at the air inlet.

[0105] Referring again to FIG. 8, the filter cloth anomaly detection device determines one anomaly detection algorithm among a plurality of anomaly detection algorithms based on the learning completion level (S820), and can determine whether the filter cloth of the horizontal filter press is abnormal based on the output value of the determined anomaly detection algorithm derived by applying the sensing information as an input value (S830).

[0106] The filter anomaly detection device may determine one anomaly detection algorithm from which to derive a conclusion based on the learning completion level of each anomaly detection algorithm among multiple anomaly detection algorithms. For example, the multiple anomaly detection algorithms may include a fuzzy inference-based anomaly detection algorithm, an ANFIS-based anomaly detection algorithm, and a 1D CNN-based anomaly detection algorithm.

[0107] The filter anomaly detection device can determine the learning completion based on the accuracy of the ANFIS-based anomaly detection algorithm and the accuracy of the 1D CNN-based anomaly detection algorithm. The accuracy of the ANFIS-based anomaly detection algorithm and the 1D CNN-based anomaly detection algorithm can be measured by comparing the model's predicted results with actual data.

[0108] In this case, the filter cloth abnormality detection device can receive an answer value regarding whether the filter cloth of the horizontal filter press is abnormal through the user interface unit. That is, the filter cloth abnormality detection device can receive an answer value regarding the actual state of the filter cloth (hardened filter cloth, torn filter cloth, or normal state) through the user interface unit. In one example, the filter cloth abnormality detection device can obtain a precise ratio of cases that match in all cases by comparing output values ​​derived based on sensing information with the answer values.

[0109] Additionally, the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm can each have their target precision preset as a first target value and a second target value. Each target value can be individually set as needed and is not limited to a specific value.

[0110] Referring to FIG. 9, the filter anomaly detection device can determine whether the current precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value (S910). If the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value, the filter anomaly detection device can determine the 1D CNN-based anomaly detection algorithm as the anomaly detection algorithm for deriving a conclusion value. That is, if the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value, the filter anomaly detection device can determine whether the filter of the horizontal filter press is abnormal based on the output value of the 1D CNN-based anomaly detection algorithm (S920).

[0111] If the precision of the 1D CNN-based anomaly detection algorithm is lower than or equal to the first target value, the filter anomaly detection device can determine whether the precision of the ANFIS-based anomaly detection algorithm is higher than the second target value (S940). If the precision of the ANFIS-based anomaly detection algorithm is higher than the second target value, the filter anomaly detection device can determine the ANFIS-based anomaly detection algorithm as the anomaly detection algorithm for deriving a conclusion value. That is, if the precision of the 1D CNN-based anomaly detection algorithm is lower than or equal to the first target value and the precision of the ANFIS-based anomaly detection algorithm is higher than the second target value, the filter anomaly detection device can determine whether the filter of the horizontal filter press is abnormal based on the output value of the ANFIS-based anomaly detection algorithm (S950).

[0112] If the precision of the 1D CNN-based anomaly detection algorithm is less than or equal to the first target value and the precision of the ANFIS-based anomaly detection algorithm is also less than or equal to the second target value, the filter bag anomaly detection device can determine the fuzzy inference-based anomaly detection algorithm as the anomaly detection algorithm for deriving a conclusion. That is, if the precision of the 1D CNN-based anomaly detection algorithm is less than or equal to the first target value and the precision of the ANFIS-based anomaly detection algorithm is less than or equal to the second target value, the filter bag anomaly detection device can determine whether the filter bag of the horizontal filter press is abnormal based on the output value of the fuzzy inference-based anomaly detection algorithm (S960).

[0113] A filter abnormality detection device can output information to the user via a user interface regarding whether a filter is abnormal. If a filter abnormality is determined to have occurred, the filter abnormality detection device can output an alarm indicating the type of filter abnormality and a filter plate replacement alarm.

[0114] That is, the filter anomaly detection device can preferentially predict the output value by sequentially applying a 1D CNN-based anomaly detection algorithm, an ANFIS-based anomaly detection algorithm, and a fuzzy inference-based anomaly detection algorithm.

[0115] As described above, the filter cloth abnormality detection device can receive an answer value regarding whether or not the filter cloth of the horizontal filter press is abnormal through the user interface unit (S970). That is, it can receive information on the actual state of the filter cloth corresponding to the output value derived through the abnormality detection algorithm. The filter cloth abnormality detection device can determine whether the sensing information used to derive the result value corresponds to abnormal data of the filter cloth based on the answer value corresponding to the actual state of the filter cloth. The filter cloth abnormality detection device can store data identified as filter cloth abnormality data and use it for future learning.

[0116] The filter anomaly detection device can train the ANFIS-based anomaly detection algorithm and the 1D CNN-based anomaly detection algorithm by reflecting the correct answer. That is, with less bias in the anomaly data, the output value for the filter anomaly data can be predicted and the learning process, which reflects the predicted results, can be repeated each time an anomaly occurs in the filter. Accordingly, as the bias in the anomaly data is resolved, the accuracy of the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm can be improved.

[0117] The filter anomaly detection device can learn linguistic variables in Layer 1 and fuzzy rules in Layer 4 for the ANFIS-based anomaly detection algorithm (S980). In Layer 1 of the ANFIS-based anomaly detection algorithm, linguistic variables of input data are defined and learned, and the linguistic variables can be used to transform the input data into a fuzzy set. In addition, in Layer 4 of the ANFIS-based anomaly detection algorithm, the weights of the fuzzy rules can be learned.

[0118] Additionally, the filter anomaly detection device can perform learning for a 1D CNN-based anomaly detection algorithm through backpropagation (S990). In this case, backpropagation can be used to learn by adjusting the weights and biases of the neural network model to minimize the loss function.

[0119] The filter anomaly detection device can distribute the weights of the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value (S930). That is, when the bias of the filter anomaly data is resolved and the performance of the 1D CNN-based anomaly detection algorithm improves to a target precision or higher, the 1D CNN-based anomaly detection algorithm model can be distributed and the presence or absence of anomalies in the filter can be determined through the algorithm. Accordingly, the 1D CNN-based anomaly detection algorithm can be applied immediately even when the filter anomaly data is biased.

[0120] This approach can minimize process downtime due to filter media abnormalities by recognizing the status of filter media installed in a filter press and suggesting corrective action to managers. Furthermore, by providing a step-by-step learning program that addresses data bias and accelerates the implementation of anomaly detection algorithms, the cost of losses resulting from process delays can be reduced.

[0121] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will readily appreciate that the present invention can be readily modified into other specific forms without altering the technical spirit or essential characteristics of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single entity may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined manner.

[0122] The scope of the present invention is indicated by the claims described below rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention.

[0123] The above-described methods and / or various embodiments may be realized by digital electronic circuits, computer hardware, firmware, software, and / or a combination thereof. Various embodiments of the present disclosure may be implemented as a computer program that is executed by a data processing device, for example, one or more programmable processors and / or one or more computing devices, or stored on a computer-readable recording medium and / or a computer-readable recording medium. The above-described computer program may be written in any form of programming language, including a compiled language or an interpreted language, and may be distributed in any form, such as a standalone program, a module, a subroutine, etc. The computer program may be distributed through a single computing device, multiple computing devices connected through the same network, and / or multiple computing devices distributed to be connected through multiple different networks.

[0124] The methods and / or various embodiments described above may be performed by one or more processors configured to execute one or more computer programs that process, store, and / or manage any function, function, etc. by operating on the basis of input data or generating output data. For example, the methods and / or various embodiments of the present disclosure may be performed by special purpose logic circuits such as Field Programmable Gate Arrays (FPGAs) or Application Specific Integrated Circuits (ASICs), and an apparatus and / or system for performing the methods and / or embodiments of the present disclosure may be implemented as special purpose logic circuits such as FPGAs or ASICs.

[0125] The one or more processors executing the computer program may include a general-purpose or special-purpose microprocessor and / or one or more processors of any type of digital computing device. The processor may receive instructions and / or data from each of read-only memory and random-access memory, or may receive instructions and / or data from the read-only memory and the random-access memory. In the present disclosure, components of a computing device performing the methods and / or embodiments may include one or more processors for executing instructions, and one or more memory devices for storing instructions and / or data.

[0126] According to one embodiment, the computing device can transmit and receive data to and from one or more mass storage devices for storing data. For example, the computing device can receive and / or transfer data from a magnetic disc or an optical disc. A computer-readable storage medium suitable for storing instructions and / or data associated with a computer program may include, but is not limited to, any form of non-volatile memory, including semiconductor memory devices such as Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable PROM (EEPROM), flash memory devices, and the like. For example, the computer-readable storage medium may include a magnetic disk such as an internal hard disk or a removable disk, a magneto-optical disk, a CD-ROM, and a DVD-ROM disk.

[0127] To provide interaction with a user, a computing device may include, but is not limited to, a display device (e.g., a cathode ray tube (CRT), a liquid crystal display (LCD), etc.) for providing or displaying information to the user, and a pointing device (e.g., a keyboard, a mouse, a trackball, etc.) for allowing the user to provide input and / or commands to the computing device. That is, the computing device may further include any other types of devices for providing interaction with the user. For example, the computing device may provide any form of sensory feedback to the user, including visual feedback, auditory feedback, and / or tactile feedback, for interaction with the user. In this regard, the user may provide input to the computing device through various gestures, such as visual, vocal, or motion.

[0128] In the present disclosure, various embodiments may be implemented in a computing system that includes a backend component (e.g., a data server), a middleware component (e.g., an application server), and / or a front-end component. In this case, the components may be interconnected via any form or medium of digital data communication, such as a communications network. For example, the communications network may include a Local Area Network (LAN), a Wide Area Network (WAN), and the like.

[0129] A computing device based on the present embodiments may be implemented using hardware and / or software configured to interact with a user, including a user device, a user interface (UI) device, a user terminal, or a client device. For example, the computing device may include a portable computing device such as a laptop computer. Additionally or alternatively, the computing device may include, but is not limited to, Personal Digital Assistants (PDAs), tablet PCs, game consoles, wearable devices, Internet of Things (IoT) devices, virtual reality (VR) devices, augmented reality (AR) devices, and the like. The computing device may further include other types of devices configured to interact with a user. In addition, the computing device may include a portable communication device (e.g., a mobile phone, a smart phone, a wireless cellular phone, etc.) suitable for wireless communication over a network such as a mobile communication network. The computing device may be configured to communicate wirelessly with a network server using wireless communication technologies and / or protocols, such as Radio Frequency (RF), Microwave Frequency (MWF), and / or Infrared Ray Frequency (IRF).

[0130] The above description is merely an illustrative example of the technical idea of ​​the present disclosure, and those skilled in the art to which the present disclosure pertains will appreciate that various modifications and variations can be made without departing from the essential characteristics of the technical idea of ​​the present disclosure. In addition, the present embodiments are not intended to limit the technical idea of ​​the present disclosure but rather to explain it, and therefore the scope of the technical idea of ​​the present disclosure is not limited by these embodiments. The scope of protection of the present disclosure should be interpreted by the claims below, and all technical ideas within a scope equivalent thereto should be interpreted as being included within the scope of the rights of the present disclosure.

[0131]

[0132] CROSS-REFERENCE TO RELATED APPLICATION

[0133] This patent application claims priority under 35 USC § 119(a) to Korean Patent Application No. 10-2023-0183080, filed December 15, 2023, the entire contents of which are incorporated herein by reference. Furthermore, this patent application claims priority in countries other than the United States for the same reasons, the entire contents of which are incorporated herein by reference.

Claims

1. In a device for detecting an abnormality in the filter cloth of a horizontal filter press, A sensing unit including a plurality of sensors that obtain sensing information from a slurry inlet, a filtrate discharge outlet, and an air inlet of a horizontal filter press; and A control unit that determines one anomaly detection algorithm among a plurality of anomaly detection algorithms based on the learning completion level, and determines whether a filter pad of the horizontal filter press is abnormal based on an output value of the determined anomaly detection algorithm derived by applying the sensing information as an input value; A filter anomaly detection device comprising:

2. In paragraph 1, The above sensing information is, A filter abnormality detection device including flow rate and pressure information detected at the slurry inlet, flow rate, pressure and turbidity information detected at the filtrate discharge port, and pressure information detected at the air inlet.

3. In paragraph 1, The above multiple anomaly detection algorithms are: A filter anomaly detection device including an anomaly detection algorithm based on a fuzzy inference engine, an anomaly detection algorithm based on an Adaptive Network-based Fuzzy Inference System (ANFIS), and an anomaly detection algorithm based on a 1D CNN (Convolutional Neural Network).

4. In paragraph 3, The above control unit, A filter anomaly detection device that judges the learning completion level based on the accuracy of an ANFIS-based anomaly detection algorithm and the accuracy of a 1D CNN-based anomaly detection algorithm.

5. In paragraph 4, The above control unit, A filter cloth anomaly detection device that determines whether a filter cloth of the horizontal filter press is abnormal based on the output value of the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value.

6. In paragraph 5, The above control unit, A filter cloth anomaly detection device that determines whether a filter cloth of the horizontal filter press is abnormal based on the output value of the ANFIS-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is lower than or equal to a first target value and the precision of the ANFIS-based anomaly detection algorithm is higher than a second target value.

7. In paragraph 6, The above control unit, A filter bag anomaly detection device that determines whether a filter bag of the horizontal filter press is abnormal based on the output value of the fuzzy inference-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is equal to or lower than a first target value and the precision of the ANFIS-based anomaly detection algorithm is equal to or lower than a second target value.

8. In paragraph 3, The above control unit, A filter bag abnormality detection device that receives an answer value regarding whether or not the filter bag of the horizontal filter press is abnormal through a user interface unit.

9. In paragraph 8, The above control unit, A filter anomaly detection device that performs learning of the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm by reflecting the above-mentioned correct answer value.

10. In paragraph 5, The above control unit, A filter anomaly detection device that distributes weights of the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value.

11. A method for detecting an abnormality in a filter cloth of a horizontal filter press, A step of acquiring sensing information from a slurry inlet, a filtrate discharge outlet, and an air inlet of a horizontal filter press; A step of determining one anomaly detection algorithm among multiple anomaly detection algorithms based on the learning completion level; and A step of determining whether there is an abnormality in the filter cloth of the horizontal filter press based on the output value of the determined abnormality detection algorithm derived by applying the sensing information as an input value; A method for detecting filter anomalies including:

12. In paragraph 11, The above sensing information is, A method for detecting a filter abnormality, the method comprising: flow rate and pressure information detected at the slurry inlet; flow rate, pressure and turbidity information detected at the filtrate discharge port; and pressure information detected at the air inlet port.

13. In paragraph 11, The above multiple anomaly detection algorithms are: A filtering anomaly detection method including a fuzzy inference engine-based anomaly detection algorithm, an Adaptive Network-based Fuzzy Inference System (ANFIS)-based anomaly detection algorithm, and a 1D CNN (Convolutional Neural Network)-based anomaly detection algorithm.

14. In paragraph 13, The step of determining the above one anomaly detection algorithm is: A filtering anomaly detection method that judges the learning completion based on the accuracy of an ANFIS-based anomaly detection algorithm and the accuracy of a 1D CNN-based anomaly detection algorithm.

15. In paragraph 14, The step for determining whether the above filter is abnormal is as follows: A filter cloth anomaly detection method for determining whether a filter cloth of the horizontal filter press is abnormal based on the output value of the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value.

16. In paragraph 15, The step for determining whether the above filter is abnormal is as follows: A filter cloth anomaly detection method for determining whether a filter cloth of the horizontal filter press is abnormal based on the output value of the ANFIS-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is lower than or equal to a first target value and the precision of the ANFIS-based anomaly detection algorithm is higher than a second target value.

17. In paragraph 16, The step for determining whether the above filter is abnormal is as follows: A filter bag anomaly detection method for determining whether a filter bag of the horizontal filter press is abnormal based on the output value of the fuzzy inference-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is equal to or lower than a first target value and the precision of the ANFIS-based anomaly detection algorithm is equal to or lower than a second target value.

18. In paragraph 13, A method for detecting filter abnormalities, further comprising: a step of receiving an answer value regarding whether or not a filter pad of the horizontal filter press is abnormal through a user interface unit; 19. In paragraph 18, A filter anomaly detection method further comprising: a step of performing learning of the 1D CNN-based anomaly detection algorithm and the ANFIS-based anomaly detection algorithm by reflecting the above correct answer value; 20. In paragraph 15, A filter anomaly detection method further comprising: a step of distributing weights of the 1D CNN-based anomaly detection algorithm when the precision of the 1D CNN-based anomaly detection algorithm is higher than the first target value;

Images