Method, apparatus, and system for predicting the timing of water change for washing.

Abstract

To provide a technique which more accurately determines the timing when washing water of sheet metal washing starts to become dirty that is notified for warning.SOLUTION: A washing water replacement timing prediction method includes the steps of: acquiring conductivity data of washing water used for sheet metal washing at a prescribed date interval by a conductivity data acquisition unit; obtaining the slope a of a conductivity graph, determining whether or not the condition that the slope a is equal to or greater than a slope increase determination value k obtained from a replacement threshold is satisfied (increase determination condition), and deciding the increase start date by a prediction calculation and determination unit; and performing alarm display on the basis of the increase start date by a display unit. In the washing water replacement timing prediction method according to Claim 1, the slope increase determination value k is obtained from the following (formula 1). k=α×(replacement threshold-stable reference value) / (replacement reference days) (formula 1). α: coefficient.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a method, apparatus, and system for predicting the timing of washing water replacement.

Background Art

[0002] In lines for manufacturing various products such as telephones and communication devices, a process of cleaning sheet metal used as a component is usually included. The cleaning process of the sheet metal mainly includes a process of cleaning the sheet metal with a cleaning liquid, rinsing it with water after cleaning, and drying it after rinsing. Here, when rinsing the sheet metal, components of the cleaning liquid adhering to the sheet metal and impurities contained in the sheet metal itself mix into the rinsing water, and the water in the rinsing tank is gradually contaminated. Therefore, it is necessary to replace the rinsing water at an appropriate time.

[0003] When replacing the rinsing water in the rinsing tank, it may affect other production lines by temporarily stopping the production line or replacing the cleaning liquid at the same time. Therefore, predicting the timing of rinsing water replacement in advance is an important issue in improving manufacturing costs and efficiency. Moreover, in the current trend of introducing IoT (Internet of Things) into the manufacturing site, instead of predicting the replacement timing based on human experience and intuition, a technology is required to continuously monitor the contamination situation, transmit measurement data via communication, accumulate the data, and then predict the replacement timing with a computer. Since the contamination of water is correlated with an increase in conductivity due to impurities, it is conceivable to predict the replacement timing using the measurement data of the conductivity of the rinsing water.

[0004] In Patent Document 1, a method is described in which the amount of change (difference) between the conductivity value at the time of measurement and the conductivity value at the previous measurement is obtained, and if this amount of change is greater than or equal to the average of the amounts of change in the past 10 times, an alarm process is executed to alert that the rinsing water has become contaminated.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

[0006] In the sheet metal cleaning process, the conductivity of the rinse water generally stabilizes at a low value immediately after the rinse water is changed (stable state), and then after a certain period, it shifts to an increasing state (increasing state). However, even during the stable state, there are temporary increases in conductivity due to the sheet metal being immersed in the rinse tank after cleaning, and temporary decreases in conductivity due to filtration by the filter, so the amount of change in conductivity fluctuates slightly in the short term.

[0007] However, Patent Document 1 assumes the use of the average value of conductivity data from the past 10 times, and therefore it is thought that the average value is calculated by focusing on short-term changes in conductivity (e.g., 1-2 days) and does not assume the use of long-term changes (e.g., 6 days). However, methods that focus on short-term changes in conductivity, such as the conventional technology, are susceptible to temporary fluctuations in conductivity, making it difficult to accurately determine the date and time when the conductivity shifts from a stable trend to an upward trend (corresponding to the date and time for alarm processing). Furthermore, since the conventional technology determines the alarm timing by comparing it with the average value of past changes, it does not take into account the conductivity (dirt) value at which the flush water should be replaced in the future, making it difficult to perform accurate alarm processing corresponding to the conductivity (dirt) that corresponds to the desired replacement level. This invention has been made in view of the above problems, and aims to provide a technology that more accurately determines the timing at which the rinse water used in sheet metal cleaning becomes contaminated, which is used to warn users. [Means for solving the problem]

[0008] To solve the above problems, one representative method of predicting the timing of water replacement for washing is a method of the present invention which includes the steps of: acquiring conductivity data of water used for washing sheet metal at predetermined time intervals in a conductivity data acquisition unit; determining the slope a of the conductivity graph in a prediction calculation and determination unit, determining whether the condition (increase determination condition) is met when the slope a is equal to or greater than the slope increase determination value k obtained from the replacement threshold, and determining the start date of the increase; and displaying an alarm in a display unit based on the start date of the increase. [Effects of the Invention]

[0009] According to the present invention, it is possible to more accurately determine the time when the rinse water used for sheet metal cleaning, which is used to warn users, begins to become contaminated. Issues, structures, and effects other than those mentioned above will be clarified by the following explanation of the implementation methods. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 is a system configuration diagram of the flush water replacement timing prediction system according to the first embodiment. [Figure 2] Figure 2 is a diagram showing the configuration of the water-washing section. [Figure 3] Figure 3 is a functional configuration diagram of the prediction device. [Figure 4] Figure 4 shows an example of conductivity data stored in the conductivity data storage unit. [Figure 5] Figure 5 shows an example of a long-term conductivity graph. [Figure 6] Figure 6 shows the effect of filter replacement on the long-term conductivity graph. [Figure 7] Figure 7 is an illustrative diagram showing how the start date of the increase is calculated according to the first embodiment. [Figure 8] Figure 8 is an illustrative diagram showing how the predicted exchange date is calculated according to the third embodiment. [Modes for carrying out the invention]

[0011] Embodiments of the present invention will be described below with reference to the drawings. However, the present invention is not limited to these embodiments. Furthermore, in the drawings, identical parts are denoted by the same reference numerals.

[0012] [First Embodiment] Figure 1 is a system configuration diagram of a water washing water change timing prediction system according to the first embodiment. In the first embodiment, sheet metal is placed in a washing basket suspended by a belt or the like and transported through each process by the belt or the like. That is, the washing basket containing sheet metal, which is brought into the washing machine 3 from the input port 4, is first washed (degreased) in the washing section 5. The washing section 5 is equipped with a washing tank and an oscillating member that shakes the washing tank. When the washing basket containing sheet metal is immersed in the washing tank, it is shaken for a predetermined time to perform washing. A circulation member may be installed together with the oscillating member, or in place of the oscillating member, to circulate the washing water in the washing tank and wash the sheet metal. An alkaline washing solution is usually used for washing (degreasing), but it is not particularly limited. When an alkaline washing solution is used, the level of washing capacity can be determined by measuring the pH value.

[0013] Next, the sheet metal-filled washing basket is washed in the washing section 6 (described later). The degree of contamination of the washing water is measured in the washing section 6. After washing, it is transported to the drying area 7 and dried with hot air. Once the series of processes in the washing machine 3 is complete, the sheet metal-filled washing basket is carried out of the washing machine 3 through the discharge port 8.

[0014] Pollution measurement data is wirelessly transmitted from the wireless communication module 9 to the gateway 10, and from the gateway 10, it is transmitted via wired LAN to the flush water change timing prediction device 1 (hereinafter also simply referred to as the "prediction device"). Although measurement data can also be transmitted via wired connection, wireless communication is used to utilize IoT communication technologies such as Wi-Fi and local 5G. Other necessary data, commands, or operations for display are input from the operation terminal 2 via wired LAN. However, transmission between the gateway 10, the operation terminal 2, and the prediction device can also be done via wireless LAN.

[0015] Figure 2 is a configuration diagram of the water washing section 6. The sheet metal washing basket is immersed in the water washing tank 20 and shaken for a predetermined time to perform water washing. A pipe 25 is connected to the water washing tank 20 and water is circulated by a circulation pump 24. By installing a filter 23 between the pipes 25, impurities in the water washing water can be filtered. Further, a conductivity sensor 21 (for example, HE-200C HORIBA) is installed between the pipes 25, and conductivity data, which is measurement data of the conductivity, is obtained by a conductivity meter 22 and sent to the wireless communication module 9. The conductivity sensor 21 can be installed at any location in the water washing section 6 as long as it can measure the conductivity of the water washing water. Also, a plurality of conductivity sensors 21 can be installed as needed to take the average of the conductivity values. The water washing water used for water washing may be pure water, but it is not particularly limited.

[0016] Figure 3 is a functional configuration diagram of the prediction device 1. The conductivity data input through the wired LAN from the gateway 10 is received by the conductivity data acquisition unit 102 through the network interface 101 and converted into the saved data format. Then, the converted data is saved in the conductivity data storage unit 103. The conductivity data storage unit 103 is composed of hardware used in a normal database such as an HD or SSD. The operation reception unit 104 has a function of receiving data and commands input from the operation terminal 2 through the network interface 101, and also receiving signals for the operation terminal 2 such as request signals and error messages from the prediction calculation and determination unit 105 and the display unit 106 described later.

[0017] The prediction calculation and determination unit 105 performs calculation and determination processes such as calculating parameters used for predicting the replacement timing of the washing water, determining the date and time for alarm processing or the predicted replacement date and time (details will be described later). The data required for calculation and determination is sent from the conductivity data storage unit 103 and the operation reception unit 104, and the results of calculation and determination are sent to the display unit 106 for display. The execution of the program process in the prediction calculation and determination unit 105 uses a normal computer arithmetic processing device such as a CPU. The display unit 106 displays a graph representing the transition of conductivity on the monitor screen (hereinafter referred to as the "conductivity graph"), an alarm for alerting the replacement of the washing water, or a regression line (described later) for predicting the replacement timing of the washing water, etc.

[0018] Figure 4 is an illustration of the conductivity data stored in the conductivity data storage unit 103. Figure 4(a) shows an excerpt of the measurement data obtained by measuring the conductivity data at intervals of 1 minute. The unit of conductivity is mS / m. The measurement interval does not necessarily have to be 1 minute. Figure 4(b) is a conductivity graph of the table in Figure 4(a), showing the transition of conductivity at a predetermined time of day (hereinafter, such a graph may also be referred to as a "short-term conductivity graph", and the data as "short-term conductivity data"). As can be seen from Figure 4(b), the initial value of the conductivity of the washing water is initially around 0.2 mS / m. However, during the time period when the washed basket with sheet metal is immersed in the washing tank 20 (with immersion), due to the influence of impurities adhering to the washing basket and the sheet metal, the water becomes dirty and the conductivity temporarily increases. Thereafter, during the time period when the washed basket with sheet metal has not been removed from the washing tank 20 (without immersion), the impurities are filtered by the filter 23 and the conductivity decreases. Furthermore, when the next washed basket with sheet metal is conveyed and immersed, the conductivity increases again, generating a cycle.

[0019] Next, we take the average of the short-term conductivity data in Figure 4 on a daily basis and consider the trend over a span of time equivalent to changing the flush water (approximately 2-3 months) (hereinafter, this graph may be referred to as the "long-term conductivity graph," and the data as "long-term conductivity data"). Figure 5 is an example of a long-term conductivity graph. Depending on the memory capacity of the conductivity data storage unit 103, it is also possible to create a long-term conductivity graph by accumulating the data from the short-term conductivity graph for a long period without taking an average. In the first embodiment, the long-term conductivity data shows a pattern in which it remains stable at a low value for a while (hereinafter referred to as the "stable trend"), and then gradually increases (hereinafter referred to as the "increasing trend"). Here, the date and time when the trend shifts from the stable trend to the increasing trend is called the increase start date. The reason for the shift from a stable trend to an upward trend is that, initially, when the sheet metal is immersed in the washing tank 20 after cleaning, almost all of the dirt in the washing water is filtered out by the filter, so there is no increase in conductivity that correlates with an increase in dirt, and the trend remains stable. However, as time passes, the filtration capacity of the filter begins to decrease, and when it can no longer adequately filter out the dirt, the amount of dirt in the washing water gradually begins to increase.

[0020] Figure 6 shows the effect of filter replacement on the long-term conductivity graph. When the filter becomes too dirty, it is replaced, but even when the filter remains the same (not replaced) before and after the rinse water change, Figure 6 shows that the conductivity remains stable for a while after the rinse water change and then starts to rise after a certain period. Therefore, it can be considered that the performance of the filter is always almost constant regardless of whether the rinse water is changed or not. Thus, even if the filter is replaced periodically regardless of the timing of the rinse water change in order to maintain filtration performance, the long-term conductivity graph will show a characteristic shift from a stable trend to an upward trend.

[0021] Figure 7 is an illustrative diagram of how the start date of the rise is calculated in the first embodiment. In the long-term conductivity graph, the slope a at date and time x is the value (slope) obtained from the change in conductivity between x-6 days and x days (span). In this disclosure, days are used when explaining long-term conductivity data and the date and time of the conductivity graph, but it is also possible to calculate parameter and prediction values ​​in units such as hours or minutes. In the first embodiment, the slope a is calculated daily. Next, the date and time that satisfies the condition (hereinafter referred to as the "rise determination condition") in which the slope a is greater than or equal to the slope rise determination value k (a≧k) described later is determined to be the start date of the rise. This calculation and determination is performed in the prediction calculation and determination unit 105. The start date of the rise is positioned as the date and time when the trend shifts from a stable trend to an upward trend, and is used as the date and time to execute an alarm process to warn about changing the flush water.

[0022] Because the operating frequency of the washing, rinsing, and drying production lines can vary from day to day, and the changes in conductivity can also change irregularly from day to day, it is difficult to accurately interpret the trend by calculating the slope 'a' over a short span (e.g., 1-2 days). On the other hand, over a long span (e.g., around 6 days), the variability in the operating frequency of the production lines is leveled out, and it is considered possible to more accurately interpret the trend of the slope 'a'. Therefore, it goes without saying that the long span for calculating the slope 'a' is not limited to 6 days, but is generally any period during which the operating frequency is leveled out.

[0023] (Slope increase detection value k) This section explains how to determine the slope rise judgment value k. First, multiple long-term conductivity data sets recorded in the past (hereinafter also referred to as "past data") are prepared in advance. The past data is stored in the conductivity data storage unit 103. Then, the start date of the rise is determined for each past data set. Here, the start date of the rise can be determined by inputting a value from the operation terminal 2 based on the operator's experience, but this is not the only way. Once the start date of the rise is determined, the average value of the conductivity up to the start date of the rise (stable transition period) is calculated for each past data set, and the average value among the past data sets is calculated (hereinafter referred to as the "stable reference value"). In addition, the number of days from the start date of the rise to the exchange date when the flush water is replaced is determined for each past data set, and the average value among the past data sets is calculated (hereinafter referred to as the "exchange reference number of days"). These calculations are performed in the prediction calculation / determination unit 105.

[0024] Examples of how to determine the stable reference value and the exchange reference period are explained below. Table 1 shows two examples of historical data, both showing a stable trend and an upward trend, obtained from March to June 2021. [Table 1] The stability reference value is determined by taking the average of the average conductivity values ​​during the stable trend period of historical data 1 (March 2 to April 9, 2021) and the average conductivity values ​​during the stable trend period of historical data 2 (April 30 to May 31, 2021). The exchange reference period can be calculated by taking the average of the 19 days from the start date of the rise in historical data 1 (2021 / 4 / 10) to the next exchange date (4 / 29) and the 29 days from the start date of the rise in historical data 2 (2021 / 6 / 1) to the next exchange date (6 / 30) ((19+29) / 2=24 days).

[0025] Once the stability threshold and replacement threshold are determined from past data in this way, the slope increase judgment value k is calculated using (Equation 1). k = α × (Replacement threshold - Stability threshold) / (Replacement reference days) ... (Equation 1) The exchange threshold is a predetermined value representing the conductivity value at which the cleaning water should be replaced. α is a coefficient. If α is too small, an alarm will be triggered even with a slight increase in conductivity, and if it is too large, the increase will be difficult to detect. Therefore, an appropriate value should be set, taking these factors into consideration. In the first embodiment, it was set to 0.5, but it is not limited to this and can be set in the range of 0.3 to 0.7, for example. For example, if the exchange threshold is 0.8 (mS / m), the stability standard value is 0.2 (mS / m), the exchange standard number of days is 30 days, and α is 0.5, then k = 0.5 × (0.8 - 0.2) / 30 = 0.01. Such calculations are also performed in the prediction calculation / determination unit 105.

[0026] (Effects / Actions) As described above, in the first embodiment, since the slope a is calculated over a long period in the long-term conductivity graph, irregular fluctuations in conductivity that depend on the operating frequency of the manufacturing line can be smoothed out. Furthermore, since the start date of the rise is determined by comparing it with the slope rise judgment value k, the detection sensitivity when the rate starts to rise can be appropriately adjusted by adjusting the coefficient α. In addition, it becomes possible to make a judgment that incorporates information on the replacement threshold, which is the conductivity value at which replacement should be made. This makes it possible to more accurately determine the timing when the rate changes from a stable trend to an upward trend, and to perform alarm processing or prepare for replacement of the wash water.

[0027] [Second Embodiment] In the first embodiment, when calculating the slope a of the long-term conductivity graph at one-day intervals, there is a risk of misdetecting the start date of the rise if outliers are included in the conductivity data. Therefore, in the second embodiment, the start date of the rise is determined when the slope a is measured to be k or higher for n consecutive days. This reduces the probability of misdetection. Otherwise, it is the same as the first embodiment. The method for determining the number of consecutive days n will be explained. Table 2 is a partial excerpt of long-term conductivity data (historical data) from the past, similar to the data used in Table 1. The start date of the increase, 2021 / 6 / 1, is assumed to be the correct data. [Table 2] Using the historical data in Table 2, and employing the same method as in the first embodiment, the slope rise determination value k is set to 0.01, and n is determined by increasing n from 1 until the correct rise start date is obtained by following the procedure below. (1) Search sequentially for dates and times that satisfy a≧k, with n=1. → May 1, 2021 is found. (2) Check if the hit date and time match 6 / 1. (3) Since there is no match, repeat the search with n=2. → 2021 / 5 / 21 is a hit. (4) Check if the hit date and time match 6 / 1. (5) Since there is no match, repeat the search with n=3. (6) While changing the value of n, increase the value of n until the hit date matches 6 / 1. Determine the value of n that matches. In the example in Table 2, n=3, and since the hit date coincided with 6 / 1, we determine that n=3. The process of determining n as described above, and the process of determining the start date of the rise based on the determined value of n, are performed in the prediction calculation and determination unit 105.

[0028] (Effects / Actions) In the second embodiment, by setting an appropriate value for n from past data and determining that the rise starts when the condition a≧k is met for n consecutive days, it is possible to avoid situations where outliers are mistakenly identified as the start date of the rise, resulting in unnecessary alarm processing.

[0029] [Third Embodiment] In the third embodiment, the process is the same as in the first and second embodiments except that a regression line is calculated using the least squares method from conductivity data (hereinafter referred to as "prediction data") up to a predetermined date and time determined from the start date of the rise (hereinafter referred to as the "regression start date") and the prediction execution date (usually corresponding to the current date and time when the prediction is made), and the date and time when the regression line intersects the exchange threshold is determined and displayed as the predicted exchange date. Figure 8 is an image diagram of how the predicted exchange date is calculated in the third embodiment. In the third embodiment, the regression start date is set to 6 days before the rise disclosure date. 6 days is the same number of days as the span used to calculate the slope a when determining the rise start date, and is included as prediction data because it is considered a date and time that exhibits the characteristics of an upward trend at an early stage. However, it is not necessary to limit it to 6 days; it may be 6 days or less, the rise start date may be used as the regression start date, and it may even be a date and time after the rise start date. The calculation of the regression line and the determination of the predicted exchange date are performed in the prediction calculation / determination unit 105. The regression line and the predicted exchange date are then displayed in the display unit 106.

[0030] (Effects / Actions) In the third embodiment, since prediction data is selected from the regression start date to the prediction execution date and used to calculate the regression line, stable conductivity data prior to the regression start date is excluded as noise, making it possible to determine an accurate predicted exchange date and enabling calculations that are cost-effective and efficient.

[0031] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above, and various modifications are possible without departing from the spirit of the present invention.

[0032] The following describes, but is not limited to, embodiments that may constitute the present invention. (Aspect 1) The conductivity data acquisition unit includes a process of acquiring conductivity data of the water used for washing sheet metal at predetermined time intervals, The prediction calculation and determination unit calculates the slope a of the conductivity graph, determines whether the condition (increase determination condition) is met such that the slope a is equal to or greater than the slope increase determination value k obtained from the exchange threshold, and determines the start date of the increase. The process of displaying an alarm on the display unit based on the date the rise started, A method for predicting the timing of water replacement for flushing. (Aspect 2) The aforementioned slope 'a' is a value obtained by calculating the change in conductivity over a long span. A method for predicting the timing of water replacement for flushing, as described in Embodiment 1. (Aspect 3) The aforementioned slope increase determination value k is obtained from the following (Equation 1): A method for predicting the timing of water replacement for flushing, as described in embodiment 1 or 2. k = α × (Replacement threshold - Stability threshold) / (Replacement reference days) ... (Equation 1) α: coefficient (Aspect 4) The coefficient α is in the range of 0.3 to 0.7, preferably 0.5. A method for predicting the timing of water replacement for flushing, as described in Embodiment 3. (Appendix 5) The process for determining the start date of the increase involves determining the first date and time when the increase determination condition is met a predetermined number of times consecutively as the start date of the increase. A method for predicting the timing of water replacement for flushing, as described in any one of embodiments 1 to 4. (Aspect 6) In the aforementioned prediction calculation and determination unit, The process involves determining a regression line between date and conductivity from conductivity data (prediction data) from a predetermined date before the start date of the increase (regression start date) to the prediction execution date, The process involves determining the date and time at which the regression line intersects the exchange threshold as the predicted exchange date for the flush water, A method for predicting the timing of water replacement for flushing, according to any one of embodiments 1 to 5. (Aspect 7) A conductivity data acquisition unit that acquires conductivity data of the water used for washing sheet metal at predetermined intervals, A prediction calculation and determination unit that determines the start date of the rise by determining whether the slope a of the conductivity graph is greater than or equal to the slope rise determination value k obtained from the exchange threshold (rise determination condition), and A display unit that displays an alarm based on the aforementioned date of the rise, A water change timing prediction device for flushing. (Pattern 8) The aforementioned slope 'a' is a value obtained by calculating the change in conductivity over a long span. A water change timing prediction device for flushing according to embodiment 7. (Aspect 9) The aforementioned slope increase determination value k is obtained from the following (Equation 1): A water change timing prediction device for flushing according to embodiment 7 or 8. k = α × (Replacement threshold - Stability threshold) / (Replacement reference days) ... (Equation 1) α: coefficient (Aspect 10) The coefficient α is in the range of 0.3 to 0.7, preferably 0.5. A water change timing prediction device for flushing according to embodiment 9. (Aspect 11) The prediction calculation and determination unit determines the date and time of the first instance in which the rise determination condition is met a predetermined number of times consecutively as the rise start date. A water change timing prediction device for flushing according to any one of embodiments 7 to 10. (Aspect 12) The prediction calculation and determination unit obtains a regression line between date and time and conductivity from conductivity data (prediction data) from a predetermined date and time before the start date of the rise (regression start date) to the prediction execution date. The date and time at which the regression line intersects the exchange threshold is determined as the predicted exchange date for flush water. A water change timing prediction device for flushing according to any one of embodiments 7 to 11. (Aspect 13) A water change timing prediction device for flushing water described in embodiments 7 to 12, A washing section having a washing tank and a conductivity meter, Wireless communication module and A flush water change timing prediction system equipped with the following features. [Explanation of symbols]

[0033] 1...Prediction device, 2...Operation terminal, 3...Washing machine, 4...Inlet, 5...Washing section 6...Washing area, 7...Drying area, 8...Discharge outlet, 9...Wireless communication module 10…Gateway, 20…Washing tub, 21…Conductivity sensor, 22…Conductivity meter 23...Filtration filter, 24...Circulation pump, 25...Pipe 101…Network interface, 102…Conductivity data acquisition unit 103... Conductivity data storage unit, 104... Operation reception unit 105…Predictive calculation / determination unit, 106…Display unit

Claims

1. The conductivity data acquisition unit includes a process of acquiring conductivity data of the water used for washing sheet metal at predetermined time intervals, The prediction calculation and determination unit determines the slope a of the conductivity graph, determines whether the condition (upgrade determination condition) is met such that the slope a is equal to or greater than the slope upgrade determination value k obtained from the exchange threshold, and determines the start date of the upgrade. The display unit has a step of displaying an alarm based on the date the rise started, The aforementioned slope rise determination value k is obtained from the following (Equation 1) as a method for predicting the timing of water change for flushing. k = α × (Replacement threshold - Stability threshold) / (Replacement reference days) ... (Equation 1) α: coefficient

2. The coefficient α is in the range of 0.3 to 0.7, preferably 0.

5. The method for predicting the timing of water change for flushing according to claim 1.

3. The process for determining the start date of the increase involves determining the first date and time when the increase determination condition is met a predetermined number of times consecutively as the start date of the increase. The method for predicting the timing of water change for flushing according to claim 1.

4. In the aforementioned prediction calculation and determination unit, The process involves determining a regression line between date and conductivity from conductivity data (prediction data) from a predetermined date before the start date of the increase (regression start date) to the prediction execution date, The process involves determining the date and time at which the regression line intersects the exchange threshold as the predicted exchange date for the flush water, A method for predicting the timing of water change for flushing according to claim 1.

5. A conductivity data acquisition unit that acquires conductivity data of the water used for washing sheet metal at predetermined intervals, A prediction calculation and determination unit that determines the start date of the rise by determining whether the slope a of the conductivity graph is greater than or equal to the slope rise determination value k obtained from the exchange threshold (rise determination condition), and It has a display unit that displays an alarm based on the aforementioned date of rise, The aforementioned slope rise determination value k is obtained from the following (Equation 1) and is used in the flush water change timing prediction device. k = α × (Replacement threshold - Stability threshold) / (Replacement reference days) ... (Equation 1) α: coefficient

6. The coefficient α is in the range of 0.3 to 0.7, preferably 0.

5. The water change timing prediction device for flushing according to claim 5.

7. The prediction calculation and determination unit determines the date and time of the first instance in which the rise determination condition is met a predetermined number of times consecutively as the rise start date. The water change timing prediction device for flushing according to claim 5.

8. The prediction calculation and determination unit obtains a regression line between date and time and conductivity from conductivity data (prediction data) from a predetermined date before the start date of the rise (regression start date) to the prediction execution date. The date and time at which the regression line intersects the exchange threshold is determined as the predicted exchange date for flush water. The water change timing prediction device for flushing according to claim 5.

9. A water change timing prediction device for flushing according to claim 5, A washing section having a washing tank and a conductivity meter, Wireless communication module and A flush water change timing prediction system equipped with the following features.

Images