Predictive maintenance system and predictive maintenance method

Abstract

To make it possible to acquire maintenance parts in time for a predicted failure occurrence and effectively utilize the acquired maintenance parts which are unnecessary afterward.SOLUTION: A predictive maintenance system is configured to perform: a stock part securing process to predict a failure of an automated teller machine and secure maintenance parts of the automated teller machine predicted to fail, at a maintenance spot when the maintenance parts are in stock at the maintenance spot in the region; an inter-spot give-away process to give, when the maintenance parts of the automated teller machine predicted to fail are not in stock at the maintenance spot in the region and when there are redundant maintenance parts in stock at another maintenance spot, the maintenance parts so as to be acquired at the maintenance spot in the region on the day of the predicted failure; and an order reservation process to place an order for the deficiency so as to be acquired on the day of the predicted failure when the maintenance parts of the automated teller machine predicted to fail are not in stock at maintenance spots at any regions.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to predictive maintenance of devices, such as automated teller machines, that are in operation at various locations throughout a city. [Background technology]

[0002] ATMs, which are used to deposit and withdraw cash and perform various other transactions, are installed in various locations throughout the city and play an important role as part of the social infrastructure. Therefore, if any malfunction occurs in such equipment, it is important to take prompt action and minimize the disruption of its function.

[0003] In addition, by preparing parts for maintenance in advance in preparation for possible future malfunctions or breakdowns, it is possible to respond quickly when a malfunction or the like occurs. Conventionally, an example of this type of invention is described in JP 2018-173793 A (Patent Document 1).

[0004] Patent document 1 describes an information processing device that includes a processing unit that identifies products that are subject to maintenance work based on product fault information, predicts the occurrence of faults in parts of the products that are subject to the maintenance work, determines a priority of the maintenance work for the parts predicted to fail in accordance with the predicted fault, and adjusts the schedule of the maintenance work for the identified products in accordance with the availability date of the parts predicted to fail and the priority. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2018-173793 A Summary of the Invention [Problem to be solved by the invention]

[0006] Using AI-based predictive technology, it is possible to predict future abnormalities and failures that may occur, as well as the timing of these occurrences, based on the current operating information of equipment in operation, and by securing maintenance parts in advance for these occurrences, it becomes possible to respond quickly when an abnormality occurs.

[0007] Although the accuracy of prediction technology using AI and other technologies has improved in recent years, it is necessary to make fault predictions on the safe side in order to ensure that maintenance parts are secured in advance. Therefore, if maintenance parts are secured based on such fault predictions on the safe side, it is expected that the secured parts will not actually occur and the parts will not be needed.

[0008] In particular, the accuracy of long-term (several months into the future) forecasts, not just the immediate future, is lower than that of the immediate future, so there is a possibility that cases where no failures actually occur will increase. If a failure is predicted based on such a long-term forecast and maintenance parts are ordered, but the subsequent forecast is revised to show that no failures will occur, there may be cases where you want to cancel the order for the maintenance parts altogether. In addition, maintenance centers that are responsible for maintenance work on installed ATMs need to store maintenance parts, but they do not necessarily have sufficient storage space, so they want to avoid storing maintenance parts for long periods of time. Therefore, it is desirable to make the maintenance parts available as close as possible to the time when the failure is predicted to occur. Also, if the failure prediction is performed on the safe side as described above, it is expected that there will be an increase in cases where the secured parts become unnecessary because no failure actually occurs, so it is desirable to make effective use of the secured parts. [Means for solving the problem]

[0009] In order to solve the above-mentioned problems, the present invention provides a predictive maintenance system comprising a failure prediction device which predicts failures in an automatic transaction device, and a maintenance parts management and ordering device which is provided for each of a plurality of regions and which prepares in advance parts necessary for failure recovery for maintenance bases which are provided for maintaining the automatic transaction devices installed in the regions, and which performs the following steps: inventory part securing process for securing a maintenance part at a maintenance base in a region for a maintenance part for an automatic transaction device which is predicted to fail, if the maintenance base in that region has stock of the maintenance part; inter-base interchange process for accommodating the maintenance parts stocked at other maintenance bases so that the maintenance parts can be obtained at the maintenance base in that region on the date of the predicted failure, if the maintenance base in that region does not have stock of the maintenance part for an automatic transaction device which is predicted to fail and there is stock unused at other maintenance bases; and order reservation process for ordering the shortage part so that the maintenance part for an automatic transaction device which is predicted to fail can be obtained on the date of the predicted failure, if none of the maintenance bases in that region has stock of the maintenance part for the automatic transaction device which is predicted to fail. Effect of the Invention

[0010] It is possible to obtain maintenance parts according to the time when a failure is predicted to occur. In addition, maintenance parts that have been secured can be effectively utilized even if they are no longer needed. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram showing a maintenance management system in an embodiment of the present invention. [Diagram 2] 1 is a diagram showing an overall configuration of a predictive maintenance system 10 according to an embodiment of the present invention. [Diagram 3] 1 is a diagram showing a configuration of a failure prediction device 200 according to an embodiment of the present invention. [Figure 4] FIG. 4 is a diagram showing the contents of operation log data in the embodiment of the present invention. [Diagram 5] FIG. 2 is a diagram showing the contents of data in a fault / maintenance record in an embodiment of the present invention. [Figure 6] 1 is a diagram showing a configuration of a maintenance parts management ordering device 300 in an embodiment of the present invention. [Figure 7]3 is a diagram showing the contents of a failure prediction management table 301 in the embodiment of the present invention. FIG. [Figure 8] FIG. 3 is a diagram showing the contents of an ATM master table 302 in the embodiment of the present invention. [Figure 9] FIG. 3 is a diagram showing the contents of a part master table 303 in the embodiment of the present invention. [Figure 10] FIG. 3 is a diagram showing the contents of a stock management table 304 in the embodiment of the present invention. [Figure 11] FIG. 3 is a diagram showing the contents of an order table 305 in the embodiment of the present invention. [Figure 12] FIG. 3 is a diagram showing the contents of a parts allocation table 306 in the embodiment of the present invention. [Figure 13] 1 is a flowchart showing the overall processing of a predictive maintenance system according to an embodiment of the present invention. [Figure 14] 10 is a flowchart showing details of a part inventory securing process in the embodiment of the present invention. [Figure 15] 11 is a flowchart showing details of an inter-base accommodation process in the embodiment of the present invention. [Figure 16] 11 is a flowchart showing details of an order reservation process in the embodiment of the present invention. [Figure 17] FIG. 4 is a diagram for explaining the transition of contents of various tables in an embodiment of the present invention. [Figure 18] FIG. 4 is a diagram for explaining the transition of contents of various tables in an embodiment of the present invention. [Figure 19] FIG. 4 is a diagram for explaining the transition of contents of various tables in an embodiment of the present invention. [Figure 20] FIG. 13 is a diagram showing a configuration of a maintenance parts management ordering device 300A in another embodiment of the present invention. [Figure 21] FIG. 13 is a diagram showing the contents of a site management table 307 in another embodiment of the present invention. [Figure 22] FIG. 13 is a diagram showing the contents of a part master table 303A in another embodiment of the present invention. [Diagram 23]13 is a flowchart showing details of an inter-base accommodation process in another embodiment of the present invention. [Figure 24] 13 is a flowchart showing details of an order reservation process in another embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, the embodiments of the present invention will be described in detail with reference to the drawings as appropriate. Note that, although the same reference numerals are used for equivalent components throughout the drawings, in some cases, different reference numerals are used for equivalent components for convenience of the drawings and description.

[0013] First, a maintenance management system in an embodiment of the present invention will be described with reference to Fig. 1. In Fig. 1, a region (area) 100 is composed of area A (101), area B (102), area C (103), and area D (104), and within each area, ATM123 (111), ATM345 (112), ATM567 (113), and ATM789 (114) are scattered as automated teller machines (ATMs) installed in bank branches, commercial facilities, stores, etc.

[0014] In addition, in order to perform maintenance and management of these ATMs, there are maintenance bases in each area where maintenance personnel are stationed and maintenance parts are stored: base A (121), base B (122), base C (123), and base D (124).

[0015] Furthermore, a maintenance center 130 is provided to supervise these maintenance bases and to perform overall management of maintenance within the area 100. The predictive maintenance system of the present invention is realized by processing information collected from each ATM and maintenance base using a computer, storage device, etc. provided in the maintenance center 130. Note that there are cases in which the functions and roles of the maintenance center 130 are shared by each of the above-mentioned maintenance bases. EXAMPLES

[0016] (1) Overall structure 2 shows the overall configuration of the predictive maintenance system 10 of this embodiment, which is made up of a failure prediction device 200 and a maintenance parts management and ordering device 300. The predictive maintenance system 10 is installed in a maintenance center 130. First, the failure prediction device 200 collects information (operation log data) in which details of the operation and internal state of each ATM are recorded. In addition, past failure information and its recovery information (information on when and what kind of failure occurred, and what kind of maintenance work was performed to recover, etc.) at each base are input to the failure prediction device 200 and learned, thereby constructing a machine learning model for failure prediction.

[0017] Next, using the constructed machine learning model, the newly collected operation log data of each ATM is used to predict future failures that may occur at each ATM, and the actions and maintenance parts required to recover from the failures are predicted, and the results are input to the maintenance parts management and ordering device 300 as failure prediction data. Then, based on the acquired failure prediction data, the maintenance parts management and ordering device 300 processes information to issue commands to each base to secure the maintenance parts it possesses, commands to exchange maintenance parts between the bases, and commands to place new orders to maintenance parts vendors, and executes these commands at the appropriate time. (2) Fault prediction device 3 shows the configuration of the failure prediction device 200. This failure prediction device 200 is composed of an ATM operation log storage unit 201, an ATM failure / maintenance record storage unit 202, a learning model construction unit 203, and a failure prediction processing unit 204.

[0018] The ATM operation log storage unit 201 collects and stores data (operation log data) on daily operation status from the ATMs 210 installed in each area. Fig. 4 shows an example of this operation log data, which includes an ATMID 400 for identifying the ATM 210, an occurrence date and time 410 of the operation event, an operation type 420 of the ATM 210 related to the operation, whether or not a fault was detected at that time 430, fault details 440, whether or not multiple error codes output from the ATM 210 were detected at that time 450, and values ​​460 detected by various sensors inside the ATM 210 at that time.

[0019] The ATM fault / maintenance record storage unit 202 stores maintenance records 220 including details of faults that have occurred in the past in each ATM 210, what maintenance work was performed at that time, and what maintenance parts were used.

[0020] FIG. 5 shows an example of this fault / maintenance record data, which includes an ATMID 500 for identifying the ATM 210, the date and time of the fault 510, details of the fault 520, details of the recovery work performed at the time 530, and the type and number of maintenance parts used at the time 540.

[0021] The learning model construction unit 203 constructs a machine learning model using the data stored in the above-mentioned ATM operation log storage unit 201 and ATM fault / maintenance record storage unit 202 as learning data. That is, based on the operation log data of the ATM 210, a machine learning model is constructed that predicts when and what kind of fault may occur in the future, what kind of maintenance work should be performed when the fault occurs, what kind of maintenance parts will be required, and in what quantities.

[0022] The constructed machine learning model is stored in the failure prediction processing unit 204. Note that the construction of this machine learning model may be performed periodically to update the model, or may be performed when deterioration of prediction performance is recognized.

[0023] The failure prediction processing unit 204 receives the operation log data of the ATM 210 described above, performs failure prediction using a machine learning model, and outputs the prediction data to the maintenance parts management ordering device 300. (3) Maintenance parts management and ordering device 6 shows the configuration of the maintenance parts management ordering device 300. The maintenance parts management ordering device 300 is made up of a failure prediction management table 301, an ATM master table 302, a parts master table 303, an inventory management table 304, an ordering table 305, a parts securing table 306, and an arithmetic processing unit 320. Each of the above configurations may be realized by the same computer, or each table may be realized by a different storage device.

[0024] 7 is an example of the failure prediction management table 301, which stores failure prediction data received from the failure prediction device 200. The predicted failure ID 3011 is information for identifying multiple predicted failures. The ATM ID 3012 is information for identifying the ATM 210 in which the failure is predicted. The occurrence prediction 3013 is the occurrence date (expected occurrence date) of the predicted failure. The failure prediction management table 301 is sorted in ascending order of the occurrence prediction 3013.

[0025] The predicted failure 3014 indicates the details of the predicted failure. The predicted recovery process 3015 indicates the results of the prediction of the process for recovering from the failure. The required parts 3016 and the required quantity 3017 indicate the parts and the quantity required for the predicted recovery process 3015, respectively.

[0026] The ordering status 3018 indicates whether or not the necessary parts 3016 and the necessary quantity 3017 have been secured at the base, and if not, whether or not the order has been arranged. In the initial state when the fault prediction management table 301 is created, basically, all of the items are "not yet" status.

[0027] FIG. 8 shows an example of an ATM master table 302, which has an ATMID 3021 for identifying the ATM 210, an installation location 3022 indicating the store or the like in which the ATM 210 is installed, a location number 3023 for the location responsible for maintenance management, etc. of the ATM 210, and a location name 3024 thereof.

[0028] FIG. 9 shows an example of a parts master table 303, which has a parts ID 3031 for identifying a maintenance part, its part name 3032, the unit price 3033 of the part, standard delivery date 3034 indicating the lead time required to obtain the part when it is newly ordered, and inter-base delivery delivery date 3035 indicating the period required to borrow the part from another base, including packaging and shipping time.

[0029] 10 shows an example of inventory management table 304, which has part number 3043 (corresponding to part ID 3031 in FIG. 9) and part name 3044 for which inventory is held at base number 3041 and base name 3042, the number of parts in stock 3045, and frequency of use 3046 of that part at that base. Frequency of use 3046 indicates how frequently a part has been used at that base in past maintenance work, etc., with "high" indicating a high probability of future use.

[0030] FIG. 11 shows an example of an order table 305, which has an order number 3051, an identification number 3052 of the ordered part (corresponding to part ID 3031 in FIG. 9), an order quantity 3053, a planned date 3054 for placing the order, a planned delivery date 3055 for the part, a storage destination warehouse 3056 indicating the base to which the part is to be delivered, and a storage source warehouse 3057 indicating the base from which maintenance parts are lent when they are lent between bases (these values ​​correspond to jurisdiction base 3023 in FIG. 8), an order amount 3058 which is also the purchase price of the maintenance parts, and a remarks column 3059.

[0031] FIG. 12 shows an example of a parts allocation table 306, which has a target location 3061, the location name 3062, the part number (identification ID) 3063 of the part to be allocated at that location, the part name 3064, the number of parts allocated 3065, and the usage 3066 of the allocated part (corresponding to the predicted failure ID 3011 in FIG. 7). (4) Processing flow FIG. 13 shows an overall process flow of the predictive maintenance system in this embodiment, and is mainly composed of failure prediction processing by failure prediction device 200 (step S1310) and maintenance parts management ordering processing by maintenance parts management ordering device 300 (step S1320). (4-1) Fault prediction process (step S1310) As described above, the failure prediction device 200 creates and updates a prediction model (step S1311) using the operation log data (FIG. 4) from each ATM 210 and the failure recovery data (FIG. 5) at each base. The prediction model may be updated periodically, for example, once a month, or when performance degradation of the prediction model is detected by a separate means or the like.

[0032] Next, using the created prediction model, fault prediction is performed at a predetermined timing, and fault prediction data (corresponding to FIG. 7) is created (step S1312). Fault prediction data may be created daily or at a predetermined interval such as once a week, but the more frequently it is created, the greater the amount of accumulated operation log data used for prediction, and therefore the greater the possibility of improving prediction accuracy. (4-2) Maintenance parts management ordering process (step S1320) The maintenance parts management ordering device 300 secures stock parts at each base based on the failure prediction data (FIG. 7) and the parts inventory information at each base (FIG. 10) (step S1321). Next, similarly based on the failure prediction data and the parts inventory information (FIG. 10), it reserves stock parts between the bases (step S1322). Then, it reserves orders for parts that are not expected to be prepared by any of the above processes (step S1323).

[0033] Furthermore, based on the order table (FIG. 11) created (updated) by the above processing, when the scheduled order date 3054 arrives, the order (supply) details are processed (step S1324). Note that this order processing may be carried out by a staff member of the maintenance center 130 visually referring to the contents of the order table and making arrangements for the order (supply).

[0034] Next, each of the above-mentioned stock part securing process (step S1321), inter-base interchange process (step S1322), and order reservation process (step S1323) will be described in detail with reference to separate flowcharts. (4-2-1) Stock parts acquisition process (step S1321) Fig. 14 is a flowchart showing details of the inventory part securing process. The predicted fault ID (3011) in the fault prediction management table 301 shown in Fig. 7 is set to "i", and "i=1" is set as the initial value (step S1401). Then, for the predicted fault with the predicted fault ID (3011) of "1", the ordering status 3018 is checked to see if it is in a state where ordering has not been completed (not yet, XX units in shortage) (step S1402).

[0035] If arrangements have been made (NO in step S1402), "i" is incremented by "1" (step S1403), and the same process is performed for the next predicted failure ID.

[0036] If the arrangement has not been made (YES in step S1402), the location where the ATMID (3012) related to the predicted failure is located is identified from the ATM master table 302 shown in FIG. 8, and further, it is confirmed by referring to the inventory management table 304 shown in FIG. 10 whether the necessary part 3016 and the necessary quantity 3017 are in stock at that location (step S1404).

[0037] If the part is in stock (YES in step S1405), the details are recorded in the parts allocation table (FIG. 12) in order to secure the part at that base (step S1406), and the ordering status 3018 in the failure prediction management table 301 and the stock quantity 3045 in the inventory management table 304 are updated. If the part is not in stock (NO in step S1405), "i" is incremented by "1" (step S1403), and the same process is performed for the next predicted failure ID.

[0038] Then, it is confirmed whether the above-mentioned processing has been completed for all predicted failure IDs, and if not (NO in step S1408), "i" is incremented by "1" (step S1403), and the same processing is performed for the next predicted failure ID. If the processing has been completed (YES in step S1408), an instruction is issued to each base to secure the part according to the contents recorded in the part allocation table (FIG. 12) (step S1409). (4-2-2) Inter-base transfer processing (step S1322) Fig. 15 is a flow chart showing details of the inter-base interchange process. As in the flow of Fig. 14, the predicted fault ID (3011) is set to "i" and the initial value "i=1" is set (step S1501). Then, for the predicted fault with the predicted fault ID (3011) of "1", the arrangement status 3018 is checked to see if it is in a state where arrangement has not been made (not yet, XX units missing) (step S1502).

[0039] If arrangements have been made (NO in step S1502), "i" is incremented by "1" (step S1503), and the same process is performed for the next predicted failure ID.

[0040] If the arrangement has not been made (YES in step S1502), the location where the ATMID (3012) related to the predicted failure is located is identified from the ATM master table 302 shown in Fig. 8, and further, it is confirmed whether the necessary part 3016 and the necessary quantity 3017 are in stock at another location by referring to the inventory management table 304 shown in Fig. 10 (step S1504). If there is no stock (NO in step S1505), "i" is incremented by "1" (step S1503), and the same process is performed for the next predicted failure ID.

[0041] If there is stock (YES in step S1505), it is checked whether there is stock at other locations, and if there is stock at multiple locations (YES in step S1506), it is determined which of the multiple locations the stock will be borrowed from (step S1507).

[0042] Here, the following criteria are used to determine the source of the part to be transferred. That is, (1) the part is transferred from the base with the largest stock of the same part, (2) if the part has a flag indicating that it has been used frequently in the past (usage frequency 3046 in the stock management table 304 in FIG. 10), the part is not transferred from that base (because it is highly likely to be used in the future), and (3) the part is transferred from the base that has other parts that can be transferred at the same time. Note that these criteria can be combined as appropriate, or applied with a priority order set. Also, the scheduled delivery date 3055 and the number of orders 3053 in the order table 305 and the predicted occurrence date 3013 in the failure prediction management table 301 can be compared, and it is also possible to take into consideration that the number of parts in stock will increase before the predicted date.

[0043] After the supplier has been determined, or if there is stock at only one base (NO in step S1506), the inter-base delivery due date 3035 in the parts master table 303 shown in Fig. 9 is referenced, and an order (delivery arrangement) date is determined by calculating backwards so that the flexible part will arrive on the predicted failure occurrence date 3013 in the failure prediction management table 301, and this is recorded (or updated) in the order table 305 in Fig. 11 (step S1508). Also, according to the above process, the ordering status 3018 in the failure prediction management table 301 and the stock quantity 3045 in the inventory management table 304 are updated (step S1509).

[0044] Then, it is confirmed whether the above-mentioned processing has been completed for all predicted failure IDs, and if not (NO in step S1510), "i" is incremented by "1" (step S1503), and the same processing is performed for the next predicted failure ID. If the processing has been completed (YES in step S1510), the inter-base accommodation reservation processing ends. (4-2-3) Order reservation processing (step S1323) Fig. 16 is a flow chart showing details of the order reservation process. As in the flow of Fig. 14, the predicted fault ID (3011) is set to "i" and the initial value "i=1" is set (step S1601). Then, for the predicted fault with the predicted fault ID (3011) of "1", the ordering status 3018 is checked to see if it is in a state where ordering has not been completed (not yet, XX units missing) (step S1602).

[0045] If arrangements have been made (NO in step S1602), "i" is incremented by "1" (step S1603), and the same process is performed for the next predicted failure ID.

[0046] If the arrangement has not been made (YES in step S1602), the location where the ATMID (3012) related to the predicted failure is located is identified from ATM master table 302 shown in Fig. 8, and further the necessary parts 3016 and the necessary quantity 3017 are confirmed, and the standard delivery date 3034 in parts master table 303 shown in Fig. 9 is referenced, and an order date is determined by calculating backwards so that the flexible parts will arrive on predicted failure occurrence date 3013 in failure prediction management table 301, and this is recorded (or updated) in order table 305 in Fig. 11 (step S1604). Also, according to the above process, the arrangement status 3018 in failure prediction management table 301 is updated (step S1605).

[0047] Then, it is confirmed whether the above-mentioned process is completed for all predicted failure IDs, and if not completed (NO in step S1606), "i" is incremented by "1" (step S1603), and the same process is performed for the next predicted failure ID. If completed (YES in step S1606), the order reservation process ends.

[0048] If the failure prediction process (step S1310) by the failure prediction device 200 is periodically performed and, for example, the result is different from the previous prediction process, and the scheduled maintenance parts are no longer needed, if the order has not yet been placed, the order reservation can be canceled and the various tables can be updated accordingly. If the order has already been placed, the parts can be secured as inventory for future maintenance. (5) Example of maintenance parts management ordering process Based on the above-mentioned processing flow, an example of maintenance parts management ordering processing will be explained, focusing on the transition of the contents of each table. Here, it is assumed that at the time when the failure prediction processing is completed by the failure prediction device 200 (hereinafter referred to as the initial state), the failure prediction management table 301 has the contents shown in Fig. 7. The ATM master table 302 and the parts master table 303 have the contents shown in Fig. 8 and Fig. 9, respectively. The inventory management table 304 has the contents shown in Fig. 10 in its initial state. Note that the ordering table shown in Fig. 11 and the parts allocation table 306 shown in Fig. 12 are all blank in their initial state. (5-1) Parts inventory acquisition process (5-1-1) About predicted fault ID=1 The case where the reserved failure ID (3011) in the failure prediction management table 301 shown in FIG. 7 is "1" will be considered.

[0049] It is understood that the maintenance base for ATM789 (ATMID(3012)=ATM789) is base D, and that the maintenance parts required are 2 units of part 3 and 24 units of part 2. By referring to inventory management table 304, it is understood that there is no inventory of part 3 at base D, but 40 units of part 2 are in stock. Therefore, it is decided to secure 24 units of part 2 at base D.

[0050] As a result of the above, part 2 corresponding to reservation problem ID=1 has been secured, so the ordering status in the table is updated from "Not yet" to "Secured." Additionally, the inventory quantity of part 2 at site D in inventory management table 304 is updated from "40" to "40-24=16." Furthermore, the parts allocation table 306 is updated so that 24 units of part 2 are secured at site D. (5-1-2) About predicted fault ID=2 The maintenance base for ATM 567 is base C, and the maintenance parts required are 2 units of part 1 and 36 units of part 2. By referring to inventory management table 304, it is determined that neither part is in stock at base C and cannot be secured. Furthermore, none of failure prediction management table 301, inventory management table 304, and part securing table 306 are updated. (5-1-3) About predicted failure ID=3 The maintenance base for ATM 123 is base A, and the predicted recovery process requires (1) module replacement and (2) board replacement. To replace the module, 2 units of part 1 and 36 units of part 2 are required, and by referring to inventory management table 304, it is determined that both parts are in stock at base A, so it is decided to secure those parts.

[0051] Four pieces of part 4 and 40 pieces of part 2 are required to replace the board. Referring to inventory management table 304, site A has no inventory of part 4 and 65 pieces of part 2 in stock, but because 36 pieces have already been reserved for the above-mentioned (1) module replacement, 65-36=29 pieces are in stock. Therefore, 29 pieces are reserved out of the required number of 40 pieces. In other words, there is a shortage of 11 pieces.

[0052] As a result of the above, in failure prediction management table 301, since both parts 1 and 2 for module replacement have been secured, the ordering status in the table is updated from "Not yet" to "Secured." Also, since part 4 for board replacement has not been secured, the ordering status is not updated, but for part 2, 29 of the required number of 40 have been secured, leaving a shortage of 11, so the ordering status is updated from "Not yet" to "11 missing."

[0053] Also, the inventory management table 304 is updated so that the inventory quantity of part 1 at site A is updated from "3" to "1", and the inventory quantity of part 2 is updated from "65" to "0". Furthermore, the part allocation table 306 is updated so that 2 units of part 1, 36 units of part 2 (for module replacement), and 29 units of part 2 (for board replacement) are allocated at site A. (5-1-4) About predicted fault ID=4 The maintenance base for ATM789 is base D, and the maintenance parts required are 2 units of part 4 and 20 units of part 2. Referring to inventory management table 304, there is no inventory of part 4 at base D, and although 40 units of part 2 were initially in stock, the above process has resulted in 16 units being in stock (a shortage of 4 units).

[0054] Therefore, it is decided to secure 16 units of part 2 at site D, and the procurement status of part 2 in failure prediction management table 301 is updated from "not yet" to "missing 4 units." Also, the inventory quantity of part 2 at site D in inventory management table 304 is updated from "16" to "0," and the parts allocation table 306 is updated to indicate that 16 units of part 2 are secured at site D.

[0055] When the above parts allocation processing is completed, the failure prediction management table 301, inventory management table 304 and parts allocation table 306 are updated to the contents of failure prediction management table 301A, inventory management table 304A and parts allocation table 306A shown in FIG. (5-2) Inter-base transfer processing (5-2-1) About predicted fault ID=1 17, for predicted failure ID=1, two units of part 3 are required at site 4, but site 4 has not been able to secure them, so it is examined whether they can be borrowed from another site. According to inventory management table 304A, it is known that the required number of units of part 3 are in stock at both site 1 and site 2. Therefore, it is examined from which site they should be borrowed.

[0056] Here, since the frequency of use of part 3 at base 1 is "high," it is decided not to accommodate from base 1, but to accommodate from base 2. Then, by referring to inter-base delivery due date 3035 in part master table 303 in FIG. 9, it is confirmed that 10 days are required as the due date, and a reservation is made to arrange delivery for December 31, 2022, calculated backwards from the due date by 10 days, so that the part will arrive on January 10, 2023, which is the predicted failure date (predicted occurrence 3013). Note that delivery may be arranged to arrive about one day to one week before the predicted occurrence date, rather than on the day of the predicted occurrence date.

[0057] As a result of the above processing, the above-mentioned order reservation details are recorded in the order table 305. Also, the ordering status of part 3 in the failure prediction management table 301A is updated from "not yet" to "reserved for exchange." Furthermore, the inventory quantity of part 3 at site 2 in the inventory management table 304A is updated from "2" to "0." (5-2-2) Predicted fault ID=2 According to the failure prediction management table 301A, for predicted failure ID=2, 2 units of part 1 and 36 units of part 2 are required at site 3, and it is examined whether these can be borrowed from other sites.

[0058] As for part 1, since there is one unit in stock at base 1, it is determined that only one unit will be lent from base 1 (one unit short), but as part 2 is not in stock at any other bases, it is determined that lent is not possible. From parts master table 303, it is confirmed that the base-to-base delivery lead time for part 1 is seven days, and a reservation is made to arrange for delivery on January 23, 2023, calculated backwards seven days from the lead time, so that part 1 will arrive on February 1, 2023, the predicted failure date (occurrence prediction 3013).

[0059] As a result of the above processing, the above-mentioned order reservation details are recorded in order table 305. Also, the ordering status of part 1 in failure prediction management table 301 is updated from "not yet" to "one missing." Furthermore, the inventory quantity of part 1 at site 1 in inventory management table 304A is updated from "1" to "0." (5-2-3) About predicted fault ID=3 We will consider whether it is possible to borrow four units of part 4 and eleven units of part 2, which base 1 has been unable to obtain, from other bases.

[0060] According to inventory management table 304, it is determined that part 4 can be transferred from base 3, but part 2 is not in stock at any of the other bases and is therefore not transferable. From parts master table 303, it is confirmed that the base-to-base delivery lead time for part 4 is five days, and a reservation is made to arrange for delivery on February 15, 2023, calculated five days backwards from the lead time, so that part 4 will arrive on February 20, 2023, the predicted failure date (occurrence prediction 3013).

[0061] As a result of the above processing, the above-mentioned order reservation details are recorded in the order table 305. Also, the ordering status of part 4 in the failure prediction management table 301 is updated from "Not yet" to "Reserved for flexibility." Furthermore, the inventory quantity of part 4 at site 3 in the inventory management table is updated from "4" to "0." (5-2-4) Predicted fault ID=4 It is considered whether two units of part 4 and four units of part 2, which cannot be procured at site 4, can be borrowed from other sites. According to the inventory management table 304, neither of the parts are in stock at any of the sites, so it is determined that they cannot be borrowed.

[0062] When the above-mentioned inter-site accommodation processing is completed, the failure prediction management table 301A and the inventory management table 304A are updated to the contents of the failure prediction management table 301B and the inventory management table 304B, as shown in Fig. 18. Also, an order table 305B is created. (5-3) Order reservation processing (5-3-1) About predicted fault ID=2 One unit of part 1 and 36 units of part 2 that cannot be procured at base 3 are ordered. The order date is calculated by counting backwards from the predicted occurrence date (February 1, 2023) and referring to the delivery date in part master table 303. Then, it is decided to make a reservation to order part 1 on December 20, 2022, and to order part 2 on January 1, 2023.

[0063] As a result of the above, the ordering status of parts 1 and 2 in failure prediction management table 301B is updated from "not yet" to "order reserved." In addition, the above order reservation for parts 1 and 2 is recorded in order table 305B. (5-3-2) About predicted failure ID=3 Similarly, 11 units of part 2 are ordered. The planned order date is January 20, 2023. The ordering status of part 1 in failure prediction management table 301B is updated from "missing 11 units" to "order reserved." The above-mentioned order reservation for part 1 is recorded in order table 305B. (5-3-3) About predicted fault ID=4 Similarly, two units of part 4 and four units of part 2 are ordered. The planned order dates are January 1, 2023 and February 1, 2023, respectively. For part 4 in failure prediction management table 301B, the ordering status is updated from "Not yet" to "Order reserved", and for part 2, the ordering status is updated from "Four units missing" to "Order reserved". The above-mentioned order reservations for parts 4 and 2 are recorded in order table 305B.

[0064] When the above inter-site accommodation processing is completed, the failure prediction management table 301B and the order table 305B are updated to the contents of the failure prediction management table 301C and the order table 305C shown in FIG. (6) Effects of this embodiment According to the embodiment described above, the failure prediction device 200 predicts the time of failure and its content, as well as the maintenance parts and quantities required to recover from the failure, and the timing of ordering is adjusted so that the maintenance parts are available at the time when they are needed by the maintenance parts management and ordering device. Since maintenance parts can be shared between maintenance bases, it becomes possible for each base to obtain the required maintenance parts at the appropriate time without delay.

[0065] In addition, by sharing maintenance parts between each base, it is possible to make effective use of the inventory of maintenance parts. Furthermore, by adjusting the timing of orders so that parts can be obtained at the time they are needed at each base, it is expected that there will be no unnecessary strain on storage space for maintenance parts. EXAMPLES

[0066] Next, a second embodiment of the present invention will be described.

[0067] In the first embodiment described above, the delivery of the maintenance parts was arranged so that they would arrive at the time when the occurrence of a failure was predicted so as not to unnecessarily consume storage space for the maintenance parts. However, if the parts were to arrive on the day of the predicted occurrence of the failure, there is a possibility that the arrival of the maintenance parts will be delayed due to a problem during delivery or that the lead time will be unexpectedly extended due to reasons of the maintenance parts manufacturer, etc., and thus the availability of the maintenance parts at the required time will be delayed and the maintenance work itself will be delayed.

[0068] Therefore, in this embodiment, when there is sufficient free space in the storage space for maintenance parts at each base, an order is immediately processed to secure the maintenance parts at that base ahead of schedule.

[0069] Fig. 20 is a diagram showing the configuration of a maintenance parts management ordering device 300A in a second embodiment of the present invention. The difference from the maintenance parts management ordering device 300 in the first embodiment shown in Fig. 6 is that a base management table 307 is added, and information on the packaging size of parts is added to the parts master table 303.

[0070] 21 shows an example of the base management table 307. It has a base ID 3071 of each base, a base name 3072, and a warehouse free capacity 3073 indicating the free space in the warehouse where maintenance parts to be used are stored. The warehouse free capacity 3073 is updated as needed according to the availability of the warehouse.

[0071] An example of a part master table 303A is shown in Fig. 22. A packing size 3036 indicating the volume of the part when packed is added to the part master table 303 in Fig. 6.

[0072] For example, looking at a module with part ID=1, the packing size 3036 is “20”, and since the warehouse space available 3073 at base A with base ID=1 is “25”, it is clear that one module part can be stored in the warehouse, but two or more cannot.

[0073] Fig. 23 is a flow chart for explaining the flow of base interchange reservation processing in this embodiment. The difference from the flow of base interchange reservation processing in the first embodiment shown in Fig. 15 is the processing from step S2318 onwards, where the processing is changed depending on the free space in the storage warehouse, and the processing before that (steps S2311 to S2317) is the same as in Fig. 15, so a description thereof will be omitted.

[0074] In step S2318 in Fig. 23, the space required for storage is calculated for the type and quantity of parts available at that base by referencing packing size 3036 in part master table 303A in Fig. 22. Then, the available space at that base is confirmed by referencing warehouse available capacity 3073 in base management table 307 in Fig. 21.

[0075] As a result, if it is determined that there is free space in the storage warehouse (YES in step S2318), the order timing is not adjusted and an order (delivery) is immediately arranged (step S2319). Then, the warehouse free space 3073 in the base management table 304 is updated accordingly (step S2320), the arrangement status 3018 in the failure prediction management table 301 is updated from "not yet" to "arranged", and the inventory management table 304 is also updated (step S2322).

[0076] If it is determined that there is no free space in the storage warehouse (NO in step S2318), similarly to steps S1508 and S1509 in the first embodiment (FIG. 5), the inter-site delivery due date 3035 in the parts master table 303A is referenced, and the order table 305C is updated to reserve the order (delivery) so that the product will arrive on the predicted failure date (step S2321).Then, the failure prediction management table 301 and the inventory management table 304 are updated accordingly (step S2322).

[0077] Then, it is confirmed whether the above-mentioned process is completed for all predicted failure IDs, and if not completed (NO in step S2324), "i" is incremented by "1" (step S2313), and the same process is performed for the next predicted failure ID. If completed (YES in step S2324), the inter-base accommodation reservation process is terminated.

[0078] Fig. 24 is a flow chart for explaining the flow of order reservation processing in this embodiment. The difference from the flow of order reservation processing in the first embodiment shown in Fig. 16 is the processing from step S2415 onwards, which is changed depending on the free space in the storage warehouse.

[0079] That is, in step S2415 in Fig. 24, the space required for storage is calculated for the type and quantity of parts lent at that base by referencing packing size 3036 in part master table 303A in Fig. 22. Then, the available space at that base is confirmed by referencing warehouse available capacity 3073 in base management table 307 in Fig. 21.

[0080] As a result, if it is determined that there is free space in the storage warehouse (YES in step S2415), the order is immediately arranged without adjusting the order timing (step S2416). Then, the warehouse free space 3073 in the base management table 304 is updated accordingly (step S2417), and the arrangement status 3018 in the failure prediction management table 301 is also updated from "Not yet" to "Order arranged" (step S2419).

[0081] If it is determined that there is no free space in the storage warehouse (NO in step S2415), similarly to steps S1604 and S1605 in the first embodiment (FIG. 6), standard delivery date 3035 in part master table 303A is referenced, and order table 305C is updated to reserve the order so that the part will arrive on the predicted failure date (step S2418). Then, failure prediction management table 301 is updated accordingly (step S2419).

[0082] Then, it is confirmed whether the above-mentioned process is completed for all predicted failure IDs, and if not completed (NO in step S2420), "i" is incremented by "1" (step S2413), and the same process is performed for the next predicted failure ID. If completed (YES in step S2420), the order reservation process ends.

[0083] As described above, according to this embodiment, when there is sufficient free storage space for maintenance parts at each base, an order is processed immediately and the maintenance parts are secured at that base ahead of schedule. Therefore, even if delivery of the maintenance parts takes longer than expected due to a problem during delivery or the lead time is unexpectedly extended due to reasons of the maintenance parts manufacturer, etc., the maintenance parts can be obtained reliably at the required time. [Explanation of symbols]

[0084] 10: Predictive maintenance system 100: Region (Area) 111:ATM123 112:ATM345 113:ATM567 114:ATM789 121: Base A 122: Base B 123: Base C 124: Base D 130: Maintenance Center 200: Fault prediction device 300: Maintenance parts management and ordering device

Claims

1. A failure prediction device that predicts a failure of an automated transaction device; a maintenance parts management and ordering device which prepares parts necessary for failure recovery in advance for a maintenance base which is provided for each of a plurality of regions and which performs maintenance on the automated transaction apparatus installed in each region; 1. A predictive maintenance system comprising: The maintenance parts management and ordering device includes: an inventory part securing process for securing the maintenance part at the maintenance base in a region in which the maintenance part for the automated transaction device predicted to have a failure is in stock; When the maintenance parts for the automated transaction device predicted to fail are not in stock at the maintenance base in the region and unused stock exists at another maintenance base, an inter-base lending process is performed to lend the maintenance parts in stock at the other maintenance base to the maintenance base in the region so that they can be obtained on the date of the predicted failure; an order reservation process for ordering the shortage of maintenance parts for the automated transaction device predicted to have a failure so that the parts can be obtained on the due date of the predicted failure, when the parts are not in stock at any of the maintenance bases in any of the regions; A predictive maintenance system comprising:

2. 2. The predictive maintenance system of claim 1, The maintenance parts management and ordering device includes: an ATM failure prediction management table for managing a predicted date of failure, a predicted type of failure, a predicted recovery process, necessary parts, the number of parts, and a status of arrangement for each of the automated transaction machines for which a failure is predicted; an inventory table in which inventory of maintenance parts for each maintenance base is recorded; A predictive maintenance system comprising:

3. 3. The predictive maintenance system of claim 2, a parts master table in which the order delivery date and the inter-site interchange delivery date are recorded for each of the maintenance parts; A predictive maintenance system, characterized in that the inter-base interchange processing and the order reservation processing are performed by referring to the parts master table and calculating backwards from the predicted failure date of the automated transaction device for which a failure is predicted.

4. 4. The predictive maintenance system of claim 3, The predictive maintenance system is characterized in that, in the case where a maintenance part for the automated transaction device for which a failure is predicted is not in stock at a maintenance base in that region and there is unused stock at multiple other maintenance bases, the maintenance part is lent from either the maintenance base with the largest amount of stock, or a maintenance base with a low future possibility of use for the maintenance part, or a maintenance base that has other maintenance parts to lent at the same time.

5. 5. The predictive maintenance system of claim 4, the maintenance parts management and ordering device has a base table in which the free space of the storage warehouse for each of the maintenance bases is recorded; The part master table records the packaging size of each part, A predictive maintenance system, characterized in that the inter-base interchange processing and the order reservation processing are immediately instructed when the packaging size of the maintenance parts to be prepared is less than the available capacity of the storage warehouse.

6. A predictive maintenance method for predicting a failure of an automated teller machine, and preparing parts necessary for failure recovery in advance at maintenance bases that are provided in a plurality of regions and that perform maintenance on the automated teller machine installed in the regions, comprising: If the maintenance parts for the automated transaction device predicted to fail are in stock at the maintenance base in the region, the maintenance parts are secured at the maintenance base; When the maintenance parts for the automated transaction device for which the failure is predicted are not in stock at the maintenance base in the region and unused stock exists at another maintenance base, the maintenance parts in stock at the other maintenance base are lent to the maintenance base in the region so that they can be obtained on the date of the failure prediction, a predictive maintenance method for predicting whether a maintenance part for an automated transaction device that is predicted to fail is not in stock at any of the maintenance bases in any region, the method comprising the steps of: ordering the shortage part so that it can be obtained on the due date of the predicted failure.

Images