JP2020019378A - Prediction method and prediction system for facility deviation amount - Google Patents

Abstract

To provide a prediction method and prediction system for facility deviation amount which can enable advance adjustment of a facility deviation amount before the occurrence of a larger than expected deviation.SOLUTION: A method for predicting a deviation amount of a facility that expands and contracts, in which more accurate predictions of facility deviation amounts after the current state are made by predicting a degree of change in the facility deviation amount from the current state based on accumulated data from the past to the present and environmental prediction data on the external environment of the facility. For example, a timing to adjust the facility deviation amount is notified before a lock deviation or the like occurs, with a point taken as the facility.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a method and system for predicting the amount of equipment shift in equipment such as a point machine.

  For example, in a switching machine as an example of equipment installed outdoors or the like, a lock piece is incorporated in a lock can as illustrated in, for example, Patent Document 1, but the lock piece and the lock can are used. If the shift of the notch (lock shift amount) becomes large, the lock piece will not be able to enter the notch of the lock can, causing lock failure, making it impossible to convert and preventing train operation. is there. Therefore, it is necessary to adjust the lock deviation amount before the lock disorder occurs. For this reason, for example, in Patent Document 1, the temperature is estimated to cope with a temperature change of a member which is one of the causes of the lock failure, and the temperature near the member is adjusted from the estimated temperature so as to be kept constant. You are. In addition, in the cited document 1, for the estimation of the temperature, the weather data near the track on which the points are installed is used.

  However, for example, there may be a case where the degree of change of the lock shift amount differs depending on the season even at the same temperature. Therefore, even if the temperature is kept constant using the weather data as in the case of Patent Document 1, this alone will increase the amount of lock deviation in the point machine and cause lock failure. Is not always possible. In particular, as described above, in the method using only weather data obtained by sensors provided near the points, even if the temperature at that time can be analyzed, the trend of the amount of lock deviation after that It is not always possible to make predictions about That is, in the case of Patent Literature 1, it is not always possible to know in advance the timing at which the lock shift amount becomes large enough to cause lock failure. In addition, the above-mentioned thing may become a problem not only in a switch but in various facilities similarly.

Japanese Patent No. 5806627

  The present invention has been made in view of the above points, and relates to an equipment shift amount in equipment installed outdoors or the like, for example, in a point machine, before a lock shift amount becomes large and lock disorder occurs. It is an object of the present invention to provide a method and system for predicting the amount of equipment shift for enabling advance adjustment such that the amount of lock shift can be adjusted in advance.

  The method of predicting the amount of equipment shift to achieve the above object is a method of predicting the amount of shift of equipment that expands and contracts, and includes accumulated data from the past to the present and environmental prediction data on the external environment of the equipment. , The degree of change of the equipment shift amount is predicted from the current state.

  In the above method of predicting the amount of equipment deviation, when predicting the degree of change in the amount of equipment deviation from the current state, based on accumulated data from the past to the present, it is simply information on the temperature at that time. In addition to this, it is possible to make predictions in consideration of, for example, differences in the degree of change in the amount of equipment shift depending on the season. In addition, based on the environment prediction data on the external environment, it is possible to make predictions in consideration of information about the equipment around the future, that is, after the present. As a result, a more accurate prediction of the amount of equipment deviation after the current state is expected, and for example, it is possible to notify a timing to adjust the amount of equipment deviation before a lock failure or the like occurs in a point machine. become.

  In a specific aspect of the present invention, weather forecast data is included as environment forecast data. In this case, it is possible to make predictions from the present onward based on the weather forecast data.

  In another aspect of the present invention, the accumulated data includes data on the amount of solar radiation with respect to the external environment of the facility. In this case, the degree of change in the equipment shift amount can be predicted based on the amount of solar radiation from the past to the present, that is, the amount of radiant energy from the sun.

  In still another aspect of the present invention, date information is included as accumulated data as an element for predicting the degree of change in the facility shift amount. In this case, it is possible to make a prediction in consideration of the difference in the degree of change in the equipment shift amount depending on the season using the information on the date.

  In still another aspect of the present invention, the accumulated data includes data on weather changes from the past to the present. In this case, the degree of change in the facility shift amount can be predicted based on weather changes from the past to the present.

  In still another aspect of the present invention, the accumulated data includes the shift of the equipment from the past to the present. In this case, the degree of change in the equipment shift amount can be predicted based on the equipment shift progress from the past to the present.

  In still another aspect of the present invention, equipment deviation due to artificial adjustment is excluded from accumulated data. In this case, the accuracy of the prediction can be increased by removing the elements that make the prediction difficult.

  In still another aspect of the present invention, the accumulated data includes a difference in the amount of equipment shift before and after the train passes through the equipment installed near the train passing position. In this case, prediction accuracy can be improved by taking into account the difference between before and after the train passes.

  In still another aspect of the present invention, an algorithm for predicting based on a difference between the comparison results is compared with a prediction result and a measurement result of the degree of change in the equipment shift amount. In this case, a more accurate prediction can be made by modifying the algorithm.

  In still another aspect of the present invention, the degree of change of the lock shift amount in a switch as equipment is predicted. In this case, it is possible to predict the amount of lock deviation in the point machine.

  The equipment shift amount prediction system for achieving the above object is a system for predicting the shift amount of equipment that expands and contracts, and includes a stored data management unit that manages stored data from the past to the present, and an external device outside the equipment. An environment prediction data acquisition unit that acquires environment prediction data about the environment; and a facility shift amount prediction unit that forecasts a degree of change in the facility shift amount from the current state based on the accumulated data and the environment forecast data. Prepare.

  In the above-described equipment shift amount prediction system, when predicting the degree of change in the subsequent equipment shift amount from the current state, based on the accumulated data from the past to the present managed by the accumulated data management unit, the The equipment shift amount prediction unit can perform not only information on the temperature at the time, but also a prediction in consideration of, for example, a difference in the degree of change in the equipment shift amount depending on the season. Further, based on the environment prediction data on the external environment acquired by the environment prediction data acquisition unit, the equipment shift amount prediction unit can make a prediction in consideration of information on the surroundings of the equipment in the future, that is, after the present. As described above, a more accurate prediction is expected for the amount of equipment deviation after the current state. For example, it is possible to notify a timing for adjusting the amount of equipment deviation before a lock failure or the like occurs in a point machine. become.

  In a specific aspect of the present invention, the equipment shift amount estimating unit estimates the degree of change in the lock shift amount in a point machine as equipment. In this case, it is possible to predict the amount of lock deviation in the point machine.

FIG. 2 is a plan view conceptually showing a point machine equipped with the equipment shift amount prediction system according to the first embodiment. It is a perspective view which shows an example of a structure about the switch in a switch. It is a perspective view showing an example of 1 composition about a lock piece and a lock can incorporated into a point machine. (A) and (B) are conceptual diagrams for explaining a lock shift amount. It is a conceptual diagram for explaining an example about detection of the lock deviation amount. It is a block diagram showing an example of composition about a prediction system in a turning machine. It is a conceptual diagram for explaining the analysis technique in a prediction system. It is a graph which shows an example about prediction of the lock shift amount based on a prediction system. It is a data table which shows an example about various data used for prediction of the lock shift amount based on a prediction system. It is a flow chart for explaining an example about a prediction method in a prediction system. 11 is a flowchart for explaining an example of an algorithm correction using the prediction created in FIG. 10. (A) is a graph showing an example of the relationship between the prediction and the actual lock shift amount in the prediction system, and (B) is a graph showing a comparative example of (A). It is a block diagram showing an example of 1 composition about a prediction system concerning a 2nd embodiment.

[First Embodiment]
Hereinafter, a system for predicting the amount of equipment shift according to the first embodiment will be described with reference to FIG. 1 and the like.

  As illustrated in FIG. 1 and FIG. 2, the equipment shift amount prediction system 100 of the present embodiment is a system for predicting the shift amount of one of the facilities that expands and contracts. The equipment for which the amount of shift is to be predicted is the point machine 50, more precisely, a part of the point machine 50. In particular, in the present embodiment, as a specific example, it is assumed that the shift amount of the equipment that undergoes expansion and contraction to be predicted is the lock shift amount of the point machine 50. Here, the lock shift amount means a shift amount of a portion where the lock piece is incorporated into the lock can, among the portions that expand and contract as a part of the point machine 50. The internal structure of each part of the point machine 50 will be described later in detail.

  As described above, the prediction system 100 according to the present embodiment predicts the degree of change of the lock shift amount. In particular, the prediction system 100 stores the accumulated data from the past to the present and the environment of the external environment of the switch 50. Based on the prediction data, the degree of change of the shift amount of the point machine 50 from the current state is predicted with higher accuracy. In order to enable the prediction, in the present embodiment, the prediction system 100 is provided inside the main body 50 </ b> A that constitutes the point machine 50, and performs a lock shift process for grasping the situation inside and performing various adjustments. And a prediction processing unit 60 provided outside the main body unit 50A to obtain information from the lock shift processing unit 40 and information on the external environment of the switch 50 and perform a prediction process based on these. It is composed of

  Hereinafter, as a premise for describing the prediction system 100, first, a configuration example of the target switching machine 50 will be described.

  In an example here, the point machine 50 is an electric point machine, and by moving the tongue rail TR among the basic rail SR and the tongue rail TR on the sleeper SP constituting the branching device P, This is a member that switches the course of the train in the branching device P.

  In order to perform the switching operation, the switching machine 50 includes a main body 50A having a motor or the like as a power source, an operation can 10 extending from the main body 50A, and a tong rail connected to the operation can 10 for example. A connection portion 20 connected to the TR or a lock can 30 for fixing the position after the switching operation is provided. In addition, the point machine 50 has the lock shift processing unit 40, so that it is possible to monitor a temporal change in the lock shift amount.

  As shown in the figure, the main body 50A is placed on one end of the first and second sleepers SP1 and SP2, which are longer than the normal sleepers SP arranged at a predetermined interval, so that a line branching device is provided. It is arranged on the side of P, and is fixed so as not to move with respect to the sleepers SP1 and SP2 by a fixing device or the like. The main body 50A has a motor, a power device for moving the motor, and the like, and generates a conversion force when the tong rail TR is moved by moving the operation can 10 back and forth. In addition, when the motor is not driven, a structure for enabling a manual switching operation may be provided.

  The operation can 10 is a plate-shaped or rod-shaped metal member extending from the inside of the main body 50A, and moves forward and backward in a horizontal direction perpendicular to the traveling direction of the train by the rail with the driving operation by the main body 50A. More specifically, the operation can 10 changes the tong rail TR from the normal position to the opposite position or from the opposite position to the normal position with respect to the basic rail SR.

  The connecting portion 20 has a configuration having other connecting devices while using the metal switch adjuster rod 21 as a main component, and the tongue rail TR is connected to the operation can 10 by the switch adjuster rod 21. It functions as a member for transmitting the force from the operation can 10 for moving. More specifically, the connecting portion 20 is connected to the operation can 10 at the distal end side of the switch adjuster rod 21, and is connected to the tong rail TR at the other end provided with another member connected to the switch adjuster rod 21. Thus, the conversion force from the operation can 10 is transmitted to the tong rail TR.

  As described above, the lock can 30 is a member for fixing the position after the switching operation. In other words, after the changeover by the operation can 10, the turning machine 50 moves the operation can 10 to the stereotactic position and the inverted position by the action on the operation can 10 accompanying the operation of the lock can 30 inside the main body 50 </ b> A. Each can be locked. For this reason, the lock can 30 is also a plate-like or rod-shaped metal member extending from the inside of the main body 50A, and a connection can 31a is provided on the tip side of the lock can 30. A front rod 31b is provided on the side. The front rod 31b is connected to the distal end side of the tongue rail TR. As the tongue rail TR is moved by the operation can 10, the lock can 30 also operates.

  Hereinafter, the structure and operation of the lock can 30 and the like will be described more specifically with reference to FIG. FIG. 3 is a perspective view showing an example of the configuration of the lock can 30 incorporated in the switch machine 50 and a pair of lock pieces LPa and LPb inserted into the cutouts NTa and NTb provided in the lock can 30. is there.

  The pair of lock pieces LPa and LPb are a pair of rod-shaped members that are advanced and retracted by a cam bar that is slidable in a horizontal direction perpendicular to the movement of the operation can 10. Locking blocks LBa, LBb, one of which protrudes downward and the other protrudes upward, are provided at the distal ends of the pair of lock pieces LPa, LPb.

  On the other hand, the lock can 30 is configured, for example, with a pair of upper and lower rod-shaped members 30a and 30b for a portion provided inside the main body 50A. The pair of bar-shaped members 30a and 30b have a notch NTa that opens upward in the lower bar-shaped member 30a and a notch NTb that opens downward in the upper bar-shaped member 30b at positions separated by a predetermined distance in the longitudinal direction. Are formed.

  With the above-described configuration, in the point machine 50, for example, the motor (not shown) of the point machine 50 rotates to operate the operation rod 10 to change from the reverse position to the normal position, and when the change is completed, The lock block LBa of the lock piece LPa is locked by fitting into the cutout NTa of the corresponding lock can 30. When switching from the normal position to the reverse position, the lock block LBb of the other lock piece LPb is locked by fitting into the cutout NTb of the corresponding lock can 30. FIG. 3 illustrates a state in which the lock block LBb is fitted into the cutout NTb.

  Here, in order to maintain the above operation, it is necessary that the positions of the cutouts NTa and NTb corresponding to the lock blocks LBa and LBb of the lock pieces LPa and LPb do not become out of order. . More specifically, as shown in FIG. 4A, for example, the gap DT on both sides of the lock piece LPa (LPb) and the notch NTa (NTb) at the reference position is about 1.5 mm. Therefore, when a gap occurs and the gaps are DT1 (= DT + α) and DT2 (= DT−α), for example, as shown in FIG. Can not be locked.

  In the present embodiment, in view of the above, for example, a normal lock cannot be performed by performing a highly accurate prediction on the shift amount between the respective parts in the lock, that is, the lock shift amount, in the prediction system 100. It is possible to take measures such as urging the user to perform maintenance before the situation occurs.

  It should be noted that the measurement of the amount of deviation when locking with the lock piece as described above, that is, the detection of the size of the gap, can be achieved by providing various sensors. Here, as illustrated in FIG. 5, the change in the distance to the position of the measuring piece 32 attached to the lock can 30 is measured by a linear sensor 54 fixed to the main body 50 </ b> A side, whereby the lock shift is performed. The amount can be calculated.

  Hereinafter, an example of a prediction method by the prediction system 100 in the switch 50 according to the present embodiment will be described with reference to FIG. 6 and the like. As described above and as illustrated, the prediction system 100 according to the present embodiment includes the lock shift processing unit 40 and the prediction processing unit 60.

  First, the lock shift processing unit 40 in the prediction system 100 will be described. The lock shift processing unit 40 is a processing unit for confirming the state of the lock shift. As the prediction system 100, the lock shift processing unit 40 indicates the components of the switch 50 and the situation in the switch 50. It functions as acquiring various types of information and transmitting the acquired information from the switch-side transmitting / receiving section 40s to the prediction processing section 60. In addition, about various processing operation | movement, you may receive the instruction | indication from the prediction processing part 60 via the switching machine side transmission / reception part 40s, and may perform it according to it.

  The lock shift processing unit 40 acquires information on various physical quantities in the switch 50 in order to confirm the state of the lock shift, and performs various processes for monitoring the lock shift based on the information. I do.

  In this example, the lock shift processing unit 40 includes a lock shift control unit 41, a lock shift storage unit 42, a temperature data processing unit 43, A vibration data processing unit 44;

  The lock shift control unit 41 receives output voltages from the N-side sensor 50N and the R-side sensor 50R provided for the lock pieces LPa and LPb (see FIG. 3) of the point machine 50, respectively. From the operation of the output voltages of the two N-side sensors 50N and the R-side sensor 50R, it is determined when the switching operation of the point machine 50 is completed, and monitoring is performed by detecting the lock shift amount at that time.

  The lock shift storage unit 42 stores the lock shift amount at the time of monitoring by the lock shift control unit 41 described above.

  The temperature data processing unit 43 transmits / receives to / from the temperature sensor TS provided in the switch 50, and monitors the temperature data around the switch 50 together with the current value data of the lock shift amount by the temperature sensor TS. Thereby, for example, the acquired temperature data can be used as weather data.

  The vibration data processing unit 44 transmits / receives to / from the vibration sensor BS provided in the switch 50, and monitors the vibration data around the switch 50 with the vibration sensor BS together with the current value data of the lock shift amount. Thereby, for example, the acquired vibration data can be used as train operation data. More specifically, when a vibration value equal to or larger than a predetermined threshold value is output, it is possible to determine that a train has passed.

  The switch-side transmitting / receiving section 40s is an interface section of the lock shift processing section 40, and transmits the data on the switch 50 acquired by each of the above sections to the prediction processing section 60. Note that the prediction processing unit 60 handles the information received via the switch-side transmitting / receiving unit 40s as accumulated data regarding the switch 50.

  Hereinafter, the prediction processing unit 60 of the prediction system 100 will be described. The prediction processing unit 60 acquires various information from the outside as necessary in addition to the information from the lock shift processing unit 40, and predicts the lock shift amount based on the obtained information. For this reason, in an example here, the prediction processing unit 60 includes, for example, a monitor display unit 62 in addition to the prediction processing transmission / reception unit 60s and the lock shift amount prediction processing unit 61 which is a main part.

  The lock shift amount prediction processing unit 61 includes a CPU and the like, and includes a main control unit 61a that performs various processes for calculating a lock shift amount, and a database that stores and manages various data, including a storage device and the like. It has a DB and functions as a facility shift amount prediction unit. In particular, here, in the database DB, a stored data management unit AD that manages stored data from the past to the present, and an environment prediction data management unit PD that manages environment prediction data regarding the external environment of the switch 50. Prepare.

  The prediction processing side transmission / reception unit 60s is an interface unit of the prediction processing unit 60, and acquires various information from outside. As described above, the prediction processing-side transmitting / receiving section 60s acquires data on the point 50 as accumulated data via the point-side transmitting / receiving section 40s. Accept environmental forecast data about the external environment. More specifically, weather forecast data is received as environment forecast data. For example, since the prediction processing unit 60 is connected to the network via the prediction processing transmission / reception unit 60s, information on weather forecast near the location where the switch 50 is installed can be obtained. .

  The prediction processing transmitting / receiving unit 60 s also receives information indicating a predetermined fact such as weather data from the past to the present to be stored data from other than the lock shift processing unit 40 as necessary.

  In the database DB that constitutes the lock shift amount prediction processing unit 61, the accumulated data management unit AD manages various types of information related to the past to present as accumulated data. Although various data can be the object of the accumulated data, here, at least, data relating to the amount of solar radiation with respect to the external environment of the switch 50, data of weather change from the past to the present, and switches from the past to the present. It is assumed that the data includes the data on the progress of the lock shift of the machine 50, the data on the equipment shift due to the artificial adjustment, and the date of the various data. In particular, here, the date information is included as an element for predicting the degree of change in the equipment shift amount, that is, the lock shift amount. Specifically, for example, based on January 1, a change in the degree of sunlight energy according to the passage of the day and time is set, and the set change value is used as an element for prediction. Can be

  On the other hand, in the database DB, the environment prediction data management unit PD manages weather forecast data as environment prediction data on the external environment of the switch 50. That is, it manages future forecast data. In addition, the environment prediction data management unit PD functions as an environment prediction data acquisition unit that acquires environment prediction data on the external environment of the switch 50 in cooperation with the prediction processing side transmission / reception unit 60s. Become.

  As described above, the database DB stores various information regarding the past to the present in the accumulated data management unit AD, and stores various information regarding the future in the environment prediction data management unit PD.

  As shown conceptually in FIG. 7, the lock shift amount prediction processing unit 61 first stores the lock shift amount, weather data, date data, operation data, work data, and the like as various types of information from the past to the present. The database DB stores and manages not only data but also environmental prediction data relating to future predictions such as weather forecast data and data relating to the surroundings of the turning point 50. Then, the lock shift amount prediction processing unit 61 analyzes these data by the main control unit 61a as an analyzer, and predicts the degree of change of the lock shift amount from the current state based on these. are doing. That is, a future lock shift amount is predicted. As a specific calculation method (calculation algorithm) for prediction in the main control unit 61a as an analyzer, various methods such as a multivariate analysis and a neural network can be used. As will be described later, the calculation method (algorithm) to be used may be modified as needed to improve the accuracy. In the above case, for example, the processing result of the prediction is transmitted to the outside of the prediction processing unit 60, particularly to the outside of the switch 50 such as the station device, via the prediction processing transmission / reception unit 60s. You may.

  In addition, in the example of FIG. 7 and the following, the date data is information relating to year, month and day and hour and minute, and means, for example, data including information of month and day. In this case, the date data is an index indicating a change in the degree of sunlight energy. The operation data refers to data relating to the operation of the train passing through the switch 50, that is, data relating to before and after the train passes through the switch 50. In addition, the work data means data relating to a lock adjustment work performed artificially by a worker.

  FIG. 8 is a graph showing an example of the prediction of the lock shift amount based on the prediction system 100. The horizontal axis of the graph is time, and the vertical axis is temperature or lock shift. Further, in the vertical axis direction, the upper limit line U1 and the lower limit line B1 indicate the limit of the amount of lock deviation, and if this line is exceeded, lock failure (unchangeable) may occur. Here, the upper and lower limits are shown at ± 1.5 mm in accordance with the above example.

  Based on the above assumptions, here, actual data up to the day, that is, actual various data, is used as accumulated data from the past to the present, and the next day's weather forecast (temperature forecast) is used as environmental forecast data relating to the future. A prediction and an example of the lock shift amount of the next day are considered and illustrated.

  For example, a solid curve T1 indicates a change in the actual temperature on the day, and a broken curve T2 following the curve T1 indicates a change in the expected temperature on the next day. On the other hand, the solid line curve C1 shows how the actual lock shift amount changes on the day, and the area D2 having a band-like width following the curve C1 shows the change in the expected lock shift amount on the next day. . That is, the location of the region D2 indicates the prediction result in the prediction system 100. For the belt-shaped region D2, for example, it is considered that an assumed curve following the curve C1 is calculated by various calculation methods, and the assumed curve is created with a certain width in the ± direction regarding the lock shift amount. Can be In this case, for example, the width in the ± direction (vertical width) is set to ± 0.25 mm in consideration of the fact that the allowable deviation amount is about 1.5 mm in the first place.

  In the case of FIG. 8 shown on the premise of the above, in the vicinity where the temperature is expected to be the highest on the next day, the prediction range shown in the area D2 exceeds the upper limit +1.5 mm, and it is impossible to convert (lock failure). It indicates that there is a possibility. By performing such a prediction on the same day, it is possible to notify the worker before the lock failure occurs on the next day, and to adjust the lock shift amount and repair the point machine 50 in advance.

  FIGS. 9A to 9C are data tables showing examples of various data used for predicting the lock shift amount based on the prediction system 100 as described above. For example, in FIG. 9A, the date data (including month and day data and even hour and minute) and the number of days from January 1 as a reference date (for example, on May 1, the number of days is 120 days ), Actual lock deviation amount data generated in the switch 50, or weather data, that is, data of weather change from the past to the present which is a result of actual occurrence, a forecast of the previous day, and the like. are doing. Of these, the number of days is defined as the number of days counted from January 1 as a reference date (counting start date) (120 days on May 1). In the case of the number of days counted as the start date (120 days on October 29), even the same 120 days can be regarded as different numerical data. More specifically, the amount of radiant energy from the sun, such as the difference in seasonal differences, that is, whether it is from winter to spring or from summer to autumn, and the change in southern hardness due to this, etc. It is possible to make predictions taking into account matters related to a certain amount of solar radiation. Thus, these data can be used in various ways in creating one prediction model.

  Further, as shown in FIG. 9 (B), it is conceivable that the data includes a lock adjustment execution date. That is, it is conceivable to include data of equipment deviation due to artificial adjustment, such as a work day on which a lock adjustment was performed by an operator. As a specific use, for example, it is conceivable to exclude the lock adjustment execution date when making the prediction. This is because the artificial work is different from the occurrence of the disorder caused by a naturally occurring matter, and if such data is included, it is possible that an appropriate prediction cannot be made. That is, it is considered that the accuracy of the prediction can be improved by eliminating such elements that make the prediction difficult. In other words, these data can be used, for example, when correcting the initially created prediction model to improve its accuracy.

  Further, as shown in FIG. 9 (C), it is conceivable that the data includes the passing time of the train at the point machine 50, that is, the operation data. When the train passes, a large force is applied to the entire switch P including the switch 50. Therefore, it is considered that the influence on the lock shift amount is very large. Therefore, for example, it is expected that a more accurate prediction can be made by performing analysis once separately before and after passing through the train and then performing comprehensive calculation.

  Note that the above is an example, and the prediction may be performed including elements other than the above, or the prediction may be performed without using a part of the above. Note that among the above, for example, the elements surrounded by a broken line Q1 shown in FIG. 9A are considered to be particularly fundamental and essential.

  Hereinafter, an example of the prediction method according to the present embodiment will be described with reference to the flowchart of FIG. 10 and the like.

  First, FIG. 10 is a flowchart for describing an example of creation of a prediction model among prediction methods in the prediction system 100. That is, in FIG. 10, a prediction model based on past data or future (next day) forecast data is created. As a result, it is possible to obtain information on the future lock shift amount prediction by the prediction system 100 as illustrated in the graph of FIG.

  First, as shown in FIG. 10, in the prediction system 100, accumulated data from the past to the present regarding the lock shift amount and the like, and environmental prediction data which is prediction data about the future such as weather forecast data are input (step). S101). Next, the main control unit 61a of the prediction system 100 checks whether or not work data for lock adjustment exists in the past in the database DB (step S102), and if it exists (step S102: Yes), excludes this. (Step S103a). On the other hand, if there is no lock adjustment work data in step S102 (step S102: No), localization and inversion are performed in a short time period during the lock adjustment work, based on how the lock deviation amount changes with time. It is checked whether there is a large number of changes in the lock shift amount in a short time because of the repetition, and if there is such data, it is excluded for that day (step S103b). This is because even if there is no lock adjustment work data, such a change in the lock deviation amount is considered to have actually been caused by the lock adjustment work. In step S103b, if there is no lock deviation amount change as described above, the process proceeds to the next process without performing the above-described exclusion process.

  In the processing of step S103a or step S103b, the main control unit 61a performs a necessary lock adjustment according to the excluded items. For example, if there is data of the lock adjustment amount artificially made in step S103a, the data is corrected (step S104). Further, for example, when there is no data of the lock adjustment amount in step S103b, the amount estimated from the difference between the lock deviation amounts before and after the lock adjustment is corrected (step S104).

  Next, the main control unit 61a checks whether or not operation data exists in the database DB (step S105). That is, it is confirmed whether there is a distinction between before the train passes and after the train passes. If there is operation data in step S105 (step S105: Yes), data is separated for passage (step S106). As an example here, data from immediately after the change by the switch 50 to before the train passes are separated from other data. If there is no operation data in step S105 (step S105: No), the process proceeds to the next process without performing any special process.

  Finally, for the data that has undergone the above processing, a prediction model for the lock adjustment amount as illustrated in FIG. 8 is created by appropriately using, for example, multivariate analysis, a neural network, or other various methods (step S107). ).

  In the processing in step S107, when there is a distinction between when the train passes in step S106 and when the train does not pass, a prediction model is created for both of them. In step S106, when there is no distinction between when passing and when the train does not pass, only the prediction model A is created. When there is a distinction between when passing and when the train does not pass, the prediction model A and the prediction model A are used. It is assumed that a model B is created.

  As described above, in the prediction system 100, a prediction model based on past data for performing prediction is created.

  Hereinafter, referring to the flowchart of FIG. 11, by performing verification on the prediction model created in FIG. 10, the prediction result and the actual measurement result regarding the degree of change of the lock shift amount are compared, and the comparison result is compared. Processing for correcting an algorithm for prediction based on a difference will be described.

  In FIG. 11, a predicted value is obtained by substituting an event that has actually occurred in the prediction model calculated in FIG. 10, and the obtained predicted value is compared with an actually occurring phenomenon, that is, an actual value. By confirming whether the prediction model is accurate, the validity of the prediction model created in FIG. 10 and the necessity of correction are considered.

  In this example, a prediction model is created for the next day (forecasting target day) in FIG. 10, and a prediction value obtained by performing the processing in FIG. The actual values shall be compared.

  Hereinafter, a method of calculating a predicted value from the prediction model will be described with reference to FIG. In FIG. 11, first, the main control unit 61a of the prediction system 100 adds actual weather data of the next day to the prediction model (prediction model A) completed through each step of FIG. 10 (step S201), and locks. A provisional predicted value of the shift amount (predicted value A) is obtained. Next, the main control unit 61a checks whether or not there is operation data on the prediction target date in the database DB (step S202). That is, it is confirmed whether there is a distinction between the time before the train passes and the time after the train passes on the prediction target date. If there is no operation data in step S202 (step S202: No), the main control unit 61a further checks whether or not there is work data for lock adjustment for the prediction target date in the database DB (step S203). If it does not exist (step S203: No), the prediction value A of step S201 is output as a prediction result as it is (step S204).

  If work data exists in step S203 (step S203: Yes), the work data is excluded, the provisional predicted value A is corrected with the corresponding lock adjustment amount, and the corrected predicted value A 'is calculated as the predicted result. (Step S205).

  On the other hand, in step S202, if there is operation data on the prediction target date (step S202: Yes), the data is separated between when the train passes and when the train does not pass, and not only the prediction model A but also the prediction model B is locked. A provisional predicted value of the shift amount (predicted value B) is obtained (step S206).

  After the processing in step S206, the main control unit 61a checks whether or not the work data of the lock adjustment for the prediction target date exists in the database DB (step S207), and if not exists (step S207: No). ), And outputs the predicted values A and B of step S201 as the prediction results as they are (step S208).

  If work schedule data exists in step S207 (step S207: Yes), provisional predicted values A and B are corrected with the corresponding lock adjustment amount in consideration of this, and predicted value A ′ as a result of the correction is corrected. , B ′ are output as prediction results (step S209).

  The results in steps S204, S205, S208, and S209 are set as prediction results (prediction values) based on the prediction model created in FIG. 10 (step S210).

  By comparing the prediction result (predicted value) of the lock shift amount obtained in step S210 as described above with the lock shift amount actually generated on the prediction target date, the validity of the prediction model, The validity of the prediction method (prediction algorithm) in 10 and the necessity of modifying the algorithm can be considered. That is, in the prediction system 100, it is possible to compare the prediction result of the degree of change of the lock shift amount with the actual measurement result, and modify the algorithm in order to perform more accurate prediction based on the difference of the comparison result. it can.

  FIG. 12A is a graph illustrating an example of a relationship between prediction in the prediction system 100 and an actual lock shift amount. The horizontal axis is time, and the vertical axis is the lock shift amount. Here, as an example, an example is shown in which data for two weeks is first accumulated as data from the past to the present, and data for the next day is predicted. In the figure, a curve DU1 indicates an upper limit in prediction by the prediction system 100, and a curve DB1 indicates a lower limit in prediction. Here, similarly to the above, considering that the allowable shift amount is about ± 1.5 mm, the vertical width is set so that the difference between the curve DU1 and the curve DB1 is ± 0.25 mm. ing. A curve R1 indicates the actual lock shift amount. In this case, the curve R1 falls between the curve DU1 and the curve DB1, and it can be said that a reasonable prediction has been made.

  On the other hand, FIG. 12B is a graph showing a comparative example of FIG. 12A. The horizontal axis represents time, and the vertical axis represents temperature and the amount of lock shift. In the example of FIG. 12B, only the change in temperature is shown as a curve X1, and the relationship between the curve X1 and a curve R1 showing the actual lock shift amount is shown. In this figure, the scale is appropriately adjusted so that the change in the temperature and the degree of change in the lock shift amount can be compared. From this figure, it can be seen that even if an attempt is made to predict the lock deviation amount based solely on the change in temperature, an accurate prediction cannot always be made. On the other hand, in the present embodiment, not only information such as the temperature at each time point and a subsequent change in temperature but also the above-described configuration enables accurate and accurate prediction.

  As described above, in the method for predicting the amount of equipment shift according to the present embodiment and the prediction system 100 using the same, when predicting the degree of change in the amount of equipment shift from the current state, the accumulated data management unit AD Based on the accumulated data managed from the past to the present, the equipment forecast is not only based on the temperature information at that point in time, but also, for example, taking into account the difference in the degree of change in the equipment shift amount due to the season. This becomes possible in the lock shift amount prediction processing unit 61 as the shift amount prediction unit. In addition, based on the environment prediction data on the external environment acquired by the environment prediction data acquisition unit configured by the environment prediction data management unit PD, the prediction in consideration of the information around the facility in the future, that is, after the present, is locked. This is possible in the shift amount prediction processing unit 61. As described above, a more accurate prediction is expected for the amount of equipment deviation after the current state, and for example, it is possible to notify the timing to adjust the amount of equipment deviation before the lock failure in the switch 50 occurs. Will be possible.

[Second embodiment]
Hereinafter, an example of a system for predicting the amount of equipment shift according to the second embodiment will be described with reference to FIG.

  The prediction system according to the present embodiment is a modified example of the prediction system illustrated in the first embodiment, and the parts other than the configuration of the processing unit that performs various processes for prediction are the same as those in the first embodiment. Therefore, the description of the entire prediction system is omitted, and only the structure of the processing unit will be described.

  FIG. 13 is a block diagram illustrating an example of a configuration of the prediction system 200 according to the present embodiment, and is a diagram corresponding to FIG.

  As illustrated, the present embodiment differs from the first embodiment in that the main part of the prediction processing unit 60 is provided inside the main body 50A.

  In the prediction system 200, various types of data relating to the points 50 acquired by the lock shift processing unit 40 are directly transmitted to the prediction processing unit 60 via the prediction processing transmitting / receiving unit 60s. In this case, similarly to the case of the first embodiment, for example, the outside of the prediction processing unit 60 via the prediction processing transmission / reception unit 60s, in particular, the outside of the switch 50 such as a station device, The processing result of the prediction may be transmitted. In this case, it is possible to adopt a mode in which the prediction is completed exclusively in the point machine 50 so that only the obtained result can be received.

  Also in the present embodiment, based on the accumulated data from the past to the present and the environmental prediction data for the future, that is, for the present and thereafter, it is possible to more accurately predict the amount of equipment shift after the current state. It becomes possible to inform the user of the timing to adjust the equipment deviation amount before the lock failure or the like occurs in the tipping machine 50.

[Others]
The present invention is not limited to the above embodiment, and can be implemented in various modes without departing from the gist of the invention.

  First, in the above description, the object to be predicted is a point machine, and in particular, regarding the lock of the point machine, the degree of change in the lock deviation amount is predicted. The method and the prediction system are not limited to this, but can be applied in various aspects. First, the prediction of the degree of change in the amount of shift as described above may be applied to, for example, a railroad crossing facility or each section constituting the railroad crossing facility as a device that is installed outdoors, in addition to the point machine. In addition to the point machine installed outdoors, the present invention is also applicable to a case where the machine is installed indoors underground such as a subway. Further, the present invention may be applied not only to the lock mechanism in the switching machine but also to other places where a similar problem may occur.

  In addition, for example, the degree of change in the shift amount may be predicted for various devices installed outdoors, other than the point switches and railroad crossings, or some of them. Further, the present invention may be applied to outdoor installations other than those related to railways.

  In the above description, the case where the next day is used for the weather forecast is exemplified. However, the present invention is not limited to this. For example, weekly forecasts, monthly forecasts, or longer forecasts may be used. The shift amount may be predicted.

  In the above description, for example, in the first embodiment, the lock shift processing unit 40 is included as an element of the prediction system 100. However, various information detected by the lock shift processing unit 40 can be obtained. If so, the lock shift processing unit 40 or a part thereof may not be included.

  DESCRIPTION OF SYMBOLS 10 ... operation | movement can, 20 ... connection part, 21 ... switch adjuster rod, 30 ... lock can, 30a, 30b ... rod-shaped member, 31a ... connection can, 31b ... front rod, 32 ... measuring piece, 40 ... lock deviation Processing unit, 40 s: Pointer-side transmitting / receiving unit, 41: Lock deviation control unit, 42: Lock deviation storage unit, 43: Temperature data processing unit, 44: Vibration data processing unit, 50: Pointing machine, 50A: Main unit, 50N: N-side sensor, 50R: R-side sensor, 54: Linear sensor, 60: Prediction processing unit, 60s: Prediction processing side transmission / reception unit, 61: Lock shift amount prediction processing unit, 61a: Main control Unit, 62 monitor display unit, 100, 200 prediction system, AD storage data management unit, B1 lower limit line, BS vibration sensor, C1 curve, D2 area, DB database, DB1 curve, D , DT1, DT2 ... clearance, DU1 ... curve, LBa, LBb ... lock block, LPa, LPb ... lock piece, NTa, NTb ... notch, P ... branching device, PD ... environment prediction data management unit, Q1 ... broken line, R1 ... Curve, SP: Sleeper, SR: Basic rail, T1, T2: Curve, TR: Tong rail, TS: Temperature sensor, U1: Upper limit line, X1: Curve, α: Deviation

Claims (12)

A method for estimating a shift amount of a facility that expands and contracts,
A method for predicting the amount of equipment shift, wherein the degree of change in the amount of equipment shift is predicted from the current state based on accumulated data from the past to the present and environmental prediction data on the external environment of the equipment.   The method for predicting the amount of equipment shift according to claim 1, wherein the environmental prediction data includes weather forecast data.   The method for predicting an equipment shift amount according to any one of claims 1 and 2, wherein the accumulated data includes data on an insolation amount regarding an external environment of the equipment.   The equipment shift amount prediction method according to any one of claims 1 to 3, wherein the accumulated data includes date information as an element for estimating a degree of change in the equipment shift amount.   The method for predicting the amount of equipment shift according to any one of claims 1 to 4, wherein the accumulated data includes data on weather changes from the past to the present.   The method of predicting the amount of equipment shift according to any one of claims 1 to 5, wherein the accumulated data includes a history of equipment shift from the past to the present.   The method of predicting the amount of equipment shift according to claim 6, wherein an equipment shift due to artificial adjustment is excluded from the accumulated data.   The facility according to any one of claims 1 to 7, wherein the accumulated data includes a difference in a facility shift amount before and after the train passes through the facility installed near the passing position of the train. How to predict the amount of shift.   The equipment deviation according to any one of claims 1 to 8, wherein a prediction result of the degree of change in the equipment deviation amount is compared with an actual measurement result, and an algorithm for performing prediction based on a difference in the comparison result is corrected. How to predict transfers.   The method for predicting an equipment shift amount according to any one of claims 1 to 9, wherein the degree of change of the lock shift amount in the switch as the equipment is estimated. A system for predicting the amount of displacement of equipment that expands and contracts,
A stored data management unit that manages stored data from the past to the present,
An environment prediction data acquisition unit that acquires environment prediction data about the external environment of the facility,
An equipment shift amount prediction system, comprising: an equipment shift amount prediction unit that predicts a degree of change of the equipment shift amount from a current state based on the accumulated data and the environment prediction data.   The equipment shift amount prediction system according to claim 11, wherein the equipment shift amount prediction unit predicts a degree of change of the lock shift amount in the switch as the equipment.

Images