State estimation device, control valve, storage medium and state estimation method

Abstract

The present invention provides a state estimation device, a control valve, a storage medium and a state estimation method. A technology for accurately estimating the state of a control valve is provided. The state estimation device (400) comprises: an acquisition unit, which acquires detection information of a data set, the data set including a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes position according to the flow rate of the working fluid, and a value of an input voltage or output voltage of a power supply circuit that supplies power to the control valve; and an estimation unit, which estimates an abnormality of the power supply circuit based on the detection information acquired by the acquisition unit.

Description

Technical Field

[0001] The present invention relates to a state estimation device for estimating the state of a control valve, a control valve, a state estimation program and a state estimation method. Background Art

[0002] In order to control the engine installed in a mobile body such as a ship, a control valve such as a hydraulic servo valve is used. By using an electrically controllable control valve to finely control the fuel supply to the engine and the exhaust gas discharged from the engine, the thermal efficiency of the engine can be improved and the fuel consumption can be suppressed.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent No. 5465365 Summary of the invention

[0006] Problem that the invention aims to solve

[0007] Ships sail on the ocean, so even if a control valve on board fails or malfunctions, it may not be possible to respond immediately. In the past, when a ship was anchored in a harbor, the control valve that had a failure or malfunction was removed and taken to a factory for inspection, and necessary repairs or replacements were performed. However, in order to prevent the ship from becoming unable to sail and causing great damage, it is very important to accurately understand the status of the control valve and take appropriate measures in advance before it becomes unable to operate due to failure or malfunction.

[0008] The present invention has been made in view of such a problem, and an object of the present invention is to provide a technology for estimating the state of a control valve more accurately.

[0009] Solutions for solving problems

[0010] In order to solve the above-mentioned problem, a state estimation device of a certain embodiment of the present invention comprises an acquisition unit and an estimation unit, wherein the acquisition unit acquires detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes its position according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve, wherein the estimation unit estimates an abnormality of the power supply circuit based on the detection information acquired by the acquisition unit.

[0011] Another aspect of the present invention is a control valve. The control valve includes a first movable part, an acquisition part, and an estimation part, wherein the position of the first movable part changes according to a control signal for specifying the position, and the first movable part controls the flow rate of a working fluid according to the position, the acquisition part acquires detection information of a data set, the data set including: a target position or actual position of the first movable part, or a target position or actual position of a second movable part whose position changes according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve, and the estimation part estimates an abnormality of the power supply circuit based on the detection information acquired by the acquisition part.

[0012] Another aspect of the present invention is a state estimation program. The state estimation program causes a computer to function as an acquisition unit and an estimation unit, the acquisition unit acquiring detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes position according to the flow rate of the working fluid; and a value of an input voltage or output voltage of a power supply circuit that supplies power to the control valve, and the estimation unit estimating an abnormality of the power supply circuit based on the detection information acquired by the acquisition unit.

[0013] Another aspect of the present invention is a state estimation method. The state estimation method causes a computer to perform the following steps: an acquisition step of acquiring detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes position according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve; and estimating an abnormality of the power supply circuit based on the detection information acquired in the acquisition step.

[0014] Another aspect of the present invention is a state estimation device. The state estimation device comprises: an acquisition unit that acquires detection information of a data set, the data set including a target position and an actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, and a combination of any two or more positions of the target position and the actual position of a second movable part that changes its position according to the flow rate of the working fluid; and an estimation unit that estimates an abnormality of the first movable part based on the detection information acquired by the acquisition unit.

[0015] Another aspect of the present invention is a control valve. The control valve includes: a first movable part whose position changes according to a control signal for specifying a position, and controls the flow rate of a working fluid according to the position; an acquisition part that acquires detection information of a data set, the data set including a target position and an actual position of the first movable part and a combination of any two or more positions of the target position and the actual position of a second movable part whose position changes according to the flow rate of the working fluid; and an estimation part that estimates an abnormality of the first movable part based on the detection information acquired by the acquisition part.

[0016] Another aspect of the present invention is a state estimation program. The state estimation program causes a computer to function as an acquisition unit and an estimation unit, the acquisition unit acquiring detection information of a data set including a target position and an actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, and a combination of any two or more positions of the target position and the actual position of a second movable part that changes its position according to the flow rate of the working fluid, and the estimation unit estimating an abnormality of the first movable part based on the detection information acquired by the acquisition unit.

[0017] Another aspect of the present invention is a state estimation method. The state estimation method causes a computer to perform the following steps: an acquisition step of acquiring detection information of a data set, the data set including a target position and an actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, and a combination of any two or more positions of the target position and the actual position of a second movable part that changes its position according to the flow rate of the working fluid; and estimating an abnormality of the first movable part based on the detection information acquired in the acquisition step.

[0018] In addition, any combination of the above-mentioned constituent elements, or a method obtained by replacing the constituent elements of the present invention or the constituent elements expressed in a method, device, program, transient or non-transient storage medium recording the program, system, etc., is also valid as a method of the present invention.

[0019] Effects of the Invention

[0020] According to the present invention, the state of the control valve can be estimated more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a diagram showing the configuration of a management system according to an embodiment of the present invention.

[0022] Figure 2 This is a diagram schematically showing the structure of the periphery of a hydraulic servo valve mounted on a ship.

[0023] Figure 3 This is a diagram schematically showing the structure of a hydraulic servo valve.

[0024] Figure 4 It is a schematic diagram schematically showing the position of the valve body of the pilot valve in the first direction and the open and closed states of the port.

[0025] Figure 5 It is a diagram showing the structure of a servo valve control device.

[0026] Figure 6 is a diagram showing the configuration of a learning device and a state estimation device.

[0027] Figure 7 : is a flowchart showing the process of the state estimation method of this embodiment.

[0028] Figure 8 This is a diagram showing the configuration of a learning device and a state estimation device according to Example 1-1.

[0029] Fig. 9 It is a diagram showing the configuration of a learning device and a state estimation device according to Embodiment 1-2.

[0030] Fig.10 It is a diagram showing the configuration of a learning device and a state estimation device according to Embodiment 1-3.

[0031] Fig.11 It is a diagram showing the configuration of a learning device and a state estimation device according to Embodiments 1 to 4.

[0032] Fig.12 It is a diagram showing the configuration of a learning device and a state estimation device according to Embodiment 1-5.

[0033] Fig.13 This is a diagram showing the configuration of a learning device and a state estimation device according to the second embodiment.

[0034] Fig.14 The diagram schematically shows the operation of the valve body of the pilot valve and the valve body of the main valve.

[0035] Fig.15 This is a diagram showing the configuration of a learning device and a state estimation device according to the third embodiment.

[0036] Fig.16 This is a diagram showing the configuration of a learning device and a state estimation device according to the fourth embodiment.

[0037] Fig.17 The diagram schematically shows the operation of the valve body of the pilot valve and the valve body of the main valve.

[0038] Fig.18 This is a diagram showing the configuration of a learning device and a state estimation device according to the fifth embodiment.

[0039] Fig.19 It is a diagram showing the state of the position sensor of the valve body of the main valve.

[0040] Fig. 20 This is a diagram showing the configuration of a learning device and a state estimation device according to the sixth embodiment.

[0041] Fig.21 This is a diagram showing the configuration of a learning device and a state estimation device according to the seventh embodiment.

[0042] Description of Reference Numerals

[0043] 1: management system; 2: ship; 10: pilot valve; 12: valve core; 16: port; 18: valve core drive unit; 19: position sensor; 20: main valve; 28: valve core; 29: position sensor; 48: working oil; 80: engine; 81: cylinder; 82: sensor; 90: log data storage device; 91: engine control device; 92: junction box; 100: hydraulic servo valve; 110: servo valve control device; 120: power supply circuit; 130: servo valve control circuit; 140: servo valve drive circuit; 150: voltage detection unit; 160: data collection circuit; 300: learning device; 301: learning data acquisition unit; 302: estimation model generation unit; 303: estimation model providing unit; 310: actual valve core position of pilot valve; 311: actual valve core position of main valve; 315: feature quantity calculation unit; 316: pilot valve actual valve core position; 317: main valve actual valve core position; 318: feature quantity calculation unit; 319: pilot valve actual valve core position; 320: main valve actual valve core position; 321: feature quantity calculation unit; 322: pilot valve actual valve core position; 323: main valve actual valve core position; 324: feature quantity calculation unit; 325: feature quantity calculation unit; 326: pilot valve actual valve core position; 327: main valve actual valve core position; 328: main valve actual valve core position; 329: main valve actual valve core position; 330: main valve actual valve core position; 331: feature quantity calculation unit; 332: main valve actual valve core position; 333: feature quantity calculation unit; 334: feature quantity calculation unit; 335: feature quantity calculation unit; 336: Target spool position of pilot valve; 320: data selection unit; 321: power supply voltage; 322: power supply failure precursor index value; 323: power supply failure judgment model generation unit; 324: power supply failure judgment model providing unit; 330: target spool position of main valve; 331: foreign matter data; 333: foreign matter detection model generation unit; 334: foreign matter detection model providing unit; 400: state estimation device; 401: detection information acquisition unit; 402: state estimation unit; 403: estimation result output unit; 410: actual spool position of pilot valve; 411: actual spool position of main valve; 414: feature quantity calculation unit; 415: target spool position of pilot valve; 417: data selection unit; 420: power supply voltage; 421: power supply failure judgment unit; 422: power supply failure judgment result output unit; 430: target spool position of main valve; 431: foreign matter detection unit. DETAILED DESCRIPTION

[0044] Below, the present invention is described with reference to the accompanying drawings based on preferred embodiments. In the embodiments and modifications, the same or equivalent components and members are marked with the same symbols, and repeated descriptions are appropriately omitted. In addition, the sizes of the components in the drawings are appropriately enlarged or reduced to facilitate understanding. In addition, in the drawings, a part of the components that are not important for explaining the embodiments is omitted.

[0045] In addition, terms including ordinal numbers such as first and second are used to describe various components, but these terms are only used to distinguish one component from other components, and the components are not limited by these terms.

[0046] Figure 1 The structure of a management system according to an embodiment of the present invention is shown. The management system 1 manages the control valve based on information associated with the operation of the control valve. The management system 1 can be used to manage any control valve, but in this embodiment, an example of managing a hydraulic servo valve for controlling an engine mounted on a ship 2 is mainly described. The management system 1 includes a learning device 300 and a state estimation device 400. The learning device 300 learns an estimation model for estimating the state of the hydraulic servo valve based on information associated with the hydraulic servo valve. The state estimation device 400 estimates the state of the hydraulic servo valve using the estimation model learned by the learning device 300.

[0047] The learning device 300 uses learning data for learning the estimation model to learn the estimation model. The learning data may be log data recorded when the hydraulic servo valve or other hydraulic servo valves of the same type as the hydraulic servo valve are actually used in the ship 2, or test data recorded when the hydraulic servo valve or other hydraulic servo valves of the same type as the hydraulic servo valve are used in a test environment other than the ship 2, or simulation data generated by a simulator for simulating the action of the hydraulic servo valve, or a combination of two or more of these data. Unless otherwise specified, the log data or test data may be log data or test data recorded when the hydraulic servo valve itself that is the object of the estimation state is used or tested, or log data or test data recorded when other hydraulic servo valves of the same type as the hydraulic servo valve are used or tested, or a combination of these.

[0048] When log data is used as learning data, the estimation model can be learned based on log data collected when the hydraulic servo valve is actually used in the same environment as the hydraulic servo valve to be estimated, so the accuracy of the estimation model can be improved. In this case, a log data storage device for storing log data can be installed in the ship 2, and the log data can be read from the log data storage device when the ship 2 is docked and supplied to the learning device 300. In addition, a communication device for ship-land communication can be installed in the ship 2, and the log data can be sent from the ship 2 to the learning device 300 via the communication network 3.

[0049] When test data or simulation data is used as learning data, a large amount of data when the hydraulic servo valve is used in various states or environments can be generated, thereby improving the accuracy and versatility of the estimation model. For example, even if it is difficult to obtain log data when a fault with a very low frequency of occurrence actually occurs, it is possible to generate an estimation model that can accurately estimate such a fault by learning the estimation model using test data when the fault occurs experimentally or simulation data when the fault is simulated.

[0050] The learning device 300 may further learn and update the learned estimation model used in the state estimation device 400 .

[0051] The state estimation device 400 acquires information related to the hydraulic servo valve mounted on the ship 2, and estimates the state of the hydraulic servo valve using the estimation model. The state estimation device 400 may also read log data from a log data storage device mounted on the ship 2, and estimate the state of the hydraulic servo valve during the past voyage of the ship 2 based on the read log data. In addition, the state estimation device 400 may also receive information related to the hydraulic servo valve from a communication device for ship-to-land communication mounted on the ship 2 via the communication network 3, and estimate the state of the hydraulic servo valve of the ship 2 during voyage based on the received information. In addition, the state estimation device 400 may also be mounted on the ship 2, acquire information related to the hydraulic servo valve in real time, and estimate the state of the hydraulic servo valve. In this case, the estimation result of the state estimation device 400 may also be sent to the owner, management body, maintenance body, etc. of the ship 2 through ship-to-land communication.

[0052] According to the technology of this embodiment, the state of the hydraulic servo valve in the past or present can be accurately estimated, so even if the hydraulic servo valve fails or fails, it is possible to take prompt and appropriate measures. In addition, even if the hydraulic servo valve does not fail or fail, the possibility of failure or failure of the hydraulic servo valve in the future voyage, the life of the hydraulic servo valve, etc. can be accurately predicted based on the current or past state of the hydraulic servo valve. As such, the technology of this embodiment is of great significance for improving the safety and efficiency of mobile objects such as ships 2.

[0053] Figure 2 The structure around the hydraulic servo valve 100 mounted on the ship 2 is schematically shown. The engine 80 mounted on the ship 2 includes a sensor 82 and a plurality of cylinders 81. The sensor 82 detects the number of revolutions, load, pressure, exhaust temperature, etc. of the engine 80. The hydraulic servo valve 100 is provided corresponding to each of the plurality of cylinders 81, and controls the injection and exhaust of fuel in each cylinder 81. In the present embodiment, the hydraulic servo valve 100 includes: a pilot valve that controls the flow rate of the hydraulic oil supplied to the actuator by electrically controlling the position of the valve core; and a main valve that is an example of an actuator controlled by the pilot valve. The hydraulic servo valve 100 controls the flow rate of the hydraulic oil supplied to other actuators provided for driving the injection valve, the exhaust valve, etc. according to the position of the valve core of the main valve. In other examples, the hydraulic servo valve 100 may directly drive the injection valve, the exhaust valve, etc. by moving the valve core of the main valve.

[0054] The engine control device 91 determines the number of revolutions of the engine 80 and the like in response to an instruction input from a control panel (not shown) for controlling the navigation of the ship 2, and inputs an instruction to the servo valve control device 110. The servo valve control device 110 calculates the target position of the valve core of the main valve of the plurality of hydraulic servo valves 100 in response to the instruction from the engine control device 91, and controls the position of the valve core of each pilot valve so that the position of the valve core of each main valve becomes the calculated target position. The servo valve control device 110 obtains information indicating the actual position of the valve core of the main valve from each of the plurality of hydraulic servo valves 100 via the junction box 92, calculates the target position of the valve core of the pilot valve based on the target position and the actual position of the valve core of the main valve, and outputs it to the hydraulic servo valve 100, thereby performing feedback control on the position of the valve core of the main valve. The servo valve control device 110 can perform feedback control on the position of the valve core of the main valve by any method such as P control, PI control, PID control, etc.

[0055] The servo valve control device 110 records input data from the engine control device 91 , output data to the hydraulic servo valve 100 , and various detection data indicating the states of the hydraulic servo valve 100 and the engine 80 acquired via the junction box 92 in the log data storage device 90 .

[0056] In this figure, an example in which the log data storage device 90 is connected to the servo valve control device 110 is shown, but the log data storage device 90 may be connected to the junction box 92, may be connected between the junction box 92 and the hydraulic servo valve 100, or may be mounted in the hydraulic servo valve 100. When the log data storage device 90 is connected to the servo valve control device 110, the log data storage device 90 can be installed at a position relatively far from the engine 80, so that the influence of vibration, heat, etc. generated in the engine 80 can be suppressed. In addition, the change of wiring etc. required when installing the log data storage device 90 in the existing ship 2 can be reduced. When the log data storage device 90 is connected to the junction box 92, even if the configuration is such that the data indicating the actual position of the valve body of the pilot valve of the hydraulic servo valve 100 is not sent to the servo valve control device 110, the data indicating the actual position of the valve body of the pilot valve can be acquired and recorded, so that the state of the hydraulic servo valve 100 can be estimated more accurately. When the log data storage device 90 is connected between the hydraulic servo valve 100 and the junction box 92, in addition to acquiring data indicating the actual position of the valve core of the pilot valve of the hydraulic servo valve 100, the value of the voltage or current supplied to the hydraulic servo valve 100 can also be acquired, so that the state of the hydraulic servo valve 100 can be estimated more accurately. The same is true when the log data storage device 90 is mounted in the hydraulic servo valve 100.

[0057] When the state estimation device 400 is mounted on the ship 2, the state estimation device 400 may obtain data from the log data storage device 90, or the state estimation device 400 may be provided in place of the log data storage device 90. In this case, as described above, the state estimation device 400 may be connected to the servo valve control device 110, may be connected to the junction box 92, may be connected between the junction box 92 and the hydraulic servo valve 100, or may be mounted in the hydraulic servo valve 100.

[0058] Figure 3 The structure of the hydraulic servo valve 100 is schematically shown. The hydraulic servo valve 100 includes a pilot valve 10 and a main valve 20. The pilot valve 10 controls the position of the valve element 28 of the main valve 20 by changing the delivery state of the hydraulic oil 48 to the main valve 20 as a controlled device based on a command from the servo valve control device 110.

[0059] The pilot valve 10 is connected to the main valve 20 by a plurality of bolts B1. The pilot valve 10 is provided with a plurality of through holes 10h for passing the bolts B1. The main valve 20 is provided with a plurality of internal threads 20h for threading with the bolts B1. The plurality of through holes 10h are arranged at positions corresponding to the positions of the plurality of internal threads 20h. The bolts B1 are threadedly coupled to the internal threads 20h through the through holes 10h, thereby connecting the pilot valve 10 to the main valve 20. The pilot valve 10 is separated from the main valve 20 by removing the bolts B1.

[0060] Figure 3 The hydraulic system of the main valve 20 includes a drain tank 44 for storing working oil 48 and a hydraulic pump 42 for pressurizing and delivering the working oil 48 from the drain tank 44. The working oil 48 delivered by the hydraulic pump 42 is supplied to the inside of the main valve 20 and the pilot valve 10 through the pump-side piping section 22p in the main valve 20. The working oil 48 discharged from the inside of the main valve 20 and the pilot valve 10 is returned to the drain tank 44 through the tank-side piping section 22t in the main valve 20. The pump-side piping section 22p and the tank-side piping section 22t are collectively referred to as the main valve piping section.

[0061] The pilot valve 10 mainly includes a main body 10b, a valve core 12, a port 16, and a valve core driving unit 18. The valve core 12 functions as a first movable unit and has a shaft 12s and a plurality of valve bodies 14 that move integrally with the shaft 12s. The valve core 12 is driven by the valve core driving unit 18 to move forward and backward in a first direction. Figure 3 The direction opposite to the extending direction is referred to as the “extending direction” or the “extending side”, and the direction opposite to the extending direction is referred to as the “extending opposite direction” or the “extending opposite side”.

[0062] A force member 12h is provided on the extension side of the valve core 12 to apply force to the valve core 12 in the opposite direction of extension. The force member 12h may be, for example, a coil spring that expands and contracts in the first direction. The valve core drive unit 18 includes an electromagnetic actuator (not shown) such as a coil that moves the shaft 12s forward and backward in the first direction. The valve core drive unit 18 moves the shaft 12s forward and backward based on the command from the servo valve control device 110, and controls the position of the valve body 14 by balancing the force applied by the force member 12h.

[0063] The valve body 14 includes a first valve body 14a, a second valve body 14b, and a third valve body 14c that are separated and arranged in the first direction. The second valve body 14b is arranged on the opposite side of the extension of the first valve body 14a, and the third valve body 14c is arranged on the extension side of the first valve body 14a. The first valve body 14a changes the communication state of the A port 16a described later according to its position in the first direction. The main body 10b has a cylindrical space 10s that extends in the first direction to accommodate the valve core 12. The cylindrical space 10s functions as a cylinder that surrounds the valve body 14 with a narrow gap.

[0064] The main body 10b is provided with a port 16. The port 16 of the present embodiment includes a P port 16p, an A port 16a, and a T port 16t. The P port 16p is connected to the pump side piping portion 22p, and is supplied with the working oil 48 pressurized by the hydraulic pump 42. The A port 16a is connected to the working oil storage portion 22a of the main valve 20. The valve core 28 of the main valve 20 moves due to the pressure of the working oil 48 supplied to the working oil storage portion 22a. The T port 16t is connected to the cabinet side piping portion 22t, and the working oil 48 flowing through the main body 10b is discharged to the drain tank 44 through the cabinet side piping portion 22t.

[0065] The main valve 20 includes a valve element 28 that functions as a second movable portion. Although details of the main valve 20 are omitted in this figure, the main valve 20 may have the same structure as the pilot valve 10 .

[0066] A position sensor 19 for detecting the position of the valve core 12 is provided at the top end of the valve core 12 of the pilot valve 10. A position sensor 29 for detecting the position of the valve core 28 is provided at the top end of the valve core 28 of the main valve 20. Data indicating the actual position of the valve core 12 of the pilot valve 10 detected by the position sensor 19 and data indicating the actual position of the valve core 28 of the main valve 20 detected by the position sensor 29 are sent to the junction box 92 via wiring.

[0067] Figure 4 This is a schematic diagram schematically showing the position of the valve body 14 in the first direction and the open / closed state of the port of the pilot valve 10. In this figure, description of elements not important for the description is omitted. Figure 4 (a) shows a state in which the valve body 14 is located in the first region that connects the A port 16a and the P port 16p. In this state, the A port 16a supplies the working oil 48 from the P port 16p to the working oil storage portion 22a (hereinafter referred to as "supply mode"). In the supply mode, the working oil 48 from the P port 16p is supplied to the working oil storage portion 22a of the main valve 20. By this action, for example, the valve core 28 of the main valve 20 moves in a direction to increase the amount of fuel supplied to the engine 80.

[0068] Figure 4 (b) shows a state in which the valve body 14 is located in a neutral region where the A port 16a is blocked and not connected to the P port 16p and the T port 16t (hereinafter, the position in the neutral region is also referred to as the "neutral position"). In this state, the A port 16a is blocked, and neither supply nor recovery is performed to the working oil storage portion 22a (hereinafter referred to as the "neutral mode"). In the neutral mode, the hydraulic pressure of the working oil storage portion 22a of the main valve 20 is maintained in a state immediately before the valve body 14 is located in the neutral region. Through this action, for example, the valve core 28 of the main valve 20 stops at the previous position, and the fuel supply amount supplied to the engine 80 is maintained at the previous state.

[0069] Figure 4 (c) shows a state in which the valve body 14 is located in the second region in which the A port 16a and the T port 16t are connected. In this state, the A port 16a recovers the working oil 48 from the working oil storage portion 22a and returns the working oil 48 to the pump-side piping portion 22p (hereinafter referred to as "recovery mode"). In the recovery mode, the working oil 48 in the working oil storage portion 22a of the main valve 20 is recovered to the drain tank 44 through the A port 16a, the T port 16t, and the tank-side piping portion 22t. By this action, for example, the valve core 28 of the main valve 20 moves in a direction in which the amount of fuel supplied to the engine 80 is reduced.

[0070] Figure 5 1 shows the structure of the servo valve control device 110. The servo valve control device 110 includes a power supply circuit 120, a servo valve control circuit 130, a servo valve drive circuit 140, a voltage detection unit 150, and a data collection circuit 160.

[0071] The power supply circuit 120 supplies power supplied from an external power supply to the servo valve control circuit 130 and the servo valve drive circuit 140. The voltage detection unit 150 detects a voltage input to the power supply circuit 120 or a voltage output from the power supply circuit 120. Instead of the voltage detection unit 150, a current detection unit that detects a current input to the power supply circuit 120 or a current output from the power supply circuit 120 may be provided, or in addition to the voltage detection unit 150, a current detection unit that detects a current input to the power supply circuit 120 or a current output from the power supply circuit 120 may be provided.

[0072] The servo valve control circuit 130 calculates the target position of the valve core 28 of the main valve 20 based on the command from the engine control device 91. The servo valve control circuit 130 calculates the target position of the valve core 12 of the pilot valve 10 for moving the valve core 28 of the main valve 20 to the calculated target position. The servo valve control circuit 130 inputs the calculated target position of the valve core 12 of the pilot valve 10 to the servo valve drive circuit 140.

[0073] The servo valve drive circuit 140 supplies power to the coil of the valve element drive unit 18 of the pilot valve 10 to move the valve element 12 according to the target position of the valve element 12 of the pilot valve 10 input from the servo valve control circuit 130. The servo valve drive circuit 140 may be provided in the hydraulic servo valve 100.

[0074] When the valve core 12 of the pilot valve 10 moves and the opening and closing state of the port 16 is changed, the valve core 28 of the main valve 20 is moved to the target position by supplying or recovering the hydraulic oil 48. When the actual position of the valve core 28 of the main valve 20 coincides with the target position, the servo valve control circuit 130 returns the valve core 12 of the pilot valve 10 to the neutral position. As a result, the valve core 28 of the main valve 20 is stationary at the target position. This series of control is repeated during the operation of the engine 80.

[0075] The data collection circuit 160 records the voltage value detected by the voltage detection unit 150 or the target position of the valve core 12 of the pilot valve 10 input from the servo valve control circuit 130 to the servo valve drive circuit 140 in the log data storage device 90. The data collection circuit 160 acquires data indicating the state of the engine 80 detected by the sensor 82, the actual position of the valve core 12 of the pilot valve 10 detected by the position sensor 19, the actual position of the valve core 28 of the main valve 20 detected by the position sensor 29, and the like via the junction box 92 and records them in the log data storage device 90.

[0076] Figure 6 The configurations of the learning device 300 and the state estimation device 400 are shown. The learning device 300 includes a learning data acquisition unit 301, an estimation model generation unit 302, and an estimation model supply unit 303. The state estimation device 400 includes a probe information acquisition unit 401, a state estimation unit 402, and an estimation result output unit 403.

[0077] The learning data acquisition unit 301 acquires learning data used for learning an estimation model for estimating the state of the hydraulic servo valve 100. The learning data includes a set of data that can be acquired in association with the operation of the hydraulic servo valve 100 and data indicating the state of the hydraulic servo valve 100. The learning data acquisition unit 301 may acquire log data stored in the log data storage device 90, may acquire test data recorded when the hydraulic servo valve 100 is used in a test environment other than the ship 2, or may acquire simulation data generated by a simulator for simulating the operation of the hydraulic servo valve 100.

[0078] The learning data acquisition unit 301 can acquire data recorded or generated under specific conditions as learning data. The learning data acquisition unit 301 can select data recorded or generated under specific conditions from the acquired data as learning data. For example, the learning data acquisition unit 301 can acquire or select log data when a specific state to be estimated by the estimation model is generated, test data when a specific state is generated in a test environment, or simulation data when a specific state is simulated by a simulator. The learning data acquisition unit 301 can acquire data recorded or generated when the hydraulic servo valve 100 is actuated under a specific environment as learning data. For example, the learning data acquisition unit 301 can classify the data according to the type of ship 2, route, sailing period, type of engine 80, number of cylinders, cumulative operating time, etc., and learn different estimation models for each of these environments.

[0079] The learning data acquisition unit 301 can pre-process the acquired data and generate learning data. For example, the learning data acquisition unit 301 can calculate a feature quantity that is correlated with the state to be estimated by the estimation model based on the acquired data as learning data. In addition, the offset time from the input target position until the actual position of the valve core 12 or the valve core 28 reaches the target position can be adjusted to correspond the target position or actual position of the valve core 12 of the pilot valve 10 and the target position or actual position of the valve core 28 of the main valve 20 to other data.

[0080] The estimation model generation unit 302 generates an estimation reference used in the state estimation device 400 to estimate the state of the hydraulic servo valve 100 using the learning data acquired by the learning data acquisition unit 301. The estimation reference may be a table or program obtained by corresponding data that can be acquired in association with the operation of the hydraulic servo valve 100 and data indicating the state of the hydraulic servo valve 100. The estimation reference may be an estimation model obtained by modeling the correspondence between data that can be acquired in association with the operation of the hydraulic servo valve 100 and data indicating the state of the hydraulic servo valve 100. The estimation model may be a formula for calculating data indicating the state of the hydraulic servo valve 100 using data that can be acquired in association with the operation of the hydraulic servo valve 100 as input variables. In this case, the estimation model generation unit 302 may generate the estimation model by a statistical method such as multivariate analysis, multiple regression analysis, and principal component analysis. The estimation model may be a neural network or the like that inputs data indicating the state of the hydraulic servo valve 100 into an input layer and outputs data indicating the state of the hydraulic servo valve 100 from an output layer. In this case, the estimation model generation unit 302 learns the estimation model by adjusting the intermediate layer of the neural network in such a manner that, when data included in the learning data and acquired in association with the action of the hydraulic servo valve 100 is input to the input layer, a value approximate to the data corresponding to the data and indicating the state of the hydraulic servo valve 100 is output from the output layer. The estimation model may be a rule-based estimation algorithm or any form of artificial intelligence.

[0081] The estimated model providing unit 303 provides the estimated model generated by the estimated model generating unit 302 to the state estimation device 400. The estimated model providing unit 303 may be provided to the state estimation device 400 when the state estimation device 400 is manufactured. In addition, when the learning device 300 learns the estimated model to update the estimated model after the state estimation device 400 is manufactured, the updated estimated model may be provided to the state estimation device 400 at a predetermined timing.

[0082] The detection information acquisition unit 401 acquires information detected in association with the operation of the hydraulic servo valve 100. The detection information acquisition unit 401 may acquire data stored in the log data storage device 90, directly acquire data from the data collection circuit 160, or acquire data from the ship 2 through ship-land communication. The detection information acquisition unit 401 may perform the same pre-processing as the learning data used when generating the estimation model on the acquired data.

[0083] The state estimation unit 402 estimates the state of the hydraulic servo valve 100 using the estimation model generated by the learning device 300. The state estimation unit 402 acquires the state of the hydraulic servo valve 100 as an estimation result by inputting the data acquired by the detection information acquisition unit 401 into the estimation model and outputting it from the estimation model.

[0084] The estimation result output unit 403 outputs the estimation result of the state estimation unit 402. When the state estimation unit 402 estimates that the hydraulic servo valve 100 is in an abnormal state, the estimation result output unit 403 may notify the state estimation unit 402 of the fact. Whether the hydraulic servo valve 100 is in an abnormal state may be determined by whether the state estimation data indicating the state of the hydraulic servo valve 100 output by the state estimation unit 402 is greater than or less than a predetermined threshold value, or is within a predetermined range, or may be determined by whether the amount of change in the state estimation data relative to the initial value is greater than a predetermined threshold value. In addition, the estimation result output unit 403 may determine whether the hydraulic servo valve 100 is in an abnormal state by comparing the state estimation results of a plurality of hydraulic servo valves 100 provided corresponding to a plurality of cylinders 81 of the same engine 80. For example, when the value of the state estimation result of a specific hydraulic servo valve 100 is greatly deviated from the average value of the state estimation results of the plurality of hydraulic servo valves 100, it may be determined that the hydraulic servo valve 100 is in an abnormal state. The deviation value of the state estimation results of each of the plurality of hydraulic servo valves 100 may be calculated, and a hydraulic servo valve 100 having a deviation value greater than or equal to a predetermined threshold value, such as 80, or less than a predetermined threshold value, such as 30, may be determined to be in an abnormal state.

[0085] Figure 7 : is a flowchart showing the process of the state estimation method of the present embodiment. First, the process of the stage of generating the estimation model is described. When the ship 2 equipped with the hydraulic servo valve 100 is sailing, the data associated with the action of the hydraulic servo valve 100 is recorded in the log data storage device 90 (step S10). The learning data acquisition unit 301 of the learning device 300 acquires the log data recorded in the log data storage device 90 as learning data (step S12). The learning data acquisition unit 301 also acquires test data or simulation data as learning data. The estimation model generation unit 302 generates an estimation model based on the acquired learning data (step S14). The estimation model providing unit 303 provides the generated estimation model to the state estimation device 400 (step S16).

[0086] Next, the process of estimating the state of the hydraulic servo valve 100 using the estimation model will be described. When the ship 2 equipped with the hydraulic servo valve 100 is sailing, data related to the operation of the hydraulic servo valve 100 is recorded in the log data storage device 90 (step S20). The detection information acquisition unit 401 of the state estimation device 400 acquires the detection information recorded in the log data storage device 90 (step S22). The state estimation unit 402 estimates the state of the hydraulic servo valve 100 using the estimation model based on the acquired detection information (step S24). The estimation result output unit 403 outputs the estimation result (step S26).

[0087] A specific example of the management system according to the present embodiment will be described.

[0088] [Example 1: Estimation of internal oil leakage]

[0089] When the pilot valve 10 is used for a long time, the valve body 14 wears and the valve articulation decreases, so that the A port 16a, the P port 16p, and the T port 16t are slightly connected to each other even in the neutral mode. When the working oil 48 leaks from the P port 16p to the A port 16a in the neutral mode, the hydraulic pressure of the working oil container 22a gradually increases and the amount of fuel supplied to the engine 80 increases, causing the fuel consumption rate of the engine 80 to deteriorate. In addition, when the working oil 48 leaks from the A port 16a to the T port 16t in the neutral mode, the hydraulic pressure of the working oil container 22a gradually decreases and the amount of fuel supplied to the engine 80 decreases, causing the output of the engine 80 to decrease. When the leakage amount of the working oil 48 exceeds the allowable amount, the hydraulic servo valve 100 no longer functions normally and causes a failure. If the leakage amount can be estimated with high accuracy, the hydraulic servo valve 100 can be replaced or repaired before the leakage amount exceeds the allowable amount, and unexpected failures can be avoided.

[0090] As described above, when the hydraulic servo valve 100 operates normally, the valve core 28 of the main valve 20 is stationary when the valve core 12 of the pilot valve 10 is in the neutral position. However, if the leakage amount of the hydraulic oil 48 increases, the valve core 28 of the main valve 20 moves even when the valve core 12 of the pilot valve 10 is in the neutral position. The present inventors have found through experiments that there is a strong correlation between the movement speed of the valve core 28 of the main valve 20 when the valve core 12 of the pilot valve 10 is in the neutral position and the leakage amount of the hydraulic oil 48. Therefore, the internal oil leakage amount in the pilot valve 10 can be estimated based on the movement speed of the valve core 28 of the main valve 20 when the valve core 12 of the pilot valve 10 is in the neutral position.

[0091] [Example 1-1]

[0092] Figure 8The configuration of the learning device and state estimation device according to Embodiment 1-1 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the pilot valve actual spool position 310, the main valve actual spool position 311, and the internal oil leakage amount actual measurement value 312 as learning data.

[0093] The pilot valve actual spool position 310 and the main valve actual spool position 311 are time series data when the hydraulic servo valve 100 is used within a specified period before or after the time point when the internal oil leakage amount is actually measured. The time series data may be time series data of one to several cycles of the rotational action of the engine 80. Since the position of the valve core 12 of the pilot valve 10 is at the neutral position at least several times in one cycle of the rotational action of the engine 80, the time series data of the pilot valve actual spool position 310 and the main valve actual spool position 311 of one cycle includes information indicating the moving speed of the valve core 28 of the main valve 20 when the valve core 12 of the pilot valve 10 is at the neutral position.

[0094] The internal oil leakage amount measured value 312 is the value of the internal oil leakage amount in the hydraulic servo valve 100. When log data or test data is used as learning data, the internal oil leakage amount measured value 312 is the value of the internal oil leakage amount actually measured by a flow meter or the like. Since the internal oil leakage amount is not considered to change rapidly during the use of the hydraulic servo valve 100, it is assumed that the hydraulic servo valve 100 is used in a state where the working oil 48 of the actually measured amount is leaked internally during a predetermined period before or after the time point when the internal oil leakage amount is actually measured, and the pilot valve actual spool position 310 and the main valve actual spool position 311 recorded during the predetermined period are associated with the internal oil leakage amount measured value 312 as learning data. The predetermined period may be a period during which the internal oil leakage amount remains constant during the use of the hydraulic servo valve 100, and may be specified by the use time of the hydraulic servo valve 100, the operation time of the engine 80, the number of revolutions of the engine 80, and the like. When simulation data is used as learning data, the internal oil leakage amount measured value 312 is the value of the internal oil leakage amount input to the simulator as a simulation condition.

[0095] The internal oil leakage amount estimation model generation unit 313 generates an internal oil leakage amount estimation model used in the state estimation device 400 to estimate the internal oil leakage amount of the hydraulic servo valve 100 using the learning data acquired by the learning data acquisition unit 301. The internal oil leakage amount estimation model may be a neural network or the like that inputs the time series data of the pilot valve actual spool position 310 and the main valve actual spool position 311 during a predetermined period into an input layer and outputs the internal oil leakage amount of the hydraulic servo valve 100 from an output layer. In this case, the internal oil leakage amount estimation model generation unit 313 learns the internal oil leakage amount estimation model by adjusting the intermediate layer of the neural network in such a manner that a value close to the internal oil leakage amount measured value 312 is output from the output layer when the time series data of the pilot valve actual spool position 310 and the main valve actual spool position 311 during a predetermined period is input into the input layer.

[0096] The internal oil leakage amount estimation model providing unit 314 provides the internal oil leakage amount estimation model generated by the internal oil leakage amount estimation model generating unit 313 to the state estimation device 400 .

[0097] The detection information acquisition unit 401 acquires, as detection information, a pilot valve actual spool position 410 and a main valve actual spool position 411. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the internal oil leakage amount estimation model.

[0098] The internal oil leakage amount estimation unit 412 estimates the internal oil leakage amount of the hydraulic servo valve 100 using the internal oil leakage amount estimation model generated by the learning device 300. The internal oil leakage amount estimation unit 412 inputs the time series data of the pilot valve actual spool position 410 and the main valve actual spool position 411 acquired by the detection information acquisition unit 401 into the internal oil leakage amount estimation model, and acquires the internal oil leakage amount output from the internal oil leakage amount estimation model as an estimation result.

[0099] The internal oil leakage amount estimated value output unit 413 outputs the estimated value of the internal oil leakage amount estimated by the internal oil leakage amount estimation unit 412. When the internal oil leakage amount estimated value output unit 413 estimates that the hydraulic servo valve 100 is in an abnormal state, it can notify the fact. Whether it is in an abnormal state can be determined by whether the value of the internal oil leakage amount output by the internal oil leakage amount estimation unit 412 is greater than a predetermined threshold value, or by whether the change amount of the internal oil leakage amount relative to the initial value is greater than a predetermined threshold value. In addition, the internal oil leakage amount estimated value output unit 413 can determine whether the hydraulic servo valve 100 is in an abnormal state by comparing the values ​​of the internal oil leakage amounts of a plurality of hydraulic servo valves 100 provided corresponding to a plurality of cylinders 81 of the same engine 80. For example, when the value of the internal oil leakage amount of a specific hydraulic servo valve 100 is greatly deviated from the average value of the internal oil leakage amounts of the plurality of hydraulic servo valves 100, it can be determined that the hydraulic servo valve 100 is in an abnormal state. A deviation value of the internal oil leakage amount of each of the plurality of hydraulic servo valves 100 may be calculated, and a hydraulic servo valve 100 having a deviation value greater than or equal to a predetermined threshold value, such as 80, or less than a predetermined threshold value, such as 30, may be determined to be in an abnormal state.

[0100] [Example 1-2]

[0101] Fig. 9 The structure of the learning device and the state estimation device involved in the embodiment 1-2 is shown. The learning device 300 of the embodiment 1-2 is Figure 8 In addition to the structure of the learning device 300 of the embodiment 1-1 shown in FIG. 1, a feature quantity calculation unit 315 is further provided. Figure 8 In addition to the configuration of the state estimation device 400 of the embodiment 1-1 shown, a feature quantity calculation unit 414 is further provided. The differences from the embodiment 1-1 will be mainly described. The other points are the same as those of the embodiment 1-1.

[0102] The characteristic quantity calculation unit 315 of the learning device 300 calculates a characteristic quantity correlated with the internal oil leakage amount based on the pilot valve actual valve core position 310 or the main valve actual valve core position 311 acquired by the learning data acquisition unit 301. As described above, the present inventors' experiments have shown that the internal oil leakage amount is correlated with the movement speed of the valve core 28 of the main valve 20 when the valve core 12 of the pilot valve 10 is in the neutral position. Therefore, the characteristic quantity calculation unit 315 calculates the movement speed of the valve core 28 of the main valve 20 based on the main valve actual valve core position 311 during the period when the pilot valve actual valve core position 310 is in the neutral position. In the case where there is a time lag from the time when the valve core 12 of the pilot valve 10 moves to the neutral position until the supply of the working oil to the main valve 20 stops and the position of the valve core 28 of the main valve 20 is stationary, the characteristic quantity calculation unit 315 can calculate the movement speed of the valve core 28 of the main valve 20 after adjusting the time lag. The adjustment amount of the time lag can be determined in advance through experiments or the like according to the environment in which the hydraulic servo valve 100 is used, such as the type and temperature of the hydraulic servo valve 100 , the pressure of the hydraulic oil 48 , the speed and load of the engine 80 , and the exhaust temperature.

[0103] The internal oil leakage amount estimation model generation unit 313 generates an internal oil leakage amount estimation model using the learning data acquired by the learning data acquisition unit 301 and the feature amount calculated by the feature amount calculation unit 315. The internal oil leakage amount estimation model generation unit 313 can generate the internal oil leakage amount estimation model using the internal oil leakage amount measured value 312 and the feature amount calculated by the feature amount calculation unit 315. For example, the internal oil leakage amount estimation model can be a formula that calculates the internal oil leakage amount using the feature amount as an input variable. In this case, the internal oil leakage amount estimation model generation unit 313 can generate the formula using a statistical method such as regression analysis. The internal oil leakage amount estimation model generation unit 313 can generate the internal oil leakage amount estimation model using the actual valve core position 310 of the pilot valve, the actual valve core position 311 of the main valve, and the feature amount. For example, the internal oil leakage amount estimation model may be a neural network that inputs the time series data of the pilot valve actual valve core position 310 and the main valve actual valve core position 311 during a specified period and the feature quantity during the specified period into an input layer and outputs the internal oil leakage amount from an output layer. In this case, the internal oil leakage amount estimation model generation unit 313 learns the internal oil leakage amount estimation model by adjusting the intermediate layer of the neural network in such a manner that a value approximate to the internal oil leakage amount measured value 312 is output from the output layer when the time series data of the pilot valve actual valve core position 310 and the main valve actual valve core position 311 during a specified period and the feature quantity during the specified period are input into the input layer.

[0104] The feature quantity calculation unit 414 of the state estimation device 400 calculates the feature quantity based on the pilot valve actual spool position 410 or the main valve actual spool position 411 acquired by the detection information acquisition unit 401 by the same method as the method in which the feature quantity calculation unit 315 calculates the feature quantity when generating the internal oil leakage amount estimation model used by the internal oil leakage amount estimation unit 412. In the case where the pilot valve actual spool position 410 and the main valve actual spool position 411 of the hydraulic servo valve 100 used in an environment different from the learning data used when generating the internal oil leakage amount estimation model are acquired by the detection information acquisition unit 401, the feature quantity calculation unit 414 can calculate the feature quantity after adjusting the difference in the environment. For example, the adjustment amount of the time lag can be changed according to the environment. In this way, the robustness of the estimation of the internal oil leakage amount can be improved.

[0105] [Examples 1-3]

[0106] Fig.10 The structures of the learning device and the state estimation device involved in the embodiments 1 to 3 are shown. Fig. 9 In the learning device 300 of the embodiment 1-2 shown in the figure, the learning data acquisition unit 301 acquires the actual valve core position 310 of the pilot valve, but in the learning device 300 of the embodiment 1-3, the learning data acquisition unit 301 acquires the target valve core position 316 of the pilot valve instead of the actual valve core position 310 of the pilot valve. Fig. 9 In the state estimation device 400 of the embodiment 1-2 shown, the detection information acquisition unit 401 acquires the actual valve core position 410 of the pilot valve, but in the state estimation device 400 of the embodiment 1-3, the detection information acquisition unit 401 acquires the target valve core position 415 of the pilot valve instead of the actual valve core position 410 of the pilot valve. The differences from the embodiment 1-2 are mainly described. The other points are the same as the embodiment 1-1 or 1-2.

[0107] As described above, depending on the connection position of the log data storage device 90, it is sometimes impossible to record the actual valve core position of the pilot valve when the hydraulic servo valve 100 is in use. In this case, the pilot valve target valve core position is used as learning data and detection information instead of the actual valve core position of the pilot valve. As a result, it is possible to generate an internal oil leakage estimation model using data that can be obtained regardless of the connection position of the log data storage device 90, and the internal oil leakage estimation model can be used to estimate the internal oil leakage. Fig.10 In the example, the feature quantity calculation unit 315 and the feature quantity calculation unit 414 are provided as in the embodiment 1-2, but the feature quantity calculation unit may not be provided as in the embodiment 1-1.

[0108] [Examples 1-4]

[0109] Fig.11The structure of the learning device and the state estimation device involved in the embodiment 1-4 is shown. The learning device 300 of the embodiment 1-4 is Fig.10 In addition to the structure of the learning device 300 of the embodiment 1-3 shown in the figure, an offset time calculation unit 319 is also provided. In addition, in the learning device 300 of the embodiment 1-4, the learning data acquisition unit 301 also acquires the coil drive voltage 317 and the pilot valve physical parameter 318. In addition, in the state estimation device 400 of the embodiment 1-4, the detection information acquisition unit 401 also acquires the coil drive voltage 416. The points different from the embodiment 1-3 are mainly described. The other points are the same as the embodiments 1-1 to 1-3.

[0110] After the command signal of the target position of the valve core 12 is inputted from the servo valve control device 110 to the pilot valve 10, a current is supplied to the coil of the valve core driving unit 18, and there is a time lag before the valve core 12 actually reaches the target position. Therefore, when the pilot valve target valve core position 316 is used as learning data instead of the pilot valve actual valve core position 310, by adjusting the offset time required until the actual position of the valve core 12 of the pilot valve 10 follows the pilot valve target valve core position 316, an internal oil leakage amount estimation model can be generated based on the position of the valve core 28 of the main valve 20 when the valve core 12 of the pilot valve 10 actually follows the neutral position, thereby improving the accuracy of the internal oil leakage amount estimation model.

[0111] The offset time calculation unit 319 calculates the offset time based on the coil drive voltage 317 and the pilot valve physical parameters 318. The pilot valve physical parameters 318 include, for example, physical quantities such as the resistance value of the elements constituting the drive circuit of the valve core drive unit 18, the inductance of the coil, the mass of the valve core 12, and the friction coefficient between the main body 10b and the valve core 12. The offset time calculation unit 319 calculates the offset time based on the motion equation including these physical parameters and the coil drive voltage 317. In the case where the offset time may differ depending on the position, movement direction, movement speed, etc. of the valve core 12 of the pilot valve 10, these data may also be used to calculate the offset time.

[0112] When the learning device 300 generates an internal oil leakage estimation model for estimating the internal oil leakage of a specific type of hydraulic servo valve 100, the pilot valve physical parameter 318 can be processed as a constant. Thus, log data in which the pilot valve physical parameter 318 is not recorded can also be used as learning data. As the pilot valve physical parameter 318, a constant predetermined according to the type of hydraulic servo valve 100 can also be used. As the pilot valve physical parameter 318, a physical parameter measured when the hydraulic servo valve 100 is shipped can also be used. Thus, the influence of the deviation of the pilot valve physical parameter caused by the individual difference of the hydraulic servo valve 100 can be suppressed, and the deviation time can be calculated more accurately. As the pilot valve physical parameter 318, a physical parameter measured during maintenance such as inspection of the hydraulic servo valve 100 can be used. Thus, the influence of the aging change of the pilot valve 10 can be suppressed, and the deviation time can be calculated more accurately.

[0113] The internal oil leakage amount estimation model generation unit 313 can generate an internal oil leakage amount estimation model that outputs the internal oil leakage amount by inputting the pilot valve actual valve core position 410 and the main valve actual valve core position 411 in the same manner as in Example 1-1 or 1-2, or can generate an internal oil leakage amount estimation model that outputs the internal oil leakage amount by inputting the pilot valve target valve core position 415 and the main valve actual valve core position 411 in the same manner as in Example 1-3, or can generate an internal oil leakage amount estimation model that outputs the internal oil leakage amount by inputting the pilot valve target valve core position 415, the main valve actual valve core position 411, and the coil drive voltage 416, or can generate an internal oil leakage amount estimation model that outputs the internal oil leakage amount by inputting the pilot valve target valve core position 415, the main valve actual valve core position 411, and the coil drive voltage 416 in addition to these. In the example of this figure, the internal oil leakage amount estimation model outputs the internal oil leakage amount by inputting the pilot valve target valve core position 415, the main valve actual valve core position 411, and the coil drive voltage 416. The coil drive voltage 416 supplied to the coil from the power supply circuit 120 of the servo valve control device 110 can change according to the number of revolutions of the engine 80, etc. Therefore, by generating an internal oil leakage estimation model that inputs the pilot valve target valve core position 415 and the main valve actual valve core position 411 and outputs the internal oil leakage amount by inputting the coil drive voltage 416, the influence of the change in offset time can be suppressed and the estimation accuracy can be improved.

[0114] The detection information acquisition unit 401 of the state estimation device 400 acquires detection information that needs to be input to the internal oil leakage estimation model used by the internal oil leakage estimation unit 412. In the example of this figure, the detection information acquisition unit 401 acquires the pilot valve target valve core position 415, the main valve actual valve core position 411, and the coil drive voltage 416. In the case where the pilot valve physical parameters need to be input to the internal oil leakage estimation model, the detection information acquisition unit 401 can also acquire the pilot valve physical parameters. Alternatively, the internal oil leakage estimation unit 412 can retain in advance the physical parameters corresponding to the type of the hydraulic servo valve 100 that is the object of estimating the internal oil leakage. In addition, in the case where the offset time needs to be input to the internal oil leakage estimation model, the detection information acquisition unit 401 can also acquire the offset time. Alternatively, the state estimation device 400 can also be further provided with an offset time calculation unit that calculates the offset time based on the coil drive voltage 416, the pilot valve physical parameters, etc.

[0115] [Examples 1-5]

[0116] Fig.12 The structure of the learning device and the state estimation device involved in the embodiment 1-5 is shown. The learning device 300 of the embodiment 1-5 is Figure 8 In addition to the structure of the learning device 300 of the embodiment 1-1 shown in FIG. 1, a data selection unit 320 is further provided. Figure 8 In addition to the configuration of the state estimation device 400 of the embodiment 1-1 shown, a data selection unit 417 is further provided. The differences from the embodiment 1-1 will be mainly described. The other points are the same as the embodiments 1-1 to 1-4.

[0117] In this embodiment, in order to improve the estimation accuracy using the internal oil leakage amount estimation model, the internal oil leakage amount estimation model is generated using log data when the hydraulic servo valve 100 is used in the same environment as the environment in which the hydraulic servo valve 100, which is the object of the internal oil leakage amount estimation by the state estimation device 400, is used. This can suppress the influence of the environment and improve the robustness of the estimation of the internal oil leakage amount.

[0118] The data selection unit 320 selects the learning data recorded when the hydraulic servo valve 100 is used in a specific environment from the learning data acquired by the learning data acquisition unit 301. The data selection unit 320 may also refer to the data indicating the use environment of the hydraulic servo valve 100 contained in the log data, test data, simulation data, etc., such as the exhaust temperature and pressure of the cylinder 81 of the engine 80, to select the learning data recorded when the hydraulic servo valve 100 is used in a specific environment. The data selection unit 320 may classify the learning data for each environment. In this case, an internal oil leakage amount estimation model may be generated for each environment. The learning data such as log data, test data, simulation data may be recorded or generated in a specific environment, and the learning data such as log data, test data, simulation data may be classified for each use environment when recorded or generated. In this case, the learning data acquisition unit 301 only needs to acquire the learning data recorded, generated, or classified in the specific environment, so the data selection unit 320 may not be provided.

[0119] The data selection unit 417 selects the detection information recorded when the hydraulic servo valve 100 is used in a specific environment from the detection information acquired by the detection information acquisition unit 401. The data selection unit 417 can select the detection information recorded in the same environment as the environment when the learning data used when the internal oil leakage amount estimation model used by the internal oil leakage amount estimation unit 412 is generated is recorded. In the case where a plurality of internal oil leakage amount estimation models generated for each use environment are retained in the state estimation device 400, the internal oil leakage amount estimation unit 412 can refer to the data indicating the use environment of the hydraulic servo valve 100 included in the detection information to select the internal oil leakage amount estimation model to be used. In this case, the data selection unit 417 does not need to be provided.

[0120] The internal oil leakage amount estimation model may be generated using learning data recorded when the hydraulic servo valve 100 is operated in a test mode in which the valve core 12 of the pilot valve 10 is operated in a specific pattern. In this case, the state estimation device 400 estimates the internal oil leakage amount using the detection information recorded when the hydraulic servo valve 100 is operated in the test mode. This can improve the robustness of the estimation of the internal oil leakage amount.

[0121] In order to further suppress the influence of the use environment, log data, test data, simulation data, etc. recorded when the engine 80 is stopped may be used as learning data. When a stop command is issued from the engine control device 91 to the servo valve control device 110 while the engine 80 is stopped but the hydraulic oil 48 is being pressurized by the hydraulic pump 42, the servo valve control device 110 repeatedly moves the valve core 12 of the pilot valve 10 in a small amplitude pattern to the extent that fuel is not supplied to the engine 80 in order to prevent the valve core 12 of the pilot valve 10 and the valve core 28 of the main valve 20 from sticking. The internal oil leakage amount estimation model is generated using the learning data recorded during the sticking prevention action, and when the state estimation device 400 estimates the internal oil leakage amount of the hydraulic servo valve 100, the influence of disturbances such as vibrations generated accompanying the operation of the engine 80 can be suppressed by using the log data recorded during the sticking prevention action, thereby improving the estimation accuracy. The pattern of the sticking prevention action may be a sine wave pattern, a sawtooth wave pattern, a dither pattern, etc. When the sticking prevention operation is performed according to a plurality of different patterns, the learning data and the detection information recorded when the sticking prevention operation is being performed according to a specific pattern may be used.

[0122] [Variation of Example 1]

[0123] In addition to the data used as learning data and detection information in Examples 1-1 to 1-5, any of other data such as engine speed, engine load, working oil pressure, working oil temperature, or any combination of these data can be used as learning data and detection information. Data that can affect data such as the actual valve core position of the pilot valve, the target valve core position of the pilot valve, and the actual valve core position of the main valve as environmental factors can also be selected and used. By generating an internal oil leakage estimation model that reflects the influence of these data, the robustness of the estimation of the internal oil leakage can be further improved.

[0124] The features described in the above embodiments 1-1 to 1-5 can be applied in any combination.

[0125] [Example 2: Failure determination of power supply]

[0126] In the second embodiment, a technique for determining failure of the power supply circuit 120 and the power supply that supplies power to the power supply circuit 120 is described. When the power supply or the power supply circuit 120 fails, the hydraulic servo valve 100 no longer operates normally, thereby causing abnormalities such as a decrease in the output of the engine 80. If the failure of the power supply or the power supply circuit 120 can be predicted with high accuracy, the power supply or the power supply circuit 120 can be replaced or repaired before the power supply or the power supply circuit 120 fails, thereby avoiding unexpected abnormalities.

[0127] The hydraulic servo valve 100 for controlling the fuel supply and exhaust to the engine 80 repeats a fixed operation during the driving of the engine 80, so a fixed pattern can be seen between the operation waveform of the valve core 12 of the pilot valve 10 or the valve core 28 of the main valve 20 and the power supply voltage waveform. Therefore, by modeling the difference between the operation waveform and the power supply voltage waveform pattern in normal state and the operation waveform and the power supply voltage waveform pattern when a sign of failure occurs, it is possible to determine a failure of the power supply or the power supply circuit 120.

[0128] Fig.13 The configuration of the learning device and the state estimation device according to Embodiment 2 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the power supply voltage 321 , the pilot valve actual spool position 310 , and the power supply failure precursor indicator value 322 as learning data.

[0129] The pilot valve actual spool position 310 is time series data when the hydraulic servo valve 100 is used within a predetermined period before or after the time point when the power failure precursor index value 322 is actually measured. The power supply voltage 321 is time series data of the power supply voltage detected by the voltage detection unit 150 when the pilot valve actual spool position 310 is recorded. The power supply voltage 321 may be an input voltage input to the power supply circuit 120, an output voltage output from the power supply circuit 120, or both. These time series data may be time series data of one to several cycles of the rotation action of the engine 80.

[0130] The power failure precursor index value 322 is calculated based on the accumulated operating time of the power supply circuit 120, the degradation state of components such as capacitors constituting the power supply circuit 120, the measured value of the emission noise released from the power supply circuit 120, etc. The calculation method of the power failure precursor index value 322 can be determined by experiments, field tests, etc.

[0131] When log data or test data is used as learning data, the power failure precursor indicator value 322 is calculated based on the actual measured values ​​of the above variables. Since the power failure precursor indicator value 322 is not considered to change rapidly during the use of the hydraulic servo valve 100, the power failure precursor indicator value 322 is considered to not change during a predetermined period before or after the time point when the variable used to calculate the power failure precursor indicator value 322 is actually measured, and the power supply voltage 321 and the actual spool position 310 of the pilot valve recorded during the predetermined period are associated with the power failure precursor indicator value 322 and set as learning data. The predetermined period can be a period during which the power failure precursor indicator value 322 remains unchanged to a certain extent during the use of the hydraulic servo valve 100, and can be specified by the use time of the hydraulic servo valve 100, the operation time of the engine 80, the number of revolutions of the engine 80, etc. When simulation data is used as learning data, the power failure precursor indicator value 322 is calculated based on the value of the variable input to the simulator as a simulation condition.

[0132] The power failure determination model generation unit 323 generates a power failure determination model used in the state estimation device 400 to determine a failure of the power supply or the power supply circuit 120, using the learning data acquired by the learning data acquisition unit 301. The power failure determination model may be a neural network or the like that inputs the time series data of the power supply voltage 321 and the actual valve core position 310 of the pilot valve during a predetermined period into an input layer and outputs a power failure precursor indicator value from an output layer. In this case, the power failure determination model generation unit 323 learns the power failure determination model by adjusting the intermediate layer of the neural network in such a manner that a value approximate to the power failure precursor indicator value 322 is output from the output layer when the time series data of the power supply voltage 321 and the actual valve core position 310 of the pilot valve during a predetermined period is input into the input layer.

[0133] The power failure determination model providing section 324 provides the power failure determination model generated by the power failure determination model generating section 323 to the state estimation device 400 .

[0134] The detection information acquisition unit 401 acquires, as detection information, a power supply voltage 420 and an actual pilot valve body position 410. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the power failure determination model.

[0135] The power failure determination unit 421 uses the power failure determination model generated by the learning device 300 to determine whether the power supply or the power supply circuit 120 of the hydraulic servo valve 100 has a fault. The power failure determination unit 421 inputs the time series data of the power supply voltage 420 and the actual spool position 410 of the pilot valve acquired by the detection information acquisition unit 401 into the power failure determination model, and acquires the power failure precursor indicator value output from the power failure determination model as a determination result. The power failure determination unit 421 can determine whether the power supply or the power supply circuit 120 has a fault based on the output power failure precursor indicator value. For example, when the output power failure precursor indicator value is greater than a predetermined threshold value, the power failure determination unit 421 can determine that the power supply or the power supply circuit 120 has a fault, and when the change amount of the power failure precursor indicator value relative to the initial value becomes greater than a predetermined threshold value, the power failure determination unit 421 can determine that the power supply or the power supply circuit 120 has a fault. The power failure determination unit 421 can estimate the period until a failure occurs in the power supply or the power supply circuit 120, that is, the life span, based on the output power failure precursor indicator value. For example, the power failure determination unit 421 can estimate the life span using a formula or the like that uses the output power failure precursor indicator value as a variable. The power failure determination model can be configured to output whether the power supply or the power supply circuit 120 has a failure or the life span of the power supply or the power supply circuit 120 instead of the power failure precursor indicator value, or to output whether the power supply or the power supply circuit 120 has a failure or the life span of the power supply or the power supply circuit 120 in addition to the power failure precursor indicator value.

[0136] When the power circuit 120 is provided in each of the plurality of hydraulic servo valves 100 provided corresponding to the plurality of cylinders 81 of the same engine 80, the power failure determination unit 421 can determine whether the power circuit 120 has failed by comparing the power failure precursor indicator values ​​of the respective hydraulic servo valves 100. For example, when the power failure precursor indicator value of a specific hydraulic servo valve 100 greatly deviates from the average value of the power failure precursor indicator values ​​of the plurality of hydraulic servo valves 100, it can be determined that the power circuit 120 of the hydraulic servo valve 100 has failed. The deviation value of the power failure precursor indicator value of each of the plurality of hydraulic servo valves 100 is calculated, and the power circuit 120 of the hydraulic servo valve 100 whose deviation value is greater than or equal to a predetermined threshold value, such as 80, or less than a predetermined threshold value, such as 30, is determined to have failed.

[0137] The power failure determination result output unit 422 outputs the result determined by the power failure determination unit 421. When the power failure determination unit 421 determines that the power supply or the power supply circuit 120 of the hydraulic servo valve 100 has failed, the power failure determination result output unit 422 may notify the failure. The power failure determination result output unit 422 may output the life of the power supply or the power supply circuit 120 determined by the power failure determination unit 421.

[0138] The present inventors have learned through experiments that the amount of drop in power supply voltage corresponding to the peak value of the load current is correlated with a failure of the power supply or the power supply circuit 120. Therefore, when an ammeter for detecting the value of the current supplied to the hydraulic servo valve 100 is provided, the current value measured by the ammeter can be used as learning data and detection information instead of the pilot valve actual spool position 310, or in addition to the pilot valve actual spool position 310. This can improve the accuracy of fault determination.

[0139] The drop in power supply voltage corresponding to the peak value of the load current can be calculated based on the power supply voltage and the actual valve core position of the pilot valve, etc. In this case, the learning device 300 may further include a feature quantity calculation unit 315 and the state estimation device 400 may further include a feature quantity calculation unit 414 as in Example 1-2.

[0140] The characteristic quantity calculation unit 315 of the learning device 300 calculates the amount of drop in the power supply voltage corresponding to the peak value of the load current as a characteristic quantity correlated with the failure of the power supply or the power supply circuit 120 based on the power supply voltage 321 and the pilot valve actual spool position 310 acquired by the learning data acquisition unit 301. The characteristic quantity calculation unit 315 can calculate the value of the current supplied to the hydraulic servo valve 100 based on the acceleration or speed of the spool 12 of the pilot valve 10 calculated based on the change of the pilot valve actual spool position 310. The characteristic quantity calculation unit 315 can calculate the amount of drop in the power supply voltage based on the value of the power supply voltage 321 when the calculated current value reaches the peak value. In the case where there is a time lag from the time when the power supply circuit 120 supplies power to the coil of the spool drive unit 18 of the pilot valve 10 until the spool 12 of the pilot valve 10 moves, the characteristic quantity calculation unit 315 can calculate the acceleration or speed of the spool 12 of the pilot valve 10 after adjusting the time lag. The adjustment amount of the time lag can be determined in advance through experiments or the like according to the environment in which the hydraulic servo valve 100 is used, such as the type and temperature of the hydraulic servo valve 100 , the pressure of the hydraulic oil 48 , the speed and load of the engine 80 , and the exhaust temperature.

[0141] The power failure determination model generation unit 323 generates a power failure determination estimation model using the learning data acquired by the learning data acquisition unit 301 and the feature quantity calculated by the feature quantity calculation unit 315. The power failure determination model generation unit 323 can generate a power failure determination model using the power failure precursor index value 322 and the feature quantity calculated by the feature quantity calculation unit 315. For example, the power failure determination model can be a formula that calculates the power failure precursor index value using the feature quantity as an input variable. In this case, the power failure determination model generation unit 323 can generate the formula by a statistical method such as regression analysis. The power failure determination model generation unit 323 can generate a power failure determination model using the power supply voltage 321, the actual valve core position 310 of the pilot valve, and the feature quantity. For example, the power failure determination model can be a neural network that inputs the time series data of the power supply voltage 321 and the actual valve core position 310 of the pilot valve in a prescribed period and the feature quantity of the prescribed period into an input layer and outputs the power failure precursor index value from an output layer. In this case, the power failure judgment model generation unit 323 learns the power failure judgment model by adjusting the intermediate layer of the neural network in such a way that a value approximate to the power failure precursor indicator value 322 is output from the output layer when the time series data of the power supply voltage 321 and the actual valve core position 310 of the pilot valve during a specified period and the feature value of the specified period are input into the input layer.

[0142] The characteristic quantity calculation unit 414 of the state estimation device 400 calculates the characteristic quantity from the power supply voltage 420 and the actual valve core position 410 of the pilot valve acquired by the detection information acquisition unit 401 by the same method as the method in which the characteristic quantity calculation unit 315 calculates the characteristic quantity when generating the power supply failure determination model used by the power supply failure determination unit 421. In the case where the power supply voltage 420 and the actual valve core position 410 of the hydraulic servo valve 100 used in an environment different from the learning data used when generating the power supply failure determination model are acquired by the detection information acquisition unit 401, the characteristic quantity calculation unit 414 can calculate the characteristic quantity after adjusting the difference in the environment. For example, the adjustment amount of the time lag can be changed according to the environment. In this way, the robustness of the power supply failure determination can be improved. In addition, since the drop amount of the power supply voltage corresponding to the peak value of the load current can be calculated as the characteristic quantity even if an ammeter is not installed, the efficiency of the circuit can be improved and the cost can be reduced.

[0143] As the characteristic quantity, a physical quantity indicating the characteristic of the operation of the valve body 12 of the pilot valve 10 may be calculated. For example, the acceleration, speed, overshoot, delay, etc. of the valve body 12 of the pilot valve 10 may be calculated as the characteristic quantity.

[0144] As learning data and detection information, the pilot valve target valve position or the main valve actual valve position may be used instead of the pilot valve actual valve core position, or the pilot valve target valve position or the main valve actual valve core position may be used in addition to the pilot valve actual valve core position. In this case, the learning device 300 may be provided with an offset time calculation unit 319 to adjust the offset time required until the actual valve core position of the pilot valve 10 follows the target valve core position of the pilot valve, as in Example 1-4. The offset time calculation unit 319 may calculate the offset time between the change in the current supplied to the hydraulic servo valve 100 and the change in the power supply voltage in order to calculate the amount of drop in the power supply voltage when the value of the current supplied to the hydraulic servo valve 100 reaches a peak value. The method of calculating the offset time may be the same as in Example 1-2.

[0145] As learning data and detection information, data recorded when the hydraulic servo valve 100 is used in a specific environment can be used. In this case, the learning device 300 can also include a data selection unit 320, and the state estimation device 400 can also include a data selection unit 417, as in Examples 1-5. The data selection unit 320 and the data selection unit 417 can select data recorded when the hydraulic servo valve 100 is actuated in a test mode in which the valve core 12 of the pilot valve 10 is actuated in a specific pattern. In this way, the robustness of fault judgment can be improved. The data selection unit 320 and the data selection unit 417 can select data recorded when the engine 80 is stopped and the valve core adhesion prevention action is being performed. In this way, the influence of interference can be suppressed and the accuracy of fault judgment can be improved.

[0146] [Example 3: Detection of foreign matter stuck]

[0147] In the third embodiment, a technique for detecting a foreign object stuck between the main body 10b and the valve core 12 of the pilot valve 10 is described. When a foreign object is stuck between the main body 10b and the valve core 12, the valve core 12 of the pilot valve 10 no longer operates normally, so an alarm is issued because the actual valve core position of the main valve 20 deviates from the target valve core position. In the past, the cause of the alarm could not be determined, but if the stuck foreign object can be detected with high accuracy, the cause of the alarm can be determined when the alarm is issued, so that appropriate measures can be taken quickly.

[0148] Fig.14 The operations of the valve element 12 of the pilot valve 10 and the valve element 28 of the main valve 20 are schematically shown. Fig.14(a) shows an example of operation when the hydraulic servo valve 100 is normally controlled. When the target valve core position of the main valve 20 is input at time t1, in order to move the valve core 28 of the main valve 20 in the positive direction, a positive value is input as the target valve core position of the pilot valve 10 at time t2. When the valve core 12 of the pilot valve 10 follows the target valve core position, the working oil is supplied to the main valve 20, and the valve core 28 of the main valve 20 starts to move in the positive direction at time t3. At time t4, the target valve core position of the main valve 20 is fixed to a certain value, and when the actual valve core position of the main valve 20 follows the target valve core position, the neutral position is input as the target valve core position of the pilot valve 10, and the valve core 28 of the main valve 20 is stationary at the target valve core position at time t5.

[0149] Fig.14 (b) shows an example of operation when a foreign object is stuck in the pilot valve 10. Assume that the operation before time t4 is the same as Fig.14 (a) is the same as that of (a), but at time t4, a foreign object is stuck in the pilot valve 10. The neutral position is input as the target valve core position of the pilot valve 10, but the valve core 12 cannot move due to the foreign object being stuck, so the working oil continues to be supplied to the main valve 20, and the valve core 28 of the main valve 20 continues to move in the positive direction. Since the feedback signal of the actual valve core position of the main valve continues to deviate from the target position, in order to return the valve core 28 of the main valve 20 to the target position, a negative value is continuously input as the target valve core position of the pilot valve 10. When the valve core 28 of the main valve 20 reaches the limit position in the positive direction, it no longer moves but stops at the limit position.

[0150] As described above, the action patterns of the valve core 12 of the pilot valve 10 and the valve core 28 of the main valve 20 are greatly different between normal and when a foreign object is stuck. Therefore, by modeling the difference between the action pattern in normal and the action pattern when a foreign object is stuck, the sticking of a foreign object can be detected. The action pattern can also be different depending on the type, size, amount, etc. of the foreign object, so it is possible to detect not only whether a foreign object is stuck, but also the type, size, amount, etc. of the stuck foreign object. Even in the case where the foreign object is naturally removed after being temporarily stuck, the history of the stuck foreign object can be detected from the log data, so it can be used as evidence of the cause of the malfunction. In the ship 2, etc., the working oil circulates not only in the hydraulic servo valve 100 for controlling the engine 80, but also in the hydraulic servo valves for controlling other devices, etc., so even if the foreign object temporarily stuck in the hydraulic servo valve 100 is naturally removed, the foreign object may be stuck in other hydraulic servo valves, etc. According to the technology of this embodiment, it is possible to detect the entrapment of foreign matter in the hydraulic servo valve 100 and perform appropriate maintenance such as replacement of hydraulic oil, thereby suppressing the entrapment of foreign matter in other devices.

[0151] Fig.15The configuration of the learning device and state estimation device according to Embodiment 3 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the main valve actual spool position 311 , the main valve target spool position 330 , the pilot valve target spool position 316 , and foreign matter data 331 as learning data.

[0152] The main valve actual spool position 311, the main valve target spool position 330, and the pilot valve target spool position 316 are time series data when the hydraulic servo valve 100 is used within a predetermined period before or after the time point when the foreign matter data 331 is actually measured. These time series data may be time series data of one to several cycles of the rotation action of the engine 80.

[0153] The foreign matter data 331 is data indicating the presence, type, size, amount, etc. of foreign matter stuck in the pilot valve 10. When log data is used as learning data, the foreign matter data 331 is the actual measured value of foreign matter stuck in the pilot valve 10. When test data is used as learning data, the foreign matter data 331 is the actual measured value of foreign matter stuck in the pilot valve 10. When simulation data is used as learning data, the foreign matter data 331 is the data of foreign matter input to the simulator as simulation conditions.

[0154] The foreign object detection model generation unit 333 generates a foreign object detection model used in the state estimation device 400 to detect the jamming of foreign objects, using the learning data acquired by the learning data acquisition unit 301. The foreign object detection model may be a neural network or the like that inputs the time series data of the actual valve core position 311 of the main valve, the target valve core position 330 of the main valve, and the target valve core position 316 of the pilot valve during a prescribed period into an input layer and outputs the foreign object data from an output layer. In this case, the foreign object detection model generation unit 333 learns the foreign object detection model by adjusting the intermediate layer of the neural network in such a way that a value approximate to the foreign object data 331 is output from the output layer when the time series data of the actual valve core position 311 of the main valve, the target valve core position 330 of the main valve, and the target valve core position 316 of the pilot valve during a prescribed period is input into the input layer.

[0155] The foreign object detection model providing unit 334 provides the foreign object detection model generated by the foreign object detection model generating unit 333 to the state estimating device 400 .

[0156] The detection information acquisition unit 401 acquires, as detection information, a main valve actual valve body position 411, a main valve target valve body position 430, and a pilot valve target valve body position 415. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the foreign object detection model.

[0157] The foreign object detection unit 431 uses the foreign object detection model generated by the learning device 300 to detect the jamming of foreign objects in the pilot valve 10. The foreign object detection unit 431 inputs the time series data of the main valve actual valve core position 411, the main valve target valve core position 430, and the pilot valve target valve core position 415 acquired by the detection information acquisition unit 401 into the foreign object detection model, and acquires the foreign object data output from the foreign object detection model as a determination result. The foreign object detection unit 431 can estimate the period until the foreign object is jammed into the pilot valve 10 based on the output foreign object data. In this case, the period until the foreign object is jammed is estimated by comparing the foreign object data of the plurality of hydraulic servo valves 100 provided corresponding to the plurality of cylinders 81 of the same engine 80.

[0158] The foreign object detection result output unit 432 outputs the result determined by the foreign object detection unit 431. When the foreign object detection unit 431 detects that a foreign object is stuck in the pilot valve 10, the foreign object detection result output unit 432 can notify this fact. The foreign object detection result output unit 432 can output the period until the foreign object is stuck estimated by the foreign object detection unit 431.

[0159] The feature quantity that indicates whether or not a foreign object is stuck, or that is correlated with the type, size, or amount of the stuck foreign object, may be calculated based on the actual valve core position of the main valve, the target valve core position of the main valve, or the target valve core position of the pilot valve. In this case, the learning device 300 may further include a feature quantity calculation unit 315, and the state estimation device 400 may further include a feature quantity calculation unit 414, as in the first embodiment 1-2. In the case where there is a time lag from the time when the power supply circuit 120 supplies power to the coil of the valve core drive unit 18 of the pilot valve 10 until the valve core 12 of the pilot valve 10 moves and the actual valve core position of the main valve follows the target valve core position of the main valve, the feature quantity calculation unit 315 may calculate the feature quantity after adjusting the time lag. The adjustment amount of the time lag may be determined in advance by experiments, etc., based on the environment in which the hydraulic servo valve 100 is used, such as the type and temperature of the hydraulic servo valve 100, the pressure of the working oil 48, the number of revolutions of the engine 80, the load, and the exhaust temperature.

[0160] As learning data and detection information, the actual pilot valve core position may be used instead of the pilot valve target core position, or the actual pilot valve core position may be used in addition to the pilot valve target core position. As learning data and detection information, a combination of any two or more positions of the pilot valve target core position, the pilot valve actual core position, the main valve target core position, and the main valve actual core position may be used.

[0161] As learning data and detection information, data recorded when the hydraulic servo valve 100 is used in a specific environment can be used. In this case, the learning device 300 can also include a data selection unit 320, and the state estimation device 400 can also include a data selection unit 417, as in Examples 1-5. The data selection unit 320 and the data selection unit 417 can select data recorded when the hydraulic servo valve 100 is actuated in a test mode in which the valve core 12 of the pilot valve 10 is actuated in a specific pattern. In this way, the robustness of the detection of foreign matter jamming can be improved. The data selection unit 320 and the data selection unit 417 can select data recorded when the engine 80 is stopped and the valve core adhesion prevention action is being performed. In this way, the influence of interference can be suppressed and the detection accuracy of foreign matter jamming can be improved.

[0162] [Example 4: Estimation of Cleanliness of Working Oil]

[0163] In the fourth embodiment, a technique for estimating the cleanliness of the hydraulic oil is described. When foreign matter is mixed into the hydraulic oil during the operation of the hydraulic servo valve 100 and the cleanliness is reduced, the operation of the valve core 12 of the pilot valve 10 is hindered due to the influence of the mixed foreign matter, or the pressure of the hydraulic oil supplied to the main valve 20 is reduced, which sometimes causes the engine 80 and other controlled objects to malfunction. In addition, the possibility of foreign matter getting stuck in the valve core 12 of the pilot valve 10 increases. By accurately estimating the cleanliness of the hydraulic oil and taking appropriate measures promptly when the cleanliness of the hydraulic oil is reduced, the hydraulic servo valve 100 and the operation of the controlled object can be well maintained, and foreign matter can be prevented from getting stuck.

[0164] When the cleanliness of the working oil is reduced due to the mixing (contamination) of a tiny amount of foreign matter, the friction force generated between the valve core 12 and the main body 10b increases when the valve core 12 of the pilot valve 10 slides, and the behavior of the valve core 12 changes. Therefore, by modeling the difference between the waveform of the current supplied to the coil of the valve core drive unit 18 of the pilot valve 10 and the pattern of the action waveform of the target valve core position and the actual valve core position of the pilot valve 10 according to the cleanliness of the working oil, the cleanliness of the working oil can be estimated.

[0165] Fig.16 The configuration of the learning device and the state estimation device according to Embodiment 4 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the coil applied current 341, the pilot valve target spool position 316, the pilot valve actual spool position 310, and the working oil cleanliness measured value 342 as learning data.

[0166] The coil applied current 341, the pilot valve target spool position 316, and the pilot valve actual spool position 310 are time series data when the hydraulic servo valve 100 is used within a predetermined period before or after the time point when the hydraulic oil cleanliness actual measured value 342 is actually measured. These time series data may be time series data of one to several cycles of the rotation operation of the engine 80.

[0167] The working oil cleanliness actual measurement value 342 is data indicating the type, quantity, size, etc. of foreign matter mixed into the working oil. When log data or test data is used as learning data, the working oil cleanliness actual measurement value 342 is a measured value obtained by measuring the cleanliness of the working oil used in the hydraulic servo valve 100 by a general liquid particle measuring device or the like. The working oil cleanliness can be measured, for example, by a weight method, a microscope method, a light scattering method, a light resistance method, a resistance method, an acoustic method, a dynamic light scattering method, or the like. Since it is believed that the cleanliness of the working oil does not change drastically during the use of the hydraulic servo valve 100, it is considered that the working oil of the actually measured cleanliness is used during a predetermined period before or after the time point when the cleanliness of the working oil is actually measured, and the coil applied current 341, the pilot valve target spool position 316, and the pilot valve actual spool position 310 recorded during the predetermined period are associated with the working oil cleanliness actual measurement value 342 as learning data. The predetermined period may be a period during which the cleanliness of the hydraulic oil is kept constant during the use of the hydraulic servo valve 100, and may be determined by the use time of the hydraulic servo valve 100, the operation time of the engine 80, the number of revolutions of the engine 80, etc. When simulation data is used as learning data, the hydraulic oil cleanliness actual measurement value 342 is a value of the cleanliness of the hydraulic oil input to the simulator as a simulation condition.

[0168] The working oil cleanliness estimation model generation unit 343 generates a working oil cleanliness estimation model used in the state estimation device 400 to estimate the cleanliness of the working oil used, using the learning data acquired by the learning data acquisition unit 301. The working oil cleanliness estimation model may be a neural network or the like that inputs the time series data of the coil applied current 341, the pilot valve target spool position 316, and the pilot valve actual spool position 310 during a predetermined period into an input layer and outputs the working oil cleanliness from an output layer. In this case, the working oil cleanliness estimation model generation unit 343 learns the working oil cleanliness estimation model by adjusting the intermediate layer of the neural network in such a manner that a value approximate to the working oil cleanliness actual measurement value 342 is output from the output layer when the time series data of the coil applied current 341, the pilot valve target spool position 316, and the pilot valve actual spool position 310 during a predetermined period is input into the input layer.

[0169] The hydraulic oil cleanliness estimation model providing unit 344 provides the hydraulic oil cleanliness estimation model generated by the hydraulic oil cleanliness estimation model generating unit 343 to the state estimation device 400 .

[0170] The detection information acquisition unit 401 acquires the coil applied current 441, the pilot valve target spool position 415, and the pilot valve actual spool position 410 as detection information. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the hydraulic oil cleanliness estimation model.

[0171] The working oil cleanliness estimation unit 442 estimates the cleanliness of the working oil using the working oil cleanliness estimation model generated by the learning device 300. The working oil cleanliness estimation unit 442 obtains the working oil cleanliness output from the working oil cleanliness estimation model by inputting the coil applied current 441, the pilot valve target spool position 415, and the pilot valve actual spool position 410 time series data acquired by the detection information acquisition unit 401 into the working oil cleanliness estimation model as an estimation result.

[0172] The working oil cleanliness estimation result output unit 443 outputs the result estimated by the working oil cleanliness estimation unit 442. The working oil cleanliness estimation result output unit 443 determines whether maintenance such as cleaning or replacement of the working oil is required based on the output working oil cleanliness, and when it is determined that the working oil needs to be maintained, it can notify this fact. For example, the working oil cleanliness estimation result output unit 443 can determine that the working oil needs to be maintained when the output working oil cleanliness is less than a predetermined threshold value, or can determine that the working oil needs to be maintained when the change in the working oil cleanliness relative to the initial value is greater than a predetermined threshold value. The working oil cleanliness estimation result output unit 443 can estimate the period until the working oil needs to be maintained and output the period. For example, the working oil cleanliness estimation result output unit 443 can estimate the period until the working oil needs to be maintained using a formula that uses the output working oil cleanliness as a variable. The hydraulic oil cleanliness estimation model may be configured to output whether hydraulic oil maintenance is required or the period until hydraulic oil maintenance becomes necessary instead of or in addition to the hydraulic oil cleanliness.

[0173] When a plurality of hydraulic servo valves 100 are provided corresponding to a plurality of cylinders 81 of the same engine 80, the hydraulic oil cleanliness estimation unit 442 may determine whether maintenance of the hydraulic oil is required based on the cleanliness of the hydraulic oil estimated in each hydraulic servo valve 100. When common hydraulic oil is circulated in a plurality of hydraulic servo valves 100, if the cleanliness of the hydraulic oil estimated in any one of the hydraulic servo valves 100 is less than a predetermined threshold value, even if the cleanliness of the hydraulic oil estimated in the other hydraulic servo valves 100 is greater than the threshold value, the circulation of the hydraulic oil with low cleanliness may affect the operation of the other hydraulic servo valves 100. Therefore, when the cleanliness of the hydraulic oil estimated in at least one hydraulic servo valve 100 is less than the predetermined threshold value, the hydraulic oil cleanliness estimation result output unit 443 may determine that maintenance of the hydraulic oil is required. The hydraulic oil cleanliness estimation result output unit 443 may also determine whether maintenance of the hydraulic oil is required based on the order in which the hydraulic oil circulates. For example, the threshold value used to determine whether the hydraulic servo valve 100 on the upstream side needs to be maintained may be higher than the threshold value used to determine whether the hydraulic servo valve 100 on the downstream side needs to be maintained. The hydraulic oil cleanliness estimation result output unit 443 may calculate a statistical value such as an average value of the cleanliness of the hydraulic oil estimated in the plurality of hydraulic servo valves 100, and determine whether the hydraulic oil needs to be maintained based on the calculated statistical value.

[0174] The characteristic quantity correlated with the cleanliness of the working oil can be calculated based on the coil applied current, the pilot valve target spool position, the pilot valve actual spool position, etc. In this case, the learning device 300 can further include a characteristic quantity calculation unit 315, and the state estimation device 400 can further include a characteristic quantity calculation unit 414, as in Example 1-2. When the cleanliness of the working oil decreases and the friction between the valve core 12 of the pilot valve 10 and the main body 10b increases, it is considered that the speed or acceleration of the valve core 12 decreases, so the characteristic quantity calculation unit can calculate the speed or acceleration of the valve core 12 of the pilot valve 10 by calculating the rate of change of the actual valve core position of the pilot valve or the rate of change of the rate of change.

[0175] The working oil cleanliness estimation model generation unit 343 generates a working oil cleanliness estimation model using the learning data acquired by the learning data acquisition unit 301 and the feature quantity calculated by the feature quantity calculation unit 315. The working oil cleanliness estimation model generation unit 343 generates a working oil cleanliness estimation model using the working oil cleanliness actual measurement value 342 and the feature quantity calculated by the feature quantity calculation unit 315. For example, the working oil cleanliness estimation model may be a formula that calculates the working oil cleanliness using the feature quantity as an input variable. In this case, the working oil cleanliness estimation model generation unit 343 may generate the formula using a statistical method such as regression analysis. The working oil cleanliness estimation model may be a formula that calculates the working oil cleanliness using the speed or acceleration of the valve core 12 when the coil applied current or the pilot valve target valve core position satisfies a prescribed condition as an input variable. For example, the speed or acceleration of the valve core 12 at the peak of the coil applied current may be used as an input variable, or the speed or acceleration of the valve core 12 when the valve core 12 moves from the neutral position may be used as an input variable. The working oil cleanliness estimation model generation unit 343 can generate a working oil cleanliness estimation model using the coil applied current 341, the pilot valve target spool position 316, the pilot valve actual spool position 310, and the feature value. For example, the working oil cleanliness estimation model can be a neural network that inputs the time series data of the coil applied current 341, the pilot valve target spool position 316, and the pilot valve actual spool position 310 during a specified period, and the feature value during the specified period into an input layer, and outputs the working oil cleanliness from an output layer. In this case, the working oil cleanliness estimation model generation unit 343 learns the working oil cleanliness estimation model by adjusting the intermediate layer of the neural network in such a way that when the time series data of the coil applied current 341, the pilot valve target spool position 316, and the pilot valve actual spool position 310 during a specified period, and the feature value during the specified period are input into the input layer, a value approximate to the working oil cleanliness actual measurement value 342 is output from the output layer.

[0176] The feature quantity calculation unit 414 of the state estimation device 400 calculates the feature quantity based on the coil applied current 441, the pilot valve target spool position 415, or the pilot valve actual spool position 410 acquired by the detection information acquisition unit 401 by the same method as the method in which the feature quantity calculation unit 315 calculates the feature quantity when generating the working oil cleanliness estimation model used by the working oil cleanliness estimation unit 442. In the case where the coil applied current 441, the pilot valve target spool position 415, and the pilot valve actual spool position 410 of the hydraulic servo valve 100 used in an environment different from the learning data used when generating the working oil cleanliness estimation model are acquired by the detection information acquisition unit 401, the feature quantity calculation unit 414 can calculate the feature quantity after adjusting the difference in the environment. In this way, the robustness of the working oil cleanliness estimation can be improved.

[0177] As the characteristic quantity, other physical quantities that indicate the characteristics of the operation of the valve body 12 of the pilot valve 10 may be calculated. For example, the overshoot amount and delay amount of the valve body 12 of the pilot valve 10 may be calculated as the characteristic quantity.

[0178] As learning data and detection information, data recorded when the hydraulic servo valve 100 is used in a specific environment can be used. In this case, the learning device 300 can also include a data selection unit 320, and the state estimation device 400 can also include a data selection unit 417, as in Examples 1-5. The data selection unit 320 and the data selection unit 417 can select data recorded when the hydraulic servo valve 100 is actuated in a test mode in which the valve core 12 of the pilot valve 10 is actuated in a specific pattern. In this way, the robustness of the estimation of the cleanliness of the working oil can be improved. The data selection unit 320 and the data selection unit 417 can select data recorded when the engine 80 is stopped and the valve core adhesion prevention action is being performed. In this way, the influence of interference can be suppressed and the estimation accuracy of the cleanliness of the working oil can be improved.

[0179] In addition to the data used as learning data and detection information, any of other data such as engine speed, engine load, pressure of the working oil pressurized by the hydraulic pump 42, pressure or flow of the working oil before and after the filter for removing foreign matter mixed in the working oil, temperature of the working oil, power supply voltage, power supply current, or any combination thereof may be used as learning data and detection information. Data that can affect data such as coil current, pilot valve target spool position, and pilot valve actual spool position as environmental factors may be selected and used. Not only the cleanliness of the working oil, but also the friction force when the valve core 12 of the pilot valve 10 slides may change due to the wear of the valve core 12 of the pilot valve 10 or the inner wall of the main body 10b, so factors that can affect the wear of the valve core 12 of the pilot valve 10 or the inner wall of the main body 10b, such as the cumulative use time of the pilot valve 10 and the cumulative movement distance of the valve core 12, may also be used as learning data and detection information. By generating a hydraulic oil cleanliness estimation model that reflects the influence of these data, the robustness of the estimation of the hydraulic oil cleanliness can be further improved.

[0180] [Example 5: Position sensor disconnection detection]

[0181] In the fifth embodiment, a technique for detecting a disconnection of the position sensor 29 for detecting the actual position of the valve core 28 of the main valve 20 is described. When the position sensor 29 is disconnected, a signal indicating the correct actual position of the valve core 28 can no longer be received from the position sensor 29, so the actual valve core position of the main valve 20 deviates from the target valve core position. As described in the third embodiment, in the past, an alarm was issued when the actual valve core position of the main valve 20 deviated from the target valve core position, but the cause of the alarm could not be determined. If the disconnection of the position sensor 29 can be detected with high accuracy, the cause of the alarm can be determined when the alarm is issued, so that appropriate measures can be taken quickly.

[0182] Fig.17 The operations of the valve element 12 of the pilot valve 10 and the valve element 28 of the main valve 20 are schematically shown. Fig.17 (a) shows an operation example when the hydraulic servo valve 100 is normally controlled. Fig.17 The action example shown in (a) is the same as Fig.14 The action example is the same as shown in (a).

[0183] Fig.17 (b) shows an example of operation when the position sensor 29 of the valve body 28 of the main valve 20 is disconnected. Fig.17 (a) is the same as that of FIG. 1 , but at time t6, the position sensor 29 of the valve core 28 of the main valve 20 is disconnected. The feedback signal of the actual valve core position of the main valve 20 changes sharply to an abnormal value (for example, zero) immediately after the disconnection occurs, and then remains at the abnormal value and is fixed. In the example of this figure, since the actual valve core position of the main valve 20 is fixed in a state of deviating from the target valve core position in the positive direction, in order to return the valve core 28 of the main valve 20 to the target valve core position, the negative maximum value is continuously input as the target valve core position of the pilot valve 10. Therefore, as shown by the dotted line, the valve core 28 of the main valve 20 continues to move in the negative direction, and when it reaches the limit position in the negative direction, it no longer moves but stops at the limit position.

[0184] As described above, the action patterns of the valve body 12 of the pilot valve 10 and the valve body 28 of the main valve 20 are greatly different between the normal state and the state when the position sensor 29 is disconnected. Therefore, by modeling the difference between the action pattern in the normal state and the action pattern when the position sensor 29 is disconnected, the disconnection of the position sensor 29 can be detected. Even when time has passed since the alarm for notifying that the actual valve body position of the main valve 20 has deviated from the target valve body position is issued, the history of the disconnection of the position sensor 29 can be detected from the log data, and thus can be used as evidence of the cause of the malfunction. According to the technology of this embodiment, when the alarm for notifying that the actual valve body position of the main valve 20 has deviated from the target valve body position is issued, it is possible to accurately determine whether it is caused by the insertion of foreign matter described in the third embodiment or the disconnection of the position sensor 29 described in this embodiment, and appropriate maintenance can be performed. In addition, the occurrence of oscillation due to poor contact of the position sensor 29 can also be accurately detected by modeling the action patterns of the actual valve body position of the main valve 20 and the target valve body position of the pilot valve 10. Thus, an abnormality can be detected before the position sensor 29 is disconnected, and appropriate maintenance can be performed, so that malfunction due to disconnection of the position sensor 29 can be prevented from occurring.

[0185] Fig.18 The configuration of the learning device and state estimation device according to Embodiment 5 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the main valve actual spool position 311 , the pilot valve target spool position 316 , and the position sensor disconnection data 351 as learning data.

[0186] The main valve actual spool position 311 and the pilot valve target spool position 316 are time series data when the hydraulic servo valve 100 is used within a predetermined period before or after the time point actually measured by the position sensor disconnection data 351. These time series data may be time series data of one to several cycles of the rotation operation of the engine 80.

[0187] The position sensor disconnection data 351 is data indicating the presence, timing, method, cause, etc. of disconnection of the position sensor 29 of the valve body 28 of the main valve 20. When log data is used as learning data, the position sensor disconnection data 351 is an actual measured value when the position sensor 29 is actually disconnected. When test data is used as learning data, the position sensor disconnection data 351 is an actual measured value when the position sensor 29 is disconnected. When simulation data is used as learning data, the position sensor disconnection data 351 is data input to the simulator as a simulation condition.

[0188] The position sensor disconnection detection model generation unit 353 generates a position sensor disconnection detection model used in the state estimation device 400 to detect disconnection of the position sensor 29, using the learning data acquired by the learning data acquisition unit 301. The position sensor disconnection detection model may be a neural network or the like that inputs the time series data of the main valve actual valve body position 311 and the pilot valve target valve body position 316 during a predetermined period into an input layer and outputs the position sensor disconnection data from an output layer. In this case, the position sensor disconnection detection model generation unit 353 learns the position sensor disconnection detection model by adjusting the intermediate layer of the neural network in such a manner that a value approximate to the position sensor disconnection data 351 is output from the output layer when the time series data of the main valve actual valve body position 311 and the pilot valve target valve body position 316 during a predetermined period is input into the input layer.

[0189] The position sensor disconnection detection model providing unit 354 provides the position sensor disconnection detection model generated by the position sensor disconnection detection model generating unit 353 to the state estimation device 400 .

[0190] The detection information acquisition unit 401 acquires, as detection information, a main valve actual spool position 411 and a pilot valve target spool position 415. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the position sensor disconnection detection model.

[0191] The position sensor disconnection detection unit 451 detects disconnection of the position sensor 29 of the valve core 28 of the main valve 20 using the position sensor disconnection detection model generated by the learning device 300. The position sensor disconnection detection unit 451 inputs the time series data of the main valve actual valve core position 411 and the pilot valve target valve core position 415 acquired by the detection information acquisition unit 401 into the position sensor disconnection detection model, and acquires the position sensor disconnection data output from the position sensor disconnection detection model as a determination result. The position sensor disconnection detection unit 451 can estimate the period until the position sensor 29 is disconnected based on the output position sensor disconnection data. In this case, the period until the position sensor 29 is disconnected can also be estimated by comparing the position sensor disconnection data of a plurality of hydraulic servo valves 100 provided corresponding to a plurality of cylinders 81 of the same engine 80.

[0192] The position sensor disconnection detection result output unit 452 outputs the result determined by the position sensor disconnection detection unit 451. When the position sensor disconnection detection unit 451 detects that the position sensor 29 is disconnected, the position sensor disconnection detection result output unit 452 may notify the position sensor disconnection detection result output unit 452. The position sensor disconnection detection result output unit 452 may also output the period until the position sensor 29 is disconnected estimated by the position sensor disconnection detection unit 451.

[0193] The feature quantity correlated with the disconnection of the position sensor 29 can be calculated based on the actual valve core position of the main valve or the target valve core position of the pilot valve. In this case, the learning device 300 may further include a feature quantity calculation unit 315, and the state estimation device 400 may further include a feature quantity calculation unit 414, as in Example 1-2. When the position sensor 29 of the valve core 28 of the main valve 20 is disconnected, the actual valve core position of the main valve changes rapidly to an abnormal value, so the feature quantity calculation unit can calculate the speed or acceleration of the valve core 28 of the main valve 20 by calculating the rate of change of the actual valve core position of the main valve or the rate of change of the rate of change.

[0194] The position sensor disconnection detection model generation unit 353 generates a position sensor disconnection detection model using the learning data acquired by the learning data acquisition unit 301 and the feature quantity calculated by the feature quantity calculation unit 315. The position sensor disconnection detection model generation unit 353 can generate a position sensor disconnection detection model using the position sensor disconnection data 351 and the feature quantity calculated by the feature quantity calculation unit 315. The position sensor disconnection detection model generation unit 353 can generate a position sensor disconnection detection model using the main valve actual valve core position 311, the pilot valve target valve core position 316, and the feature quantity. For example, the position sensor disconnection detection model can be a neural network that inputs time series data of the main valve actual valve core position 311 and the pilot valve target valve core position 316 during a specified period and the feature quantity of the specified period into an input layer and outputs the position sensor disconnection data from an output layer. In this case, the position sensor disconnection detection model generation unit 353 learns the position sensor disconnection detection model by adjusting the intermediate layer of the neural network in such a way that a value approximate to the position sensor disconnection data 351 is output from the output layer when the time series data of the actual valve core position 311 of the main valve and the target valve core position 316 of the pilot valve during a specified period and the feature value of the specified period are input into the input layer.

[0195] The feature quantity calculation unit 414 of the state estimation device 400 calculates the feature quantity based on the main valve actual spool position 411 or the pilot valve target spool position 415 acquired by the detection information acquisition unit 401 by the same method as the method in which the feature quantity calculation unit 315 calculates the feature quantity when generating the position sensor disconnection detection model used by the position sensor disconnection detection unit 451. In the case where the main valve actual spool position 411 and the pilot valve target spool position 415 of the hydraulic servo valve 100 used in an environment different from the learning data used when generating the position sensor disconnection detection model are acquired by the detection information acquisition unit 401, the feature quantity calculation unit 414 can calculate the feature quantity after adjusting the difference in the environment. In this way, the robustness of the position sensor disconnection detection can be improved.

[0196] As the characteristic amount, other physical quantities that indicate the characteristics of the operation of the valve body 28 of the main valve 20 may be calculated. For example, the overshoot amount and the delay amount of the valve body 28 of the main valve 20 may be calculated as the characteristic amount.

[0197] As learning data and detection information, the actual pilot valve core position may be used instead of the pilot valve target core position, or the actual pilot valve core position may be used in addition to the pilot valve target core position. In this case, the learning device 300 may be provided with an offset time calculation unit 319 to adjust the offset time required until the actual valve core position of the pilot valve 10 follows the target valve core position of the pilot valve, as in Example 1-4. The method for calculating the offset time may be the same as in Example 1-2. As learning data and detection information, a combination of any two or more positions of the pilot valve target core position, the pilot valve actual core position, the main valve target core position, and the main valve actual core position may be used.

[0198] As learning data and detection information, data recorded when the hydraulic servo valve 100 is used in a specific environment can be used. In this case, the learning device 300 can also include a data selection unit 320, and the state estimation device 400 can also include a data selection unit 417, as in Examples 1-5. The data selection unit 320 and the data selection unit 417 can select data recorded when the hydraulic servo valve 100 is operated in a test mode in which the valve core 12 of the pilot valve 10 is operated in a specific pattern. In this way, the robustness of the disconnection detection of the position sensor 29 can be improved. The data selection unit 320 and the data selection unit 417 can select data recorded when the engine 80 is stopped and the valve core adhesion prevention action is being performed. In this way, the influence of interference can be suppressed and the detection accuracy of the disconnection of the position sensor 29 can be improved.

[0199] [Example 6: Position sensor looseness detection]

[0200] In the sixth embodiment, a technique for detecting a situation in which the position sensor 29 for detecting the actual position of the valve core 28 of the main valve 20 is displaced due to loosening of a screw for fixing the position sensor 29 for detecting the actual position of the valve core 28 of the main valve 20, etc. is described. When the position sensor 29 is displaced, a feedback signal of the actual valve core position of the main valve is offset from the actual position of the valve core 28 of the main valve 20, so that the valve core 28 of the main valve 20 can no longer be normally controlled, which may cause malfunction of a controlled object such as the engine 80. If the situation in which the position sensor 29 is displaced can be detected with high accuracy, appropriate measures such as tightening the screws or replacing the position sensor 29 can be taken quickly.

[0201] Fig.19 The state of the position sensor 29 of the valve element 28 of the main valve 20 is shown. Fig.19 (a) shows a state where the position sensor 29 is normally installed. The position sensor 29 is screwed into the internal thread 30 cut out at the upper end of the valve core 28. When the screw is loosened, as shown in FIG. Fig.19 As shown in (b), a gap is generated between the position sensor 29 and the valve core 28, so the position of the valve core 28 is shifted downward by an amount corresponding to the gap d. However, since the position of the valve core 28 is detected by the position sensor 29, Fig.19 (a) and Fig.19 In the case of (b), the same position is detected, and the actual position of the valve core 28 which is offset downward cannot be detected. Fig.19 In the state (b), although the actual main valve spool position seems to correctly follow the main valve target spool position, the actual position of the spool 28 is also offset below the main valve target spool position, so the operation of the controlled object is problematic. In this embodiment, since the amount of control oil supplied to the fuel injection actuator for supplying fuel to the cylinder 81 of the engine 80 is reduced, the amount of fuel supplied to the cylinder 81 is reduced, so the pressure in the cylinder 81 or the temperature of the exhaust gas discharged from the cylinder 81 is reduced, and an alarm is issued.

[0202] In this way, when the position sensor 29 is loose and the position is shifted, the actual valve core position of the main valve follows the target valve core position of the main valve, but the pressure and exhaust temperature of the cylinder 81 are reduced. Therefore, when an alarm is issued due to an abnormality in the pressure or exhaust temperature of the cylinder 81, by confirming whether the actual valve core position of the main valve follows the target valve core position of the main valve, it is possible to determine whether the cause is an abnormality in the position sensor 29. Even if time has passed since the alarm for notifying the abnormality of the engine 80 was issued, it is possible to detect the looseness of the position sensor 29 based on the log data, and thus use it as evidence of the cause of the malfunction. As a result, the looseness of the position sensor 29 can be accurately detected to implement appropriate maintenance, so that the cost and man-hours of maintenance required to improve the malfunction of the engine 80 can be reduced.

[0203] Fig. 20 The configuration of the learning device and the state estimation device according to Embodiment 6 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the main valve actual spool position 311, the main valve target spool position 330, the cylinder exhaust temperature 361, and the position sensor looseness data 362 as learning data.

[0204] The main valve actual spool position 311, the main valve target spool position 330, and the cylinder exhaust temperature 361 are time series data when the hydraulic servo valve 100 is used within a predetermined period before or after the time point when the position sensor looseness data 362 is actually measured. These time series data may be time series data of one to several cycles of the rotation action of the engine 80.

[0205] The position sensor looseness data 362 is data indicating the presence, degree, method, cause, etc. of the position deviation of the position sensor 29 of the valve core 28 of the main valve 20. When log data is used as learning data, the position sensor looseness data 362 is an actual measured value when the position sensor 29 is actually deviated. When test data is used as learning data, the position sensor looseness data 362 is an actual measured value when the position sensor 29 is deviated. When simulation data is used as learning data, the position sensor looseness data 362 is data input to the simulator as a simulation condition.

[0206] The position sensor looseness detection model generation unit 363 generates a position sensor looseness detection model used in the state estimation device 400 to detect the looseness of the position sensor 29 using the learning data acquired by the learning data acquisition unit 301. The position sensor looseness detection model may be a neural network or the like that inputs the time series data of the main valve actual spool position 311, the main valve target spool position 330, and the cylinder exhaust temperature 361 during a prescribed period into an input layer and outputs the position sensor looseness data from an output layer. In this case, the position sensor looseness detection model generation unit 363 learns the position sensor looseness detection model by adjusting the intermediate layer of the neural network in such a manner that a value approximate to the position sensor looseness data 362 is output from the output layer when the time series data of the main valve actual spool position 311, the main valve target spool position 330, and the cylinder exhaust temperature 361 during a prescribed period is input into the input layer.

[0207] The position sensor looseness detection model providing section 364 provides the position sensor looseness detection model generated by the position sensor looseness detection model generating section 363 to the state estimating device 400 .

[0208] The detection information acquisition unit 401 acquires the main valve actual spool position 411, the main valve target spool position 430, and the cylinder exhaust temperature 461 as detection information. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the position sensor looseness detection model.

[0209] The position sensor looseness detection unit 462 detects the looseness of the position sensor 29 of the valve core 28 of the main valve 20 using the position sensor looseness detection model generated by the learning device 300. The position sensor looseness detection unit 462 inputs the time series data of the main valve actual valve core position 411, the main valve target valve core position 430 and the cylinder exhaust temperature 461 acquired by the detection information acquisition unit 401 into the position sensor looseness detection model, and acquires the position sensor looseness data output from the position sensor looseness detection model as the determination result. The position sensor looseness detection unit 462 can estimate the period until the position sensor 29 deviates by more than the specified value based on the output position sensor looseness data. In this case, the period until the position sensor 29 deviates by more than the specified value can also be estimated by comparing the position sensor looseness data of a plurality of hydraulic servo valves 100 provided corresponding to a plurality of cylinders 81 of the same engine 80.

[0210] In the case where a plurality of hydraulic servo valves 100 are provided corresponding to a plurality of cylinders 81 of the same engine 80, the position sensor looseness detection unit 462 can detect the looseness of the position sensor 29 based on the position sensor looseness data estimated in each hydraulic servo valve 100. For example, the position sensor looseness detection unit 462 can detect the looseness of the position sensor 29 by comparing the exhaust temperatures of each cylinder 81. The position sensor looseness detection unit 462 can determine that the position sensor 29 of the hydraulic servo valve 100 provided in the cylinder 81 is loose when the exhaust temperature of a cylinder 81 is lower than the average value of the exhaust temperatures of the other cylinders 81 by a predetermined value or more.

[0211] The position sensor looseness detection result output unit 463 outputs the result determined by the position sensor looseness detection unit 462. When the position sensor looseness detection unit 462 detects that the position sensor 29 is loose, the position sensor looseness detection result output unit 463 may notify the fact. The position sensor looseness detection result output unit 463 may output the period estimated by the position sensor looseness detection unit 462 until the position sensor 29 deviates by more than a specified value.

[0212] A feature quantity correlated with the looseness of the position sensor 29 can be calculated based on the actual valve core position of the main valve, the target valve core position of the main valve, or the cylinder exhaust temperature. In this case, the learning device 300 can also include a feature quantity calculation unit 315, and the state estimation device 400 can also include a feature quantity calculation unit 414, as in Example 1-2. The greater the degree of looseness of the position sensor 29 of the valve core 28 of the main valve 20, the greater the deviation between the actual valve core position of the main valve and the actual position of the valve core 28, and therefore it is considered that the exhaust temperature or pressure of the cylinder has decreased relative to the normal value. Therefore, the feature quantity calculation unit can calculate the decrease in the exhaust temperature or pressure of the cylinder relative to the normal value, the rate of change, the rate of change of the rate of change, etc. as the feature quantity.

[0213] The position sensor looseness detection model generation unit 363 generates a position sensor looseness detection model using the learning data acquired by the learning data acquisition unit 301 and the feature quantity calculated by the feature quantity calculation unit 315. For example, the position sensor looseness detection model may be a formula that calculates the position sensor looseness data using the feature quantity as an input variable. In this case, the position sensor looseness detection model generation unit 363 may generate the formula by a statistical method such as regression analysis. The position sensor looseness detection model generation unit 363 may generate a position sensor looseness detection model using the position sensor looseness data 362 and the feature quantity calculated by the feature quantity calculation unit 315. The position sensor looseness detection model generation unit 363 may generate a position sensor looseness detection model using the actual valve core position 311 of the main valve, the target valve core position 330 of the main valve, and the feature quantity. For example, the position sensor looseness detection model may be a neural network that inputs the time series data of the actual valve core position 311 of the main valve and the target valve core position 330 of the main valve during a prescribed period and the feature quantity of the prescribed period into an input layer and outputs the position sensor looseness data from an output layer. In this case, the position sensor looseness detection model generation unit 363 learns the position sensor looseness detection model by adjusting the intermediate layer of the neural network in such a way that a value approximate to the position sensor looseness data 362 is output from the output layer when the time series data of the main valve actual valve core position 311 and the main valve target valve core position 330 during a specified period and the feature value of the specified period are input into the input layer.

[0214] The feature quantity calculation unit 414 of the state estimation device 400 calculates the feature quantity based on the main valve actual spool position 411, the main valve target spool position 430, or the cylinder exhaust temperature 461 acquired by the detection information acquisition unit 401 by the same method as the method in which the feature quantity calculation unit 315 calculates the feature quantity when generating the position sensor looseness detection model used by the position sensor looseness detection unit 462. In the case where the main valve actual spool position 411, the main valve target spool position 430, or the cylinder exhaust temperature 461 of the hydraulic servo valve 100 used in an environment different from the learning data used when generating the position sensor looseness detection model is acquired by the detection information acquisition unit 401, the feature quantity calculation unit 414 can calculate the feature quantity after adjusting the difference in the environment. In this way, the robustness of the position sensor looseness detection can be improved.

[0215] As the characteristic amount, other physical quantities that indicate the characteristics of the operation of the valve body 28 of the main valve 20 may be calculated. For example, the overshoot amount and the delay amount of the valve body 28 of the main valve 20 may be calculated as the characteristic amount.

[0216] As learning data and detection information, the pressure of the cylinder 81 or the maximum value of the pressure may be used instead of the cylinder exhaust temperature, or the pressure of the cylinder 81 or the maximum value of the pressure may be used in addition to the cylinder exhaust temperature. In addition, as learning data and detection information, various data indicating the state of the control object controlled by the hydraulic servo valve 100, whether the operation is good, etc. may be used.

[0217] As learning data and detection information, data recorded when the hydraulic servo valve 100 is used in a specific environment can be used. In this case, the learning device 300 can also include a data selection unit 320, and the state estimation device 400 can also include a data selection unit 417, as in Examples 1-5. The data selection unit 320 and the data selection unit 417 can select data recorded when the hydraulic servo valve 100 is actuated in a test mode in which the valve core 12 of the pilot valve 10 is actuated in a specific pattern. In this way, the robustness of the position sensor loose detection can be improved. The data selection unit 320 and the data selection unit 417 can select data recorded when the engine 80 is stopped and the valve core adhesion prevention action is being performed. In this way, the influence of interference can be suppressed and the detection accuracy of the position sensor looseness can be improved.

[0218] [Example 7: Detection of Neutral Position Deviation of Pilot Valve]

[0219] In the seventh embodiment, a technique for detecting that the neutral position of the valve core 12 of the pilot valve 10 has shifted is described. When the neutral position is indicated as the pilot valve target valve core position, when the pilot valve 10 operates normally, the valve core 12 moves to the neutral position, the supply and recovery of the working oil to the main valve 20 are stopped, and the valve core 28 of the main valve 20 is stationary at the main valve target valve core position. However, if the neutral position of the valve core 12 shifts due to some reasons, the supply and recovery of the working oil to the main valve 20 can no longer be normally controlled, sometimes causing malfunction of the controlled object such as the engine 80. If the neutral position shift of the valve core 12 of the pilot valve 10 can be detected with high accuracy, appropriate measures such as maintenance of the control system of the pilot valve 10 or replacement of the pilot valve 10 can be taken quickly.

[0220] There are two main reasons for the deviation of the neutral position of the valve core 12 of the pilot valve 10. One is an electrical reason that the valve core 12 cannot be accurately moved to the neutral position due to an abnormality in the position sensor 19 for detecting the position of the valve core 12 of the pilot valve 10, an abnormality in the electrical system for driving the valve core 12, etc. The other is a mechanical reason that the valve body 14 cannot block the A port 16a from communicating with both the P port 16p and the T port 16t even if the valve core 12 is in the neutral position due to uneven wear of the valve core 12 or the main body 10b. In either case, when the neutral position is designated as the pilot valve target valve core position, the valve core 28 of the main valve 20 moves and does not stop, so the feedback signal of the actual valve core position of the main valve deviates from the main valve target valve core position, and the pilot valve 10 is controlled to correct the deviation. As described above, the action patterns of the valve body 12 of the pilot valve 10 and the valve body 28 of the main valve 20 are greatly different between the normal state and the state when the valve body 12 of the pilot valve 10 is deviated from the neutral position. Therefore, by modeling the difference between the action pattern in the normal state and the action pattern when the valve body 12 is deviated from the neutral position, it is possible to detect the neutral position deviation of the valve body 12. The action pattern can also differ depending on the degree or cause of the deviation from the neutral position of the valve body 12, so it is possible to detect not only the presence or absence of the neutral position deviation of the valve body 12 but also the cause of the neutral position deviation, the magnitude of the deviation, etc.

[0221] In the case of uneven wear of the valve core 12, even if the valve core 12 is in the correct neutral position, the A port 16a cannot be blocked by the valve body 14, so the working oil leaks from the A port 16a. By actually measuring the internal oil leakage, it can be determined whether the neutral position deviation of the valve core 12 is caused by uneven wear or electrical reasons. The internal oil leakage can be measured by providing a flow meter for the pilot valve 10, or by an external flow meter when the pilot valve 10 is maintained, or it can be estimated by the technology of Example 1.

[0222] When the neutral position deviation of the valve core 12 of the pilot valve 10 occurs, in the engine 80 as the control object, the pressure of the cylinder 81, the exhaust temperature, etc. are abnormal and an alarm is issued. When the alarm is issued, by confirming the action pattern of the valve core 12 of the pilot valve 10 and the valve core 28 of the main valve 20, it can be determined whether the cause is an abnormality of the position sensor 29 described in Example 6 or the cause is the neutral position deviation of the valve core 12 of the pilot valve 10. In addition, when the cause is the neutral position deviation of the valve core 12 of the pilot valve 10, by confirming the internal oil leakage, it can be determined whether the neutral position deviation is caused by electrical reasons or mechanical reasons. As a result, the cause of the malfunction of the engine 80 can be accurately determined, and appropriate maintenance can be performed according to the cause, so that the cost or man-hours of maintenance required to improve the malfunction of the engine 80 can be reduced. Furthermore, when PID control is performed in the feedback control of the hydraulic servo valve 100, a certain degree of offset is corrected by the integral component (I component), but according to the technology of this embodiment, even when the neutral position offset is corrected by PID control, the neutral position offset of the valve body 12 of the pilot valve 10 can be accurately detected. In addition, when P control or PD control is performed, since the offset is not corrected, the technology of this embodiment is particularly effective.

[0223] Fig.21 The structure of the learning device and the state estimation device involved in Example 7 is shown. In the learning device 300 of this embodiment, the learning data acquisition unit 301 acquires the main valve actual valve core position 311, the main valve target valve core position 330, the pilot valve actual valve core position 310, the cylinder exhaust temperature 361 and the neutral position deviation data 371 as learning data.

[0224] The main valve actual spool position 311, the main valve target spool position 330, the pilot valve actual spool position 310, and the cylinder exhaust temperature 361 are time series data when the hydraulic servo valve 100 is used within a predetermined period before or after the time point when the neutral position deviation data 371 is actually measured. These time series data may be time series data of one to several cycles of the rotation action of the engine 80.

[0225] The neutral position deviation data 371 is data indicating the presence, degree, cause, etc. of the neutral position deviation of the valve body 12 of the pilot valve 10. When log data is used as learning data, the neutral position deviation data 371 is an actual measured value when the neutral position deviation of the valve body 12 actually occurs. When test data is used as learning data, the neutral position deviation data 371 is an actual measured value when the neutral position deviation of the valve body 12 occurs. When simulation data is used as learning data, the neutral position deviation data 371 is data input to the simulator as a simulation condition.

[0226] The neutral position shift detection model generation unit 373 generates a neutral position shift detection model used in the state estimation device 400 to detect the neutral position shift of the valve body 12 of the pilot valve 10, using the learning data acquired by the learning data acquisition unit 301. The neutral position shift detection model may be a neural network or the like that inputs the time series data of the main valve actual valve body position 311, the main valve target valve body position 330, the pilot valve actual valve body position 310, and the cylinder exhaust temperature 361 during a predetermined period into an input layer and outputs the neutral position shift data from an output layer. In this case, the neutral position shift detection model generation unit 373 learns the neutral position shift detection model by adjusting the intermediate layer of the neural network in such a manner that a value close to the neutral position shift data 371 is output from the output layer when the time series data of the main valve actual valve body position 311, the main valve target valve body position 330, the pilot valve actual valve body position 310, and the cylinder exhaust temperature 361 during a predetermined period is input into the input layer.

[0227] The neutral position shift detection model providing section 374 provides the neutral position shift detection model generated by the neutral position shift detection model generating section 373 to the state estimating device 400 .

[0228] The detection information acquisition unit 401 acquires the main valve actual spool position 411, the main valve target spool position 430, the pilot valve actual spool position 410 and the cylinder exhaust temperature 461 as detection information. The detection information is time series data of a predetermined period, which is the same as the learning data used to generate the neutral position shift detection model.

[0229] The neutral position shift detection unit 472 detects the neutral position shift of the valve core 12 of the pilot valve 10 using the neutral position shift detection model generated by the learning device 300. The neutral position shift detection unit 472 inputs the time series data of the main valve actual valve core position 411, the main valve target valve core position 430, the pilot valve actual valve core position 410, and the cylinder exhaust temperature 461 acquired by the detection information acquisition unit 401 into the neutral position shift detection model, and acquires the neutral position shift data output from the neutral position shift detection model as a determination result. The neutral position shift detection unit 472 can estimate the period until the neutral position of the valve core 12 of the pilot valve 10 deviates by a predetermined value or more relative to the actual neutral position based on the output neutral position shift data. In this case, by comparing the neutral position deviation data of the plurality of hydraulic servo valves 100 provided corresponding to the plurality of cylinders 81 of the same engine 80, the period until the neutral position of the valve body 12 of the pilot valve 10 deviates by more than a specified value from the actual neutral position can be estimated.

[0230] When a plurality of hydraulic servo valves 100 are provided corresponding to a plurality of cylinders 81 of the same engine 80, the neutral position deviation detector 472 can detect the neutral position deviation of the valve body 12 of the pilot valve 10 based on the neutral position deviation data estimated in each hydraulic servo valve 100. For example, the neutral position deviation detector 472 can detect the neutral position deviation of the valve body 12 of the pilot valve 10 by comparing the exhaust gas temperatures of each cylinder 81. When the exhaust gas temperature of a certain cylinder 81 is lower than the average value of the exhaust gas temperatures of other cylinders 81 by a predetermined value or more, the neutral position deviation detector 472 can determine that the neutral position of the valve body 12 of the pilot valve 10 of the hydraulic servo valve 100 provided in the cylinder 81 has been deviated.

[0231] The neutral position deviation detection result output unit 473 outputs the result determined by the neutral position deviation detection unit 472. When the neutral position deviation detection unit 472 detects that the neutral position of the valve body 12 of the pilot valve 10 has deviated, the neutral position deviation detection result output unit 473 can notify the result. The neutral position deviation detection result output unit 473 can output the period until the neutral position of the valve body 12 of the pilot valve 10 deviates by more than a predetermined value, which is estimated by the neutral position deviation detection unit 472.

[0232] A feature quantity correlated with the neutral position deviation of the valve core 12 of the pilot valve 10 can be calculated based on the actual valve core position of the main valve, the target valve core position of the main valve, the actual valve core position of the pilot valve, or the cylinder exhaust temperature. In this case, the learning device 300 can further include a feature quantity calculation unit 315, and the state estimation device 400 can further include a feature quantity calculation unit 414, as in Example 1-2. The greater the degree of the neutral position deviation of the valve core 12 of the pilot valve 10, the greater the degree of deviation of the actual valve core position of the main valve relative to the target valve core position of the main valve, and therefore it is considered that the change amount of the exhaust temperature or pressure of the cylinder relative to the normal value is greater. Therefore, the feature quantity calculation unit can calculate the change amount, change rate, change rate of the change rate, etc. of the exhaust temperature or pressure of the cylinder relative to the normal value as a feature quantity.

[0233] The neutral position shift detection model generation unit 373 generates a neutral position shift detection model using the learning data acquired by the learning data acquisition unit 301 and the feature quantity calculated by the feature quantity calculation unit 315. For example, the neutral position shift detection model may be a formula that calculates the neutral position shift data using the feature quantity as an input variable. In this case, the neutral position shift detection model generation unit 373 may generate the formula by a statistical method such as regression analysis. The neutral position shift detection model generation unit 373 may generate a neutral position shift detection model using the neutral position shift data 371 and the feature quantity calculated by the feature quantity calculation unit 315. The neutral position shift detection model generation unit 373 may generate a neutral position shift detection model using the actual spool position 311 of the main valve, the target spool position 330 of the main valve, the actual spool position 310 of the pilot valve, and the feature quantity. For example, the neutral position shift detection model may be a neural network or the like that inputs the time series data of the main valve actual valve core position 311, the main valve target valve core position 330, and the pilot valve actual valve core position 310 during a prescribed period, and the feature quantity during the prescribed period into an input layer, and outputs the neutral position shift data from an output layer. In this case, the neutral position shift detection model generation unit 373 learns the neutral position shift detection model by adjusting the intermediate layer of the neural network in such a manner that a value approximate to the neutral position shift data 371 is output from the output layer when the time series data of the main valve actual valve core position 311, the main valve target valve core position 330, and the pilot valve actual valve core position 310 during a prescribed period, and the feature quantity during the prescribed period are input into the input layer.

[0234] The feature quantity calculation unit 414 of the state estimation device 400 calculates the feature quantity based on the main valve actual spool position 411, the main valve target spool position 430, the pilot valve actual spool position 410, or the cylinder exhaust temperature 461 acquired by the detection information acquisition unit 401 by the same method as the method in which the feature quantity calculation unit 315 calculates the feature quantity when generating the neutral position shift detection model used by the neutral position shift detection unit 472. When the main valve actual spool position 411, the main valve target spool position 430, the pilot valve actual spool position 410, or the cylinder exhaust temperature 461 of the hydraulic servo valve 100 used in an environment different from the learning data used when generating the neutral position shift detection model is acquired by the detection information acquisition unit 401, the feature quantity calculation unit 414 can calculate the feature quantity after adjusting the difference in the environment. Thereby, the robustness of the neutral position shift detection can be improved.

[0235] As the characteristic quantity, other physical quantities that represent the characteristics of the action of the valve core 12 of the pilot valve 10 or the valve core 28 of the main valve 20 can be calculated. For example, the acceleration, speed, overshoot, delay, etc. of the valve core 12 of the pilot valve 10 or the valve core 28 of the main valve 20 can be calculated as the characteristic quantity.

[0236] As learning data and detection information, the pressure of the cylinder 81 or the maximum value of the pressure may be used instead of the cylinder exhaust temperature, or the pressure of the cylinder 81 or the maximum value of the pressure may be used in addition to the cylinder exhaust temperature. In addition, as learning data and detection information, various data indicating the state of the control object controlled by the hydraulic servo valve 100, whether the operation is good, etc. may be used.

[0237] As learning data and detection information, the pilot valve target valve position may be used instead of the pilot valve actual valve core position, or the pilot valve target valve core position may be used in addition to the pilot valve actual valve core position. In this case, the learning device 300 may be provided with an offset time calculation unit 319 to adjust the offset time required until the actual valve core position of the pilot valve 10 follows the target valve core position of the pilot valve, as in Example 1-4. The method for calculating the offset time may be the same as in Example 1-2. As learning data and detection information, a combination of any two or more positions of the pilot valve target valve core position, the pilot valve actual valve core position, the main valve target valve core position, and the main valve actual valve core position may be used.

[0238] As learning data and detection information, data recorded when the hydraulic servo valve 100 is used in a specific environment can be used. In this case, the learning device 300 can also include a data selection unit 320, and the state estimation device 400 can also include a data selection unit 417, as in Examples 1-5. The data selection unit 320 and the data selection unit 417 can select data recorded when the hydraulic servo valve 100 is operated in a test mode in which the valve core 12 of the pilot valve 10 is operated in a specific pattern. In this way, the robustness of the detection of the neutral position deviation can be improved. The data selection unit 320 and the data selection unit 417 can select data recorded when the engine 80 is stopped and the valve core adhesion prevention action is being performed. In this way, the influence of interference can be suppressed and the detection accuracy of the neutral position deviation can be improved.

[0239] The features described in the above embodiments 1 to 7 can be applied in any combination.

[0240] The above describes in detail examples of the embodiments of the present invention. The above embodiments are only specific examples of how to implement the present invention. The contents of the embodiments do not limit the scope of protection of the present invention, and many design changes such as changes, additions, and deletions of constituent elements can be made without departing from the scope of the idea of ​​the invention specified in the claims. In the above embodiments, the contents that can be subjected to such design changes are explained by marking expressions such as "embodiment mode" and "in the embodiment mode", but it does not mean that the contents without such expressions are not allowed to be subjected to design changes.

[0241] [Modifications]

[0242] Next, a modification example is described. In the drawings and description of the modification example, the same reference numerals are given to the same or equivalent components and members as those in the embodiment. The description repeated with the embodiment is appropriately omitted, and the structure different from the embodiment is described emphatically.

[0243] In the description of the embodiment, an example is shown in which the pilot valve 10 of the hydraulic servo valve 100 is a three-port valve, but the present invention is not limited to this. The hydraulic servo valve 100 may be a valve of another type.

[0244] In the description of the embodiment, an example is shown in which the pilot valve 10 of the hydraulic servo valve 100 controls the main valve 20, but the present invention is not limited to this. The pilot valve 10 may control any other actuator.

[0245] In the description of the embodiment, an example is shown in which the hydraulic servo valve 100 is used to control the engine 80 of the ship 2, but the present invention is not limited to this. The hydraulic servo valve 100 can be used to control any control object.

[0246] The above-mentioned modification examples achieve the same functions and effects as those of the embodiment.

[0247] Any combination of the above-described embodiment and modification examples is also useful as an embodiment of the present invention. A new embodiment generated by the combination has the effects of the combined embodiment and modification examples.

[0248] [Modes of the Invention]

[0249] A certain embodiment of the present invention is a state estimation device. The state estimation device includes an acquisition unit and an estimation unit, the acquisition unit acquires detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes its position according to the flow rate of the working fluid; and a value of an input voltage or output voltage of a power supply circuit that supplies power to the control valve, and the estimation unit estimates an abnormality of the power supply circuit based on the detection information acquired by the acquisition unit. According to this embodiment, an abnormality of the power supply circuit in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0250] The estimation unit may estimate the abnormality of the power supply circuit using an estimation reference for estimating the abnormality of the power supply circuit based on the detection information of the data set. According to this aspect, the estimation accuracy of the abnormality of the power supply circuit can be improved.

[0251] The estimation reference may be generated by the relationship between the value of a data set when the control valve or a control valve of the same type as the control valve is used within a specified period and the index value indicating a precursor of an abnormality in the power supply circuit calculated within the specified period. According to this method, the estimation accuracy of the abnormality in the power supply circuit can be improved.

[0252] The index value may be calculated based on the accumulated operation time of the power supply circuit, the physical quantity of the elements constituting the power supply circuit, or the value of the emission noise released from the power supply circuit. According to this aspect, the estimation accuracy of the abnormality of the power supply circuit can be improved.

[0253] The estimation reference is generated based on the relationship between the feature quantity calculated based on the value of the data set and the index value, the estimation unit calculates the feature quantity based on the detection information acquired by the acquisition unit, and estimates the abnormality of the power supply circuit based on the calculated feature quantity. According to this method, the estimation accuracy of the abnormality of the power supply circuit can be improved.

[0254] The characteristic quantity may be a drop in the input voltage or output voltage of the power supply circuit corresponding to the peak value of the current supplied to the control valve. According to this method, the abnormality of the power supply circuit can be estimated using the characteristic quantity having a high correlation with the abnormality of the power supply circuit, thereby improving the estimation accuracy of the abnormality of the power supply circuit.

[0255] The acquisition unit may acquire detection information of a data set in a usage environment that is the same as the usage environment of a control valve or a control valve of the same type as the control valve when the value of the data set used to generate the estimation reference is recorded. According to this method, the robustness of the estimation of the abnormality of the power supply circuit can be improved.

[0256] The use environment may be a situation where the control valve or a control object of the same type as the control valve is stopped and the first movable part and the second movable part are performing a sticking prevention operation. According to this aspect, the robustness of the estimation of abnormality of the power supply circuit can be improved.

[0257] The state estimation device may further include a notification unit, which notifies the estimation unit of the abnormality in the power circuit when the estimation unit estimates that the abnormality exists in the power circuit. According to this embodiment, the abnormality of the power circuit of the control valve can be accurately grasped, and appropriate measures can be taken according to the state of the control valve.

[0258] The notification unit may estimate the time until an abnormality occurs in the power supply circuit and make the notification. According to this aspect, an abnormality in the power supply circuit of the control valve can be accurately predicted, and appropriate measures can be taken according to the state of the control valve.

[0259] The control valve may be a control valve for adjusting the amount of fuel supplied to the engine. According to this aspect, an abnormality of the control valve for controlling the engine can be accurately grasped, and appropriate measures can be taken according to the state of the control valve.

[0260] Another embodiment of the present invention is a control valve. The control valve includes a first movable part, an acquisition part, and an estimation part. The position of the first movable part changes according to a control signal for specifying a position. The first movable part controls the flow rate of a working fluid according to the position. The acquisition part acquires detection information of a data set, and the data set includes: a target position or actual position of the first movable part, or a target position or actual position of a second movable part whose position changes according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve. The estimation part estimates an abnormality of the power supply circuit based on the detection information acquired by the acquisition part. According to this embodiment, an abnormality of the power supply circuit in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0261] Another aspect of the present invention is a state estimation program. The state estimation program causes a computer to function as an acquisition unit and an estimation unit, the acquisition unit acquires detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes its position according to the flow rate of the working fluid, and a value of an input voltage or output voltage of a power supply circuit that supplies power to the control valve, and the estimation unit estimates an abnormality of the power supply circuit based on the detection information acquired by the acquisition unit. According to this aspect, an abnormality of the power supply circuit in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0262] Another embodiment of the present invention is a state estimation method. The state estimation method causes a computer to perform the following steps: an acquisition step to acquire detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes its position according to the flow rate of the working fluid; and a value of an input voltage or output voltage of a power supply circuit that supplies power to the control valve; and based on the detection information acquired in the acquisition step, an abnormality of the power supply circuit is estimated. According to this embodiment, an abnormality of the power supply circuit in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0263] A certain embodiment of the present invention is a state estimation device. The state estimation device comprises: an acquisition unit that acquires detection information of a data set, the data set including a target position and an actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, and a combination of any two or more positions of the target position and the actual position of a second movable part that changes its position according to the flow rate of the working fluid; and an estimation unit that estimates an abnormality of the first movable part based on the detection information acquired by the acquisition unit. According to this embodiment, an abnormality of the first movable part in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0264] The estimation unit may estimate the abnormality of the first movable unit using an estimation reference for estimating the abnormality of the first movable unit based on the detection information of the data set. According to this aspect, the estimation accuracy of the abnormality of the first movable unit can be improved.

[0265] The estimation reference may be generated based on the relationship between the value of a data set when a control valve or a control valve of the same type as the control valve is used after or within a specified period of time when the foreign object is stuck in the first movable part, and information related to the foreign object stuck in the first movable part. According to this method, the estimation accuracy of the abnormality of the first movable part can be improved.

[0266] The information on the foreign matter may be the presence, type, size, or amount of the foreign matter. According to this aspect, the accuracy of estimating the abnormality of the first movable unit can be improved.

[0267] The acquisition unit may acquire detection information of a data set in a usage environment that is the same as the usage environment of a control valve or a control valve of the same type as the control valve when the value of the data set used to generate the estimation reference is recorded. According to this method, the robustness of the estimation of the abnormality of the first movable part can be improved.

[0268] The use environment may be a situation where the control valve or a control object of the same type as the control valve is stopped and the first movable part and the second movable part are performing a sticking prevention operation. According to this aspect, the robustness of the estimation of the abnormality of the first movable part can be improved.

[0269] The state estimation device may further include a notification unit, which notifies the estimation unit of the abnormality of the first movable part when the estimation unit estimates that the first movable part has an abnormality. According to this embodiment, the abnormality of the first movable part of the control valve can be accurately grasped, and appropriate measures can be taken according to the state of the control valve.

[0270] The notification unit may estimate the time until the abnormality occurs in the first movable part and make the notification. According to this aspect, the abnormality of the first movable part of the control valve can be accurately predicted, and appropriate measures can be taken according to the state of the control valve.

[0271] The control valve may be a control valve for adjusting the amount of fuel supplied to the engine. According to this aspect, an abnormality of the control valve for controlling the engine can be accurately grasped, and appropriate measures can be taken according to the state of the control valve.

[0272] Another aspect of the present invention is a control valve. The control valve includes: a first movable part, whose position changes according to a control signal for specifying a position, and controls the flow rate of a working fluid according to the position; an acquisition part, which acquires detection information of a data set, the data set including a target position and an actual position of the first movable part and a combination of any two or more positions of the target position and the actual position of a second movable part whose position changes according to the flow rate of the working fluid; and an estimation part, which estimates an abnormality of the first movable part based on the detection information acquired by the acquisition part. According to this aspect, an abnormality of the first movable part in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0273] Another aspect of the present invention is a state estimation program. The state estimation program is a program that causes a computer to function as an acquisition unit and an estimation unit, the acquisition unit acquires detection information of a data set, the data set including a target position and an actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, and a combination of any two or more positions of the target position and the actual position of a second movable part that changes its position according to the flow rate of the working fluid, and the estimation unit estimates an abnormality of the first movable part based on the detection information acquired by the acquisition unit. According to this aspect, an abnormality of the first movable part in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

[0274] Another aspect of the present invention is a state estimation method. The state estimation method causes a computer to perform the following steps: an acquisition step of acquiring detection information of a data set, the data set including a target position and an actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, and a combination of any two or more positions of a target position and an actual position of a second movable part that changes its position according to the flow rate of the working fluid; and estimating an abnormality of the first movable part based on the detection information acquired in the acquisition step. According to this aspect, an abnormality of the first movable part in the control valve can be estimated, so that appropriate measures can be taken according to the state of the control valve.

Claims

1. A state estimation device comprising an acquisition unit and an estimation unit, The acquisition unit acquires detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes position according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve, The estimation unit estimates abnormality of the power supply circuit based on the detection information acquired by the acquisition unit.

2. The state estimation device according to claim 1, wherein: The estimation section estimates abnormality of the power supply circuit using an estimation reference for estimating abnormality of the power supply circuit based on the detection information of the data set.

3. The state estimation device according to claim 2, wherein: The estimation reference is generated based on a relationship between a value of the data set when the control valve or a control valve of the same type as the control valve is used within a predetermined period and an index value indicating a precursor of abnormality of the power supply circuit calculated within the predetermined period.

4. The state estimation device according to claim 3, wherein: The index value is calculated based on the accumulated operation time of the power supply circuit, the physical quantity of an element constituting the power supply circuit, or the value of emission noise released from the power supply circuit.

5. The state estimation device according to claim 3 or 4, wherein: The estimation reference is generated based on the relationship between the feature quantity calculated based on the value of the data set and the index value, The estimation unit calculates the feature amount from the detection information acquired by the acquisition unit, and estimates abnormality of the power supply circuit based on the calculated feature amount.

6. The state estimation device according to claim 5, wherein: The characteristic amount is an amount of drop in input voltage or output voltage of the power supply circuit corresponding to a peak value of the current supplied to the control valve.

7. The state estimation device according to any one of claims 2 to 4, wherein: The acquisition unit acquires the detection information of the data set in a usage environment that is the same as the usage environment of the control valve or a control valve of the same type as the control valve when the value of the data set used to generate the estimation reference is recorded.

8. The state estimation device according to claim 7, wherein: The use environment is a situation in which a control target of the control valve or a control valve of the same type as the control valve is stopped and a sticking prevention operation of the first movable part and the second movable part is being performed.

9. The state estimation device according to any one of claims 1 to 4, wherein: The device further includes a notification unit configured to notify that the power supply circuit has an abnormality when the estimation unit estimates that the power supply circuit has an abnormality.

10. The state estimation device according to claim 9, wherein: The notification unit estimates a time until an abnormality occurs in the power supply circuit and makes a notification.

11. The state estimation device according to any one of claims 1 to 4, wherein: The control valve is a control valve for adjusting the amount of fuel supplied to the engine.

12. A control valve comprising a first movable portion, an acquisition portion, and an estimation portion, The position of the first movable part changes according to a control signal for specifying the position, and the first movable part controls the flow rate of the working fluid according to the position. The acquisition unit acquires detection information of a data set, the data set including: a target position or an actual position of the first movable part, or a target position or an actual position of a second movable part whose position changes according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve, The estimation unit estimates abnormality of the power supply circuit based on the detection information acquired by the acquisition unit.

13. A state estimation program product, comprising a state estimation program for causing a computer to function as an acquisition unit and an estimation unit, The acquisition unit acquires detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls the flow rate of a working fluid according to the position of the first movable part, or a target position or actual position of a second movable part that changes position according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve, The estimation unit estimates abnormality of the power supply circuit based on the detection information acquired by the acquisition unit.

14. A state estimation method, causing a computer to perform the following steps: an acquisition step of acquiring detection information of a data set, the data set including: a target position or actual position of a first movable part in a control valve that controls a flow rate of a working fluid according to a position of the first movable part, or a target position or actual position of a second movable part that changes position according to the flow rate of the working fluid; and a value of an input voltage or an output voltage of a power supply circuit that supplies power to the control valve; and Based on the detection information acquired in the acquisition step, an abnormality of the power supply circuit is estimated.

15. A state estimation device comprising: an acquisition unit that acquires detection information of a data set including a combination of any two or more positions of a target position and an actual position of a first movable part in a control valve that controls a flow rate of a working fluid according to a position of the first movable part and a target position and an actual position of a second movable part that changes position according to the flow rate of the working fluid; and an estimating unit that estimates abnormality of the first movable unit based on the detection information acquired by the acquiring unit, The estimation unit estimates the abnormality of the first movable unit using an estimation reference for estimating the abnormality of the first movable unit based on the detection information of the data set, The estimation benchmark is generated based on the relationship between the value of the data set when the control valve or a control valve of the same type as the control valve is used within a specified period before or after the time when the foreign object is stuck in the first movable part, and the information related to the foreign object stuck in the first movable part.

16. The state estimation device according to claim 15, wherein: The information related to the foreign matter is the presence, type, size or amount of the foreign matter.

17. The state estimation device according to claim 15, wherein: The acquisition unit acquires the detection information of the data set in a usage environment that is the same as the usage environment of the control valve or a control valve of the same type as the control valve when the value of the data set used to generate the estimation reference is recorded.

18. A state estimation device comprising: an acquisition unit that acquires detection information of a data set including a combination of any two or more positions of a target position and an actual position of a first movable part in a control valve that controls a flow rate of a working fluid according to a position of the first movable part and a target position and an actual position of a second movable part that changes position according to the flow rate of the working fluid; and an estimating unit that estimates abnormality of the first movable unit based on the detection information acquired by the acquiring unit, The estimation unit estimates the abnormality of the first movable unit using an estimation reference for estimating the abnormality of the first movable unit based on the detection information of the data set, The acquisition unit acquires the detection information of the data set in the same usage environment as the usage environment of the control valve or the control valve of the same type as the control valve when the value of the data set used to generate the estimation reference is recorded, The use environment is a situation in which a control target of the control valve or a control valve of the same type as the control valve is stopped and a sticking prevention operation of the first movable part and the second movable part is being performed.

19. The state estimation device according to any one of claims 15 to 18, wherein: The present invention further includes a notification unit configured to notify that the first movable unit has an abnormality when the estimation unit estimates that the first movable unit has an abnormality.

20. A state estimation device comprising: an acquisition unit that acquires detection information of a data set including a combination of any two or more positions of a target position and an actual position of a first movable part in a control valve that controls a flow rate of a working fluid according to a position of the first movable part and a target position and an actual position of a second movable part that changes position according to the flow rate of the working fluid; an estimating unit that estimates abnormality of the first movable unit based on the detection information acquired by the acquiring unit; as well as a notification unit configured to notify that the first movable unit has an abnormality when the estimation unit estimates that the first movable unit has an abnormality, The notification unit estimates a time until abnormality occurs in the first movable unit and makes a notification.

21. The state estimation device according to any one of claims 15 to 18, wherein: The control valve is a control valve for adjusting the amount of fuel supplied to the engine.

22. A control valve comprising: a first movable portion of the control valve, the position of which changes according to a control signal for specifying the position, and the flow rate of the working fluid is controlled according to the position; an acquisition unit that acquires detection information of a data set including a combination of any two or more of a target position and an actual position of the first movable part and a target position and an actual position of a second movable part whose position changes according to a flow rate of the working fluid; and an estimating unit that estimates abnormality of the first movable unit based on the detection information acquired by the acquiring unit, The estimation unit estimates the abnormality of the first movable unit using an estimation reference for estimating the abnormality of the first movable unit based on the detection information of the data set, The estimation benchmark is generated based on the relationship between the value of the data set when the control valve or a control valve of the same type as the control valve is used within a specified period before or after the time when the foreign object is stuck in the first movable part, and the information related to the foreign object stuck in the first movable part.

23. A state estimation program product, comprising a state estimation program for causing a computer to function as an acquisition unit and an estimation unit, The acquisition unit acquires detection information of a data set, the data set including a target position and an actual position of a first movable part in a control valve that controls a flow rate of a working fluid according to a position of the first movable part, and a combination of any two or more positions of a target position and an actual position of a second movable part that changes position according to the flow rate of the working fluid, The estimating unit estimates abnormality of the first movable unit based on the detection information acquired by the acquiring unit. The estimation unit estimates the abnormality of the first movable unit using an estimation reference for estimating the abnormality of the first movable unit based on the detection information of the data set, The estimation benchmark is generated based on the relationship between the value of the data set when the control valve or a control valve of the same type as the control valve is used within a specified period before or after the time when the foreign object is stuck in the first movable part, and the information related to the foreign object stuck in the first movable part.

24. A state estimation method, causing a computer to perform the following steps: an acquisition step of acquiring detection information of a data set including a combination of any two or more positions of a target position and an actual position of a first movable part in a control valve that controls a flow rate of a working fluid according to a position of the first movable part and a target position and an actual position of a second movable part that changes position according to the flow rate of the working fluid; and estimating abnormality of the first movable part based on the detection information acquired in the acquisition step, The estimating step includes the following steps: estimating the abnormality of the first movable part using an estimation reference for estimating the abnormality of the first movable part based on the detection information of the data set, The estimation benchmark is generated based on the relationship between the value of the data set when the control valve or a control valve of the same type as the control valve is used within a specified period before or after the time when the foreign object is stuck in the first movable part, and the information related to the foreign object stuck in the first movable part.

Images