Measurement device, estimation device, and measurement method

Abstract

To estimate a deterioration level of a machine auxiliary agent to be supplied to machines by a simple structure or a simple method.SOLUTION: A measurement device 10 comprises: a spectrometer 13; and a processing device 14. The spectrometer 13 spectroscopically measures a machine auxiliary agent, using a visible light ray, near infrared-ray, middle infrared-ray or ultraviolet ray to acquire spectral data. The processing device 14 outputs a deterioration level of the machine auxiliary agent on the basis of the spectral data. A model is configured by machine-learning using additive consumption rate as second learning data, or a deterioration level converting the additive consumption rate by use of the spectral data as first learning data. The processing device 14 is configured to output the deterioration level of the machine auxiliary agent estimated based on the additive consumption rate the model outputs, or the deterioration level of the machine auxiliary agent the model outputs.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present application primarily relates to an apparatus or method for providing degradation levels of machine aids. [Background technology]

[0002] Patent Document 1 discloses a deterioration diagnosis device for a machine. The machine is equipped with a power transmission mechanism that transmits the power of a motor via gears. Grease is supplied to the gears, and additives are added to the grease. As the grease is used in the power transmission mechanism, the additives in the grease are consumed. The deterioration diagnosis device stores in advance the change trend of the additive consumption rate that accompanies the operation of the machine. Based on the stored change trend, the deterioration diagnosis device predicts the remaining life, which is the period until the additive consumption rate reaches a preset value. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2019 / 103091 Summary of the Invention [Problem to be solved by the invention]

[0004] Patent Document 1 describes a method of measuring the amount of additives that should have been dissolved and precipitated as sludge using X-ray fluorescence analysis, and calculating the additive consumption rate based on the measurement results. However, when using X-ray fluorescence analysis, the device tends to become large due to the need for a shielding structure to prevent X-rays from being irradiated to the outside. Furthermore, since a process for extracting sludge is required, the additive consumption rate cannot be calculated easily. Therefore, with the method of Patent Document 1, it is difficult to calculate the additive consumption rate with a simple structure or simple method. In addition, this type of problem is not limited to the estimation of the remaining life of grease. For example, a similar problem exists in deterioration diagnosis using various machine auxiliary agents such as lubricating oil and fuel oil.

[0005] The present application has been made in consideration of the above circumstances, and a main object of the application is to estimate the deterioration level of a machine auxiliary agent supplied to a machine using a simple structure or a simple method. [Means for solving the problem]

[0006] The problem to be solved by the present application is as described above. Next, the means for solving this problem and the effects thereof will be described.

[0007] According to a first aspect of the present application, there is provided a measuring device having the following configuration. That is, the measuring device outputs a deterioration level of a machine auxiliary. The machine auxiliary is an object that contains an additive, is liquid or semi-solid, and is supplied to a machine to maintain the operation of the machine. The measuring device includes a spectrometer and a processing device. The spectrometer performs spectroscopic measurement of the machine auxiliary using visible light, near infrared light, mid-infrared light, or ultraviolet light to obtain spectral data. The processing device outputs the deterioration level of the machine auxiliary based on the spectral data. The processing device inputs the spectral data obtained by the spectrometer as input data into a model. When creating a model, the model is constructed by machine learning using the spectral data as first learning data and the additive consumption rate or the deterioration level converted from the additive consumption rate as second learning data. When the additive consumption rate is used as the second learning data, when the model is used, the processing device outputs the deterioration level of the machine auxiliary estimated based on the estimated additive consumption rate, which is the output data of the model. When the deterioration level is used as the second learning data, during use of the model, the processing device outputs the output data of the model as the estimated deterioration level of the machine auxiliary.

[0008] According to a second aspect of the present application, there is provided an estimation device having the following configuration. The estimation device includes a communication device and a processing device. The communication device receives spectral data acquired by spectroscopically measuring the machine auxiliary using visible light, near infrared light, mid-infrared light, or ultraviolet light. The processing device estimates a deterioration level of the machine auxiliary based on the spectral data. The processing device inputs the spectral data received by the communication device as input data into a model. When creating a model, the model is constructed by machine learning using the spectral data as first learning data and an additive consumption rate or a deterioration level converted from the additive consumption rate as second learning data. When the additive consumption rate is used as the second learning data, when using the model, the processing device estimates a deterioration level of the machine auxiliary based on the additive consumption rate, which is output data of the model. In the case where the deterioration level is used as the second learning data, when the model is used, the processing device sets the output data of the model as the estimation result of the deterioration level of the machine auxiliary.

[0009] According to a third aspect of the present application, the following estimation method is provided. That is, in the estimation method, the mechanical auxiliary is spectroscopically measured using visible light, near infrared light, mid-infrared light, or ultraviolet light to obtain spectral data. The spectral data is input to a model as input data. A deterioration level of the mechanical auxiliary is estimated and output based on the additive consumption rate, which is output data of the model. When creating a model, the model is constructed by machine learning using the spectral data as first learning data and the additive consumption rate or a deterioration level converted from the additive consumption rate as second learning data. When the additive consumption rate is used as the second learning data, when the model is used, an estimated deterioration level of the mechanical auxiliary is output based on the estimated additive consumption rate, which is output data of the model. When the deterioration level is used as the second learning data, when the model is used, the output data of the model is output as the estimated deterioration level of the mechanical auxiliary. Effect of the Invention

[0010] According to the present application, the deterioration level of a machine auxiliary agent supplied to a machine can be estimated using a simple structure or a simple method. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a perspective view of a measurement device according to a first embodiment. [Diagram 2] Graph showing daily changes in additive consumption rate and iron powder concentration. [Diagram 3] FIG. 2 is an explanatory diagram showing learning data when creating a model, and input / output data when using the model. [Figure 4] 5 is a flowchart showing a process in which the measuring device measures grease and outputs a grease deterioration level. [Diagram 5] FIG. 13 is a perspective view of a measurement device according to a second embodiment. [Figure 6] FIG. 11 is a block diagram of a measurement device according to a third embodiment. [Figure 7] FIG. 13 is an explanatory diagram showing learning data when creating a model according to the fourth embodiment, and input / output data when the model is used. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Next, an embodiment of the present application will be described with reference to the drawings.

[0013] The measuring device 10 is a device that measures a machine auxiliary agent supplied to a machine and estimates the deterioration level of the machine auxiliary agent. In this embodiment, the machine auxiliary agent is an object that is supplied to the machine in order to maintain the operation of the machine. To maintain the operation of the machine means that the machine auxiliary agent is necessary for the machine to continue to operate. In other words, the machine auxiliary agent is not disposable, but remains in the machine or circulates within the machine. The machine auxiliary agent is, for example, grease, lubricating oil, fuel oil, coolant, and cleaning liquid. Grease and lubricating oil can also be collectively referred to as a lubricant.

[0014] Moreover, the mechanical aid of this embodiment is liquid or semi-solid and contains additives. The additives are substances added to the mechanical aid for the purpose of improving the quality of the mechanical aid, suppressing deterioration of the quality of the mechanical aid, or other purposes. As the mechanical aid is used in a machine, at least one type of additive is consumed. Consumption of an additive includes cases where the additive disappears due to volatilization or the like with use, and cases where the composition or properties of the additive change with use of the additive.

[0015] In this embodiment, a measuring device 10 is described that measures grease, which is an example of a mechanical auxiliary, and estimates the deterioration level of the grease. When the mechanical auxiliary is grease, additives such as friction modifiers, extreme pressure materials, or wear-resistant materials are added. Specifically, ZnDTP (zinc dialkyldithiophosphate) or MoDTP (molybdenum dialkyldithiophosphate) are used as additives. Note that even when a mechanical auxiliary other than grease is used, the same effect can be achieved using the same principle and device as when grease is used.

[0016] In this embodiment, multiple types of grease are supplied to the robot 30. The robot 30 is a robot that performs work in a work area such as a factory or a warehouse. The robot is a teaching playback type. The teaching playback type means that the robot is taught the operation of the robot in advance, and the robot repeats the operation according to the taught content. The robot is, for example, a vertical multi-joint or horizontal multi-joint arm robot. However, the robot may be a robot other than an arm robot, such as a parallel link robot. The work performed by the robot is, for example, assembly, welding, painting, machining, or transportation. Note that the technical content of this embodiment is not limited to a machine auxiliary agent supplied to a robot, but can also be applied to a machine auxiliary agent supplied to other machines. The other machines are, for example, machine tools, hydraulic machines, transport machines, construction machines, etc.

[0017] The robot 30 includes a plurality of arms 31, a working tool 32, and a plurality of reducers 33. The arms 31 are rotatably connected to each other. The working tool 32 is attached to the tip of the arm 31. The working tool 32 is also called a manipulator or an end effector. When the work performed by the robot 30 is assembly or transportation, the working tool 32 is a hand that holds a workpiece. The reducer 33 is disposed on each of the arms 31. The reducer 33 reduces the speed of the rotational power generated by the motor. The reduced rotational power is used to drive the arm 31. The reducer 33 has a plurality of gears. Grease is supplied to the reducer 33 to reduce wear of the gears.

[0018] The grease deteriorates as the robot 30 operates. If the robot 30 is operated with deteriorated grease, wear between the gears of the reduction gear 33 will become severe, significantly shortening the life of the reduction gear 33 or causing the reduction gear 33 to break down. In the workshop, the grease is replaced before it deteriorates to the extent that it causes problems for the reduction gear 33.

[0019] Specifically, the worker removes old grease from the reducer 33 of the robot 30, and then supplies new grease to the reducer 33. However, since some of the old grease remains on each part of the reducer 33, it is difficult to completely remove the old grease. Naturally, it is also difficult to accurately estimate the remaining amount of old grease. Therefore, after the grease is replaced, the ratio of old grease to new grease is unknown. Therefore, in order to know the deterioration level of the grease, it is necessary to remove the grease and measure it.

[0020] A measuring device 10 and a cell 20 shown in Fig. 1 are used to estimate the deterioration level of the grease. The measuring device 10 and the cell 20 are assumed to be used in a workplace. An operator fills the cell 20 with grease taken from a robot 30, and sets the cell 20 in the measuring device 10. Next, the operator operates the measuring device 10 to cause it to estimate the deterioration level of the grease. The measuring device 10 and the cell 20 will be described in detail below.

[0021] The measuring device 10 includes a housing 11. The housing 11 is a substantially box-shaped object and constitutes the exterior of the measuring device 10. The housing 11 is provided with a setting section 12. The setting section 12 is a section for setting the cell 20. The setting section 12 in this embodiment is an opening for inserting and setting the cell 20. However, the setting section 12 may have a configuration different from that in this embodiment. For example, the setting section 12 may be configured to be able to place the cell 20. In addition, although the setting section 12 in this embodiment is configured to be able to set one cell 20, it may be able to set multiple cells 20 (for example, three cells 20). In this case, the housing 11 or the setting section 12 may have a holding structure (for example, a clamp structure) that holds and fixes multiple cells.

[0022] As described above, the measuring device 10 is used in a work site. Therefore, the housing 11 has portability. Portability means that the size is such that the worker can carry it by himself. For example, the height, width, and depth of the housing 11 are preferably 1 m or less, and more preferably 0.5 m or less, 0.3 m or less, or 0.2 m or less. In order to improve the handleability of the measuring device 10, the devices constituting the measuring device 10 of this embodiment are disposed inside or on the outer surface of the housing 11. As shown in FIG. 1, a spectrometer 13, a processing device 14, and a battery 15 are disposed inside the housing 11. A display 16 and a measurement button 17 are disposed on the outer surface of the housing 11.

[0023] The spectrometer 13 measures visible light or near infrared light. The spectrometer 13 irradiates the grease of the cell 20, which is the measurement target, with reference light of visible light or near infrared light. After passing through the grease, the reference light is appropriately reflected by optical components arranged in the housing 11 and is again incident on the spectrometer 13. The spectrometer 13 separates the reference light that has passed through the grease, measures the light intensity for each wavelength, and acquires spectral data. When the reference light passes through the grease, the reference light in a predetermined wavelength range is absorbed by the grease. For example, when an additive is consumed, the amount of absorption of the reference light in a predetermined wavelength range decreases. Therefore, the light intensity of the predetermined wavelength range in the spectral data changes significantly. In this way, there is a correlation between the spectral data and the additive consumption rate. Note that the measurement target of the spectrometer 13 is not limited to visible light or near infrared light. For example, the measurement target of the spectrometer 13 may be ultraviolet light or mid-infrared light. In summary, the spectrometer 13 may be configured to perform measurements using visible light, near infrared light, mid-infrared light, or ultraviolet light, in other words, electromagnetic waves having a wavelength of 10 nm or more and 2.5 μm or less.

[0024] The processing device 14 includes an arithmetic device such as a CPU, a storage device such as an HDD, an SSD, or a flash memory, and an input / output device. The storage device stores a program and various parameters for estimating the deterioration level of the grease. The arithmetic device executes the program stored in the storage device to perform various processes for estimating the deterioration level of the grease. The arithmetic device performs, for example, a process of causing the spectrometer 13 to start measurement, a process of estimating the deterioration level of the grease based on the spectrum data acquired from the spectrometer 13, and a process of displaying text indicating the deterioration level of the grease on the display 16. The input / output device performs signal processing of data input to the processing device 14, and signal processing of data created and output by the processing device 14. Details of the processes performed by the processing device 14 will be described later.

[0025] The battery 15 stores power. The battery 15 is connected to the spectrometer 13 and the processing device 14, and supplies power to them. The housing 11 is provided with a connector for charging the battery 15. The measuring device 10 is not limited to being battery-powered, and may be configured to receive power from an external source via a wire.

[0026] The display 16 is a liquid crystal display or an organic electroluminescence display. The display 16 may be a dot matrix type or a segment type. The display 16 displays information based on the data output by the processing device 14. The information displayed by the display 16 is, for example, the progress of the measurement and the deterioration level of the grease. Note that the display 16 is not an essential component. The measuring device 10 may display the deterioration level of the grease on the display of an external terminal.

[0027] The measurement button 17 is a button that allows an operator to instruct the start of measurement of the deterioration level of the grease. The measurement button 17 is connected to the processing device 14 by an electric wire or the like. When the measurement button 17 is operated, a predetermined electric signal is supplied to the processing device 14. This allows the processing device 14 to detect the operation state of the measurement button 17. When the processing device 14 detects that the measurement button 17 has been operated, it instructs the spectrometer 13 to start measuring the grease. The functions of the display 16 and the measurement button 17 may be combined into a touch panel.

[0028] The cell 20 is made of a material that is transmissive to light of a wavelength handled by the spectrometer. The cell 20 is made of a lower case 21 and an upper case 22. The lower case 21 and the upper case 22 are engageable. A groove 21a is formed in the lower case 21. The groove 21a has a constant depth. The grease taken out from the robot 30 is filled into the groove 21a. Since the grease is semi-solid, it is difficult for the thickness to become constant due to gravity. In this regard, by filling the groove 21a, which has a constant depth, with the grease, the thickness of the grease can be made constant. This makes it possible to increase the accuracy of measurements by the spectrometer 13.

[0029] Next, the relationship between the additive consumption rate of the grease and the deterioration level of the grease will be described with reference to FIG.

[0030] FIG. 2 is a graph showing the results of the durability test shown in Patent Document 1. The horizontal axis of the graph in FIG. 2 is the number of days, and the vertical axis is the additive consumption rate and the iron powder concentration. The iron powder concentration is the concentration of iron powder contained in the grease. The iron powder concentration increases as the grease deteriorates and the wear between the gears becomes severe. As shown in the graph in FIG. 2, the additive consumption rate increases linearly with the number of days. Therefore, if the slope and intercept can be specified, the future change tendency of the additive consumption rate can be estimated. For example, by measuring the grease twice with an interval of several days, the slope and intercept of the additive consumption rate can be specified, so that the future change tendency of the additive consumption rate can be estimated. Alternatively, if the conditions such as the type of robot 30, the type of work performed by the robot 30, and the type of grease are the same, it is considered that the additive consumption rate will show the same change tendency. Therefore, by measuring and recording the additive consumption rate of the grease according to the number of days in advance, the future change tendency of the additive consumption rate can be estimated with a single measurement.

[0031] When the number of days or additive consumption rate is low, the iron powder concentration remains close to 0%. After that, when a certain number of days is exceeded, in other words, when the additive consumption rate exceeds a certain percentage, the iron powder concentration and additive consumption rate begin to increase, and then the iron powder concentration increases rapidly. The reason for the sudden increase in iron powder concentration is that the grease has reached the end of its life and is no longer functioning. In this way, there is a high correlation between the additive consumption rate and the grease life, and by identifying the additive consumption rate, the deterioration level of the grease can be accurately estimated.

[0032] In summary, the deterioration level of the grease can be estimated by identifying the additive consumption rate of the grease removed from the robot 30. Furthermore, the lifespan of the grease can be estimated by comparing the additive consumption rate of the grease with the change trend of the additive consumption rate stored in the processing device 14 (i.e., the graph of black circles in FIG. 2).

[0033] Next, a method for estimating the additive consumption rate of the grease will be described with reference to FIG.

[0034] The processing device 14 estimates the additive consumption rate using a model constructed by machine learning. FIG. 3 shows the learning data at the time of creating the model. The learning data is a set of spectral data obtained by measuring the grease and the additive consumption rate of the grease. In this embodiment, the first learning data is the spectral data, and the second learning data is the additive consumption rate. As described above, the spectral data and the additive consumption rate are correlated. The learning data includes a plurality of sets of data with different additive consumption rates. In addition, the temperature of the place where the grease is measured or the temperature of the grease itself is called the environmental temperature. When the environmental temperature is different, the amount of absorption of light in a predetermined wavelength range changes, so that the spectral data is different even if the additive consumption rate is the same. Therefore, in this embodiment, the learning data includes spectral data with different environmental temperatures. However, the environmental temperature does not have to be included in the learning data.

[0035] Multiple types of grease are used in the workshop. Therefore, the spectral data used as training data includes spectral data of multiple types of grease. Although there are slight differences in the spectral data depending on the type of grease, a model with sufficient accuracy can be constructed by enriching the training data. By training the spectral data of multiple types of grease together, a model that can be used for multiple types of grease can be constructed, thereby reducing the work of constructing the model.

[0036] The additive consumption rate for learning can be measured by the following method. That is, as described in Patent Document 1, by extracting the sludge of the additive of the grease and performing the fluorescent X-ray analysis, the amount of the additive that should have been dissolved and precipitated as sludge can be measured. The additive consumption rate can be calculated based on the measurement result. That is, the additive consumption rate for learning is an additive consumption rate measured by an instrument other than the measuring instrument 10. The instrument used in the fluorescent X-ray analysis requires a shielding structure to prevent the X-ray from being irradiated to the outside, and the instrument size tends to be large. In addition, instead of the fluorescent X-ray analysis, a spectroscopic analysis using Fourier transform infrared spectroscopy may be performed. Hereinafter, Fourier transform infrared spectroscopy is referred to as FT-IR. By performing FT-IR, the spectral peak caused by the additive can be measured. Therefore, the additive consumption rate can be calculated based on the intensity of the spectral peak caused by the additive. In addition, since mid-infrared rays, which have a longer wavelength than visible light or near-infrared rays, are used in FT-IR, the optical system tends to be complicated or large, and the instrument size tends to be large. The learning spectrum data is obtained by measuring the grease used to measure the learning additive consumption rate with the measuring device 10.

[0037] Machine learning is performed using the above learning data. Since the additive consumption rate corresponds to the correct answer, this machine learning is supervised learning. By performing machine learning, the relationship between the spectrum data and the additive consumption rate is learned. Various types of machine learning can be used, and for example, partial least squares method or principal component analysis method can be used. The machine learning used in this embodiment is not deep learning, but deep learning may be used.

[0038] When the above-mentioned FT-IR is used, the spectral peak caused by the additive can be measured. In contrast, since the measurement device 10 performs spectroscopic analysis using visible light or near-infrared light, it is difficult to accurately measure the spectral peak caused by the additive. However, as described above, the spectral data that can be measured by the measurement device 10 has a correlation with the additive consumption rate. Then, by using machine learning, the relationship between the spectral data that can be measured by the measurement device 10 and the additive consumption rate is learned. As a result, the additive consumption rate can be estimated with sufficient accuracy using the spectral data that can be measured by the measurement device 10.

[0039] In addition, the effect of the environmental temperature on the spectral data is different from the effect of the additive consumption rate on the spectral data. In other words, even if the environmental temperature is different, if the additive consumption rate is the same, the spectral data will have common characteristics. Therefore, it is not necessary to include the environmental temperature in the training data.

[0040] In this embodiment, the constructed model is stored by the processing device 14. As shown in FIG. 3, when the model is used, spectral data is input to the model. The input spectral data is spectral data measured and acquired by the spectrometer 13. Here, the model is a model that has learned the relationship between the spectral data and the additive consumption rate. Therefore, by inputting the spectral data to the model, the model can output an estimated additive consumption rate. In other words, the additive consumption rate output when the model is used is an additive consumption rate estimated using the processing device 14 of the measuring device 10. As described above, the correlation between the deterioration level of the grease and the additive consumption rate is high. Therefore, by acquiring data in which the deterioration level and the additive consumption rate correspond to each other in advance, the additive consumption rate can be converted to the deterioration level. The configuration that performs this process is called a converter. In this embodiment, the deterioration level of the grease is obtained by inputting the additive consumption rate estimated by the model to the converter. The measuring device 10 outputs the deterioration level thus obtained.

[0041] Next, a process in which the measuring device 10 measures the grease and estimates and outputs the grease deterioration level will be described with reference to Fig. 4. The flowchart in Fig. 4 is executed by the processing device 14.

[0042] After replacing the grease, the worker mixes the old grease with the new grease, for example by slightly moving the gear of the reducer 33. Next, the worker removes a portion of the grease from the reducer 33 and fills it into the cell 20. The worker sets the cell 20 filled with grease in the setting unit 12 and operates the measurement button 17. Note that the timing for measuring the grease is not limited to when the grease is replaced, and may be, for example, after a predetermined period of time has passed since the grease was replaced.

[0043] The processing device 14 determines whether or not a measurement instruction has been issued (S101). This determination can be made based on whether or not the measurement button 17 has been operated. When the processing device 14 determines that a measurement instruction has been issued, it instructs the spectrometer 13 to measure the grease and obtain spectrum data (S102).

[0044] Next, the processing device 14 inputs the acquired spectral data into the model (S103) and acquires the additive consumption rate output by the model (S104).

[0045] Next, the processing device 14 estimates the deterioration level of the grease based on the additive consumption rate (S105). The above-mentioned converter is used for this process. The deterioration level of the grease is an index that indirectly or directly indicates the degree to which the grease has deteriorated. The deterioration level may be expressed, for example, as a percentage, a stage, or a remaining life.

[0046] When degradation is expressed as a percentage, the degradation level of the grease can be determined based on the estimated additive consumption rate, because the iron concentration (and therefore the risk of failure) increases as the additive consumption rate increases. Note that the additive consumption rate varies linearly, making it easy to interpolate where data is missing.

[0047] When the deterioration level is expressed in stages, the processing device 14 divides the additive consumption rate into multiple stages and sets a deterioration level for each stage. For example, the additive consumption rate may be divided into 10 stages in 10% increments, with each stage being set as stage 1 from 0% to 10% and stage 2 from 10% to 20%. The processing device 14 then uses the stage of the stage that includes the estimated additive consumption rate as the deterioration level.

[0048] When the deterioration level is expressed as the remaining life, the processing device 14 performs the following process. That is, the processing device 14 stores the change trend of the additive consumption rate in advance. The processing device 14 calculates the remaining life, which is the number of days until the additive consumption rate reaches a predetermined value, based on the estimated additive consumption rate and the change trend of the additive consumption rate. The processing device 14 uses the calculated remaining life as the deterioration level. Note that the expression of the deterioration level is not limited to the above example, and may be, for example, the iron powder concentration. Specifically, the processing device 14 may calculate an estimated value of the iron powder concentration based on the relationship between the additive consumption rate and the iron powder concentration shown in the graph of FIG. 2, and express the calculated estimated value of the iron powder concentration as the deterioration level.

[0049] Next, the processing device 14 outputs the deterioration level of the grease estimated in step S105 to the display 16 (S106). The display 16 displays the data output by the processing device 14.

[0050] In this manner, in this embodiment, the processing device 14 estimates the additive consumption rate and estimates the deterioration level of the grease by itself. Therefore, there is no need to connect the measuring device 10 to an external device, which reduces the burden on the operator.

[0051] In this embodiment, the grease deterioration level is estimated by actually measuring the grease. Therefore, the deterioration level taking into account the influence of the remaining old grease can be notified to the worker. In addition, when the grease is measured after a predetermined period of time has passed since the grease was replaced, the deterioration level taking into account the operation actually performed by the robot 30, the environmental temperature, and the like can be notified to the worker. Therefore, compared to a method of determining the deterioration of the grease based only on the operation history of the robot 30, the deterioration level of the grease can be notified to the worker with high accuracy.

[0052] As described above, a large measuring device is required to perform measurements using X-ray fluorescence analysis or FT-IR. Therefore, X-ray fluorescence analysis or FT-IR is generally performed at an analysis test site rather than at a work site. If the grease is removed from the robot 30 and taken to an analysis test site or the like located away from the work site, a time lag occurs before the deterioration level of the grease is determined. In addition, it is time-consuming to transport the grease.

[0053] On the other hand, spectroscopic measurement using visible light or near infrared rays does not require a shielding member required in X-ray fluorescence analysis, and has a high signal-to-noise ratio compared to FT-IR, so the optical system is simple or small. In addition, in recent years, MEMS-based spectroscopes using visible light or near infrared rays have been developed, so the size of the device is small. Therefore, if the additive consumption rate can be measured by spectroscopic measurement using visible light or near infrared rays, the deterioration level of the grease can be estimated at the work site without bringing the grease to an analysis and testing site. However, a problem with spectroscopic measurement using visible light or near infrared rays is that the spectral peak caused by the additive cannot be measured with high accuracy. In this regard, in this embodiment, machine learning is performed using the additive consumption rate obtained using X-ray fluorescence analysis or FT-IR as learning data. This makes it possible to accurately measure the additive consumption rate based on the spectral data obtained by spectroscopic measurement using visible light or near infrared rays. In other words, by learning data obtained using a measurement method with higher accuracy than spectroscopic measurement using visible light or near infrared rays as learning data, it is possible to realize the advantage of simplicity of spectroscopic measurement using visible light or near infrared rays while performing measurement with accuracy comparable to that of a highly accurate measurement method. Therefore, by using the measurement device 10 of this embodiment, the deterioration level of grease can be estimated at the work site. As a result, the time lag until the deterioration level of grease is estimated is significantly shortened, and the effort of transporting grease can be eliminated.

[0054] Next, a second embodiment will be described with reference to FIG.

[0055] In the measurement device 10 of the first embodiment, all necessary devices are arranged in a single housing 11. In contrast, in the measurement device 10 of the second embodiment, the housing 11 is connected to a PC 40.

[0056] The housing 11 is formed with a setting section 12 for setting the cell 20. The setting section 12 in the second embodiment is a surface for placing the cell 20. A window section that is light-transmitting is formed in the setting section 12. In this embodiment, the upper case 22 is engaged with the lower case 21, and then the holder 23 is engaged with the upper case 22. The holder 23 has openings on the upper and lower surfaces, and a standard reflector 24 is attached to the opening. The standard reflector 24 has known reflection characteristics and is used as a reference.

[0057] The spectroscope 13 is disposed inside the housing 11. The spectroscope 13 irradiates the cell 20 set in the setting unit 12 with reference light through an opening on the bottom surface of the holder 23, and measures the light intensity for each wavelength by dispersing the reflected light to obtain spectral data. Note that instead of forming an opening on the bottom surface of the holder 23, the bottom surface of the holder 23 may be omitted. In this case, for example, the holder 23 is provided with a guide into which the cell 20 can be inserted and which can hold the inserted cell 20.

[0058] A PC 40 is connected to the housing 11. The PC 40 functions as an estimation device that estimates the deterioration level of the grease based on the spectrum data. The PC 40 includes a communication device 41, a processing device 42, and a display 43. The communication device 41 is a wired communication module or a wireless communication module, and communicates with an external device. Specifically, the communication device 41 performs a process of receiving spectrum data from the spectrometer 13. The processing device 42 has a configuration similar to that of the processing device 14 of the first embodiment, and performs the same process. That is, the processing device 42 stores the above-mentioned model, and estimates the additive consumption rate based on the spectrum data. Furthermore, the processing device 42 estimates the deterioration level of the grease based on the estimated additive consumption rate, and outputs it to the display 43. The display 43 has the same use and function as the display 16 of the first embodiment.

[0059] In the second embodiment, an operator issues an instruction to measure the grease by operating the PC 40. Spectral data acquired by the spectrometer 13 measuring the grease is output to a processing device 42 of the PC 40. The processing device 42 estimates the deterioration level of the grease in the same manner as in the first embodiment, and displays it on a display 43.

[0060] In the second embodiment, the PC 40, which is used for general purposes, constitutes a part of the measurement device 10. Therefore, the cost of the measurement device 10 can be reduced.

[0061] Next, a third embodiment will be described with reference to FIG.

[0062] In the third embodiment, a part of the processing performed by the processing device 14 in the first embodiment is performed by a server 52 connected via a wide area network such as the Internet. The measuring device 10 in the third embodiment includes a communicator 18 for network communication. The communicator 18 is a wired communication module or a wireless communication module. The communicator 18 can communicate with the server 52 via a router 51 arranged in the same LAN.

[0063] In the third embodiment, the measurement device 10 transmits to the server 52 the spectrum data acquired by the spectrometer 13 measuring the grease.

[0064] The server 52 functions as an estimation device that estimates the deterioration level of the grease based on the spectrum data. The server 52 includes a communication device 52a and a processing device 52b. The communication device 52a is a wired communication module or a wireless communication module, and communicates with the measurement device 10. The processing device 52b stores the above-mentioned model and estimates the additive consumption rate based on the spectrum data. Furthermore, the processing device 52b estimates the deterioration level of the grease based on the estimated additive consumption rate. The server 52 transmits the estimated deterioration level of the grease to the measurement device 10.

[0065] The processing device 14 outputs the deterioration level of the grease received from the server 52 to the display 16 .

[0066] In the third embodiment, the model is provided in the server 52. Therefore, it is possible to suppress the specifications required for the processing device 14. In addition, by providing the model in the server 52, the model can be shared among a plurality of measuring devices 10. Furthermore, when the model is updated by additional learning or the like, it is only necessary to update the model in the server 52.

[0067] Next, a fourth embodiment will be described with reference to FIG.

[0068] In the fourth embodiment, the first learning data is spectrum data, and the second learning data is a deterioration level. As shown in Fig. 7, when creating a model, the learning deterioration level is a value obtained by converting an additive consumption rate measured by an instrument other than the measuring instrument 10 using the above-mentioned converter. The only difference between the processing when creating a model in the first and second embodiments is the second learning data, and the other processing is the same as in the first embodiment.

[0069] As shown in Fig. 7, when the model is used, spectral data is input to the model. The input spectral data is spectral data measured and acquired by the spectrometer 13. Here, the model has learned the relationship between the spectral data and the degradation level. Therefore, by inputting the spectral data to the model, the model outputs an estimated degradation level. Therefore, in the fourth embodiment, it is not necessary to use a converter when the model is used.

[0070] The features of the above-mentioned first to fourth embodiments can be combined as appropriate. For example, the feature of the third embodiment in which the model is provided in the server 52 can be combined with the second embodiment. Also, the feature of the fourth embodiment in which the model outputs a degradation level can be combined with the second or third embodiment.

[0071] As described above, the measuring device 10 of the above embodiment outputs the deterioration level of the machine auxiliary. The machine auxiliary is an object that contains an additive, is liquid or semi-solid, and is supplied to a machine to maintain the operation of the machine. The measuring device 10 includes a spectrometer 13 and a processing device 14, 42. The spectrometer 13 performs spectroscopic measurement of the machine auxiliary using visible light, near infrared light, mid-infrared light, or ultraviolet light to obtain spectral data. The processing device 14, 42 outputs the deterioration level of the machine auxiliary based on the spectral data. The processing device 14, 42 inputs the spectral data obtained by the spectrometer 13 into the model as input data. In the first embodiment, when creating the model, the model is constructed by machine learning using the spectral data as the first learning data and the additive consumption rate as the second learning data. Then, when using the model, the processing device 14 of the first embodiment outputs the deterioration level of the grease estimated based on the estimated additive consumption rate, which is the output data of the model. In the fourth embodiment, when creating a model, the model is constructed by machine learning using spectrum data as the first learning data and the deterioration level as the second learning data. When using the model, the processing device 14 of the fourth embodiment outputs the output data of the model as the estimated deterioration level of the grease. This is characteristic 1. In addition, the case where the additive consumption rate is used as the second learning data is characteristic 2, and the case where the deterioration level is used as the second learning data is characteristic 3.

[0072] By performing spectroscopic measurement using visible light, near infrared light, mid-infrared light, or ultraviolet light, it is possible to perform the measurement with a simpler structure or method than when performing X-ray fluorescence analysis.

[0073] In addition, when performing spectroscopic measurement using visible light or near infrared light, the measurement can be performed with a simple structure or method compared to FT-IR. In particular, spectroscopic measurement using visible light or near infrared light may not be able to directly detect the change in the spectrum associated with additive consumption. However, by using a model constructed by machine learning, the additive consumption rate can be estimated with sufficient accuracy.

[0074] In the measurement device 10 of the above embodiment, the model is constructed by machine learning multiple sets of spectrum data and additive consumption rates at different environmental temperatures as learning data. This is characteristic 4.

[0075] This makes it possible to estimate the additive consumption rate while reducing the influence of the environmental temperature when the model is used.

[0076] In the measurement device 10 of the above embodiment, the input data for the model includes the spectrum data, but does not include the environmental temperature.

[0077] This allows the learning data to be simplified. Since there is no need to input the environmental temperature when using the model, the processing when using the model can be simplified.

[0078] The measurement device 10 of this embodiment includes a portable housing 11 that houses a spectrometer 13 therein.

[0079] This makes it possible to realize a measuring device 10 suitable for measurements in a workplace.

[0080] The measuring device 10 of this embodiment includes a housing 11 that houses a spectrometer 13. The housing 11 includes a setting unit 12 for setting a cell 20 filled with a machine auxiliary agent removed from a machine. This is the seventh feature.

[0081] This improves the efficiency of the worker's work of setting up the cell 20.

[0082] The measuring device 10 of this embodiment is disposed on the outer surface of the housing 11 and includes a display 16, 43 that displays the deterioration level of the mechanical auxiliary output by the processing device 14. The processing device 14 is disposed inside the housing 11. The processing device 14 independently estimates the additive consumption rate using a model, and estimates the deterioration level of the mechanical auxiliary based on the additive consumption rate. This is characteristic 8.

[0083] This allows the deterioration level of machine aids to be estimated without external communication.

[0084] In the measurement device 10 of this embodiment, the spectrometer 13 obtains a spectroscopic measurement of the mechanical auxiliary agent using visible light or near infrared light.

[0085] By using visible light or near-infrared light, the optical system can be simplified or made smaller, and the size of the device can be reduced, compared to FT-IR, which uses mid-infrared light. When FT-IR, etc. is used to create training data, both high measurement accuracy and simple device configuration can be achieved.

[0086] The above-mentioned features 1 to 8 can be appropriately combined as long as no contradiction occurs. For example, feature N (N=1, 2, . . . , 9) can be appropriately combined with at least one of features 1 to N-1.

[0087] Although the preferred embodiment of the present application has been described above, the above configuration can be modified, for example, as follows.

[0088] The flowcharts shown in the above embodiment are merely examples, and some processes may be omitted, some processes may be changed, or new processes may be added. For example, instead of the process of outputting the deterioration level of the grease to a display, the deterioration level of the grease may be output to a terminal carried by the worker.

[0089] The method of creating the model shown in the above embodiment is merely an example, and a model may be created using different learning data. For example, the environmental temperature may be divided into a plurality of levels, and a model may be created by learning learning data for each environmental temperature of the same level. Also, the spectrum data of one type of grease may be learned, and a model may be constructed for each type of grease.

[0090] The functions of the elements disclosed herein can be performed using circuits or processing circuits, including general purpose processors, special purpose processors, integrated circuits, Application Specific Integrated Circuits (ASICs), conventional circuits, and / or combinations thereof, configured or programmed to perform the disclosed functions. Processors are considered processing circuits or circuits because they include transistors and other circuits. In this disclosure, a circuit, unit, or means is hardware that performs the recited functions or hardware that is programmed to perform the recited functions. The hardware may be hardware disclosed herein or other known hardware that is programmed or configured to perform the recited functions. Where the hardware is a processor, which is considered a type of circuit, the circuit, means, or unit is a combination of hardware and software, and the software is used to configure the hardware and / or the processor. [Explanation of symbols]

[0091] 10. Measurement Equipment 11. Cabinet 12 Set section 13 Spectrometer 14 Processing equipment 15 Battery 16 Display 17 Measurement button 40 PC (estimation device) 52 Server (Estimation device)

Claims

1. A measuring device that outputs a deterioration level of a machine auxiliary agent, which includes an additive, is liquid or semi-solid, and is supplied to a machine to maintain the operation of the machine, A spectrometer that measures the mechanical auxiliary by using visible light, near infrared light, mid-infrared light, or ultraviolet light to obtain spectral data; a processor that outputs a degradation level of the machine aid based on the spectral data; Equipped with The processing device inputs the spectral data acquired by the spectrometer into a model as input data; At the time of creating the model, the model is constructed by machine learning using the spectrum data as first learning data and an additive consumption rate or a deterioration level converted from the additive consumption rate as second learning data, When the additive consumption rate is used as the second learning data, during use of the model, the processing device outputs a deterioration level of the machine auxiliary estimated based on the estimated additive consumption rate, which is output data of the model; A measuring device in which, when the deterioration level is used as second learning data, when a model is used, the processing device outputs the output data of the model as the estimated deterioration level of the machine auxiliary agent.

2. The measurement device according to claim 1 , At the time of creating the model, the model is constructed by machine learning using the spectrum data as first learning data and the additive consumption rate as second learning data, When using a model, the processor outputs an estimated degradation level of the machine aid based on an estimated additive consumption rate, which is the output data of the model.

3. The measurement device according to claim 1 , At the time of creating the model, the model is constructed by machine learning using the spectrum data as first learning data and a deterioration level converted from an additive consumption rate as second learning data, When using a model, the processing device outputs the output data of the model as an estimated degradation level of the machine auxiliary.

4. The measurement device according to claim 1 , The model is constructed by machine learning multiple sets of the spectral data and the additive consumption rate at different environmental temperatures as learning data.

5. The measurement device according to claim 4, The input data for the model includes the spectral data and does not include the environmental temperature.

6. The measurement device according to claim 1 , A measurement device comprising a portable housing in which the spectrometer is housed.

7. The measurement device according to claim 1 , a housing for accommodating the spectrometer therein; The measuring device, wherein the housing is provided with a setting section for setting a cell into which the machine auxiliary agent removed from the machine is filled.

8. The measurement device according to claim 7, a display disposed on an outer surface of the housing and displaying a deterioration level of the machine auxiliary agent output by the processing device; The processing device is disposed inside the housing, The processor is configured to independently estimate the additive consumption rate using the model and to estimate a degradation level of the machine aid based on the additive consumption rate.

9. The measurement device according to claim 1 , The spectrometer is a measuring device that performs spectroscopic measurement of the mechanical auxiliary using visible light or near infrared light to obtain spectral data.

10. 1. An estimation device for estimating a deterioration level of a machine auxiliary, the machine auxiliary including an additive, being liquid or semi-solid, and being supplied to a machine for maintaining the operation of the machine, comprising: A communication device for receiving spectral data obtained by spectroscopically measuring the mechanical auxiliary using visible light, near infrared light, mid-infrared light, or ultraviolet light; a processing device for estimating a degradation level of the machine aid based on the spectral data; Equipped with The processing device inputs the spectrum data received by the communication device as input data into a model; At the time of creating the model, the model is constructed by machine learning using the spectrum data as first learning data and an additive consumption rate or a deterioration level converted from the additive consumption rate as second learning data, When the additive consumption rate is used as the second learning data, during use of the model, the processing device estimates a deterioration level of the machine auxiliary agent based on the additive consumption rate, which is output data of the model; When the deterioration level is used as second learning data, when the model is used, the processing device sets the output data of the model as an estimation result of the deterioration level of the machine auxiliary.

11. 1. A method for estimating a deterioration level of a machine auxiliary, the machine auxiliary including an additive, being liquid or semi-solid, and being supplied to a machine for maintaining the operation of the machine, comprising: Obtaining spectral data by spectroscopically measuring the mechanical auxiliary using visible light, near infrared light, mid-infrared light or ultraviolet light; inputting the spectral data as input data into a model; Estimating and outputting a deterioration level of the machine auxiliary agent based on the output data of the model; At the time of creating the model, the model is constructed by machine learning using the spectrum data as first learning data and an additive consumption rate or a deterioration level converted from the additive consumption rate as second learning data, When the additive consumption rate is used as the second learning data, when the model is used, a deterioration level of the machine auxiliary estimated based on the estimated additive consumption rate, which is output data of the model, is output; The estimation method, in which, when the deterioration level is used as second learning data, output data of the model is output as the estimated deterioration level of the machine auxiliary when the model is used.

Images