WO2024067824A1 - 气缸劣化趋势确定方法及装置、电子设备、存储介质 - Google Patents

气缸劣化趋势确定方法及装置、电子设备、存储介质 Download PDF

Info

Publication number
WO2024067824A1
WO2024067824A1 PCT/CN2023/122730 CN2023122730W WO2024067824A1 WO 2024067824 A1 WO2024067824 A1 WO 2024067824A1 CN 2023122730 W CN2023122730 W CN 2023122730W WO 2024067824 A1 WO2024067824 A1 WO 2024067824A1
Authority
WO
WIPO (PCT)
Prior art keywords
cylinder
score
value
future
preset time
Prior art date
Application number
PCT/CN2023/122730
Other languages
English (en)
French (fr)
Inventor
左志军
江伟乐
陈旻琪
贺毅
张凯
Original Assignee
广州明珞装备股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州明珞装备股份有限公司 filed Critical 广州明珞装备股份有限公司
Publication of WO2024067824A1 publication Critical patent/WO2024067824A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance

Definitions

  • the present application relates to the technical field of cylinder detection, and in particular to a method and device for determining a cylinder degradation trend, an electronic device, and a storage medium.
  • Cylinder equipment maintenance is generally done by prompting equipment managers through real-time early warnings, but they can only give real-time alarms and cannot determine the status of the equipment, so the equipment cannot be maintained in advance.
  • the present application provides a method and device for determining the degradation trend of a cylinder, an electronic device, and a storage medium to solve the problem in the related art that the cylinder equipment is in a passive state in which maintenance is performed only after a fault is found, and maintenance work cannot be arranged in advance.
  • an embodiment of the present application provides a method for determining a cylinder degradation trend, comprising: obtaining cylinder score data within a preset time range, wherein the cylinder score data is used to characterize the score of the cylinder action within the preset time range; predicting the cylinder score at a certain point in the future based on the cylinder score data and a score prediction model; and determining the degradation trend of the cylinder based on the cylinder score at the certain point in the future.
  • the score prediction model includes a first prediction algorithm, and the first prediction algorithm is implemented by the following formula:
  • St ⁇ ( yt - Lt )+(1- ⁇ ) Sts ;
  • ⁇ is the first smoothing coefficient
  • ⁇ is the second smoothing coefficient
  • ⁇ is the third smoothing coefficient
  • the value range of ⁇ , ⁇ , and ⁇ is between 0 and 1
  • s is the length of the seasonal cycle
  • k is the length from time t to the certain future time point within the preset time range
  • y t is the cylinder score at time t
  • L t is the smoothed cylinder score at time t
  • L t-1 is the smoothed cylinder score at time (t-1) within the preset time range
  • b t is the trend value of the cylinder score at time t
  • b t-1 is the trend value of the cylinder score at time (t-1)
  • S t is the seasonal cycle value of the cylinder score at time t
  • S ts is the seasonal cycle value of the cylinder score at time (ts)
  • S t+ks is the seasonal cycle value of the cylinder score at time (t+ks)
  • the first smoothing coefficient, the second smoothing coefficient and the third smoothing coefficient have the same value.
  • the method for determining the cylinder degradation trend further includes: dividing the cylinder score data within the preset time range into a sample data set and a verification data set, wherein the sample data set is the cylinder score data of a first time period within the preset time range, and the verification data set is the cylinder score data of a second time period within the preset time range, and the second time period is located after the first time period; using the sample data set, determining the value of the first smoothing coefficient, the value of the second smoothing coefficient, and the value of the third smoothing coefficient; according to the determined values of the first smoothing coefficient, the value of the second smoothing coefficient, and the value of the third smoothing coefficient, using the first prediction algorithm, predicting the cylinder score of the second time period to obtain a predicted value of the cylinder score of the second time period; The cylinder score prediction value is compared with the cylinder score data in the verification data value of the second time period to determine the prediction deviation value of the first prediction algorithm.
  • the sample data set is the first X% of historical data within the preset time range
  • the verification data set is the last 1-X% of historical data within the preset time range, where the value of X is between 0 and 100.
  • b is the slope of the regression equation
  • a is the intercept of the regression equation
  • X is the time value at a certain time point in the future
  • Y is the cylinder score at the certain time point in the future predicted by the score prediction model.
  • the slope of the regression equation is calculated as follows:
  • xi is the time value at any time point within the preset time range
  • yi is the cylinder score at any time point within the preset time range
  • n is the length of the preset time range.
  • predicting the cylinder score at a certain point in the future based on the cylinder score data and the score prediction model includes: when the prediction deviation value of the first prediction algorithm is less than or equal to the preset deviation value, using the first prediction algorithm to determine the cylinder score at the certain point in the future; when the prediction deviation value of the first prediction algorithm is greater than the preset deviation value, using the second prediction algorithm to determine the cylinder score at the certain point in the future.
  • the obtaining of cylinder score data within a preset time range includes: obtaining action data of the cylinder within the preset time range, the action data including action duration of the cylinder action, the number of cylinder actions and the cumulative action duration of the cylinder within the preset time range; determining the cylinder score of at least one time point within the preset time range as the cylinder score data based on the action duration of the cylinder action, the number of cylinder actions and the cumulative action duration of the cylinder, wherein one time point corresponds to one cylinder score.
  • the cylinder score at any time point within the preset time range is obtained by the following formula:
  • C is the cylinder score at any time point within the preset time range
  • X is the upper limit alarm value of the action duration of the cylinder action
  • Ttotal is the cumulative working time of the cylinder at any time point within the preset time range
  • Y is the number of cylinder actions at any time point within the preset time range.
  • determining the degradation trend of the cylinder according to the cylinder score at a certain point in the future includes: determining a magnitude relationship between the cylinder score at a certain point in the future and different preset scores; determining the degradation trend of the cylinder according to the magnitude relationship;
  • the different preset scores include a normal score and a warning score, wherein the degradation trend of the cylinder is determined based on the size relationship, including: when the cylinder score at the certain time point in the future is higher than the normal score, it is determined that the cylinder has no degradation trend; when the cylinder score at the certain time point in the future is between the warning score and the normal score, it is determined that the cylinder has a slow degradation trend; when the cylinder score at the certain time point in the future is lower than the warning score, it is determined that the cylinder has a rapid degradation trend.
  • an embodiment of the present application further provides a cylinder degradation determination device, comprising: a collection module for acquiring cylinder score data within a preset time range, wherein the cylinder score data is used to characterize the cylinder degradation within the preset time range. The score of the cylinder action within the range; a prediction module, used to predict the cylinder score at a certain point in the future according to the cylinder score data and the score prediction model; a processing module, determining the degradation trend of the cylinder according to the cylinder score at the certain point in the future.
  • an embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the cylinder degradation trend determination method described in any embodiment of the first aspect is implemented.
  • an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the cylinder degradation trend determination method described in any embodiment of the first aspect above is implemented.
  • the cylinder degradation trend determination method and device, electronic device, and storage medium provided in the above embodiments of the present application have the following beneficial effects:
  • the cylinder score at a certain point in the future is predicted in the score prediction model.
  • the traditional acquisition device can be used for data acquisition, and there is no need to add new acquisition equipment, and there will be no cost for new equipment.
  • the degradation trend of the cylinder is determined by the cylinder score at a certain point in the future. If there is no degradation trend or a degradation trend, the degradation trend can be divided into a rapid degradation trend and a slow degradation trend.
  • cylinders that do not show a degradation trend they can continue to be used without additional maintenance; for cylinders that show a degradation trend, it is determined whether they are in a rapid degradation trend or a slow degradation trend, and the cylinders in the rapid degradation trend are maintained first, and then the cylinders in the slow degradation trend are processed. In this way, multiple cylinders can be compared horizontally and arranged in order of cylinder maintenance.
  • FIG1 is a schematic diagram of the structure of a cylinder degradation trend determination system provided in one embodiment of the present application.
  • FIG2 is a flow chart of a method for determining a cylinder degradation trend according to an embodiment of the present application.
  • FIG3 is a schematic diagram of the structure of a cylinder to be tested provided in one embodiment of the present application.
  • FIG. 4 is a schematic diagram of cylinder scores changing over time according to an embodiment of the present application.
  • FIG5 is a flow chart of a method for determining a cylinder degradation trend according to an embodiment of the present application.
  • FIG6 is a flow chart of a method for determining a cylinder degradation trend according to an embodiment of the present application.
  • FIG. 7 is a schematic diagram of the structure of a cylinder degradation trend determination device provided in an embodiment of the present application.
  • FIG8 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.
  • the directional indications are only used to explain the relative position relationship, movement status, etc. between the components in a certain specific posture. If the specific posture changes, the directional indication will also change accordingly.
  • the data of the Programmable Logic Controller (PLC) on the production line is collected by a collector, and the action data and status data in the PLC data are distinguished according to the rules pre-set by the operator, and the action data in the PLC data is monitored.
  • the collected data such as the action data of the cylinder
  • the action data of the cylinder can be processed and converted, and the future failure risk probability of the cylinder can be predicted according to the set judgment rules, so as to timely discover hidden dangers and troubleshoot problem cylinders in a timely manner.
  • PLC The program written by PLC is the PLC control program.
  • PLC is a digital computing electronic system specially designed for use in industrial environments. It uses a programmable memory to store instructions for performing logical operations, sequential control, timing, counting, and arithmetic operations. It controls various types of mechanical equipment or production processes through digital or analog input and output.
  • Fig. 1 is a schematic diagram of a cylinder degradation trend determination system provided by an embodiment of the present application. As shown in Fig. 1 , the system includes: a server 110 and a plurality of cylinders 120, 121, ..., 12n (n is an integer greater than 0) and a bus 130.
  • Each of the plurality of cylinders 120, 121, ..., 12n may be the cylinder 300 described in FIG. 3.
  • the server 110 is a server, or is composed of a plurality of servers, or is a virtualization platform, or is a cloud computing service center.
  • Each of the plurality of cylinders 120, 121, ..., 12n and the server 110 transmit data to each other via a bus 130.
  • the server 110 obtains cylinder score data within a preset time range from any cylinder 12x through the bus 130; the server 110 predicts the cylinder score at a certain point in the future based on the cylinder score data and the score prediction model; the server 110 determines the motion stability index of the first device based on the statistical value; the server 110 determines the degradation trend of the cylinder based on the cylinder score predicted at the certain point in the future.
  • the server 110 can also send the degradation trend of the cylinder to the terminal device so that the user at the terminal device can know what state the cylinder is in.
  • the server 110 can exist independently of the multiple cylinders 120, 121...12n, and the two transmit data through the bus 130, but this is just an example to illustrate the application scenario of the present application.
  • each of the multiple cylinders 120, 121...12n can also be equipped with a server, and each cylinder executes the cylinder degradation trend determination method of the present application through its own server.
  • an embodiment of the present application provides a method for determining a cylinder degradation trend.
  • the method may be executed by the server 110 shown in FIG. 1 .
  • the method may include the following steps:
  • the process of acquiring cylinder score data within a preset time range includes: acquiring cylinder action data within the preset time range, wherein the cylinder action data includes action duration of the cylinder action, number of cylinder actions, and cumulative duration of the cylinder action.
  • the cylinder score at at least one time point within the preset time range is determined as the cylinder score data.
  • One time point corresponds to one cylinder score.
  • the preset time range can be in units of years, months, weeks, or days.
  • the preset time range is a year, a month, a week, or a day. The value of the preset time range can be adjusted according to actual needs.
  • the cylinder 300 includes a cylinder body 310 and a piston 320 disposed inside the cylinder body 310.
  • the piston 320 drives the piston rod 330 to reciprocate in the cylinder body 310, thereby realizing the function of the cylinder 300.
  • the cylinder body 310 is provided with a top dead center sensor 311 and a bottom dead center sensor 312.
  • the top dead center sensor 311 is disposed at the top dead center of the cylinder, i.e., at point B.
  • the top dead center sensor 311 will receive the position information of the piston 320 and send the information that the piston 320 is at the top dead center to the controller.
  • the bottom dead center sensor 312 is disposed at the bottom dead center of the cylinder, i.e., at point A.
  • the bottom dead center sensor 312 will receive the position information of the piston 320 and send the information that the piston 320 is at the bottom dead center to the controller. Therefore, by setting the top dead center sensor 311 and the bottom dead center sensor 312, the movement state of the piston 320 in the cylinder 300 can be effectively detected, thereby providing analysis data for the subsequent working state of the cylinder.
  • the top dead center sensor 311 and the bottom dead center sensor 312 are Hall sensors. A magnet is provided in the piston 320.
  • the Hall sensor can detect the position of the piston 320.
  • the piston 320 of the cylinder 300 can be restored to its original position under the action of a spring, gravity or other external force.
  • the action of the cylinder 300 can be opening or clamping.
  • the piston 320 moves from point A to point B for opening, and the time from collecting the signal from the bottom dead center sensor 312 to the signal from the top dead center sensor 311 is the action duration of the opening action of the cylinder 300; the piston 320 moves from point B to point A for clamping, and the time from collecting the signal from the top dead center sensor 311 to the signal from the bottom dead center sensor 312 is the action duration of the clamping action of the cylinder 300.
  • the cylinder score at any time point within the preset time range is calculated according to the following formula:
  • C is the cylinder score at any time point within the preset time range
  • X is the upper limit alarm value of the action duration of the cylinder action
  • Y is the number of cylinder actions at any time point within the preset time range
  • Ttotal is the cumulative working time of the cylinder at any time point within the preset time range.
  • the product of the upper limit alarm value X of the action duration of the cylinder and the number of cylinder actions Y is the maximum cumulative action duration of the cylinder under normal working conditions.
  • the ratio of the total action duration to the actual cumulative cylinder working duration T is multiplied by 100 to obtain the cylinder score C at any time point within the preset time range.
  • the cylinder may include multiple cylinder actions at any time point within the preset time range, for example, n cylinder actions, and the upper limit alarm values of the action durations of the n cylinder actions are X1 , X2 ... Xn , and the cylinder scores corresponding to the n cylinder actions at any time point within the preset time range are C1 , C2 ... Cn , and the average value of C1 , C2 ... Cn is determined as the cylinder score C of the cylinder 300 at any time point within the preset time range.
  • the cylinder score within the preset time range is determined by using the action duration of the cylinder action within the preset time range, and the cylinder score within the preset time range is used as a target to predict the cylinder score at a certain time point in the future in the score prediction model.
  • the certain time point in the future can be a month, a week, a day or an hour, and accordingly, the cylinder score at a certain time point in the future refers to the cylinder score in the next month, the cylinder score in the next week, the cylinder score in the next day or the cylinder score in the next hour.
  • the cylinder degradation trend is determined, such as no degradation trend or degradation trend.
  • the degradation trend can be divided into rapid degradation trend and slow degradation trend. For cylinders that do not show degradation trend, they can continue to be used without additional maintenance; for cylinders that show degradation trend, it is determined whether the cylinder degradation trend is in a rapid degradation trend or a slow degradation trend, and the cylinders in the rapid degradation trend are maintained first, and then the cylinders in the slow degradation trend are processed. It should be understood that the judgment rule can be pre-set in the score prediction model.
  • the judgment rule is to judge the size relationship between the cylinder score at a certain point in the future and the normal score and the warning score.
  • the judgment rule is to judge the size relationship between the cylinder score at a certain point in the future and the normal score and the warning score.
  • the cylinder score at a certain point in the future is between the warning score and the normal score, it can be determined that the cylinder has a slow deterioration trend; when the cylinder score at a certain point in the future is lower than the warning score, it can be determined that the cylinder has a rapid deterioration trend.
  • the cylinder score at any time point can be obtained, for example, the cylinder score for the next 7 days.
  • the cylinder scores at multiple future time points can be obtained, for example, the cylinder score for the next 5 days, the cylinder score for the next 6 days, and the cylinder score for the next 7 days.
  • the judgment rule is to judge the relationship between the cylinder scores at multiple future time points and the historical cylinder scores.
  • multiple cylinders can be compared horizontally and in parallel, that is, according to the time dimension, only the degradation trend of one cylinder can be determined at one time, or whether multiple cylinders have degradation trends can be determined at the same time, so as to determine which cylinder with degradation trends should be maintained first, so that maintenance work can be carried out in an orderly manner, and maintenance preparations can be made in advance with sufficient preparation time, so as to minimize the impact on production work and ensure normal operation of the production line. Avoid the production of defective products due to the degradation of one of the cylinders, which affects the quality of production, causes waste of materials, and reduces production efficiency.
  • the score prediction model includes a first prediction algorithm, and the first prediction algorithm includes the following formula:
  • St ⁇ ( yt - Lt )+(1- ⁇ ) Sts ;
  • ⁇ is the first smoothing coefficient
  • ⁇ is the second smoothing coefficient
  • ⁇ is the third smoothing coefficient
  • the value range of ⁇ , ⁇ , and ⁇ is between 0 and 1
  • s is the length of the seasonal cycle
  • k is the length from time t within the preset time range to the future time point.
  • y t is the cylinder score at time t;
  • L t is the smoothed cylinder score at time t,
  • L t-1 is the smoothed cylinder score at time (t-1) within the preset time range,
  • b t is the trend value of the cylinder score at time t,
  • b t-1 is the trend value of the cylinder score at time (t-1),
  • S t is the seasonal periodic value of the cylinder score at time t,
  • S ts is the seasonal periodic value of the cylinder score at time (ts), and
  • S t+ks is the seasonal periodic value of the cylinder score at time (t+ks);
  • F t+k is the cylinder score at time (t+k) predicted by the score prediction model, wherein the time (t+k) is a certain time point in the future.
  • L t ⁇ (y t -S ts )+(1- ⁇ )(L t-1 +b t-1 ) is the level function.
  • the weighted average of the seasonally adjusted observed value i.e., y t -S ts
  • the previous period's non-seasonal forecast value i.e., L t-1 +b t-1
  • b t ⁇ (L t -L t-1 )+(1- ⁇ )b t-1 is a trend function.
  • the trend of the cylinder score can be determined by using the trend function. From a sufficient number of cylinder scores, the trend of the cylinder score at a certain time point in the future can be predicted.
  • St ⁇ ( yt - Lt ) + (1- ⁇ ) Sts is a seasonal function.
  • the sliding average of the current seasonal coefficient (i.e., St ) and the seasonal coefficient of the same period of the previous cycle (i.e., Sts ) can be obtained, which combines the cyclic changes of the same action of the cylinder or the same product.
  • the first prediction algorithm can realize cumulative prediction through the above formula.
  • the first prediction algorithm can also use a cumulative prediction method for prediction.
  • the additive prediction method is used.
  • the prediction results obtained by using the multiplicative prediction method i.e., the cumulative method
  • the prediction results obtained by using the multiplicative prediction method are more accurate.
  • the prediction results obtained by using the multiplicative prediction method are more accurate.
  • a sufficient number of cylinder scores are substituted into the first prediction algorithm for calculation, and a sufficient number of smoothing points of the cylinder scores are obtained through the level function.
  • the value corresponding to the smoothing point is called the cylinder smoothing value; the smoothing point is substituted into the trend function to obtain a sufficient number of smoothing lines of the cylinder scores, in other words, a sufficient number of trend values of the cylinder smoothing values are obtained; the smoothing line is then substituted into the seasonal function, and combined with the actual periodic changes, the cylinder score at a certain point in the future that fits the actual production and changes is obtained.
  • the difference between the cylinder score and the smoothing point and the average value of the difference in the same period of the previous cycle are substituted into the seasonal function to obtain a sufficient number of seasonal coefficients, and the smoothing point, trend value and seasonal coefficient of the closest predicted future cycle are substituted into the forecast function to obtain the cylinder score at a certain point in the future that fits the actual production and changes.
  • a sufficient number of cylinder scores are cylinder scores corresponding to at least 7 time points, and the cylinder score at a future time point can be obtained through the cylinder scores corresponding to at least 7 time points.
  • the cylinder score of the 8th day i.e., the cylinder score at a future time point
  • the cylinder score of the 8th day can only be obtained after determining the cylinder scores of the previous 7 days.
  • the first smoothing coefficient ⁇ , the second smoothing coefficient ⁇ and the third smoothing coefficient ⁇ have the same value.
  • the first smoothing coefficient ⁇ , the second smoothing coefficient ⁇ and the third smoothing coefficient ⁇ have a value range between 0 and 1.
  • a combination of multiple values can be selected from the values between 0 and 1 to predict the cylinder score at a certain time point in the future. It has been verified that when the values of the three smoothing coefficients are the same, the cylinder score at a specific time node in the future can be predicted more accurately.
  • the seasonal cycle value S can be understood as a constraint value, which is used to constrain the cylinder scores of future time points predicted by the score prediction model to change periodically. When the cylinder scores of future time points predicted by the score prediction model do not change periodically, the value of S is close to 0. When the cylinder scores of future time points predicted by the score prediction model change periodically, the seasonal function is used to calculate the value of S at the corresponding moment.
  • the length s of the seasonal cycle refers to the length of a cycle. For example, if the prediction cycle is days, the length of the seasonal cycle is 7 days, if the prediction cycle is weeks, the length of the seasonal cycle is 4 weeks, and if the prediction cycle is months, the length of the seasonal cycle is 12 months.
  • the preset time range may include multiple moments, and moment t refers to the end time node of the preset time range.
  • moment t refers to the end time node of the preset time range.
  • there are multiple moments before moment t for example, moment t-1, moment t-2, etc.
  • the cylinder score data within the preset time range may include the cylinder score at at least one time point, so the cylinder score at time t refers to the cylinder score at the end of the preset time range.
  • the preset time range is 14 days, time t is the 14th day, k is 7 days, then the cylinder score at time t refers to the cylinder score on the 14th day, and F t+k refers to the cylinder score for the next 7 days predicted by the score prediction model.
  • the method further includes: determining the accuracy of the first prediction algorithm.
  • determining the accuracy of the first prediction algorithm includes the following steps:
  • S540 Compare the cylinder score prediction value with the cylinder score data in the verification data value of the second time period to determine a prediction deviation value of the first prediction algorithm.
  • the cylinder score data within a preset time range is divided into a sample data set and a verification data set, wherein the sample data set is the historical data of the first time period within the preset time range (i.e., the cylinder score data), and the verification data set is It is the historical data of the second time period within the preset time range (i.e., cylinder score data), and the second time period is located after the first time period.
  • the first 90% of the historical data within the preset time range is used as the sample data set
  • the last 10% of the historical data within the preset time range is used as the verification data set.
  • the sample data set i.e., the data from the first 328 days
  • the verification data set i.e., the data from the last 37 days
  • the first time period is the first 328 days
  • the second time period is the last 37 days.
  • the values of the first time period and the second time period can be determined according to the actual situation of the production line.
  • the sample data set can be used to obtain the values of the first smoothing coefficient ⁇ , the second smoothing coefficient ⁇ , and the third smoothing coefficient ⁇ .
  • the sample data set can be used for model training, and the values of the first smoothing coefficient ⁇ , the second smoothing coefficient ⁇ , and the third smoothing coefficient ⁇ can be obtained through a deep learning method.
  • the first prediction algorithm is used to predict the cylinder score of the second time period to obtain a predicted cylinder score value of the second time period; the predicted cylinder score value is compared with the cylinder score data in the verification data value of the second time period to obtain a prediction deviation value of the first prediction algorithm.
  • the cylinder score prediction value obtained by the first prediction algorithm is compared with the actual cylinder score, and the difference between the two is the prediction deviation value.
  • the Symmetric Mean Absolute Percentage Error (SMAPE) method is used to calculate the difference between the cylinder score prediction value and the actual cylinder score, that is, the prediction deviation value of the first prediction algorithm.
  • SMAPE Symmetric Mean Absolute Percentage Error
  • ⁇ is the prediction deviation value
  • R is the cylinder score prediction value
  • F is the actual cylinder score.
  • the prediction deviation value ranges from 0% to 200%, and the closer to 0%, the more accurate the prediction.
  • the cylinder score data within a preset time range is divided into a sample data set and a verification data set;
  • the value of the smoothing coefficient is obtained by substituting the known historical data (i.e., the sample data set) into the first prediction algorithm;
  • the smoothing coefficient is substituted into the first prediction algorithm to improve the first prediction algorithm;
  • the cylinder scores in the sample data set are substituted into the improved first prediction algorithm to obtain the cylinder score prediction value;
  • a sufficient number of cylinder score prediction scores are obtained by repeating the fourth step; in the sixth step, each cylinder score prediction score is compared with the corresponding cylinder score in the verification data set through the SMAPE method to obtain the prediction deviation value.
  • the cylinder scores of the sample data set are substituted into the first prediction algorithm to obtain the values of the first smoothing coefficient ⁇ , the second smoothing coefficient ⁇ , and the third smoothing coefficient ⁇ to improve the first prediction algorithm.
  • the cylinder scores of the sample data set are then substituted into the improved first prediction algorithm to obtain a set of cylinder score prediction values, and each cylinder score prediction value is compared with the corresponding cylinder score of the validation data set.
  • the sample data set is the first X% of historical data within the preset time range
  • the verification data set is the last 1-X% of historical data within the preset time range.
  • the value of X is between 0 and 100.
  • the value of X can be determined according to actual needs.
  • X can be 90, in which case the sample data set is the first 90% of historical data within the preset time range, and the verification data set is the last 10% of historical data within the preset time range.
  • X can be 70, in which case the sample data set is the first 70% of historical data within the preset time range, and the verification data set is the last 30% of historical data within the preset time range.
  • the score prediction model also includes a second prediction algorithm.
  • the first prediction algorithm is used to predict the cylinder score at a certain point in the future.
  • the second prediction algorithm is used to predict the cylinder score at a certain point in the future.
  • the method includes the following steps:
  • step S620 When the prediction deviation value of the first prediction algorithm is less than or equal to the preset deviation value, executing step S620, using the first prediction algorithm, and executing step S640 to predict the cylinder score at a certain time point in the future;
  • step S630 is executed to adopt the second prediction algorithm
  • step S640 is executed to predict the cylinder score at a certain time point in the future.
  • the predicted deviation value exceeds the preset deviation value, it proves that the error of the cylinder score predicted by the first prediction algorithm is too large, and the prediction is judged to be invalid.
  • the first prediction algorithm is not applicable, and the second prediction algorithm is switched for prediction. For example, if the preset deviation value is 20% and the predicted deviation value is 23%, then the first prediction algorithm is invalid, and the second prediction algorithm is used for prediction. For another example, if the preset deviation value is 20% and the predicted deviation value is 10%, then it is determined that the first prediction algorithm is accurate, and the first prediction algorithm continues to be used for prediction. That is, by switching between the first prediction algorithm and the second prediction algorithm, the prediction result of the score prediction model is made more accurate.
  • b is the slope of the regression equation
  • a is the intercept of the regression equation
  • X is the time value at a certain time point in the future
  • Y is the cylinder score at the certain time point in the future predicted by the score prediction model.
  • the cylinder score Y at any future time node can be obtained by the time value X of the time node.
  • the prediction data sample is all historical data within the query time range. For example, when using the data of July 2022 to calculate the slope and intercept of the regression equation, the slope b and intercept a are calculated using the number of days X in July and the cylinder score Y of each day, so that the specific linear regression equation is obtained through the slope b and intercept a.
  • the calculation method of the slope and intercept of the regression equation in the second prediction algorithm includes:
  • xi is the time value at any time point within the preset time range
  • yi is the cylinder score at any time point within the preset time range
  • It is the average value of the time values at all time points within the preset time range.
  • n is the length of the preset time range.
  • the time value x i at any time point within the preset time range, the cylinder score y i at any time point within the preset time range, and the average value of the time values at all time points within the preset time range are used. and the average of all cylinder scores within a preset time range Calculate the slope b of the regression equation. After obtaining the slope b, calculate the average value of the time values of all time points within the preset time range. and the average of all cylinder scores within a preset time range Calculate the intercept a of the regression equation. For example, if the cylinder score data for July 2022 is used to calculate the slope and intercept of the regression equation, then:
  • a cylinder degradation determination device 700 includes:
  • the acquisition module 710 is used to obtain cylinder score data within a preset time range, wherein the cylinder score data is used to represent the score of the cylinder action within the preset time range;
  • a prediction module 720 configured to predict the cylinder score at a certain time point in the future according to the cylinder score data and the score prediction model;
  • the processing module 730 determines the degradation trend of the cylinder according to the cylinder score at the future time point.
  • the cylinder score at a certain point in the future is predicted in the score prediction model.
  • the traditional acquisition device can be used for data acquisition, and there is no need to add new acquisition equipment, and there will be no cost for new equipment.
  • the degradation trend of the cylinder is determined by the cylinder score at a certain point in the future. If there is no degradation trend or a degradation trend, the degradation trend can be divided into a rapid degradation trend and a slow degradation trend.
  • cylinders that do not show a degradation trend they can continue to be used without additional maintenance; for cylinders that show a degradation trend, it is determined whether they are in a rapid degradation trend or a slow degradation trend, and the cylinders in the rapid degradation trend are maintained first, and then the cylinders in the slow degradation trend are processed. In this way, multiple cylinders can be compared horizontally and arranged in order of cylinder maintenance.
  • the score prediction model includes a first prediction algorithm, and the first prediction algorithm is implemented by the following formula:
  • St ⁇ ( yt - Lt )+(1- ⁇ ) Sts ;
  • ⁇ is the first smoothing coefficient
  • ⁇ is the second smoothing coefficient
  • ⁇ is the third smoothing coefficient
  • the value range of ⁇ , ⁇ , and ⁇ is between 0 and 1
  • s is the length of the seasonal cycle
  • k is the length from time t to the certain future time point within the preset time range
  • y t is the cylinder score at time t
  • L t is the smoothed cylinder score at time t
  • L t-1 is the smoothed cylinder score at time (t-1) within the preset time range
  • b t is the trend value of the cylinder score at time t
  • b t-1 is the trend value of the cylinder score at time (t-1)
  • S t is the seasonal cycle value of the cylinder score at time t
  • S ts is the seasonal cycle value of the cylinder score at time (ts)
  • S t+ks is the seasonal cycle value of the cylinder score at time (t+ks)
  • the first smoothing coefficient, the second smoothing coefficient, and the third smoothing coefficient have the same value.
  • the prediction module 720 is also used to: divide the cylinder score data within the preset time range into a sample data set and a verification data set, wherein the sample data set is the cylinder score data of the first time period within the preset time range, and the verification data set is the cylinder score data of the second time period within the preset time range, and the second time period is located after the first time period; using the sample data set, determine the value of the first smoothing coefficient, the value of the second smoothing coefficient, and the value of the third smoothing coefficient; based on the determined value of the first smoothing coefficient, the value of the second smoothing coefficient, and the value of the third smoothing coefficient, predict the cylinder score of the second time period through the first prediction algorithm to obtain a predicted value of the cylinder score of the second time period; compare the predicted value of the cylinder score with the cylinder score data in the verification data value of the second time period to determine the prediction deviation value of the first prediction algorithm.
  • the sample data set is the first X% of historical data within the preset time range
  • the verification data set is the last 1-X% of historical data within the preset time range, where the value of X is between 0 and 100.
  • b is the slope of the regression equation
  • a is the intercept of the regression equation
  • X is the time value at a certain time point in the future
  • Y is the cylinder score at the certain time point in the future predicted by the score prediction model.
  • the slope of the regression equation is calculated as follows:
  • xi is the time value at any time point within the preset time range
  • yi is the cylinder score at any time point within the preset time range
  • n is the length of the preset time range.
  • the prediction module 720 predicts the cylinder score at a certain point in the future based on the cylinder score data and the score prediction model, and is specifically used to: when the prediction deviation value of the first prediction algorithm is less than or equal to the preset deviation value, use the first prediction algorithm to determine the cylinder score at the certain point in the future; when the prediction deviation value of the first prediction algorithm is greater than the preset deviation value, use the second prediction algorithm to determine the cylinder score at the certain point in the future.
  • the acquisition module 710 when the acquisition module 710 acquires the cylinder score data within a preset time range, it is specifically used to: acquire the action data of the cylinder within the preset time range, the action data including the action duration of the cylinder action, the number of cylinder actions and the cumulative action duration of the cylinder within the preset time range; determine the cylinder score of at least one time point within the preset time range as the cylinder score data based on the action duration of the cylinder action, the number of cylinder actions and the cumulative action duration of the cylinder, wherein one time point corresponds to one cylinder score.
  • the cylinder score at any time point within the preset time range is obtained by the following formula:
  • C is the cylinder score at any time point within the preset time range
  • X is the upper limit alarm value of the action duration of the cylinder action
  • Ttotal is the cumulative working time of the cylinder at any time point within the preset time range
  • Y is the number of cylinder actions at any time point within the preset time range.
  • the processing module 730 is specifically used to: determine the magnitude relationship between the cylinder score at a certain time point in the future and different preset scores; determine the degradation trend of the cylinder according to the magnitude relationship;
  • different preset scores include a normal score and a warning score.
  • the processing module 730 determines the degradation trend of the cylinder based on the size relationship, it is specifically used to: when the cylinder score at the certain point in the future is higher than the normal score, determine that the cylinder has no degradation trend; when the cylinder score at the certain point in the future is between the warning score and the normal score, determine that the cylinder has a slow degradation trend; when the cylinder score at the certain point in the future is lower than the warning score, determine that the cylinder has a rapid degradation trend.
  • an electronic device 800 includes a memory 820 and a processor 810 .
  • the memory 820 stores a computer program 821 .
  • the processor 820 executes the computer program 821 , the method for determining the cylinder degradation trend described in any of the above embodiments is implemented.
  • the processor 810 and the memory 820 transmit data via a data bus 830 .
  • the electronic device 800 may also include a network interface, and data exchange between the electronic device 800 and an external device may be realized through the network interface.
  • a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the cylinder degradation trend determination method described in any of the above embodiments is implemented.
  • An embodiment of the present application further provides a computer program product comprising instructions, which, when executed by a computer, causes the computer to execute the cylinder degradation trend determination method described in any of the above embodiments.
  • the processor in the embodiment of the present application can be an integrated circuit chip with signal processing capabilities.
  • each step of the above method implementation method can be completed by the hardware integrated logic circuit or software instruction in the processor.
  • the above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field Programmable Gate Array
  • the methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general processor can be a microprocessor or the processor can also be any conventional processor, etc.
  • the steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor are combined and performed.
  • the software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, and other mature storage media in the art.
  • the storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.
  • the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories.
  • the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (programmable ROM, PROM), an erasable programmable read-only memory (erasable PROM, EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory.
  • the volatile memory may be a random access memory (RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device implementation described above is only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present implementation scheme.
  • each functional unit in each embodiment of the present specification may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • the technical solution of this specification, or the part that contributes to the relevant technology or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Development Economics (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

本申请提供了一种气缸劣化趋势确定方法及装置、电子设备、存储介质。所述气缸劣化趋势确定方法包括：获取在预设时间范围内的气缸分值数据；根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；根据未来某一时间点的气缸分值，确定气缸的劣化趋势。通过在分值预测模型中，预测未来某一时间点的气缸分值，并根据未来某一时间点的气缸分值，确定气缸的劣化趋势，可以横向并列地比较多个气缸，可以根据时间的维度，确定同一时间的一个气缸的劣化趋势，也可以确定同一时间多个气缸是否出现劣化趋势，并且能够提前做好维护准备，且准备时间较充分，尽可能的减少对生产工作的影响，保证生产线上正常运行工作。

Description

气缸劣化趋势确定方法及装置、电子设备、存储介质 技术领域
本申请涉及气缸检测技术领域，具体涉及一种气缸劣化趋势确定方法及装置、电子设备、存储介质。
申请背景
自动化生产线上，应用的气缸种类数量繁多且分布范围广，气缸设备维护一般是通过实时预警来提示设备管理者，但是只能实时报警，无法确定设备的状态，所以不能提前对设备进行维护。
发明内容
为了克服相关技术的不足，本申请提供一种气缸劣化趋势确定方法及装置、电子设备、存储介质，以解决相关技术中的气缸设备发现故障后才进行维护的被动状态，且不能提前做好维护工作安排的问题。
第一方面，本申请一实施例提供了一种气缸劣化趋势确定方法，包括：获取在预设时间范围内的气缸分值数据，其中，所述气缸分值数据用于表征在所述预设时间范围内的气缸动作的得分；根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势。
在本申请一些实施例中，所述分值预测模型包括第一预测算法，所述第一预测算法通过以下公式实现：
Level函数:Lt＝α(yt-St-s)+(1-α)(Lt-1+bt-1)；
Trend函数:bt＝β(Lt-Lt-1)+(1-β)bt-1；
Seasonal函数:St＝γ(yt-Lt)+(1-γ)St-s；
Forecast函数:Ft+k＝Lt+kbt+St+k-s，
其中，α为第一平滑系数，β为第二平滑系数，γ为第三平滑系数，且α、β、γ的取值范围在0到1之间；s为季节性周期的长度，k为所述预设时间范围内的t时刻至所述未来某一时间点的长度，yt为所述t时刻的气缸分值，Lt为所述t时刻的平滑后的气缸分值，Lt-1为所述预设时间范围内的(t-1)时刻的平滑后的气缸分值，bt为所述t时刻的气缸分值的趋势值，bt-1为所述(t-1)时刻的气缸分值的趋势值，St为所述t时刻的气缸分值的季节性周期值，St-s为(t-s)时刻的气缸分值的季节性周期值，St+k-s为(t+k-s)时刻的气缸分值的季节性周期值；Ft+k为所述分值预测模型预测的(t+k)时刻的气缸分值，其中，所述(t+k)时刻为所述未来某一时间点。
在本申请一些实施例中，所述第一平滑系数、所述第二平滑系数以及所述第三平滑系数的取值相同。
在本申请一些实施例中，气缸劣化趋势确定方法还包括：将所述预设时间范围内的气缸分值数据划分为样本数据集和验证数据集，其中，所述样本数据集为所述预设时间范围内的第一时间段的气缸分值数据，所述验证数据集为所述预设时间范围内的第二时间段的气缸分值数据，所述第二时间段位于所述第一时间段之后；利用所述样本数据集，确定出所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值；根据确定的所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值，通过所述第一预测算法，对所述第二时间段的气缸分值进行预测，得到所述第二时间段的气缸分值预测值； 将所述气缸分值预测值与所述第二时间段的验证数据值中的气缸分值数据相比较，确定所述第一预测算法的预测偏差值。
在本申请一些实施例中，所述样本数据集为所述预设时间范围内的前X％的历史数据，所述验证数据集为所述预设时间范围内的后1-X％的历史数据，其中，X的取值为0到100之间。
在本申请一些实施例中，所述分值预测模型还包括第二预测算法，所述第二预测算法通过以下公式实现：
Y＝bX+a；
其中，b为回归方程的斜率，a为所述回归方程的截距，X为未来某一时间点的时间数值，Y为所述分值预测模型预测的所述未来某一时间点的气缸分值。
在本申请一些实施例中，所述回归方程的斜率的计算方式为：
所述回归方程的截距的计算方式为：
其中，xi为所述预设时间范围内的任一时间点的时间数值，yi为所述预设时间范围内的任一时间点的气缸分值，为所述预设时间范围内的所有时间点的时间数值的平均值，为所述预设时间范围内的所有气缸分值的平均值，n为所述预设时间范围的长度。
在本申请一些实施例中，所述根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值，包括：当所述第一预测算法的预测偏差值小于或者等于预设偏差值时，采用所述第一预测算法，确定所述未来某一时间点的气缸分值；当所述第一预测算法的预测偏差值大于预设偏差值时，采用所述第二预测算法，确定所述未来某一时间点的气缸分值。
在本申请一些实施例中，所述获取在预设时间范围内的气缸分值数据，包括：获取所述预设时间范围内的气缸的动作数据，所述动作数据包括所述设时间范围内的，气缸动作的动作时长、气缸动作的次数以及气缸累计动作时长；根据所述气缸动作的动作时长、所述气缸动作的次数以及所述气缸累计动作时长，确定所述预设时间范围内的至少一个时间点的气缸分值为所述气缸分值数据，其中，一个时间点对应一个气缸分值。
在本申请一些实施例中，所述预设时间范围内的任一时间点的气缸分值通过以下公式获得：
其中，C为所述预设时间范围内的任一时间点的气缸分值，X为气缸动作的动作时长的上限报警值，T总为所述预设时间范围内的任一时间点的气缸累计工作时长，Y为所述预设时间范围内的任一时间点的气缸动作的次数。
在本申请一些实施例中，所述根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势，包括：确定所述未来某一时间点的气缸分值与不同预设分值之间的大小关系；根据所述大小关系，确定所述气缸的劣化趋势；
在本申请一些实施例中，所述不同预设分值包括正常分值和警告分值，其中，根据所述大小关系，确定所述气缸的劣化趋势，包括：当所述未来某一时间点的气缸分值高于所述正常分值，确定所述气缸未出现劣化趋势；当所述未来某一时间点的气缸分值位于所述警告分值与所述正常分值之间时，确定所述气缸出现缓慢劣化趋势；当所述未来某一时间点的气缸分值低于所述警告分值，确定所述气缸出现急速劣化趋势。
第二方面，本申请一个实施例还提供了一种气缸劣化确定装置，包括：采集模块，用于获取在预设时间范围内的气缸分值数据，其中，所述气缸分值数据用于表征在所述预设时间 范围内的气缸动作的得分；预测模块，用于根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；处理模块，根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势。
第三方面，本申请一个实施例还提供了一种电子设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时，实现上述第一方面中任一实施例所述的气缸劣化趋势确定方法。
第四方面，本申请一个实施例还提供了一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时，实现上述第一方面中任一实施例所述的气缸劣化趋势确定方法。
本申请以上实施例所提供的气缸劣化趋势确定方法及装置、电子设备、存储介质具有以下有益效果：
以预设时间范围内的气缸分值数据为目标，在分值预测模型中预测未来某一时间点的气缸分值。可以沿用传统的采集装置进行数据采集，不需要另外增加新的采集设备，不会出现新设备的成本。通过未来某一时间点的气缸分值，确定气缸的劣化趋势，如未出现劣化趋势、出现劣化趋势，出现劣化趋势可分为急速劣化趋势、缓慢劣化趋势。对于未出现劣化趋势的气缸，则能够继续使用，不需要另外的维护；对于出现劣化趋势的气缸，判断是处于急速劣化趋势还是处于缓慢劣化趋势，优先对处于急速劣化趋势的气缸进行维护，再处理缓慢劣化趋势的气缸。这样可以横向并列地比较多个气缸，安排气缸维护的先后次序。还可以根据时间的维度，确定一个时间的一个气缸的劣化趋势，也可以同一时间确定多个气缸是否出现劣化趋势，优先对出现劣化趋势的哪一个气缸进行维护，使维护工作有序进行，并且能够提前做好维护准备，且准备时间较充分，尽可能的减少对生产工作的影响，保证生产线上正常运行工作。
附图简要说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图示出的结构获得其他的附图。
图1为本申请一个实施例提供的气缸劣化趋势确定系统的结构示意图。
图2为本申请一个实施例提供的气缸劣化趋势确定方法的流程示意图。
图3为本申请一个实施例提供的待检测气缸的结构示意图。
图4为本申请一个实施例提供的气缸分值随时间变化的示意图。
图5为本申请一个实施例提供的气缸劣化趋势确定方法的流程示意图。
图6为本申请一实施例提供的气缸劣化趋势确定方法的流程示意图。
图7为本申请一实施例提供的气缸劣化趋势确定装置的结构示意图。
图8为本申请一实施例提供的电子设备的结构示意图。
实施本发明的方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请的一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
需要说明，若本申请实施例中有涉及方向性指示(诸如上、下、左、右、前、后……)，则该方向性指示仅用于解释在某一特定姿态下各部件之间的相对位置关系、运动情况等，如果该特定姿态发生改变时，则该方向性指示也相应地随之改变。
另外，若本申请实施例中有涉及“第一”、“第二”等的描述，则该“第一”、“第二”等的描述仅用于描述目的，而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。另外，若全文中出现的“和/或”或者“及/或”，其含义包括三个并列的方案，以“A和/或B”为例，包括A方案、或B方案、或A和B同时满足的方案。另外，各个实施例之间的技术方案可以相互结合，但是必须是以本领域普通技术人员能够实现为基础，当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在，也不在本申请要求的保护范围之内。
在生产车间中，生产线上分布有数量繁多的各种气缸设备，气缸设备维护一般是通过设备动作时长的阈值，向设备管理者进行实时地报警，以提示设备管理者，但是只能进行实时地报警，无法确定设备的状态，例如，无法从大量的历史数据获得设备的运行状态是处于劣化状态还是良好的状态，也无法获得未来一周或者未来特定时间，设备可能发生的状态，因此，不能提前对设备进行维护。
在相关技术中，通过采集器采集生产线上的可编程逻辑控制器(Programmable Logic Controller，PLC)的数据，按照操作人员预先设置的规则，区分PLC的数据中的动作数据和状态数据，并监听PLC的数据中的动作数据。然而，仅将采集到的数据(例如气缸的动作数据)经过简单的处理，就去确定未来某一时间点气缸是否出现劣化，缺乏以数据为依据的判断规则。而本申请中，能够将气缸的动作数据进行处理转换，并根据设定的判断规则，预测气缸未来故障风险概率，从而及时发现隐患，及时排查问题气缸。
PLC所编写的程序为PLC控制程序，PLC是一种专门为在工业环境下应用而设计的数字运算操作电子系统，其采用一种可编程的存储器，在其内部存储执行逻辑运算、顺序控制、定时、计数和算术运算等操作的指令，通过数字式或模拟式的输入输出来控制各种类型的机械设备或生产过程。
图1是本申请一实施例提供的一种气缸劣化趋势确定系统的结构示意图。如图1所示，该系统包括：服务器110和多个气缸120、121……12n(n为大于0的整数)以及总线130。
多个气缸120、121……12n中的每个气缸可以为图3所述的气缸300。服务器110是一台服务器，或者由若干台服务器组成，或者是一个虚拟化平台，或者是一个云计算服务中心。多个气缸120、121……12n中的每个气缸与服务器110之间通过总线130相互传输数据。
下面结合本申请方案和图1所示的气缸劣化趋势确定系统，对气缸劣化趋势确定的过程进行详细说明。
服务器110通过总线130，从任一气缸12x处获取在预设时间范围内的气缸分值数据；服务器110根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；服务器110根据所述统计数值，确定所述第一设备的动作稳定性指数；服务器110根据所述未来某一时间点预测的气缸分值，确定气缸的劣化趋势。服务器110还可以向终端设备发送气缸的劣化趋势，以便于终端设备处的用户能够获知气缸处于何种状态。
应理解，如图1所示，服务器110可以独立于多个气缸120、121……12n而存在，二者通过总线130进行数据的传输，但这只是一个示例，以对本申请的应用场景进行举例说明，除此之外，多个气缸120、121……12n中的每个气缸上还可以配备有一个服务器，每个气缸通过各自的服务器执行本申请的气缸劣化趋势确定方法。
下面结合图2至6，对本申请实施例提供的气缸劣化趋势确定方法进行更为详细的举例说明。
请参见图2，本申请一实施例提供了一种气缸劣化趋势确定方法，该方法可以由图1所述的服务器110来执行，该方法可以包括如下步骤：
S210，获取在预设时间范围内的气缸分值数据，其中，所述气缸分值数据用于表征在所述预设时间范围内的气缸动作的得分；
S220，根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；
S230，根据预测的气缸分值，确定气缸的劣化趋势。
获取在预设时间范围内的气缸分值数据的过程包括：获取在所述预设时间范围内的气缸的动作数据，所述气缸的动作数据包括气缸动作的动作时长、气缸动作的次数、以及气缸动作的累计时长。
根据所述气缸动作的动作时长、所述气缸动作的次数以及所述气缸动作的累计时长，确定预设时间范围内的至少一个时间点的气缸分值为所述气缸分值数据。一个时间点对应一个气缸分值。需要说明的，预设时间范围内可以以年、月、周或日为单位，例如，预设时间范围为一年内的时间、一个月内的时间、一个星期内的时间或者一天的时间。根据实际需要，预设时间范围的取值可以进行调整。
请参见图3，气缸300包括缸体310以及设置在缸体310内部的活塞320。所述活塞320带动活塞杆330在所述缸体310内作往复运动，从而实现气缸300的功能。所述缸体310上设置有上止点传感器311和下止点传感器312。所述上止点传感器311设置在气缸的上止点上即B点的位置，当所述活塞320运动到气缸的上止点时，所述上止点传感器311将会接收到所述活塞320的位置信息，并将所述活塞320位于上止点的信息发送至控制器。所述下止点传感器312设置在气缸的下止点上即A的位置，当所述活塞320运动到气缸的下止点时，所述下止点传感器312将会接收到所述活塞320的位置信息，并将所述活塞320位于下止点的信息发送至控制器。因此，通过设置上止点传感器311和下止点传感器312，即可有效地检测出气缸300中活塞320的运动状况，从而为后续的气缸的工作状态提供分析数据。在本实施例中，所述上止点传感器311和所述下止点传感器312为霍尔传感器。所述活塞320中设置有磁铁。当所述活塞320运动至气缸的上止点(即，B点)或者下止点(即，A点)时，所述霍尔传感器即可检测出所述活塞320所处的位置。在本实施例中，气缸300的活塞320可在弹簧、重力或其他外力的作用下恢复到原来的位置。例如，气缸300的动作可以为打开或夹紧。活塞320从A点到B点为打开动作，采集所述下止点传感器312的信号到所述上止点传感器311的信号的时间为气缸300打开动作的动作时长；活塞320从B点到A点为夹紧，采集所述上止点传感器311的信号到所述下止点传感器312的信号的时间为气缸300夹紧动作的动作时长。
所述预设时间范围内的任一时间点的气缸分值根据以下公式计算得到：
其中，C为预设时间范围内的任一时间点的气缸分值，X为气缸动作的动作时长的上限报警值，Y为预设时间范围内的任一时间点的气缸动作的次数，T总为预设时间范围内的任一时间点的气缸累计工作时长。
也就是说，气缸动作的动作时长的上限报警值X和气缸动作的次数Y的乘积为气缸在正常工作状态下所累计的最大动作时长，其与实际的气缸累计工作时长T总的比例乘以100，得到预设时间范围内的任一时间点的气缸分值C。
需要说明的，气缸在预设时间范围内的任一时间点可以包括多种气缸动作，例如，n种气缸动作，n种气缸动作的动作时长的上限报警值为X1、X2···Xn，n种气缸动作在预设时间范围内的任一时间点对应的气缸分值为C1、C2···Cn，确定C1、C2···Cn的平均值为所述气缸300在预设时间范围内的任一时间点的气缸分值C。
利用预设时间范围内的气缸动作的动作时长，确定预设时间范围内的气缸分值，以预设时间范围内的气缸分值为目标，在分值预测模型中预测未来某一时间点的气缸分值。未来某一时间点可以是一个月，一个星期，一天或者一个小时，相应地，未来某一时间点的气缸分值就是指未来一个月的气缸分值，未来一个星期的气缸分值，未来一天的气缸分值或者未来一个小时的气缸分值。
通过判断规则和未来某一时间点的气缸分值，确定气缸劣化趋势，如未出现劣化趋势、出现劣化趋势，出现劣化趋势可分为急速劣化趋势、缓慢劣化趋势。对于未出现劣化趋势的气缸，则能够继续使用，不需要另外的维护；对于出现劣化趋势的气缸，判断气缸劣化趋势是处于急速劣化趋势还是处于缓慢劣化趋势，优先对处于急速劣化趋势的气缸进行维护，再处理缓慢劣化趋势的气缸。应理解，该判断规则可以是在分值预测模型中预先设定的。
例如，该判断规则为判断未来某一时间点的气缸分值分别与正常分值以及警告分值之间的大小关系，当判断未来某一时间点的气缸分值高于正常分值，则可以确定气缸未出现劣化趋势；当未来某一时间点的气缸分值低于正常分值，则可以确定气缸出现劣化趋势。尤其是当未来某一时间点的气缸分值位于警告分值以及正常分值之间时，则可以确定气缸出现缓慢劣化趋势；当未来某一时间点的气缸分值低于警告分值，则可以确定气缸出现急速劣化趋势。
再例如，通过单次执行上述步骤S210至S230，可以获得任一时间点的气缸分值，例如，未来7天的气缸分值，那么通过重复多次执行上述步骤S210至S230，可以获得未来多个时间点的气缸分值，例如，未来5天的气缸分值，未来6天的气缸分值，未来7天的气缸分值，此时，该判断规则为判断未来多个时间点的气缸分值与历史气缸分值之间的大小关系，当未来多个时间点的气缸分值持续小于历史气缸分值，则可以确定气缸出现劣化趋势。
通过本申请实施例所述的气缸劣势确定方法，可以横向并列地比较多个气缸，即，可以根据时间的维度，一个时间只确定一个气缸的劣化趋势，也可以同一时间确定多个气缸是否出现劣化趋势，以确定优先对出现劣化趋势的哪一个气缸进行维护，使维护工作有序进行，并且能够提前做好维护准备，且准备时间较充分，尽可能的减少对生产工作的影响，保证生产线上正常运行工作。避免出现由于其中一个气缸出现劣化而生产残次产品，影响生产的质量，导致造成物资的浪费，降低生产效益。
在一个实施例中，所述分值预测模型包括第一预测算法，所述第一预测算法包括以下公式：
Level函数:Lt＝α(yt-St-s)+(1-α)(Lt-1+bt-1)；
Trend函数:bt＝β(Lt-Lt-1)+(1-β)bt-1；
Seasonal函数:St＝γ(yt-Lt)+(1-γ)St-s；
Forecast函数:Ft+k＝Lt+kbt+St+k-s，
其中，α为第一平滑系数，β为第二平滑系数，γ为第三平滑系数，且α、β、γ的取值范围在0到1之间；s为季节性周期的长度，k为所述预设时间范围内的t时刻至所述未来某一时间点的长度，
yt为t时刻的气缸分值；Lt为t时刻的平滑后的气缸分值，Lt-1为所述预设时间范围内的(t-1)时刻的平滑后的气缸分值，bt为t时刻的气缸分值的趋势值，bt-1为所述(t-1)时刻的气缸分值的趋势值，St为t时刻的气缸分值的季节性周期值，St-s为(t-s)时刻的气缸分值的季节性周期值，St+k-s为(t+k-s)时刻的气缸分值的季节性周期值；Ft+k为分值预测模型预测的(t+k)时刻的气缸分值，其中，所述(t+k)时刻为所述未来某一时间点。
Lt＝α(yt-St-s)+(1-α)(Lt-1+bt-1)为level函数，利用level函数，能够获得季节性调整后的观测值(即，yt-St-s)和上一期不含季节性的预测值(即，Lt-1+bt-1)的加权平均值。
bt＝β(Lt-Lt-1)+(1-β)bt-1为trend函数，利用trend函数，能够确定气缸分值的趋势，从足够多数量的气缸分值中，预测未来某一时间点的气缸分值的趋势。
St＝γ(yt-Lt)+(1-γ)St-s为seasonal函数，利用seasonal函数，能够获得当前的季节性系数(即，St)和上一个周期同期的季节性系数(即，St-s)的滑动平均值，其结合了气缸的同一动作或同一产品的周期的变化。可以理解的，为了简化以便于理解，将第一预测算法中的初始值设为L0＝y0，b0＝y1-y0，基于初始值，预测未来某一时间点的气缸分值。所述第一预测算法通过上述公式，可以实现累加方式的预测。根据需要，所述第一预测算法还可以采用累乘方式进行预测。一般来说，在季节性变化基本固定的时间序列上，采用加性预测的方法 (即，累加方式)所得出的预测结果较为准确。在季节性变化和当前水平成比例变化的时间序列上，采用乘性预测的方法(即，累乘方式)所得出的预测结果较为准确。
实际上，将足够多数量的气缸分值代入第一预测算法中计算，通过level函数得到足够多数量的气缸分值的平滑点，该平滑点对应的值称为气缸平滑值；将平滑点代入trend函数中，得到足够多数量的气缸分值的平滑线，换句话说，得到足够多数量的气缸平滑值的趋势值；再将平滑线代入seasonal函数，并结合实际的周期变化，得到贴合实际生产的且变化的未来某一时间点的气缸分值，换句话说，将气缸分值与平滑点的差异值和上一周期同期的差异平均值代入seasonal函数中，得到足够多数量的季节性系数，将最接近的预测未来周期的平滑点、趋势值和季节性系数代入forecast函数，得到贴合实际生产的且变化的未来某一时间点的气缸分值。
可以理解的，足够多数量的气缸分值为至少7个时间点对应的气缸分值，通过至少7个时间点对应的气缸分值，才能得出未来某一时间点的气缸分值。请参见图4，例如，未来某一时间点为一天，则在确定了前7天的气缸分值，才能得到第8天的气缸分值(即，未来某一时间点的气缸分值)。
在一个实施例中，所述第一平滑系数α、所述第二平滑系数β以及所述第三平滑系数γ的取值相同。第一平滑系数α、所述第二平滑系数β、第三平滑系数γ的取值范围在0到1之间，可以在0到1之间的数值中选取多种数值的组合，预测未来某一时间点的气缸分值。经过验证，将三个平滑系数得数值取值相同时，能够更准确的预测未来特定时间节点的气缸分值。
季节性周期值S可以理解为是一个约束值，用于约束分值预测模型预测的未来时间点的气缸分值呈周期变化，当分值预测模型预测的未来时间点的气缸分值未呈周期变化时，S的取值接近于0，当分值预测模型预测的未来时间点的气缸分值呈周期变化时，利用seasonal函数，计算对应时刻的S的取值。季节性周期的长度s是指一个周期的长度，例如，预测周期为天，季节性周期的长度就是指7天，预测周期为周，季节性周期的长度就是指4周，预测周期为月，季节性周期的长度就是指12个月。
应理解，按照预测周期，在预设时间范围内可以包括多个时刻，t时刻就是指预设时间范围的截至时间节点，在预设时间范围内，t时刻之前还包括多个时刻，例如，t-1时刻，t-2时刻等等。将初始值L0和b0代入上述公式，并迭代执行上述公式，能够得到上述公式中所提及的t时刻的各个值。
如上所述，预设时间范围内的气缸分值数据可以包括至少一个时间点的气缸分值，那么t时刻的气缸分值就是指预设时间范围内的截至时间点的气缸分值，例如，预设时间范围为14天，t时刻为第14天，k等于7天，那么t时刻的气缸分值就是指第14天的气缸分值，Ft+k就是指分值预测模型预测的未来7天的气缸分值。
在一个实施例中，该方法还包括：判断所述第一预测算法的准确度。在一示例中，请一并参见图5，判断所述第一预测算法的准确度，包括以下步骤：
S510，将所述预设时间范围内的气缸分值数据划分为样本数据集和验证数据集，其中，所述样本数据集为所述预设时间范围内的第一时间段的气缸分值数据，所述验证数据集为所述预设时间范围内的第二时间段的气缸分值数据，所述第二时间段位于所述第一时间段之后；
S520，利用所述样本数据集，确定出所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值；
S530，根据确定的所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值，通过所述第一预测算法，对所述第二时间段的气缸分值进行预测，得到所述第二时间段的气缸分值预测值；
S540，将所述气缸分值预测值与所述第二时间段的验证数据值中的气缸分值数据相比较，确定所述第一预测算法的预测偏差值。
将预设时间范围内的气缸分值数据划分为样本数据集和验证数据集，其中，所述样本数据集为所述预设时间范围内的第一时间段的历史数据(即，气缸分值数据)，所述验证数据集 为所述预设时间范围内的第二时间段的历史数据(即，气缸分值数据)，所述第二时间段位于所述第一时间段之后。例如，所述预设时间范围内的前90％的历史数据作为样本数据集，所述预设时间范围内的后10％的历史数据作为验证数据集。例如，采用2021年全年数据，则利用样本数据集(即，前328天的数据)进行气缸分值的预测，利用验证数据集(即，后37天的数据)，验证算法模型是否准确。此时，第一时间段即为前328天，第二时间段即为后37天。根据需要，第一时间段和第二时间段的取值可以根据产线的实际情况确定。
通过所述样本数据集，可以得到所述第一平滑系数α、所述第二平滑系数β、所述第三平滑系数γ的数值。在一个实施例中，可以利用所述样本数据集进行模型训练，通过深度学习的方法，得出所述第一平滑系数α、所述第二平滑系数β、所述第三平滑系数γ的数值。
根据所述第一平滑系数α、所述第二平滑系数β、所述第三平滑系数γ的数值，利用所述第一预测算法，对所述第二时间段的气缸分值进行预测，得到所述第二时间段的气缸分值预测值；将所述气缸分值预测值与所述第二时间段的验证数据值中的气缸分值数据相比较，得出所述第一预测算法的预测偏差值。
也就是说，为了验证第一预测算法的准确性，将通过第一预测算法得到的气缸分值预测值与实际的气缸分值进行比较，二者的差距为预测偏差值。采用对称平均绝对百分比误差(Symmetric Mean Absolute Percentage Error，SMAPE)的方式，计算气缸分值预测值和实际的气缸分值之间的差距，即，第一预测算法的预测偏差值。SMAPE公式为：
其中，η为预测偏差值，R为气缸分值预测值，F为实际的气缸分值。一般来说，预测偏差值的范围为0％至200％，越接近0％，说明预测得越准确。
可以理解的，第一步，将预设时间范围内的气缸分值数据划分为样本数据集和验证数据集；第二步，通过将已知历史数据(即，样本数据集)代入第一预测算法中，得到平滑系数的取值；第三步，将平滑系数代入第一预测算法中，以完善第一预测算法；第四步，将样本数据集中的气缸分值代入完善的第一预测算法中，得到气缸分值预测值；第五步，通过重复执行第四步，得到足够多的气缸分值预测分值；第六步，通过SMAPE方式，将每个气缸分值预测分值与验证数据集中对应的气缸分值进行比较，得到预测偏差值。
换句话说，将样本数据集的气缸分值代入第一预测算法中，得到第一平滑系数α、所述第二平滑系数β、第三平滑系数γ的取值，以完善所述第一预测算法。再将样本数据集的气缸分值代入完善后的所述第一预测算法，得到一组气缸分值预测值，将每个气缸分值预测值与验证数据集的对应的气缸分值进行比较。
在一个实施例中，所述样本数据集为所述预设时间范围内的前X％的历史数据，所述验证数据集为所述预设时间范围内的后1-X％的历史数据。其中，X的取值为0到100之间。在本实施例中，X的取值可以根据实际需要确定。例如，X可以取90，此时，所述样本数据集为所述预设时间范围内的前90％的历史数据，所述验证数据集为所述预设时间范围内的后10％的历史数据。又例如，X可以取70，此时，所述样本数据集为所述预设时间范围内的前70％的历史数据，所述验证数据集为所述预设时间范围内的后30％的历史数据。
在一个实施例中，所述分值预测模型还包括第二预测算法，当所述第一预测算法的预测偏差值小于或者等于预设偏差值时，采用所述第一预测算法对未来某一时间点的气缸分值进行预测，当所述第一预测算法的预测偏差值大于预设偏差值时，采用所述第二预测算法对未来某一时间点的气缸分值进行预测。
请一并参考图6，该方法包括如下步骤：
S610，判断所述第一预测算法的预测偏差值是否小于或者等于预设偏差值；
在所述第一预测算法的预测偏差值小于或者等于预设偏差值时，执行步骤S620，采用所述第一预测算法，并执行步骤S640，对未来某一时间点的气缸分值进行预测；
在所述第一预测算法的预测偏差值大于预设偏差值时，执行步骤S630，采用所述第二预测算法，并执行步骤S640，对未来某一时间点的气缸分值进行预测。
需要说明的，当预测偏差值超出了预设偏差值，则证明通过第一预测算法预测的气缸分值的误差过大，判断预测无效，第一预测算法不适用，转换第二预测算法进行预测。例如，预设偏差值为20％，而预测偏差值为23％，那么第一预测算法无效，采用第二预测算法进行预测。再例如，预设偏差值为20％，而预测偏差值为10％，那么确定第一预测算法准确，继续采用第一预测算法进行预测。即，通过第一预测算法和第二预测算法之间的切换，从而使所述分值预测模型的预测结果更加准确。
在一个实施例中，所述第二预测算法包括以下公式：
Y＝bX+a；
其中，b为回归方程的斜率，a为回归方程的截距，X为未来某一时间点的时间数值，Y为所述分值预测模型预测的所述未来某一时间点的气缸分值。
在确定好回归方程的斜率b和回归方程的截距a后，未来任意时间节点的气缸分值Y都可以通过时间节点的时间数值X求出。在本实施例中，预测数据样本为查询时间范围内的所有历史数据。例如，当利用2022年7月的数据，计算回归方程的斜率和截距时，利用7月份的天数X和每一天的气缸分值Y，计算斜率b和截距a，从而通过斜率b和截距a，得到具体的线性回归方程。
在一个实施例中，所述第二预测算法中的回归方程的斜率和回归方程的截距的计算方式包括：

其中，xi为预设时间范围内的任一时间点的时间数值，yi为预设时间范围内的任一时间点的气缸分值，为预设时间范围内的所有时间点的时间数值的平均值，为预设时间范围内的所有气缸分值的平均值，n为预设时间范围的长度。
在本实施例中，通过预设时间范围内的任一时间点的时间数值xi，预设时间范围内的任一时间点的气缸分值yi，预设时间范围内的所有时间点的时间数值的平均值以及预设时间范围内的所有气缸分值的平均值计算回归方程的斜率b。在得到斜率b之后，再根据预设时间范围内的所有时间点的时间数值的平均值和预设时间范围内的所有气缸分值的平均值计算回归方程的截距a。例如，如果利用2022年7月的气缸分值数据，计算回归方程的斜率和截距的话，则：
其中，
上文结合图1至6，详细描述了本申请的装置实施例，下面结合图7和8，详细描述本申请的方法实施例。应理解，方法实施例的描述与装置实施例的描述相互对应，因此，未详细描述的部分可以参见前面装置实施例。
请一并参见图7，一种气缸劣化确定装置700，包括：
采集模块710，用于获取在预设时间范围内的气缸分值数据，其中，所述气缸分值数据用于表征在所述预设时间范围内的气缸动作的得分；
预测模块720，用于根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；
处理模块730，根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势。
以预设时间范围内的气缸分值数据为目标，在分值预测模型中预测未来某一时间点的气缸分值。可以沿用传统的采集装置进行数据采集，不需要另外增加新的采集设备，不会出现新设备的成本。通过未来某一时间点的气缸分值，确定气缸的劣化趋势，如未出现劣化趋势、出现劣化趋势，出现劣化趋势可分为急速劣化趋势、缓慢劣化趋势。对于未出现劣化趋势的气缸，则能够继续使用，不需要另外的维护；对于出现劣化趋势的气缸，判断是处于急速劣化趋势还是处于缓慢劣化趋势，优先对处于急速劣化趋势的气缸进行维护，再处理缓慢劣化趋势的气缸。这样可以横向并列地比较多个气缸，安排气缸维护的先后次序。还可以根据时间的维度，确定一个时间的一个气缸的劣化趋势，也可以同一时间确定多个气缸是否出现劣化趋势，优先对出现劣化趋势的哪一个气缸进行维护，使维护工作有序进行，并且能够提前做好维护准备，且准备时间较充分，尽可能的减少对生产工作的影响，保证生产线上正常运行工作，避免出现由于其中一个气缸出现劣化，而生产残次产品，影响生产的质量，导致造成物资的浪费，降低生产效益。
在本申请的一些实施例中，所述分值预测模型包括第一预测算法，所述第一预测算法通过以下公式实现：
Level函数:Lt＝α(yt-St-s)+(1-α)(Lt-1+bt-1)；
Trend函数:bt＝β(Lt-Lt-1)+(1-β)bt-1；
Seasonal函数:St＝γ(yt-Lt)+(1-γ)St-s；
Forecast函数:Ft+k＝Lt+kbt+St+k-s，
其中，α为第一平滑系数，β为第二平滑系数，γ为第三平滑系数，且α、β、γ的取值范围在0到1之间；s为季节性周期的长度，k为所述预设时间范围内的t时刻至所述未来某一时间点的长度，yt为所述t时刻的气缸分值，Lt为所述t时刻的平滑后的气缸分值，Lt-1为所述预设时间范围内的(t-1)时刻的平滑后的气缸分值，bt为所述t时刻的气缸分值的趋势值，bt-1为所述(t-1)时刻的气缸分值的趋势值，St为所述t时刻的气缸分值的季节性周期值，St-s为(t-s)时刻的气缸分值的季节性周期值，St+k-s为(t+k-s)时刻的气缸分值的季节性周期值；Ft+k为所述分值预测模型预测的(t+k)时刻的气缸分值，其中，所述(t+k)时刻为所述未来某一时间点。
在本申请的一些实施例中，所述第一平滑系数、所述第二平滑系数以及所述第三平滑系数的取值相同。
在本申请的一些实施例中，预测模块720还用于：将所述预设时间范围内的气缸分值数据划分为样本数据集和验证数据集，其中，所述样本数据集为所述预设时间范围内的第一时间段的气缸分值数据，所述验证数据集为所述预设时间范围内的第二时间段的气缸分值数据，所述第二时间段位于所述第一时间段之后；利用所述样本数据集，确定出所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值；根据确定的所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值，通过所述第一预测算法，对所述第二时间段的气缸分值进行预测，得到所述第二时间段的气缸分值预测值；将所述气缸分值预测值与所述第二时间段的验证数据值中的气缸分值数据相比较，确定所述第一预测算法的预测偏差值。
在本申请的一些实施例中，所述样本数据集为所述预设时间范围内的前X％的历史数据，所述验证数据集为所述预设时间范围内的后1-X％的历史数据，其中，X的取值为0到100之间。
在本申请的一些实施例中，所述分值预测模型还包括第二预测算法，所述第二预测算法通过以下公式实现：
Y＝bX+a；
其中，b为回归方程的斜率，a为所述回归方程的截距，X为未来某一时间点的时间数值，Y为所述分值预测模型预测的所述未来某一时间点的气缸分值。
在本申请的一些实施例中，所述回归方程的斜率的计算方式为：
所述回归方程的截距的计算方式为：
其中，xi为所述预设时间范围内的任一时间点的时间数值，yi为所述预设时间范围内的任一时间点的气缸分值，为所述预设时间范围内的所有时间点的时间数值的平均值，为所述预设时间范围内的所有气缸分值的平均值，n为所述预设时间范围的长度。
在本申请的一些实施例中，预测模块720在所述根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值，具体用于：当所述第一预测算法的预测偏差值小于或者等于预设偏差值时，采用所述第一预测算法，确定所述未来某一时间点的气缸分值；当所述第一预测算法的预测偏差值大于预设偏差值时，采用所述第二预测算法，确定所述未来某一时间点的气缸分值。
在本申请的一些实施例中，采集模块710在获取在预设时间范围内的气缸分值数据时，具体用于：获取所述预设时间范围内的气缸的动作数据，所述动作数据包括所述设时间范围内的，气缸动作的动作时长、气缸动作的次数以及气缸累计动作时长；根据所述气缸动作的动作时长、所述气缸动作的次数以及所述气缸累计动作时长，确定所述预设时间范围内的至少一个时间点的气缸分值为所述气缸分值数据，其中，一个时间点对应一个气缸分值。
在本申请的一些实施例中，所述预设时间范围内的任一时间点的气缸分值通过以下公式获得：
其中，C为所述预设时间范围内的任一时间点的气缸分值，X为气缸动作的动作时长的上限报警值，T总为所述预设时间范围内的任一时间点的气缸累计工作时长，Y为所述预设时间范围内的任一时间点的气缸动作的次数。
在本申请的一些实施例中，处理模块730具体用于：确定所述未来某一时间点的气缸分值与不同预设分值之间的大小关系；根据所述大小关系，确定所述气缸的劣化趋势；
在本申请的一些实施例中，不同预设分值包括正常分值和警告分值，处理模块730在根据所述大小关系，确定所述气缸的劣化趋势时，具体用于：当所述未来某一时间点的气缸分值高于所述正常分值，确定所述气缸未出现劣化趋势；当所述未来某一时间点的气缸分值位于所述警告分值与所述正常分值之间时，确定所述气缸出现缓慢劣化趋势；当所述未来某一时间点的气缸分值低于所述警告分值，确定所述气缸出现急速劣化趋势。
请参见图8，一种电子设备800，包括存储器820和处理器810，所述存储器820存储有计算机程序821，所述处理器820执行所述计算机程序821时，实现上述任一实施例所述的气缸劣化趋势确定方法。在本实施例中，处理器810及存储器820通过数据总线830传输数据。
在本实施例中，该电子设备800还可以包括网络接口，电子设备800与外部设备的数据交换可以通过该网络接口实现。一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述任一实施例所述的气缸劣化趋势确定方法。
本申请一实施例还提供一种包含指令的计算机程序产品，该指令被计算机执行时，使得计算机执行上述任一实施例所述的气缸劣化趋势确定方法。
可以理解，本申请实施例的处理器可以是一种集成电路芯片，具有信号的处理能力。在 实现过程中，上述方法实施方式的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器可以是通用处理器、数字信号处理器(Digital SignalProcessor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。
可以理解，本申请实施例中的存储器可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(ROM)、可编程只读存储器(programmable ROM，PROM)、可擦除可编程只读存储器(erasable PROM，EPROM)、电可擦除可编程只读存储器(EEPROM)或闪存。易失性存储器可以是随机存取存储器(RAM)。应注意，本文描述的系统和方法的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
本领域普通技术人员可以意识到，结合本文中所公开的实施方式描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本说明书的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施方式中的对应过程，在此不再赘述。
在本说明书所提供的几个实施方式中，应所述理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施方式仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施方式方案的目的。
另外，在本说明书各个实施方式中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本说明书的技术方案本质上或者说对相关技术做出贡献的部分或者所述技术方案的部分可以以软件产品的形式体现出来，所述计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM)、随机存取存储器(RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是在本申请的申请构思下，利用本申请说明书及附图内容所作的等效结构变换，或直接/间接运用在其他相关的技术领域均包括在本申请的专利保护范围内。

Claims (15)

  1. 一种气缸劣化趋势确定方法，其特征在于，包括：
    获取在预设时间范围内的气缸分值数据，其中，所述气缸分值数据用于表征在所述预设时间范围内的气缸动作的得分；
    根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；
    根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势。
  2. 如权利要求1所述的气缸劣化趋势确定方法，其特征在于，所述分值预测模型包括第一预测算法，所述第一预测算法通过以下公式实现：
    Level函数:Lt＝α(yt-St-s)+(1-α)(Lt-1+bt-1)；
    Trend函数:bt＝β(Lt-Lt-1)+(1-β)bt-1；
    Seasonal函数:St＝Υ(yt-Lt)+(1-Υ)St-s；
    Forecast函数:Ft+k＝Lt+kbt+St+k-s，
    其中，α为第一平滑系数，β为第二平滑系数，Υ为第三平滑系数，且α、β、Υ的取值范围在0到1之间；s为季节性周期的长度，k为所述预设时间范围内的t时刻至所述未来某一时间点的长度，
    yt为所述t时刻的气缸分值，Lt为所述t时刻的平滑后的气缸分值，Lt-1为所述预设时间范围内的(t-1)时刻的平滑后的气缸分值，bt为所述t时刻的气缸分值的趋势值，bt-1为所述(t-1)时刻的气缸分值的趋势值，St为所述t时刻的气缸分值的季节性周期值，St-s为(t-s)时刻的气缸分值的季节性周期值，St+k-s为(t+k-s)时刻的气缸分值的季节性周期值；Ft+k为所述分值预测模型预测的(t+k)时刻的气缸分值，其中，所述(t+k)时刻为所述未来某一时间点。
  3. 如权利要求2所述的气缸劣化趋势确定方法，其特征在于，所述第一平滑系数、所述第二平滑系数以及所述第三平滑系数的取值相同。
  4. 如权利要求2或3所述的气缸劣化趋势确定方法，其特征在于，还包括：
    将所述预设时间范围内的气缸分值数据划分为样本数据集和验证数据集，其中，所述样本数据集为所述预设时间范围内的第一时间段的气缸分值数据，所述验证数据集为所述预设时间范围内的第二时间段的气缸分值数据，所述第二时间段位于所述第一时间段之后；
    利用所述样本数据集，确定出所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值；
    根据确定的所述第一平滑系数的数值、所述第二平滑系数的数值、所述第三平滑系数的数值，通过所述第一预测算法，对所述第二时间段的气缸分值进行预测，得到所述第二时间段的气缸分值预测值；
    将所述气缸分值预测值与所述第二时间段的验证数据值中的气缸分值数据相比较，确定所述第一预测算法的预测偏差值。
  5. 如权利要求4所述的气缸劣化趋势确定方法，其特征在于，所述样本数据集为所述预设时间范围内的前X％的历史数据，所述验证数据集为所述预设时间范围内的后1-X％的历史数据，其中，X的取值为0到100之间。
  6. 如权利要求4或5所述的气缸劣化趋势确定方法，其特征在于，所述分值预测模型还包括第二预测算法，所述第二预测算法通过以下公式实现：
    Y＝bX+a；
    其中，b为回归方程的斜率，a为所述回归方程的截距，X为未来某一时间点的时间数值，Y为所述分值预测模型预测的所述未来某一时间点的气缸分值。
  7. 如权利要求6所述的气缸劣化趋势确定方法，其特征在于，所述回归方程的斜率的计算方式为：
    所述回归方程的截距的计算方式为：
    其中，xi为所述预设时间范围内的任一时间点的时间数值，yi为所述预设时间范围内的任一时间点的气缸分值，为所述预设时间范围内的所有时间点的时间数值的平均值，为所述预设时间范围内的所有气缸分值的平均值，n为所述预设时间范围的长度。
  8. 如权利要求6或7所述的气缸劣化趋势确定方法，其特征在于，所述根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值，包括：
    当所述第一预测算法的预测偏差值小于或者等于预设偏差值时，采用所述第一预测算法，确定所述未来某一时间点的气缸分值；
    当所述第一预测算法的预测偏差值大于预设偏差值时，采用所述第二预测算法，确定所述未来某一时间点的气缸分值。
  9. 如权利要求1至8中任一项所述的气缸劣化趋势确定方法，其特征在于，所述获取在预设时间范围内的气缸分值数据，包括：
    获取所述预设时间范围内的气缸的动作数据，所述动作数据包括所述设时间范围内的，气缸动作的动作时长、气缸动作的次数以及气缸累计动作时长；
    根据所述气缸动作的动作时长、所述气缸动作的次数以及所述气缸累计动作时长，确定所述预设时间范围内的至少一个时间点的气缸分值为所述气缸分值数据，其中，一个时间点对应一个气缸分值。
  10. 如权利要求9所述的气缸劣化趋势确定方法，其特征在于，所述预设时间范围内的任一时间点的气缸分值通过以下公式获得：
    其中，C为所述预设时间范围内的任一时间点的气缸分值，X为气缸动作的动作时长的上限报警值，T总为所述预设时间范围内的任一时间点的气缸累计工作时长，Y为所述预设时间范围内的任一时间点的气缸动作的次数。
  11. 如权利要求1至10中任一项所述的气缸劣化趋势确定方法，其特征在于，所述根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势，包括：
    确定所述未来某一时间点的气缸分值与不同预设分值之间的大小关系；
    根据所述大小关系，确定所述气缸的劣化趋势；
  12. 如权利要求11所述的气缸劣化趋势确定方法，其特征在于，所述不同预设分值包括正常分值和警告分值，其中，根据所述大小关系，确定所述气缸的劣化趋势，包括：
    当所述未来某一时间点的气缸分值高于所述正常分值，确定所述气缸未出现劣化趋势；
    当所述未来某一时间点的气缸分值位于所述警告分值与所述正常分值之间时，确定所述气缸出现缓慢劣化趋势；
    当所述未来某一时间点的气缸分值低于所述警告分值，确定所述气缸出现急速劣化趋势。
  13. 一种气缸劣化确定装置，其特征在于，包括：
    采集模块，用于获取在预设时间范围内的气缸分值数据，其中，所述气缸分值数据用于表征在所述预设时间范围内的气缸动作的得分；
    预测模块，用于根据所述气缸分值数据以及分值预测模型，预测未来某一时间点的气缸分值；
    处理模块，根据所述未来某一时间点的气缸分值，确定气缸的劣化趋势。
  14. 一种电子设备，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时，实现权利要求1至12中任一项所述的气缸劣化趋势确定方法。
  15. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时，实现权利要求1至12中任一项所述的气缸劣化趋势确定方法。
PCT/CN2023/122730 2022-09-30 2023-09-28 气缸劣化趋势确定方法及装置、电子设备、存储介质 WO2024067824A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211206947.9A CN115511188A (zh) 2022-09-30 2022-09-30 一种气缸劣化分析方法及系统、设备、存储介质
CN202211206947.9 2022-09-30

Publications (1)

Publication Number Publication Date
WO2024067824A1 true WO2024067824A1 (zh) 2024-04-04

Family

ID=84508784

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/122730 WO2024067824A1 (zh) 2022-09-30 2023-09-28 气缸劣化趋势确定方法及装置、电子设备、存储介质

Country Status (2)

Country Link
CN (1) CN115511188A (zh)
WO (1) WO2024067824A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115511188A (zh) * 2022-09-30 2022-12-23 广州明珞装备股份有限公司 一种气缸劣化分析方法及系统、设备、存储介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6006154A (en) * 1998-03-02 1999-12-21 Cummins Engine Company, Inc. System and method for cylinder power imbalance prognostics and diagnostics
CN104295373A (zh) * 2014-10-08 2015-01-21 哈尔滨工程大学 基于三点模型的柴油机故障预测方法
CN104794283A (zh) * 2015-04-22 2015-07-22 哈尔滨工程大学 基于离群特征分析的柴油机故障灰预测方法
CN108386273A (zh) * 2018-02-09 2018-08-10 安徽江淮汽车集团股份有限公司 一种汽油机爆震发生时刻的预测方法及装置
US20180281805A1 (en) * 2017-04-04 2018-10-04 Ford Global Technologies, Llc Systems and methods for active engine mount diagnostics
CN113806346A (zh) * 2021-08-25 2021-12-17 浙江浙能台州第二发电有限责任公司 一种基于大数据分析的汽轮机劣化趋势测量方法及终端机
CN115511188A (zh) * 2022-09-30 2022-12-23 广州明珞装备股份有限公司 一种气缸劣化分析方法及系统、设备、存储介质

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6006154A (en) * 1998-03-02 1999-12-21 Cummins Engine Company, Inc. System and method for cylinder power imbalance prognostics and diagnostics
CN104295373A (zh) * 2014-10-08 2015-01-21 哈尔滨工程大学 基于三点模型的柴油机故障预测方法
CN104794283A (zh) * 2015-04-22 2015-07-22 哈尔滨工程大学 基于离群特征分析的柴油机故障灰预测方法
US20180281805A1 (en) * 2017-04-04 2018-10-04 Ford Global Technologies, Llc Systems and methods for active engine mount diagnostics
CN108386273A (zh) * 2018-02-09 2018-08-10 安徽江淮汽车集团股份有限公司 一种汽油机爆震发生时刻的预测方法及装置
CN113806346A (zh) * 2021-08-25 2021-12-17 浙江浙能台州第二发电有限责任公司 一种基于大数据分析的汽轮机劣化趋势测量方法及终端机
CN115511188A (zh) * 2022-09-30 2022-12-23 广州明珞装备股份有限公司 一种气缸劣化分析方法及系统、设备、存储介质

Also Published As

Publication number Publication date
CN115511188A (zh) 2022-12-23

Similar Documents

Publication Publication Date Title
CN104809051B (zh) 用于预测计算机应用中的异常和故障的方法和装置
WO2024067824A1 (zh) 气缸劣化趋势确定方法及装置、电子设备、存储介质
CN102763048B (zh) 在虚拟量测中使用适应性预测算法及决定何时使用适应性预测算法的方法及设备
CN104350471B (zh) 在处理环境中实时地检测异常的方法和系统
WO2023284887A1 (zh) 长期降水预测模型构建方法、长期降水预测方法及装置
JP5962190B2 (ja) 短期電力負荷を予測する方法及び装置
CN108038040A (zh) 计算机集群性能指标检测方法、电子设备及存储介质
WO2021004324A1 (zh) 资源数据的处理方法、装置、计算机设备和存储介质
CN105956734A (zh) 动态设置it设备的性能的指标阈值的方法及系统
WO2021038960A1 (ja) 部分放電判定装置及び方法
WO2022252454A1 (zh) 异常数据检测方法、装置、计算机设备和可读存储介质
CN114240243A (zh) 一种基于动态系统辨识的精馏塔产品质量预测方法及装置
CN107591001B (zh) 基于在线标定的快速路交通流数据填补方法及系统
US20220253214A1 (en) Data processing method, apparatus, device, and readable storage medium
CN103336433A (zh) 基于反步法的混合自适应预测控制系统及其预测控制方法
CN108664603A (zh) 一种修复时序数据的异常聚合值的方法及装置
CN114155072B (zh) 基于大数据分析的财务预测模型构建方法及系统
WO2024065988A1 (zh) 设备监测方法和系统、电子设备、存储介质
JP2023518771A (ja) 機械学習に基づくデータ・モニタリング
WO2021056708A1 (zh) 一种大批量数据导入方法、装置、计算机设备及存储介质
WO2023241461A1 (zh) 污水检查井的液位预测方法、装置、电子设备及存储介质
CN111949498B (zh) 应用服务器异常预测方法及系统
US20190104028A1 (en) Performance monitoring at edge of communication networks using hybrid multi-granular computation with learning feedback
CN115269108A (zh) 一种数据处理方法、装置及设备
CN111835536B (zh) 一种流量预测方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23871049

Country of ref document: EP

Kind code of ref document: A1