CN111751508A - Performance evaluation prediction method and system for life cycle of water quality sensor - Google Patents
Performance evaluation prediction method and system for life cycle of water quality sensor Download PDFInfo
- Publication number
- CN111751508A CN111751508A CN202010396203.2A CN202010396203A CN111751508A CN 111751508 A CN111751508 A CN 111751508A CN 202010396203 A CN202010396203 A CN 202010396203A CN 111751508 A CN111751508 A CN 111751508A
- Authority
- CN
- China
- Prior art keywords
- water quality
- quality sensor
- life
- model
- performance
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 146
- 238000000034 method Methods 0.000 title claims abstract description 85
- 238000011156 evaluation Methods 0.000 title claims abstract description 15
- 238000012423 maintenance Methods 0.000 claims abstract description 40
- 238000009826 distribution Methods 0.000 claims abstract description 39
- 238000000611 regression analysis Methods 0.000 claims abstract description 31
- 230000032683 aging Effects 0.000 claims abstract description 17
- 230000008569 process Effects 0.000 claims description 19
- 238000009499 grossing Methods 0.000 claims description 10
- 238000012935 Averaging Methods 0.000 claims description 8
- 238000011161 development Methods 0.000 claims description 5
- 230000018109 developmental process Effects 0.000 claims description 5
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 230000035772 mutation Effects 0.000 abstract description 4
- 238000005259 measurement Methods 0.000 description 7
- 230000003044 adaptive effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000004140 cleaning Methods 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 239000003792 electrolyte Substances 0.000 description 3
- 239000011521 glass Substances 0.000 description 3
- 239000012528 membrane Substances 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000009897 systematic effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000007500 overflow downdraw method Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 235000008733 Citrus aurantifolia Nutrition 0.000 description 1
- 235000011941 Tilia x europaea Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000003925 fat Substances 0.000 description 1
- 229910052602 gypsum Inorganic materials 0.000 description 1
- 239000010440 gypsum Substances 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 150000002500 ions Chemical class 0.000 description 1
- 239000004571 lime Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000013028 medium composition Substances 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 239000002574 poison Substances 0.000 description 1
- 231100000614 poison Toxicity 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- -1 steam Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000035882 stress Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/18—Water
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Engineering & Computer Science (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Physics & Mathematics (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
The invention discloses a performance evaluation and prediction method and a system for a life cycle of a water quality sensor, wherein the method comprises the following steps: s100, collecting and storing state parameters for reflecting the performance of the water quality sensor; s200, performing regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor, and predicting the residual service life, the next calibration time and the next maintenance time of the water quality sensor. The method provided by the invention is used for systematically and comprehensively evaluating and predicting the performance of the sensor. The problems of complex service life distribution and easy mutation of states are solved by self-adaptive regression; and the integrated model solves the problems of low prediction precision and poor stability.
Description
Technical Field
The invention relates to a performance evaluation and prediction method and system for a life cycle of a water quality sensor.
Background
In the process industries of petrochemical industry and the like, environmental parameters such as water, steam, oil and the like of raw materials in the storage or processing process need to be monitored to ensure the product quality, improve the process safety and control the operation cost.
Sensors are windows that know the state of the process, and their effectiveness is a prerequisite and basis for implementing process control, optimization. The sensors must maintain optimum measurement performance, require proper cleaning and maintenance, and often have natural drawbacks in terms of reliability. Optimizing the performance of these sensors can enhance their reliability, improve process integrity and reduce downtime, avoiding longer interruptions greatly improving productivity. Statistically up to 60% of the sensor maintenance work is unnecessary maintenance. Maintenance is typically scheduled and the sensors may still be forced to calibrate when not needed. The sensor typically performs periodic calibrations, with two or more point calibrations established using appropriate criteria. These calibration data are stored inside the transmitter or sensor for measurement value estimation. Besides the reference calibration data, the performance of the sensors is analyzed, and the state attribute of the sensors, the temperature of the measured medium, extreme parameter data and the like are also provided. For example, the service life of a pH sensor is affected by the working environment, on the one hand, since the glass film is not durable, the film surface often forms deposits from lime, gypsum, fat, protein, etc., but the effect on the film surface can be mitigated by cleaning. On the other hand, the aging process of the pH sensor cannot be compensated, and especially when the pH sensor works at an extreme pH value or a high temperature, the aging of the glass electrode can be greatly accelerated.
The glass membrane and the reference half-cell gradually decompose under the influence of the medium and thus increase the probability of failure, the reference electrode having a membrane which allows electrolytic contact between the reference electrolyte and the measurement solution, membrane variations which may poison and degrade the reference electrolyte, penetrating foreign ions in the reference electrolyte, so that variations in the cell voltage (half-cell voltage) of the reference electrode cause significant variations in the zero point. Therefore, the self-state attribute data such as the internal resistance of the measuring electrode and the reference electrode can be used for evaluating the performance of the pH sensor.
The ph sensor experiences continuous wear due to the media. The rate of abrasion depends on the media composition, temperature, pressure, concentration, etc. Thus, data that analyzes sensor performance may also be, for example, total run time, run time under certain conditions; cleaning, calibration, maintenance, number of times of a particular process, etc. In particular, the performance of the sensor is estimated by adding stress sensing elements to the sensor surface or by comparing the response data of the sensor at different performance states by changing one or more operating conditions.
The process data such as the slope generated by the pH sensor in periodic calibration can also be used as important reference data. Some of these data reflecting the status information of the pH sensor may be obtained during normal measurements, while others may only be obtained during interrupted measurement operations, such as calibration.
Through analysis of transformation rules of the data or comparison with reference values and the like, the performance of the pH sensor can be tracked, and the failure mode and the functional state can be predicted.
The present research has mostly determined the wear characteristics by determining threshold values of the parameters, estimating the corresponding performances of said parameters and determining that the remaining service life exceeds a limit value representative of the time interval between extrapolated parameter values. If these parameters have an allowable range of values, measurements with a certain error range can be performed, prompting replacement of the sensor. A systematic and omnibearing sensor performance evaluation and prediction method is not formed, the induction and the utilization of state data are not sufficient, and the accuracy and the real-time performance are lacked.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a method and a system for evaluating and predicting the performance of a life cycle of a water quality sensor, which are used for evaluating and predicting the performance of the sensor systematically and comprehensively.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a performance evaluation prediction method for a life cycle of a water quality sensor comprises the following steps:
(1) collecting and storing state parameters for reflecting the performance of the water quality sensor;
(2) and performing regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor, and predicting the residual service life, the next calibration time and the next maintenance time of the water quality sensor.
Further, the method for estimating and predicting the performance of the life cycle of the water quality sensor as described above, wherein the predicting the remaining life of the water quality sensor comprises:
(2.1) estimating the development trend of historical data of each state parameter according to the state parameters of the water quality sensor, obtaining a curve of the electrode life of the water quality sensor and time, namely life distribution, defining an electrode wear limit, performing parameter estimation on the life distribution according to a total sample of a full life cycle of each state parameter to obtain an initial life model, and optimizing a regression coefficient of the initial life model;
and (2.2) improving the initial life model by using an incremental learning and integrated learning method according to the current state parameters of the water quality sensor, and predicting the life by using the improved model.
Further, the performance evaluation and prediction method for the life cycle of the water quality sensor as described above includes the following steps (2.2):
I. estimating the prediction result of the initial life model in real time according to the current state parameters of the water quality sensor, if the prediction result exceeds a preset range, performing incremental learning on the current state parameters to obtain an incremental model, using the incremental model to perform life prediction, and if the difference between the prediction result and the actual result is within the preset range, using the initial life model to perform life prediction;
II. And if the state parameters of the initial life model are multiple parameters, respectively establishing a regression model for each parameter, performing regression analysis to obtain a life prediction result, and finally fusing the life prediction results of each parameter to obtain a final life prediction result.
Further, the method for evaluating and predicting the performance of the life cycle of the water quality sensor comprises the steps of obtaining a water quality sensor life cycle, wherein the life cycle comprises a Weibull distribution, an exponential distribution and a polynomial distribution; the method for optimizing the regression coefficient of the initial life model comprises a gradient descent method, a Newton method and a least square method; the method for fusing the life prediction result of each parameter comprises an averaging method and a weighted averaging method.
Further, the method for predicting the performance evaluation of the life cycle of the water quality sensor as described above, wherein the predicting the next calibration time of the water quality sensor comprises:
analyzing the current aging rule of the water quality sensor before the water quality sensor is calibrated in a regression manner, and predicting the next calibration time;
when the water quality sensor is to be calibrated, prompting a user to calibrate the current state parameters of the water quality sensor to obtain calibration parameters;
after each calibration, regression analyzes the historical trend of the calibration parameters of the water quality sensor, predicts the calibration parameter value at any given time, and calculates the remaining time interval until the next calibration.
Further, according to the performance evaluation and prediction method for the life cycle of the water quality sensor, before the regression analysis is performed on the calibration parameters of the water quality sensor, expert experience is added to the prior distribution of the parameters to be estimated.
Further, the method for predicting the performance evaluation of the life cycle of the water quality sensor as described above, wherein the predicting the next maintenance time of the water quality sensor comprises:
regression analyzing the current aging rule of the water quality sensor before maintenance, defining a maintenance limit and a method, and predicting the next maintenance time;
and when the water quality sensor is required to be maintained, prompting a user to maintain the water quality sensor and giving corresponding maintenance measures.
Further, the method for evaluating and predicting the performance of the life cycle of the water quality sensor further comprises the following steps:
and S300, repeatedly executing S100-S200 for each water quality sensor until the residual service life reaches the use limit, and prompting a user to replace the water quality sensor.
Further, the method for evaluating and predicting the performance of the life cycle of the water quality sensor further comprises the following steps:
and identifying invalid parameter values in the running process of the water quality sensor, and respectively carrying out smoothing treatment before and after regression analysis.
The invention also provides a performance evaluation and prediction system for the life cycle of the water quality sensor, which comprises the following components:
the acquisition module is used for acquiring and storing state parameters for reflecting the performance of the water quality sensor;
and the prediction module is used for carrying out regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor and predicting the residual service life, the next calibration time and the next maintenance time of the water quality sensor.
The invention has the beneficial effects that: according to the design principle, the use process, the operation environment and the like of the water quality sensor, the state parameters which can reflect the performance of the water quality sensor and can be obtained are analyzed, the residual service life, the next calibration time and the next maintenance time of the water quality sensor are predicted based on a self-adaptive integrated model, and the systematic life cycle management of the water quality sensor is realized.
Compared with the prior art, the method provided by the invention is used for evaluating and predicting the performance of the sensor systematically and comprehensively. The problems of complex service life distribution and easy mutation of states are solved by self-adaptive regression; and the integrated model solves the problems of low prediction precision and poor stability.
Drawings
Fig. 1 is a schematic flow chart of a method for evaluating and predicting the performance of a life cycle of a water quality sensor according to an embodiment of the present invention;
FIG. 2 is a line graph of state parameters versus time during incremental learning provided in an embodiment of the present invention;
FIG. 3 is a schematic flow chart of a model for improving learning thought by adaptive learning provided in an embodiment of the present invention;
FIG. 4 is a schematic flow chart of an improved model for applying an ensemble learning concept provided in an embodiment of the present invention;
fig. 5 is a comparison graph of the original data of the state parameters and the predicted model values of the remaining life provided in the embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the drawings and the detailed description.
As shown in fig. 1, a method for estimating and predicting the performance of a life cycle of a water quality sensor includes:
s100, collecting and storing state parameters for reflecting the performance of the water quality sensor;
s200, performing regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor, and predicting the residual service life, next calibration time and next maintenance time of the water quality sensor.
In S200, predicting the remaining life of the water quality sensor includes:
s201, estimating the development trend of historical data of each state parameter according to the state parameters of the water quality sensor, obtaining a curve of the electrode life of the water quality sensor and time, namely life distribution, defining an electrode wear limit, performing parameter estimation on the life distribution according to a total sample of a full life cycle of each state parameter, obtaining an initial life model, and optimizing a regression coefficient of the initial life model;
the lifetime distribution includes a weibull distribution, an exponential distribution, and a polynomial distribution. Optimization methods include gradient descent methods, newton's methods, and least squares methods.
S202, improving an initial life model by using an incremental learning and integrated learning method according to the current state parameters of the water quality sensor, and predicting the life by using the improved model.
S202 comprises the following steps:
I. estimating a prediction result of an initial life model in real time according to current state parameters of the water quality sensor, if the prediction result exceeds a preset range, performing incremental learning on the current state parameters to obtain an incremental model, performing life prediction by using the incremental model, and if the difference between the prediction result and an actual result is within the preset range, performing life prediction by using the initial life model;
and evaluating the prediction effect of the regression analysis initial model in real time, if the prediction effect is poor, performing incremental learning on the real-time parameters of the regression analysis initial model, updating the parameters of the regression analysis initial model, and predicting by using the updated model before the next parameter value arrives.
II. And if the state parameters of the initial life model are multiple parameters, respectively establishing a regression model for each parameter, performing regression analysis to obtain a life prediction result, and finally fusing the life prediction results of each parameter to obtain a final life prediction result.
The fusion method includes an averaging method and a weighted averaging method.
In S200, predicting the next calibration time of the water quality sensor includes:
and (3) performing regression analysis on the current aging rule of the water quality sensor before the water quality sensor is calibrated, defining a maintenance limit and a method, and predicting the next calibration time.
And when the water quality sensor is to be calibrated, prompting a user to calibrate the water quality sensor.
Before carrying out regression analysis on the calibration parameters of the water quality sensor, the expert experience is added as the prior distribution of the parameters to be estimated. After each calibration, the historical trend of the calibration parameters of the water quality sensor is regression analyzed, the calibration parameter value at any given time is predicted, and the remaining time interval until the next calibration is calculated.
In S200, predicting the next maintenance time of the water quality sensor includes:
and (4) performing regression analysis on the current aging rule of the water quality sensor before the water quality sensor is required to be maintained, and predicting the next maintenance time.
And when the water quality sensor is required to be maintained, prompting a user to maintain the water quality sensor and giving a specific maintenance measure.
Such as pH sensors, require maintenance, e.g., cleaning, activation, etc., at the appropriate point in time. The time interval of maintenance is highly probable depending on the state parameters of the electrode itself and the environmental impact during use. This is basically quite the same as the predicted data of remaining life.
And adjusting the recommended pre-maintenance measures and time according to the sensor life dynamic information such as the residual life, the next calibration time and the like. Defining the maintenance limit on the basis of expert experience, fusing the remaining life and the next calibration time, and calculating the remaining time interval until the next due maintenance.
Further comprising:
and S300, repeatedly executing S100-S200 for each water quality sensor until the residual service life reaches the use limit, and prompting a user to replace the water quality sensor.
And repeating the acquisition state parameters, the regression analysis residual life, the next calibration time and the next maintenance time for each sensor until the residual life reaches the use limit, and prompting a user to replace the sensors.
Further comprising:
and identifying invalid parameter values in the running process of the water quality sensor, and respectively carrying out smoothing treatment before and after regression analysis.
The invention respectively carries out smoothing treatment before and after regression analysis aiming at parameter abnormity possibly occurring in the running process of the water quality sensor. Such as sensors damaged or specially handled during calibration maintenance, and therefore invalid parameter values, should be identified and passed through a smoothing process and then into various regression models.
The invention also provides a performance prediction system of the life cycle of the water quality sensor, which comprises the following components:
the acquisition module is used for acquiring and storing state parameters for reflecting the performance of the water quality sensor;
and the prediction module is used for carrying out regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor and predicting the residual service life, the next calibration time and the next maintenance time of the water quality sensor.
Predicting the remaining life of the water quality sensor includes:
(2.1) estimating the development trend of historical data of each state parameter according to the state parameters of the water quality sensor, obtaining a curve of the electrode life of the water quality sensor and time, namely life distribution, defining an electrode wear limit, performing parameter estimation on the life distribution according to a total sample of a full life cycle of each state parameter to obtain an initial life model, and optimizing a regression coefficient of the initial life model;
and (2.2) improving an initial life model by using an incremental learning and integrated learning method according to the current state parameters of the water quality sensor, and predicting the life by using the improved model.
(2.2) comprising:
I. estimating a prediction result of an initial life model in real time according to current state parameters of the water quality sensor, if the prediction result exceeds a preset range, performing incremental learning on the current state parameters to obtain an incremental model, performing life prediction by using the incremental model, and if the difference between the prediction result and an actual result is within the preset range, performing life prediction by using the initial life model;
II. And if the state parameters of the initial life model are multiple parameters, respectively establishing a regression model for each parameter, performing regression analysis to obtain a life prediction result, and finally fusing the life prediction results of each parameter to obtain a final life prediction result.
The lifetime distribution includes a weibull distribution, an exponential distribution, and a polynomial distribution; methods for optimizing the regression coefficients of the initial life model include a gradient descent method and a newton method; methods for fusing the life prediction result of each parameter include an averaging method, a weighted averaging method, and a voting method.
Predicting the next calibration time for the water quality sensor includes:
performing regression analysis on the current aging rule of the water quality sensor before the water quality sensor is calibrated, and predicting the next calibration time;
when the water quality sensor is to be calibrated, prompting a user to calibrate the current state parameters of the water quality sensor to obtain calibration parameters;
after each calibration, the historical trend of the calibration parameters of the water quality sensor is regression analyzed, the calibration parameter value at any given time is predicted, and the remaining time interval until the next calibration is calculated.
Before carrying out regression analysis on the calibration parameters of the water quality sensor, the expert experience is added as the prior distribution of the parameters to be estimated.
Predicting the next maintenance time of the water quality sensor comprises:
and (4) performing regression analysis on the current aging rule of the water quality sensor before the water quality sensor is required to be maintained, and predicting the next maintenance time.
Further comprising:
and (5) repeatedly executing S100-S200 for each water quality sensor until the residual service life reaches the use limit, and prompting the user to replace the water quality sensor.
Further comprising:
and identifying invalid parameter values in the running process of the water quality sensor, and respectively carrying out smoothing treatment before and after regression analysis.
Example one
Assuming that the water quality sensor is a pH sensor, acquiring state parameter data of the pH sensor, estimating the development trend of historical data through mathematical modeling, obtaining a curve of electrode service life and time, namely service life distribution, and defining an electrode wear limit, thereby predicting the time for the sensor to continuously run to reach the service limit. The performance of the electrode can be described sufficiently well by the model, assuming that the properties of the electrode are substantially known per se, medium and environmental conditions.
(1) Model improvement using adaptive learning thought
As shown in fig. 2-3, based on the overall samples of the full life cycle of each parameter, the life distribution is subjected to parameter estimation to obtain an initial life model, and meanwhile, regression coefficients are optimized, and the optimization method includes gradient descent, newton method and the like. And evaluating the regression prediction effect of the initial life model in real time, if the effect is poor, performing incremental learning on the current historical data, updating the model parameters to obtain an incremental model, and predicting by using the updated incremental model before the next parameter value comes, so that the problems of complex life distribution and easy mutation of the state are solved. The lifetime distribution includes, but is not limited to, weibull distribution, exponential distribution, polynomial distribution, and the like.
It should be noted that the remaining life is mainly affected by some expected parameter, and the change of the expected parameter satisfies the adaptive prediction assumption, i.e. the historical rule before the next parameter value arrives, so the change of the remaining life can be described by the adaptive prediction model.
Incremental learning means that every time data is newly added, all knowledge bases do not need to be reconstructed, and only updating caused by the newly added data is performed on the basis of the original knowledge base.
The regression model for predicting remaining life is better and better matched to specific environmental requirements in the adaptive process, and also statistically considers fluctuations between individual electrode samples. The problems that real data samples are difficult to obtain and the failure period is long are solved to a certain extent, and the dependence of the model on the data is relieved.
(2) Model improvement using integrated learning thought
As shown in fig. 4, for multiple parameters of the pH sensor, a regression model is established for each parameter, and then regression results are fused. Aiming at the characteristics of more state parameters, small assumed space and the like, regression analysis needs to be respectively carried out, regression results of different parameter characteristics are fused, and the fusion method can be a mean value, a weighted average, a voting method and the like. The accuracy and stability of the method are improved, and meanwhile, the variance of the result is reduced, so that the occurrence of overfitting is avoided.
(3) Smoothing process
And smoothing is respectively carried out before and after regression analysis aiming at parameter abnormity possibly occurring in the operation process. Such as sensors that are damaged or specially handled during calibration maintenance, and therefore invalid parameter values, should be identified and passed through the smoothing process and then into the model.
(4) Introducing expert prior knowledge
Before regression analysis is carried out on the calibration parameters, expert experience is added to the prior distribution of the parameters to be estimated, and reliability of prediction estimation is improved. After each calibration, regression analyzes the historical trend of the calibration parameters, predicts the calibration parameter value at any given time, and calculates the time interval left until the next due check.
According to the design principle, the use process, the operation environment and the like of the water quality sensor, the state parameters which can reflect the performance of the water quality sensor and can be obtained are analyzed, the residual service life, the next calibration time and the next maintenance time of the water quality sensor are predicted based on a self-adaptive integrated model, and the systematic life cycle management of the water quality sensor is realized.
Compared with the prior art, the method provided by the invention is used for evaluating and predicting the performance of the sensor systematically and comprehensively. On one hand, the problems of complex service life distribution and easy mutation of states are solved by the self-adaptive regression; the integrated model solves the problems of more state parameters and small single model assumed space; on the other hand, expert experience is introduced, prior distribution of the parameters to be estimated is added, and reliability of prediction estimation is improved. And the possible abnormal or invalid parameters are processed by means of data smoothing processing, sensor replacement automatic detection and the like, so that the robustness of the model is improved.
The method provided by the invention can enable the user to make the most reasonable judgment and decision on the pre-maintenance and the process optimization. Helping to maintain high production efficiency and tailor spare part plans.
The method provided by the invention increases the measurement reliability of the sensor, reduces the potential safety hazard, reduces the labor cost and improves the operation and maintenance efficiency.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is intended to include such modifications and variations.
Claims (10)
1. A performance evaluation prediction method for a life cycle of a water quality sensor is characterized by comprising the following steps:
(1) collecting and storing state parameters for reflecting the performance of the water quality sensor;
(2) and performing regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor, and predicting the residual service life, the next calibration time and the next maintenance time of the water quality sensor.
2. The method of claim 1, wherein predicting the remaining life of the water quality sensor comprises:
(2.1) estimating the development trend of historical data of each state parameter according to the state parameters of the water quality sensor, obtaining a curve of the electrode life of the water quality sensor and time, namely life distribution, defining an electrode wear limit, performing parameter estimation on the life distribution according to a total sample of a full life cycle of each state parameter to obtain an initial life model, and optimizing a regression coefficient of the initial life model;
and (2.2) improving the initial life model by using an incremental learning and integrated learning method according to the current state parameters of the water quality sensor, and predicting the life by using the improved model.
3. The method for evaluating and predicting the performance of the life cycle of the water quality sensor as claimed in claim 2, wherein the step (2.2) comprises the following steps:
I. estimating the prediction result of the initial life model in real time according to the current state parameters of the water quality sensor, if the prediction result exceeds a preset range, performing incremental learning on the current state parameters to obtain an incremental model, using the incremental model to perform life prediction, and if the difference between the prediction result and the actual result is within the preset range, using the initial life model to perform life prediction;
II. And if the state parameters of the initial life model are multiple parameters, respectively establishing a regression model for each parameter, performing regression analysis to obtain a life prediction result, and finally fusing the life prediction results of each parameter to obtain a final life prediction result.
4. The method of claim 3, wherein the lifetime distribution comprises a Weibull distribution, an exponential distribution, and a polynomial distribution; the method for optimizing the regression coefficient of the initial life model comprises a gradient descent method, a Newton method and a least square method; the method for fusing the life prediction result of each parameter comprises an averaging method and a weighted averaging method.
5. The method of claim 1, wherein predicting the next calibration time of the water quality sensor comprises:
analyzing the current aging rule of the water quality sensor before the water quality sensor is calibrated in a regression manner, and predicting the next calibration time;
when the water quality sensor is to be calibrated, prompting a user to calibrate the current state parameters of the water quality sensor to obtain calibration parameters;
after each calibration, regression analyzes the historical trend of the calibration parameters of the water quality sensor, predicts the calibration parameter value at any given time, and calculates the remaining time interval until the next calibration.
6. The method as claimed in claim 5, wherein expert experience is added as a prior distribution of the parameters to be estimated before regression analysis of the calibration parameters of the water quality sensor.
7. The method of claim 1, wherein predicting the next maintenance time of the water quality sensor comprises:
regression analyzing the current aging rule of the water quality sensor before maintenance, defining a maintenance limit and a method, and predicting the next maintenance time;
and when the water quality sensor is required to be maintained, prompting a user to maintain the water quality sensor and giving corresponding maintenance measures.
8. The method for performance evaluation and prediction of a water quality sensor lifecycle according to any of claims 1-7, further comprising:
and S300, repeatedly executing S100-S200 for each water quality sensor until the residual service life reaches the use limit, and prompting a user to replace the water quality sensor.
9. The method for performance evaluation and prediction of a water quality sensor lifecycle according to any of claims 1-7, further comprising:
and identifying invalid parameter values in the running process of the water quality sensor, and respectively carrying out smoothing treatment before and after regression analysis.
10. A performance evaluation prediction system for a life cycle of a water quality sensor, comprising:
the acquisition module is used for acquiring and storing state parameters for reflecting the performance of the water quality sensor;
and the prediction module is used for carrying out regression analysis on the performance, the aging rule and the residual service life of the water quality sensor based on the state parameters and the self-adaptive integrated model of the water quality sensor, evaluating the performance of the water quality sensor and predicting the residual service life, the next calibration time and the next maintenance time of the water quality sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010396203.2A CN111751508A (en) | 2020-05-12 | 2020-05-12 | Performance evaluation prediction method and system for life cycle of water quality sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010396203.2A CN111751508A (en) | 2020-05-12 | 2020-05-12 | Performance evaluation prediction method and system for life cycle of water quality sensor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111751508A true CN111751508A (en) | 2020-10-09 |
Family
ID=72673250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010396203.2A Pending CN111751508A (en) | 2020-05-12 | 2020-05-12 | Performance evaluation prediction method and system for life cycle of water quality sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111751508A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113029998A (en) * | 2021-03-31 | 2021-06-25 | 河南中烟工业有限责任公司 | Method, device and equipment for measuring checking period during online infrared moisture meter |
| CN113917103A (en) * | 2021-10-15 | 2022-01-11 | 贵州师范学院 | Multi-water-quality parameter detection device capable of automatically adjusting depth |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1678885A (en) * | 2002-08-29 | 2005-10-05 | 恩德莱斯和豪瑟尔分析仪表两合公司 | Method for monitoring sensor function |
| CN202710094U (en) * | 2012-07-31 | 2013-01-30 | 高博 | Sensor service life prediction system for automobile |
| CN104268367A (en) * | 2014-08-06 | 2015-01-07 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | Substation equipment state evaluation weight correction method and system based on multiple linear regression |
| CN104412247A (en) * | 2012-07-24 | 2015-03-11 | 通用电气公司 | Systems and methods for improving control system reliability |
| CN105488101A (en) * | 2015-11-13 | 2016-04-13 | 信阳师范学院华锐学院 | System for obtaining spatial geographic environment information |
| CN108133326A (en) * | 2017-12-22 | 2018-06-08 | 华润电力（菏泽）有限公司 | A kind of status early warning method and system based on thermal power generating equipment |
| CN108896312A (en) * | 2018-08-08 | 2018-11-27 | 国电联合动力技术有限公司 | A kind of wind power principal axis holds failure predication and service life evaluation system and method |
| CN109033569A (en) * | 2018-07-09 | 2018-12-18 | 哈尔滨理工大学 | A method of optimize for carrier-borne aircraft sensing system prophylactic repair threshold intensity and number |
| CN109238339A (en) * | 2018-09-30 | 2019-01-18 | 北京英视睿达科技有限公司 | The intelligence of the enviromental monitoring equipment of ginseng more than a kind of determines longevity method and device |
| CN110059412A (en) * | 2019-04-19 | 2019-07-26 | 上海工程技术大学 | Service life of metal material prediction technique |
| CN110456199A (en) * | 2019-08-14 | 2019-11-15 | 四川大学 | A kind of method for predicting residual useful life of multisensor syste |
| US20200096477A1 (en) * | 2018-09-21 | 2020-03-26 | Rosemount Inc. | Predictive lifespan of analytical sensors |
| CN110956268A (en) * | 2019-10-16 | 2020-04-03 | 中国石化青岛液化天然气有限责任公司 | Compressor fault diagnosis method based on big data |
-
2020
- 2020-05-12 CN CN202010396203.2A patent/CN111751508A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1678885A (en) * | 2002-08-29 | 2005-10-05 | 恩德莱斯和豪瑟尔分析仪表两合公司 | Method for monitoring sensor function |
| CN104412247A (en) * | 2012-07-24 | 2015-03-11 | 通用电气公司 | Systems and methods for improving control system reliability |
| CN202710094U (en) * | 2012-07-31 | 2013-01-30 | 高博 | Sensor service life prediction system for automobile |
| CN104268367A (en) * | 2014-08-06 | 2015-01-07 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | Substation equipment state evaluation weight correction method and system based on multiple linear regression |
| CN105488101A (en) * | 2015-11-13 | 2016-04-13 | 信阳师范学院华锐学院 | System for obtaining spatial geographic environment information |
| CN108133326A (en) * | 2017-12-22 | 2018-06-08 | 华润电力（菏泽）有限公司 | A kind of status early warning method and system based on thermal power generating equipment |
| CN109033569A (en) * | 2018-07-09 | 2018-12-18 | 哈尔滨理工大学 | A method of optimize for carrier-borne aircraft sensing system prophylactic repair threshold intensity and number |
| CN108896312A (en) * | 2018-08-08 | 2018-11-27 | 国电联合动力技术有限公司 | A kind of wind power principal axis holds failure predication and service life evaluation system and method |
| US20200096477A1 (en) * | 2018-09-21 | 2020-03-26 | Rosemount Inc. | Predictive lifespan of analytical sensors |
| CN110940714A (en) * | 2018-09-21 | 2020-03-31 | 罗斯蒙特公司 | Analyzing predicted lifetime of a sensor |
| CN109238339A (en) * | 2018-09-30 | 2019-01-18 | 北京英视睿达科技有限公司 | The intelligence of the enviromental monitoring equipment of ginseng more than a kind of determines longevity method and device |
| CN110059412A (en) * | 2019-04-19 | 2019-07-26 | 上海工程技术大学 | Service life of metal material prediction technique |
| CN110456199A (en) * | 2019-08-14 | 2019-11-15 | 四川大学 | A kind of method for predicting residual useful life of multisensor syste |
| CN110956268A (en) * | 2019-10-16 | 2020-04-03 | 中国石化青岛液化天然气有限责任公司 | Compressor fault diagnosis method based on big data |
Non-Patent Citations (2)
| Title |
|---|
| 张雪霞 等: "基于数据驱动的质子交换膜燃料电池寿命预测", 《西南交通大学学报》, vol. 55, no. 2, 30 April 2020 (2020-04-30), pages 417 * |
| 梅子行: "《智能风控 原理、算法与工程实践》", 机械工业出版社, pages: 145 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113029998A (en) * | 2021-03-31 | 2021-06-25 | 河南中烟工业有限责任公司 | Method, device and equipment for measuring checking period during online infrared moisture meter |
| CN113029998B (en) * | 2021-03-31 | 2024-01-09 | 河南中烟工业有限责任公司 | Method, device and equipment for measuring checking period during online infrared moisture meter |
| CN113917103A (en) * | 2021-10-15 | 2022-01-11 | 贵州师范学院 | Multi-water-quality parameter detection device capable of automatically adjusting depth |
| CN113917103B (en) * | 2021-10-15 | 2023-05-19 | 贵州师范学院 | Multi-water quality parameter detection device capable of automatically adjusting depth |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111508216B (en) | Intelligent early warning method for dam safety monitoring data | |
| US10521490B2 (en) | Equipment maintenance management system and equipment maintenance management method | |
| JP4601492B2 (en) | Quality prediction system and method for production process | |
| Coble et al. | Identifying optimal prognostic parameters from data: a genetic algorithms approach | |
| US20060155511A1 (en) | Method for monitoring sensor function | |
| US9164057B2 (en) | Method for operating a measuring point | |
| CN111751508A (en) | Performance evaluation prediction method and system for life cycle of water quality sensor | |
| CN113252544B (en) | Corrosion monitoring system and method for oil refining device | |
| WO2011002803A2 (en) | Methods and apparatus for predictive preventive maintenance of processing chambers | |
| CN108133326A (en) | A kind of status early warning method and system based on thermal power generating equipment | |
| CN116391159A (en) | Quality control of products for mass production | |
| CN107992991A (en) | Annual electricity sales amount Forecasting Methodology based on external environmental factor and Co-integration Theory | |
| CN112100574A (en) | Resampling-based AAKR model uncertainty calculation method and system | |
| CN114970688A (en) | Landslide monitoring data preprocessing method based on LSTMAD algorithm and Hermite interpolation method | |
| CN117665231A (en) | Method for monitoring industrial enterprise wastewater quality | |
| US11994488B2 (en) | Method of predictive monitoring of a variable of a medium and of a measurement accuracy of a measurement device measuring this variable | |
| Chi et al. | Online identification of a link function degradation model for solid oxide fuel cells under varying-load operation | |
| CN116957120A (en) | Device state history trend anomaly prediction method based on data analysis | |
| CN113191506B (en) | Aperiodic condition-based maintenance method considering equipment detection uncertainty | |
| Zhou et al. | Quantification of value of information associated with optimal observation actions within partially observable Markov decision processes | |
| CN110889186A (en) | Consistency inspection method for degradation data of accelerated storage and natural storage based on equal degradation increment time interval gray correlation analysis | |
| CN118052082B (en) | Heating armoured wire pipe orifice sealing control method and system | |
| CN118024035B (en) | Machine learning-based polishing solution replacement early warning method and device | |
| CN118482817B (en) | Digital textile equipment management method and system based on data analysis | |
| CN114757054B (en) | Petroleum, petrochemical and natural gas industrial equipment residual life prediction method based on three-parameter Weibull distribution algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201009 |
|
| RJ01 | Rejection of invention patent application after publication |