Data-driven piston life prediction method

Abstract

The invention relates to the technical field of piston life prediction, and particularly discloses a data-driven piston life prediction method, which comprises the following steps: establishing a piston life prediction model, and periodically inputting real-time characteristics into the piston life prediction model to obtain the residual life W of a piston; if the residual life W is smaller than a preset residual life threshold value, obtaining a numerical value of the characteristic in the verification period and recording the numerical value as a historical value; sorting the historical values of the single types according to a time axis sequence to obtain a first sequence, and calculating a verification rate; determining a positive characteristic and a negative characteristic according to a positive correlation or negative correlation relationship between the characteristic and the residual life of the piston; determining whether the positive feature and the negative feature are abnormal features based on the verification rate and the historical value; and judging whether to send piston life early warning or not according to the total number of the abnormal characteristics. According to the invention, the accuracy of sending the piston life early warning is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of piston life prediction, and particularly relates to a data-driven piston life prediction method. BACKGROUND

[0002] The piston is a key component in the automobile engine, located inside the cylinder, connected with the connecting rod and the crankshaft. Its basic function is to convert the high-temperature and high-pressure gas energy generated by the combustion in the combustion chamber into mechanical energy to push the crankshaft to rotate, thereby driving the vehicle forward.

[0003] The piston needs to reciprocate repeatedly in a high-temperature, high-pressure and high-speed environment, and is more prone to wear, deformation, cracking and other problems, which will gradually reduce its working performance and shorten its service life. When the remaining life of the piston reaches the limit, it often leads to a decrease in sealing performance, a decrease in combustion efficiency, and even power loss or engine failure.

[0004] The conventional piston life prediction method is mostly based on experimental data under fixed working conditions, and the remaining life is inferred by collecting temperature, pressure, vibration and other signals and training an artificial intelligence model. However, in the real application environment, the working conditions of the piston are complex and changeable, and there may be two pistons with different actual remaining lives. The same operating data is collected under two different working conditions, and then the same remaining life is output, that is, there is a deviation between the model prediction result and the actual life, and then when the remaining life output by the model is lower than expected and an alarm message is sent to the driver, the actual life of the piston may be higher than expected, which affects the normal use of the piston. SUMMARY

[0005] The purpose of the present application is to provide a data-driven piston life prediction method to solve the above technical problems.

[0006] The purpose of the present application can be achieved by the following technical solutions: A data-driven piston life prediction method, comprising the following steps: Collecting operating data of the piston and extracting features affecting the remaining life of the piston therefrom, establishing a piston life prediction model according to the features, periodically collecting the features of the piston and recording them as real-time features, the time interval between adjacent two times of collecting the real-time features being t, and inputting the real-time features into the piston life prediction model to obtain the remaining life W of the piston; If the remaining life W is less than a preset remaining life threshold, the values of the features in a verification period are obtained and recorded as historical values, the verification period being [Tsta-T, Tsta], Tsta representing the time point at which the remaining life W is obtained, and T representing the preset length of the verification period; The historical values of a single type are sorted in chronological order to obtain a first sorting, and a verification rate AC is calculateda = (C a -C a+1 ) / t, C a , C a+1 respectively represent the historical values of the a-th and the a+1-th in the first sorting; If the k-th feature is positively correlated with the remaining life of the piston, the k-th feature is recorded as a positive feature; if the k-th feature is negatively correlated with the remaining life of the piston, the k-th feature is recorded as a negative feature; whether the positive feature and the negative feature are abnormal features is determined based on the verification rate and the historical value respectively; The total number N of abnormal features is counted, if the judgment value Y=N / N1 is greater than a preset judgment threshold, N1 represents the number of feature categories, then a piston life warning is sent to a display terminal in the cockpit, the piston life warning includes the remaining life W and the abnormal feature.

[0007] As a further scheme of the application: establishing a piston life prediction model comprises: Features of the same piston obtained at the same time are added with labels, and the labels are the remaining life of the piston at this time; A piston life prediction model is established based on deep learning, the piston life prediction model is trained and verified based on the features added with labels, in the training process, a back propagation algorithm is used to minimize a loss function, and k-fold cross-validation is used for verification.

[0008] As a further scheme of the application: whether the positive feature and the negative feature are abnormal features is determined based on the verification rate and the historical value respectively, comprising: For the positive feature Z1: If C a is less than a preset historical value first threshold and / or ΔC a is less than a preset verification rate first threshold, the historical value of the a-th in the first sorting is recorded as a positive abnormal value; The proportion of the number of positive abnormal values to the number of historical values is counted, if the proportion is greater than 0.7, the positive feature Z1 is recorded as an abnormal feature; For the negative feature Z2: If C a is greater than a preset historical value second threshold and / or ΔC a is greater than a preset verification rate second threshold, the historical value of the a-th in the first sorting is recorded as a negative abnormal value; The proportion of the number of negative abnormal values to the number of historical values is counted, if the proportion is greater than 0.7, the negative feature Z2 is recorded as an abnormal feature; Wherein, the historical value first threshold is greater than the historical value second threshold, and the verification rate second threshold is greater than the verification rate first threshold.

[0009] As a further scheme of the present application: different positive features correspond to different historical value first threshold values and verification rate first threshold values, and different negative features correspond to different historical value second threshold values and verification rate second threshold values.

[0010] As a further scheme of the present application: based on the verification rate and the historical value, respectively determining whether the positive feature and the negative feature are abnormal features further comprises: The abnormal feature is marked as a continuous attention feature, n time periods are set, the time interval between adjacent two time periods is one day, whether the continuous attention feature is an abnormal feature is determined in each time period, and n is a preset number. A proportion of the number of time periods in which the continuous attention feature is an abnormal feature to n is taken as a continuous proportion, and if the continuous proportion is greater than 0.5, a state abnormality early warning is generated to prompt the driver that the corresponding continuous attention feature is abnormal.

[0011] As a further scheme of the present application: establishing the piston life prediction model further comprises: A performance index of the piston life prediction model is acquired, the performance index includes an accuracy rate, a recall rate and a determination coefficient, and a performance score P of the piston life prediction model is calculated based on the performance index and a non-dominated solution distance method. If the performance score P is greater than a preset performance score threshold value, it is determined that the verification is passed.

[0012] As a further scheme of the present application: in the process of collecting the operation data of the piston, further comprising: the collected operation data is subjected to cleaning, denoising and normalization processing, and for missing data, a mean value, a median value or a predicted value based on a machine learning algorithm is used for filling The present application has the following advantages: (1) The present application introduces a comparison mechanism of historical value and verification rate in the life prediction process, which can perform secondary verification when the prediction result is close to the threshold value, avoiding triggering the early warning directly due to single abnormal working condition; when the piston shows data fluctuation under different working conditions, it can be identified that it belongs to local change driven by working condition rather than material itself degradation, thereby effectively reducing the false alarm situation; the early warning information received by the driver is more reliable, ensuring that the piston will not be prematurely determined as abnormal when the remaining life is sufficient, improving the stability and use efficiency of the vehicle operation. (2) The present application can accurately distinguish between "local abnormality caused by working condition" and "overall degradation caused by aging" by globally statistically analyzing and identifying a plurality of features; when the piston indeed enters the aging or degradation stage, the damage presents consistency across working conditions, that is, multiple features such as temperature, pressure, vibration, and combustion efficiency will simultaneously appear universal decline or abnormality; the present application uses such multi-dimensional consistency features for judgment, thereby ensuring that a real aging can be timely warned, and a local working condition change will not be misjudged, improving the accuracy and reliability of life prediction; (3) The present application can dynamically monitor the health state of the piston and judge whether the abnormality is persistent by continuously paying attention to the performance of the features in multiple time periods; this can not only capture occasional abnormal phenomena in the short term, but also identify long-term existing degradation trends; when a feature repeatedly shows abnormality in a long time range, a state abnormality warning will be generated, prompting the driver that the feature has appeared persistent degradation; this method can distinguish between transient fluctuations and structural recession, so that the piston life warning has both sensitivity and stability and reliability. BRIEF DESCRIPTION OF DRAWINGS

[0013] The present application will be further described below in conjunction with the accompanying drawings.

[0014] Figure 1 is a flowchart of a data-driven piston life prediction method of the present application. DETAILED DESCRIPTION

[0015] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0016] Please refer to Figure 1 The present application is a data-driven piston life prediction method, which comprises the following steps: Collecting the running data of the piston and extracting features affecting the remaining life of the piston therefrom, establishing a piston life prediction model according to the features, periodically collecting the features of the piston and recording them as real-time features, the time interval between adjacent two times of collecting the real-time features is t, and inputting the real-time features into the piston life prediction model to obtain the remaining life W of the piston; For example, when collecting piston operating data, multi-source signals related to its working state are acquired during its operation, including but not limited to temperature signals, pressure signals, vibration signals, combustion state signals, and kinematic signals. The acquired raw data is cleaned and denoised, and data from different sources are aligned to ensure temporal consistency. Based on this, the data is normalized for subsequent processing. Then, features that can reflect the remaining life of the piston are extracted. The feature extraction methods include statistical analysis, frequency domain analysis, and trend analysis of the raw signals. For example, temperature stability and fluctuation amplitude are extracted from temperature signals, pressure peak value and pressure change rate are extracted from pressure signals, vibration acceleration amplitude and frequency distribution features are extracted from vibration signals, combustion efficiency and combustion instability indicators are extracted from combustion state signals, and stroke displacement law and velocity curve shape are extracted from kinematic signals. In a preferred embodiment of the present invention, establishing a piston life prediction model includes: Add labels to the features of the same piston acquired at the same time, with the label indicating the remaining lifespan of the piston at that time; A piston life prediction model was established based on deep learning. The piston life prediction model was trained and validated based on labeled features. During the training process, the backpropagation algorithm was used to minimize the loss function, and k-fold cross-validation was used for validation. It should be noted that by adding the remaining lifespan as a label to the features acquired at the same time, a direct correspondence can be established between the piston's operating state and its actual lifespan. This allows the model to learn the mapping rules between different feature combinations and lifespan length during training. Using deep learning methods can automatically extract potential nonlinear relationships in a high-dimensional feature space, thus avoiding the limitations of relying solely on human experience to select features. Continuously adjusting parameters using the backpropagation algorithm allows the model to gradually approach the true correspondence between features and lifespan. Introducing cross-validation ensures that the model has stable performance under different data partitions. By using a large number of labeled samples, the model can master the true degradation rules of pistons under complex operating conditions, making the prediction results closer to the actual lifespan state, thereby providing a reliable basis for subsequent lifespan warnings.

[0017] In a preferred embodiment, establishing the piston life prediction model further includes: Obtain the performance metrics of the piston life prediction model, including precision, recall, and coefficient of determination. Calculate the performance score P of the piston life prediction model based on the performance metrics and the best-to-worst solution distance method. If the performance score P is greater than the preset performance score threshold, the verification is considered successful. It is worth noting that after the establishment of the piston life prediction model, the performance indicators are obtained and evaluated, which can reflect the accuracy, integrity and fitting degree of the model for life prediction from different angles. The precision, recall and determination coefficient are used as comprehensive indicators to avoid the deviation caused by single evaluation standard. The multiple indicators are unified into a performance score by the optimal solution distance method, so that the evaluation result is more objective and comprehensive. When the performance score exceeds the preset threshold, it means that the model has strong generalization ability and stability, and can maintain reliable prediction performance under different working conditions.

[0018] If the remaining life W is less than the preset remaining life threshold, the values of the features in the verification period are obtained and recorded as historical values, and the verification period is [Tsta-T, Tsta], Tsta represents the time point of obtaining the remaining life W, and T represents the preset length of the verification period; Sort the historical values of a single category in chronological order to obtain a first sorting, and calculate the verification rate AC a = (C a -C a+1 ) / t, C a , C a+1 represent the historical values of the a-th and a+1-th positions in the first sorting, respectively; If the kth feature is positively correlated with the remaining life of the piston, the kth feature is recorded as a positive feature; if the kth feature is negatively correlated with the remaining life of the piston, the kth feature is recorded as a negative feature; whether the positive feature and the negative feature are abnormal features is determined based on the verification rate and the historical value, respectively; In another preferred embodiment of the present application, whether the positive feature and the negative feature are abnormal features is determined based on the verification rate and the historical value, respectively, including: For the positive feature Z1: If C a is less than the preset historical value first threshold and / or AC a is less than the preset verification rate first threshold, the historical value of the a-th position in the first sorting is recorded as a positive abnormal value; The proportion of the number of positive abnormal values to the number of historical values is calculated, and if the proportion is greater than 0.7, the positive feature Z1 is recorded as an abnormal feature; For the negative feature Z2: If C a is greater than the preset historical value second threshold and / or AC a is greater than the preset verification rate second threshold, the historical value of the a-th position in the first sorting is recorded as a negative abnormal value; The proportion of the number of negative abnormal values to the number of historical values is calculated, and if the proportion is greater than 0.7, the negative feature Z2 is recorded as an abnormal feature; Wherein, the historical value first threshold is greater than the historical value second threshold, and the verification rate second threshold is greater than the verification rate first threshold.

[0019] Understandably, by distinguishing features into positive and negative correlations and using historical values ​​and validation rates for threshold judgments, reasonable judgment criteria can be set for the relationship between different features and lifespan, thus avoiding the situation where the same rule applies to all features. Positive features are more likely to reflect lifespan decline when the values ​​or rates of change are low, while negative features are more likely to reflect degradation trends when the values ​​or rates of change are high. Using proportional statistics to distinguish between occasional anomalies and widespread anomalies ensures that the judgment results reflect the overall trend rather than local random fluctuations. By using quantitative comparison and threshold setting to reveal the changing patterns of features with lifespan decline, it is ensured that signals identified as anomalous features are indeed consistent with lifespan degradation. This ensures that subsequent lifespan warnings are based on the consistency of multi-dimensional features, thereby improving the accuracy and reliability of the warnings. It is understandable that different features have different dimensions, value ranges, baseline levels, and noise distributions. The same value and the same magnitude of change do not represent equivalent lifetime information on different features. Therefore, it is necessary to set historical value thresholds and verification rate thresholds that match the scale of each positive and negative feature, so that the judgment is based on the relative deviation of the feature's normal range and rate of change rather than a uniform scale. Therefore, when setting thresholds, the following should be met: different positive features correspond to different first historical value thresholds and first verification rate thresholds, and different negative features correspond to different second historical value thresholds and second verification rate thresholds.

[0020] In another preferred embodiment of the present invention, determining whether positive and negative features are anomalous features based on the verification rate and historical values ​​respectively further includes: The abnormal features are marked as features of continuous attention. n time periods are set, and the time interval between two adjacent time periods is one day. In each time period, it is determined whether the features of continuous attention are abnormal features. n is a preset number. The proportion of the number of time periods in which the continuously monitored feature is an abnormal feature to n is taken as the continuous proportion. If the continuous proportion is greater than 0.5, an abnormal status warning is generated to alert the driver that the corresponding continuously monitored feature is abnormal. It can be understood that in feature monitoring, a single occurrence of an anomaly can only be an occasional fluctuation or environmental interference, and cannot directly represent that the piston state has substantially degraded, so it is necessary to repeatedly judge the same feature in multiple consecutive time periods, and confirm whether the feature is truly deviated from the normal range by the statistical proportion of the continuous occurrence of the anomaly; the consistency in the time dimension is used to distinguish between temporary fluctuations and long-term anomalies, so that the judgment result is more stable and reliable, and when the continuous proportion exceeds the threshold value, it is considered that there is a reliable anomaly trend for the feature, thereby generating a state anomaly warning to prompt the driver to pay attention to the continuous abnormal performance of the feature, and providing timely information support for state monitoring in the running process.

[0021] The total number N of statistical abnormal features is counted, and if the judgment value Y=N / N1 is greater than a preset judgment threshold, N1 represents the number of feature categories, then a piston life warning is sent to a display terminal in the cockpit, and the piston life warning includes residual life W and abnormal features; It is worth noting that by counting the number of abnormal features and calculating the proportion thereof in all features, the occasional deviation of a single / few features can be distinguished from the overall degradation trend; when the piston really enters the aging or degradation stage, the damage performance is global, regardless of the working condition, the temperature stability, the pressure fluctuation, the vibration amplitude, the combustion efficiency and other types of features will simultaneously appear universal decline or anomaly, so in this case the proportion of abnormal features will significantly increase, and the overall judgment value will also increase; on the contrary, if it is only affected by a certain working condition change, it will only cause one or a few types of features to be abnormal, while other features remain in the normal range, and the abnormal proportion will not be greatly improved. By this phenomenon, it is determined whether to send a warning message, thereby providing a reliable trigger basis in life prediction, so that the warning can truly reflect the overall damage features of the piston entering the degradation stage; In order to cover different working conditions, the length T of the verification period should be relatively long, and the specific length can be set by experts and technical personnel based on experience.

[0022] The above formulas are dimensionless values calculated, the formulas are obtained by collecting a large amount of data to simulate a formula of the latest real situation, and the preset parameters and threshold values in the formula are set by technicians in the art according to the actual situation; meanwhile, the preset related threshold values in the present application, including but not limited to the first threshold value of the historical value, the second threshold value of the historical value, the second threshold value of the verification rate, and the first threshold value of the verification rate, can be set by technicians in the art according to the actual situation and / or experience.

[0023] The above has been described in detail one embodiment of the present application, but the content is only the preferred embodiment of the present application, cannot be considered for limiting the scope of the present application. Any equivalent changes and improvements made in the scope of the present application, should still belong to the scope of the present application.

Claims

1. A data-driven piston life prediction method, characterized by, The method comprises the following steps: Collecting operation data of the piston and extracting features affecting the remaining life of the piston from the operation data, establishing a piston life prediction model according to the features, periodically collecting the features of the piston and recording the features as real-time features, the time interval between adjacent two times of collecting the real-time features being t, and inputting the real-time features into the piston life prediction model to obtain the remaining life W of the piston; If the remaining life W is less than a preset remaining life threshold, obtaining the values of the features in a verification period and recording the values as historical values, the verification period being [Tsta-T, Tsta], Tsta representing the time point at which the remaining life W is obtained, and T representing a preset length of the verification period; Sort the historical values of a single category in chronological order to obtain a first ordering, and calculate the verification rate ΔC a = (C a - C a+1 ) / t, C a , C a+1 represent the historical values of the a-th and a+1-th positions in the first ordering, respectively. If the kth feature is positively correlated with the remaining life of the piston, the kth feature is recorded as a positive feature; if the kth feature is negatively correlated with the remaining life of the piston, the kth feature is recorded as a negative feature; Determining whether the positive feature and the negative feature are abnormal features based on the verification rate and the historical values, respectively; Counting the total number N of the abnormal features, and if a judgment value Y=N / N1 is greater than a preset judgment threshold, N1 representing the number of feature types, sending a piston life warning to a display terminal in the cockpit, the piston life warning including the remaining life W and the abnormal features.

2. The data-driven piston life prediction method of claim 1, wherein, The method for establishing the piston life prediction model comprises: Adding a label to the features of the same piston obtained at the same time, the label being the remaining life of the piston at this time; Establishing the piston life prediction model based on deep learning, training and verifying the piston life prediction model based on the features with the label, using a back propagation algorithm to minimize a loss function in the process of training, and verifying by using k-fold cross-validation.

3. The data-driven piston life prediction method of claim 1, wherein, The method for determining whether the positive feature and the negative feature are abnormal features based on the verification rate and the historical values respectively comprises: For the positive feature Z1: If C a is less than a preset historical value first threshold and / or ΔC a is less than a preset verification rate first threshold, the historical value of the a-th position in the first ranking is recorded as a positive abnormal value. Counting the proportion of the number of positive abnormal values to the number of historical values, and if the proportion is greater than 0.7, the positive feature Z1 is recorded as an abnormal feature; For the negative feature Z2: If C a greater than a preset historical value second threshold and / or ΔC a greater than a preset verification rate second threshold, the historical value of the a-th position in the first ranking is recorded as a negative abnormal value. Counting the proportion of the number of negative abnormal values to the number of historical values, and if the proportion is greater than 0.7, the negative feature Z2 is recorded as an abnormal feature; The first threshold value of the historical values is greater than the second threshold value of the historical values, and the second threshold value of the verification rate is greater than the first threshold value of the verification rate.

4. The data-driven piston life prediction method of claim 3, wherein, Different positive features correspond to different first threshold values of the historical values and different first threshold values of the verification rate, and different negative features correspond to different second threshold values of the historical values and different second threshold values of the verification rate.

5. The data-driven piston life prediction method of claim 1, wherein, The method for determining whether the positive feature and the negative feature are abnormal features based on the verification rate and the historical values respectively further comprises: Marking the abnormal features as continuously concerned features, setting n time periods, the time interval between adjacent two time periods being one day, determining whether the continuously concerned features are abnormal features in each time period, and n being a preset number; Taking the proportion of the number of time periods in which the continuously concerned features are abnormal features to n as a continuous proportion, and if the continuous proportion is greater than 0.5, generating a state abnormality warning for prompting the driver that the corresponding continuously concerned features are abnormal.

6. The data-driven piston life prediction method of claim 2, wherein, The method for establishing the piston life prediction model further comprises: Obtaining a performance index of the piston life prediction model, the performance index including an accuracy rate, a recall rate and a determination coefficient, calculating a performance score P of the piston life prediction model based on the performance index and a distance method of superior and inferior solutions. If the performance score P is greater than a preset performance score threshold, it is determined that the verification is passed.

7. The data-driven piston life prediction method of claim 1, wherein, In the process of collecting the operation data of the piston, further comprising: cleaning, denoising and normalizing the collected operation data, and filling the missing data with the mean value, the median value or the predicted value based on a machine learning algorithm.

Images