Engine combustion efficiency detection method

Abstract

The invention relates to the technical field of data processing, in particular to an engine combustion efficiency detection method, which comprises the following steps of: acquiring an output power sequence, a tail gas efficiency sequence, a pressure efficiency sequence and a combustion efficiency sequence of an engine in a plurality of historical time periods, calculating similarities among the data sequences, and clustering the similarities to obtain a clustering result; a plurality of clustering clusters are obtained; acquiring an output power sequence, a tail gas efficiency sequence and a pressure efficiency sequence of the engine in real time, calculating the similarity among the data sequences, comparing the similarity with the clustering cluster center, and selecting the nearest clustering cluster as a target clustering cluster; calculating the real-time combustion efficiency based on the accuracy of the two clusters, the real-time tail gas efficiency and the pressure efficiency; and if the real-time combustion efficiency exceeds the set threshold value, it is judged that the engine breaks down, and early warning is triggered. The problem that the reliability of the engine is not high is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, to a method for detecting engine combustion efficiency. Background Art

[0002] With the continuous advancement of industrialization, engines are increasingly used in various types of equipment and transportation, especially in the fields of automobiles, ships, aircraft, and generator sets. Engines play a vital role as a power source. The combustion efficiency of the engine directly affects the operating performance, energy utilization and emission level of the equipment. Improving the combustion efficiency of the engine not only helps to reduce energy consumption and reduce environmental pollution, but also prolongs the service life of the equipment and improves the overall economic benefits. Therefore, how to accurately detect and evaluate the combustion efficiency of the engine has become one of the research hotspots in recent years.

[0003] Traditional methods for testing engine combustion efficiency mainly rely on combustion tests and sensor data collection in laboratory environments, using methods such as exhaust gas analysis, temperature monitoring, and exhaust gas concentration measurement. These traditional methods are effective in specific environments, but due to the complex and changeable combustion state during engine operation and the interference of various external factors such as load fluctuations, changes in fuel quality, and ambient temperature, traditional testing methods often cannot reflect the engine's combustion efficiency in real time and accurately, and the testing cost is high and the cycle is long, making it difficult to achieve rapid and continuous online monitoring.

[0004] To address the above problems, a method for detecting engine combustion efficiency based on similarity calculation and cluster analysis was proposed. By introducing accuracy calculation and real-time monitoring mechanism, the shortcomings of the existing technology were addressed and a more accurate and reliable solution was provided in practical applications.

[0005] Existing technologies only rely on directly measuring engine exhaust gas, temperature, pressure and other parameters to infer combustion efficiency. They often only focus on a single parameter and ignore the complex correlation between various factors and parameters during engine operation. This makes it difficult to fully reflect the comprehensive performance of the engine and has certain limitations, which makes it impossible to fully reveal changes in the overall performance of the engine, leading to the problem of low engine reliability. Summary of the invention

[0006] In order to solve the problems of low accuracy and insufficient reliability of detection results raised in the above background technology, the present invention provides the following solutions.

[0007] The present invention provides an engine combustion efficiency detection method, comprising: obtaining an output power sequence, an exhaust gas efficiency sequence, a pressure efficiency sequence and a combustion efficiency sequence in a plurality of historical time periods of an engine; calculating the similarity between the output power sequence and the exhaust gas efficiency sequence in each historical time period and clustering them to obtain a plurality of first clustering clusters; calculating the similarity between the output power sequence and the pressure efficiency sequence in each historical time period and clustering them to obtain a plurality of second clustering clusters; calculating the accuracy of each clustering cluster, wherein the accuracy is positively correlated with the exhaust gas efficiency and negatively correlated with the combustion efficiency; respectively calculating the similarity between the acquired real-time output power sequence and the real-time exhaust gas efficiency sequence and the real-time pressure efficiency sequence; calculating the distance between the similarity between the real-time output power sequence and the real-time exhaust gas efficiency sequence and the similarity between the centers of the plurality of first clustering clusters, and selecting the clustering cluster with the smallest distance as the first target clustering cluster; calculating the distance between the similarity between the real-time output power sequence and the real-time pressure efficiency sequence and the similarity between the centers of the plurality of second clustering clusters, and selecting the clustering cluster with the smallest distance as the second target clustering cluster; calculating the real-time combustion efficiency , , is the accuracy of the first target cluster, is the accuracy of the second target clustering, is the real-time exhaust efficiency, It is the real-time pressure efficiency; when the real-time combustion efficiency is greater than the set threshold, it is an engine failure.

[0008] The accuracy of each cluster obtained by clustering in the above technical solution provides a reliable reference for the real-time evaluation of combustion efficiency, and triggers an alarm mechanism when the combustion efficiency exceeds the set threshold, thereby achieving early warning of potential faults and improving the reliability of the engine.

[0009] Furthermore, the similarity is, , where is the similarity between the output power sequence and the exhaust efficiency sequence, , Respectively The output power sequence , The output power value, , Respectively The first in the exhaust efficiency series , The exhaust efficiency value is is the length of the sequence, is an exponential function with the natural constant e as base, is the normalization function.

[0010] The above technical solution not only considers the direct difference of each parameter, but also combines the changes between adjacent parameters, thereby more comprehensively reflecting the correlation between sequences. Through this similarity calculation, it is possible to more accurately identify the working state similar to the real-time data in the historical time period, thereby improving the accuracy of cluster analysis and making fault warning more accurate. This intelligent detection method based on time series similarity improves the ability of engine fault diagnosis, can respond to potential anomalies in real time, and ensure the efficient and safe operation of the engine.

[0011] Furthermore, the similarity is calculated using a dynamic time warping algorithm.

[0012] Furthermore, calculate the The accuracy of clusters , , where is the total number of similarities in the clusters, For cluster The first in the exhaust efficiency sequence corresponding to the similarity Exhaust efficiency, For the cluster The output power sequence corresponding to the similarity The combustion efficiency corresponding to the output power is is the normalization function.

[0013] The above technical solution calculates the accuracy of clustering clusters and comprehensively considers the relationship between exhaust efficiency sequence and combustion efficiency, thereby realizing accurate quantification of clustering cluster similarity. The calculation of accuracy not only reflects the matching degree of internal elements of the sequence, but also further improves its adaptability to different data scales through normalization processing. By evaluating the accuracy of clustering clusters, the matching degree between historical data and real-time data can be more accurately measured, thereby improving the reliability of fault detection and prediction.

[0014] Furthermore, calculate the The accuracy of clusters , , where is the total number of similarities in the clusters, For the cluster The length of the output power sequence corresponding to the similarity is For the cluster The first in the exhaust efficiency sequence corresponding to the similarity elements, For cluster The output power sequence corresponding to the similarity The combustion efficiency corresponding to each element is is the normalization function.

[0015] The above technical solution fully considers the difference between the output power sequence and the exhaust efficiency sequence and their relationship with combustion efficiency by calculating the accuracy of the clustering cluster, and then accurately evaluates the similarity of the data within the clustering cluster. By introducing normalization processing and the application of the maximum value function, the solution can eliminate the influence of the dimensional differences between different sequences and improve the stability and precision of the accuracy calculation. This enables cluster analysis to better reflect the essential similarities between data, thereby improving the sensitivity and accuracy of fault detection.

[0016] Furthermore, the K-Means clustering algorithm is used to cluster the similarities between each output power sequence and the exhaust efficiency sequence and pressure efficiency sequence.

[0017] The above technical solution uses the K-Means clustering algorithm to cluster the similarities between the output power sequence and the exhaust efficiency sequence and pressure efficiency sequence, effectively classifying similar operating states into the same group, thereby achieving fine classification of different engine operating conditions. Through cluster analysis, normal and abnormal operating modes can be more clearly identified, providing more accurate basic data for subsequent fault detection. This method can fully explore the internal connections between various efficiency indicators and improve the sensitivity and prediction ability of engine state changes.

[0018] Furthermore, the distance is Manhattan distance.

[0019] Furthermore, it also includes standardizing the output power sequence, the exhaust efficiency sequence, the pressure efficiency sequence and the combustion efficiency sequence.

[0020] Furthermore, it also includes triggering an alarm signal through an audible and visual alarm when an engine failure occurs.

[0021] Furthermore, the set threshold is 0.5.

[0022] The beneficial effects of the present invention are: The present invention provides a method for monitoring the combustion efficiency of an engine by calculating the similarity between real-time and historical data. The method combines cluster analysis with accurate similarity calculation, and can effectively monitor and classify different working states of the engine in real time. By standardizing and clustering multi-dimensional data, the present invention can accurately capture the changing trend of combustion efficiency and compare it with historical data to timely discover potential failure risks. By utilizing this method, when abnormal engine performance occurs, accurate judgment can be made through the calculation of real-time combustion efficiency, thereby triggering an alarm, ensuring rapid response and processing of engine failures, and improving the safety and reliability of the engine. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart schematically illustrating a method for detecting engine combustion efficiency according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] An embodiment of a method for detecting engine combustion efficiency.

[0025] like Figure 1 As shown, a flow chart of an engine combustion efficiency detection method according to an embodiment of the present invention comprises the following steps: S1: Calculate the similarity between each output power sequence and the exhaust efficiency sequence and pressure efficiency sequence in multiple historical time periods and perform clustering.

[0026] In one embodiment, the output power sequence, exhaust efficiency sequence, pressure efficiency sequence and combustion efficiency sequence of the engine in multiple historical time periods are collected.

[0027] The specific method of obtaining the engine's output power, exhaust efficiency, pressure efficiency and combustion efficiency is an existing technology and will not be described in detail in this solution. The output power sequence, exhaust efficiency sequence, pressure efficiency sequence and combustion efficiency sequence are standardized.

[0028] For example, the Z-score standardization method is used to standardize the output power series, exhaust efficiency series, pressure efficiency series, and combustion efficiency series, ensuring that all types of data are analyzed on the same scale, avoiding the deviation caused by different dimensions, and making the subsequent clustering analysis, similarity calculation, and fault diagnosis more accurate and reliable. The standardization process can fully reflect the relative changes and potential operating rules between parameters, thereby greatly improving the performance of the data analysis model, especially when predicting faults based on historical data, it can more accurately capture abnormal fluctuations and deviations.

[0029] In one embodiment, the K-Means clustering algorithm can be used to cluster the similarities between the output power sequence and the exhaust efficiency sequence in each historical time period to obtain multiple first cluster clusters, and the similarities between the output power sequence and the pressure efficiency sequence in each historical time period can be clustered to obtain multiple second cluster clusters.

[0030] The similarity is, , where is the similarity between the output power sequence and the exhaust efficiency sequence, , Respectively The output power sequence , The output power value, , Respectively The first in the exhaust efficiency series , The exhaust efficiency value is is the length of the sequence, is an exponential function with the natural constant e as base, is the normalization function.

[0031] By introducing a specific similarity calculation formula, the similarity between the output power sequence and the exhaust efficiency sequence is quantified, providing a more accurate input for the K-Means clustering algorithm; by considering the differences and change trends between the points in the sequence, the similarity metric can better reflect the dynamic relationship between the two. This method can effectively identify the similarity between output power and exhaust efficiency in different historical time periods, thereby performing more reasonable clustering and improving the accuracy and interpretability of the clustering results.

[0032] In another embodiment, the similarity is calculated using a dynamic time warping algorithm.

[0033] S2: Calculate the accuracy of each cluster.

[0034] In one embodiment, the calculation The accuracy of clusters , , where is the total number of similarities in the clusters, For the cluster The first in the exhaust efficiency sequence corresponding to the similarity Exhaust efficiency, For the cluster The output power sequence corresponding to the similarity The combustion efficiency corresponding to the output power is is the normalization function.

[0035] By introducing a sophisticated accuracy calculation method, the representativeness and similarity consistency of each cluster are quantified, and the similarity and combustion efficiency between the sequences within the cluster are comprehensively considered. The accuracy calculation can effectively reflect the compactness and consistency of the sequences within the cluster, and through normalization processing, the accuracy between different clusters is ensured to be comparable, avoiding the impact caused by differences in sequence length or data size. It provides a reliable quantitative indicator for the quality of clustering results, helps to identify more representative and high-quality clusters in cluster analysis, and thus provides more accurate data support for subsequent performance analysis and fault prediction.

[0036] In another embodiment, calculating the The accuracy of clusters , , where is the total number of similarities in the clusters, For the cluster The length of the output power sequence corresponding to the similarity is For cluster The first in the exhaust efficiency sequence corresponding to the similarity elements, For the cluster The output power sequence corresponding to the similarity The combustion efficiency corresponding to each element is is the normalization function.

[0037] The improved accuracy calculation method further improves the consistency of data within the cluster and the accuracy of similarity evaluation. By normalizing the differences of each sequence element and considering the relative change of the maximum value, the accuracy calculation can not only reflect the similarity between sequences within the cluster, but also effectively overcome the influence of different data ranges and amplitudes on the calculation results. This method improves the robustness and representativeness of the clustering results by quantifying the relationship between each sequence within the cluster and the combustion efficiency.

[0038] S3: Calculate the similarity between the real-time output power sequence and the real-time exhaust efficiency sequence and the real-time pressure efficiency sequence and determine the target cluster.

[0039] In one embodiment, the similarity between the real-time output power sequence and the real-time exhaust efficiency sequence and the real-time pressure efficiency sequence is first calculated to fully capture the relationship between the various parameters under real-time working conditions. Through the similarity measurement, the matching degree between the real-time data and the different clusters in the historical data can be clearly reflected, providing a basis for the subsequent cluster selection. Then, the distance from the similarity of the real-time output power sequence and the real-time exhaust efficiency sequence to the similarity of the centers of multiple first clusters is calculated, and the cluster with the smallest distance is selected as the first target cluster; similarly, the Manhattan distance from the similarity of the real-time output power sequence and the real-time pressure efficiency sequence to the similarity of the centers of multiple second clusters is calculated, and the cluster with the smallest distance is selected as the second target cluster. This method accurately matches the centers of different clusters through the Manhattan distance, ensuring that the selected target cluster has the greatest consistency in similarity, thereby avoiding erroneous clustering due to data fluctuations or anomalies.

[0040] By comparing the distance between real-time data and the centers of multiple historical clusters, it is possible to dynamically determine the similarity between the current operating conditions and historical data, and quickly and accurately match the most appropriate cluster to the engine's operating state. This process enhances the adaptability and real-time nature of fault prediction, and can respond to potential operating abnormalities in a short period of time, providing timely and effective data support for subsequent performance adjustments and fault warnings. In addition, the use of Manhattan distance ensures the intuitiveness of cluster selection and the simplicity of calculation.

[0041] S4: Calculate the real-time combustion efficiency according to the target clustering accuracy, the real-time exhaust efficiency, and the real-time pressure efficiency, and respond that the real-time combustion efficiency is greater than a set threshold value as an engine failure.

[0042] The above threshold value may be set to 0.5, and of course, it may be determined according to actual conditions.

[0043] In one embodiment, the real-time combustion efficiency is calculated , , is the accuracy of the first target cluster, is the accuracy of the second target clustering, is the real-time exhaust efficiency, For real-time pressure efficiency; The real-time combustion efficiency is calculated by combining the accuracy of different target clusters with the weighted average of real-time exhaust efficiency and pressure efficiency. This method can comprehensively consider the performance of the first target cluster and the second target cluster, and make the calculation of combustion efficiency more accurate through weight adjustment, reflecting the relative importance of the two efficiency indicators. Through this weighted method, the real-time combustion efficiency can be dynamically adjusted according to the accuracy of the cluster and the real-time operating conditions, providing more accurate feedback.

[0044] When the real-time combustion efficiency exceeds the set threshold, it will automatically determine that the engine may have a fault and quickly trigger the sound and light alarm to send an alarm signal. This judgment is based on the real-time calculated combustion efficiency, which reflects the operating status of the engine under the current working conditions. When the combustion efficiency increases abnormally, it often means that there is a problem in the combustion process, such as incomplete combustion of the fuel or other performance degradation problems, which may cause the engine to overheat or damage. Through this real-time monitoring and threshold setting, potential hidden faults can be identified in a timely manner to prevent the engine from operating under poor working conditions for a long time, avoiding more serious damage and accidents. By combining real-time data with threshold judgment, an active early warning mechanism is provided, which effectively enhances the early diagnosis capability of engine faults. With the feedback of the sound and light alarm, operators can quickly receive alarm signals and take necessary intervention measures, thereby reducing downtime and repair costs, and ensuring production efficiency and equipment safety.

[0045] The solution of the present invention effectively extracts key features under different working conditions by performing similarity calculation and cluster analysis on the output power sequence, exhaust efficiency sequence, pressure efficiency sequence and combustion efficiency sequence in the historical time period, and accurately evaluates the real-time operating status of the engine by calculating the accuracy of each cluster. Combining real-time data with set thresholds, potential engine failures can be identified in a timely manner, improving the accuracy and response speed of fault detection. In addition, the combination of standardization processing and Manhattan distance calculation optimizes the comparison and classification effects of data, improves the ability of intelligent fault diagnosis and early warning, and provides strong support for the efficient and stable operation of the engine.

[0046] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0047] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for detecting engine combustion efficiency, characterized in that: include: Obtain the output power sequence, exhaust efficiency sequence, pressure efficiency sequence and combustion efficiency sequence of the engine in multiple historical time periods; Calculate the similarity between the output power sequence and the exhaust efficiency sequence in each historical time period and perform clustering to obtain a plurality of first clusters; calculate the similarity between the output power sequence and the pressure efficiency sequence in each historical time period and perform clustering to obtain a plurality of second clusters; Calculating the accuracy of each cluster, wherein the accuracy is positively correlated with the exhaust efficiency and negatively correlated with the combustion efficiency; Calculate the similarity between the acquired real-time output power sequence and the real-time exhaust efficiency sequence and the real-time pressure efficiency sequence respectively; calculate the distance between the similarity between the real-time output power sequence and the real-time exhaust efficiency sequence and the similarity between the centers of multiple first clusters, and select the cluster with the smallest distance as the first target cluster; calculate the distance between the similarity between the real-time output power sequence and the real-time pressure efficiency sequence and the similarity between the centers of multiple second clusters, and select the cluster with the smallest distance as the second target cluster; Calculate real-time combustion efficiency , , is the accuracy of the first target cluster, is the accuracy of the second target clustering, is the real-time exhaust efficiency, It is the real-time pressure efficiency; when the real-time combustion efficiency is greater than the set threshold, it is an engine failure.

2. The method for detecting engine combustion efficiency according to claim 1, characterized in that: The similarity is, , where is the similarity between the output power sequence and the exhaust efficiency sequence, , Respectively The output power sequence , The output power value, , Respectively The first in the exhaust efficiency series , The exhaust efficiency value is is the length of the sequence, is an exponential function with the natural constant e as base, is the normalization function.

3. The method for detecting engine combustion efficiency according to claim 1, characterized in that: The similarity is calculated using a dynamic time warping algorithm.

4. The method for detecting engine combustion efficiency according to claim 1, characterized in that: Calculate the The accuracy of clusters , , where is the total number of similarities in the clusters, For cluster The first in the exhaust efficiency sequence corresponding to the similarity Exhaust efficiency, For cluster The output power sequence corresponding to the similarity The combustion efficiency corresponding to the output power is is the normalization function.

5. The method for detecting engine combustion efficiency according to claim 1, characterized in that: Calculate the The accuracy of clusters , , where is the total number of similarities in the clusters, For cluster The length of the output power sequence corresponding to the similarity is For cluster The first in the exhaust efficiency sequence corresponding to the similarity elements, For cluster The output power sequence corresponding to the similarity The combustion efficiency corresponding to each element is is the normalization function.

6. The method for detecting engine combustion efficiency according to claim 1, characterized in that: The K-Means clustering algorithm is used to cluster the similarities between each output power sequence and the exhaust efficiency sequence and pressure efficiency sequence.

7. The method for detecting engine combustion efficiency according to claim 1, characterized in that: The distance is the Manhattan distance.

8. The method for detecting engine combustion efficiency according to claim 1, characterized in that: It also includes standardizing the output power sequence, exhaust efficiency sequence, pressure efficiency sequence and combustion efficiency sequence.

9. The method for detecting engine combustion efficiency according to claim 1, characterized in that: It also includes triggering an alarm signal through an audible and visual alarm when an engine failure occurs.

10. The method for detecting engine combustion efficiency according to claim 1, characterized in that: The threshold is set to 0.5.

Images