Unit valve flow monitoring optimization method and system based on data mining

Abstract

The invention belongs to the technical field of unit valve monitoring optimization, and discloses a unit valve flow monitoring optimization method and system based on data mining, and the method comprises the steps: obtaining the operation data in the operation process of a unit valve, constructing a flow model based on the structural characteristics of the valve and the flow state of fluid, and carrying out the recognition of the flow state of the fluid based on a characteristic through-flow area recognition method; analyzing the linearity of the flow in the valve; based on a data mining clustering algorithm, performing data extraction on a linearity analysis result, and clustering valve flow characteristics according to different modes; a valve flow characteristic curve is divided into a plurality of sections, independent linearity optimization is carried out on each section, the linear error of each section is analyzed, and a key area of valve performance optimization is identified. According to the method, the flow model is accurately constructed by combining the structural characteristics of the valve and the flow state of the fluid, and the linearity of the flow is analyzed by adopting a characteristic through-flow area identification method, so that the precision and the response capability in the valve adjusting process are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unit valve monitoring and optimization, and in particular to a unit valve flow monitoring and optimization method and system based on data mining. Background Art

[0002] Thermal power units (thermal power generating units) refer to power generation equipment that uses fossil fuels such as coal, oil, and natural gas as the main energy source, heats water to generate steam through the heat energy generated by combustion, and drives the steam turbine to generate electricity. Thermal power units are usually composed of boilers, steam turbines, generators, regulating valves, condensers, etc., and are currently one of the most common power generation methods in the world.

[0003] A regulating valve is a valve used to control the flow, pressure or temperature of fluid in a system. In a thermal power unit, it is mainly used to adjust the flow of boiler water, steam and gas flow in related systems to ensure the operating stability and efficiency of the thermal power unit. The flow characteristic curve of the regulating valve describes the control effect of the valve on the fluid flow at different openings. Specifically, the flow characteristic curve of the regulating valve is usually expressed as the relationship between the valve opening (i.e. the degree of opening of the valve) and the flow. According to different flow requirements, the valve opening will also change accordingly. The valve flow characteristic curve can be divided into linear, equal percentage, fast opening and other types, and each type of valve characteristic curve is suitable for different adjustment requirements. For example, linear characteristics are usually suitable for occasions that require smooth adjustment, while equal percentage characteristics are suitable for occasions that require a larger control range.

[0004] In order to ensure that the thermal power unit can operate within the optimal efficiency range, improve power generation efficiency, and reduce energy waste, it is necessary to monitor the flow characteristic curve of the thermal power unit's regulating valve. Since the flow characteristics of the thermal power unit's regulating valve are directly related to the changes in steam flow and pressure, the stability and accuracy of these parameters are crucial to the unit's thermal efficiency, fuel consumption, and emission control. If the flow characteristic curve of the regulating valve deviates from the optimal state, it will lead to inaccurate flow control, thereby affecting the thermal balance of the system, which will in turn cause safety hazards in the operation of the unit, and may even cause equipment wear, reduced energy efficiency, and economic losses.

[0005] At present, the monitoring and optimization technology of the flow characteristic curve of the regulating valve of thermal power units has made certain progress, but the current technology still has certain shortcomings. First, the performance monitoring of the regulating valves of many thermal power units still relies on manual inspections, and the degree of automation is low, resulting in slow response speed and inability to achieve real-time and accurate adjustment. Secondly, in the process of optimizing the flow characteristic curve of the regulating valve, although the model can be optimized through existing algorithms, due to the complex structure of the thermal power unit itself and many influencing factors, the optimization effect may be limited by the accuracy of the model and the integrity of the data. Furthermore, with the aging of the equipment of the thermal power unit, the flow characteristics of the regulating valve may deviate greatly. Traditional optimization methods often fail to provide sufficiently effective response solutions when facing aging or damage of equipment.

[0006] Therefore, how to provide a method and system for optimizing the flow monitoring of unit valves based on data mining is a problem that needs to be solved urgently. Summary of the invention

[0007] The embodiment of the present invention provides a method and system for optimizing the flow monitoring of valves of a unit based on data mining, so as to solve the above-mentioned technical problems existing in the prior art.

[0008] In order to have a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended to be a general review, nor is it intended to identify key / important components or to delineate the scope of protection of these embodiments. Its only purpose is to present some concepts in a simple form as a preface to the detailed description that follows.

[0009] According to a first aspect of an embodiment of the present invention, a method for optimizing unit valve flow monitoring based on data mining is provided.

[0010] In one embodiment, the unit valve flow monitoring optimization method based on data mining includes:

[0011] Obtain the operating data of the unit valve during operation, build a flow model based on the valve structure characteristics and fluid flow state, and analyze the linearity of the flow in the valve based on the characteristic flow area identification method;

[0012] Based on the data mining clustering algorithm, the linearity analysis results are extracted, the valve flow characteristics are clustered according to different modes, and the flow characteristics under different operating conditions are identified;

[0013] The valve flow characteristic curve is divided into multiple segments, and independent linearity optimization is performed on each segment. The linear error of each segment is analyzed to identify the key areas for valve performance optimization. Based on the results of the piecewise linear analysis, the flow characteristic parts with abnormal areas are optimized and adjusted.

[0014] In one embodiment, the operation data of the valve of the unit is obtained during operation, a flow model based on the valve structure characteristics and the fluid flow state is constructed, and the linearity of the flow in the valve is analyzed based on the characteristic flow area identification method, including:

[0015] The operating data of the valve is collected through the sensor, and the operating data is denoised, outliers are eliminated and standardized, and the timestamp information of the collection time of the operating data is recorded;

[0016] Obtain the structural characteristics and fluid flow state of the unit valve, build a flow model of the valve body and internal fluid based on the valve structural characteristics and fluid flow state, and use the flow model to display the characteristic flow area and flow behavior of the valve under different opening degrees and working conditions;

[0017] Based on the characteristic flow area of ​​the valve, the flow rate at different openings is calculated, the linear relationship between different flow values ​​and openings is analyzed, the linearity analysis results of the valve flow are obtained, and the existing linearity errors are identified.

[0018] In one embodiment, based on the characteristic flow area of ​​the valve, the flow rate at different openings is calculated, and the linear relationship between different flow values ​​and the opening is analyzed to obtain the linearity analysis result of the valve flow rate. The identification of the existing linearity error includes:

[0019] The relationship between different flow rates and valve openings is obtained through the flow model, the flow characteristic curves under different openings are drawn, and the flow relationship expression including the characteristic flow area and the fluid pressure difference is established;

[0020] By measuring the corresponding data of actual flow rate and opening, minimizing the error between actual flow rate and theoretical flow rate is taken as the correction target, and the characteristic flow area identification method is used to correct the actual flow characteristics. By fitting the relationship between valve opening and actual flow rate, the characteristic flow area and flow characteristics of the valve are obtained.

[0021] The goodness of fit between the valve opening and the flow rate is calculated, and the linearity of the flow curve is judged by the goodness of fit. If the goodness of fit value is closer to 1, it means that the relationship between the flow rate and the opening is close to linear. If the goodness of fit value is smaller, it means that the relationship between the flow rate and the opening is close to linearity, and there is a linearity error.

[0022] In one embodiment, the calculation formula for goodness of fit is:

[0023]

[0024] In the formula, R 2 Indicates goodness of fit; Q i Indicates the actual flow rate; represents the fitted value; represents the average value of the flow rate; n represents the number of data points.

[0025] In one embodiment, based on a data mining clustering algorithm, data is extracted from the linearity analysis results, valve flow characteristics are clustered according to different modes, and flow characteristic patterns under different operating conditions are identified, including:

[0026] Perform linearity analysis on each operating condition of the valve to evaluate whether the relationship between different openings and flow rates is close to linearity, and combine the linearity analysis results with the original operating data to form a characteristic data set;

[0027] Based on the clustering algorithm optimized by the gap statistics algorithm, the gap statistics of different cluster numbers are calculated, and the cluster number that maximizes the gap statistics is selected as the optimal cluster number;

[0028] Hierarchical clustering is used to cluster the feature data in the feature data set. The data is divided according to the optimal number of clusters. Each data point is assigned to a cluster. Different clusters represent different traffic characteristic patterns. The cluster center represents the typical traffic characteristics of the cluster, and the traffic characteristic rules of different clusters are identified.

[0029] In one embodiment, based on a clustering algorithm optimized by a gap statistics algorithm, calculating gap statistics of different cluster numbers, and selecting the cluster number that maximizes the gap statistics as the optimal cluster number includes:

[0030] Select a range of cluster numbers, randomly select different cluster numbers within the range, and for each cluster number, use the K-means clustering algorithm to cluster the feature data set and calculate the total clustering error;

[0031] Randomly generate a reference data set with the same features as the feature data set, cluster the reference data set, repeatedly generate multiple reference data sets, and obtain the clustering errors of the multiple reference data sets;

[0032] The gap statistics of each reference data set are calculated, and the number of clusters that maximizes the gap statistics is selected as the optimal number of clusters.

[0033] In one embodiment, the calculation formula of the gap statistic is:

[0034]

[0035] Where Gap(k) represents the gap statistic; B represents the number of reference data sets; SSE k Represents the clustering error of the feature data set; SSE b,k represents the clustering error of the b-th reference dataset.

[0036] In one embodiment, the valve flow characteristic curve is divided into multiple segments, and independent linearity optimization is performed on each segment. The linear error of each segment is analyzed to identify the key area for valve performance optimization, and the flow characteristic part with abnormal areas is optimized and adjusted according to the segmented linear analysis results, including:

[0037] Based on the clustering results of valve flow characteristics, the valve flow characteristic curve is divided into multiple segments using the opening range and flow change law. Each segment contains a group of points with similar flow characteristics, and the corresponding division point is selected according to the coordinates of the cluster center point.

[0038] Perform linear regression analysis on the relationship between the flow rate and the opening of each segment, use the linear regression model to fit the data points of each segment, obtain the fitting equation, and calculate the fitting goodness of fit and linear error of each segment;

[0039] The segments where both the goodness of fit and the linear error meet the preset thresholds are marked as key areas, and a feedback adjustment mechanism is introduced to adjust the valve opening control signal according to the linear error size of each segment. The valve opening is controlled using a PID controller.

[0040] Any segment that does not meet the preset threshold in the goodness of fit and linear error is marked as an abnormal area. The nonlinear relationship between flow and opening in the abnormal area is compensated by local optimization and nonlinear compensation. The linearity of each segment is gradually optimized through multiple iterations.

[0041] In one embodiment, any segment that does not meet the preset threshold value in the goodness of fit and the linear error is marked as an abnormal area, and the nonlinear relationship between the flow rate and the opening in the abnormal area is compensated by using local optimization and nonlinear compensation. The linearity of each segment is gradually optimized through multiple iterations, including:

[0042] Perform local optimization on the abnormal area and replace the linear model with a nonlinear regression model to fit the flow and opening relationship of the corresponding section in the abnormal area, thus reducing the fitting error in the abnormal area;

[0043] For nonlinear problems that cannot be solved by linear or polynomial regression, nonlinear compensation method is introduced to correct nonlinear errors caused by external factors by adding dynamic compensation mechanism;

[0044] The fitting error and nonlinear error are optimized through iteration. After each iteration, the goodness of fit and residual are recalculated until the preset value is reached and the iteration is stopped.

[0045] According to a second aspect of an embodiment of the present invention, a unit valve flow monitoring and optimization system based on data mining is provided.

[0046] In one embodiment, the unit valve flow monitoring and optimization system based on data mining includes:

[0047] The data acquisition and processing module is used to obtain the operating data of the unit valve during operation, build a flow model based on the valve structure characteristics and fluid flow state, and analyze the linearity of the flow in the valve based on the characteristic flow area identification method;

[0048] Cluster analysis and identification module, which is used to extract data from linearity analysis results based on data mining clustering algorithm, cluster valve flow characteristics according to different modes, and identify flow characteristics under different operating conditions;

[0049] The segmented optimization and adjustment module is used to divide the valve flow characteristic curve into multiple segments, perform independent linear optimization on each segment, analyze the linear error of each segment, identify the key areas for valve performance optimization, and optimize and adjust the flow characteristic parts in abnormal areas based on the segmented linear analysis results.

[0050] The technical solution provided by the embodiment of the present invention may include the following beneficial effects: by combining the valve structure characteristics and the fluid flow state, the flow model is accurately constructed, and the characteristic flow area identification method is used to analyze the linearity of the flow, which effectively improves the accuracy and responsiveness of the valve adjustment process; based on the clustering algorithm of data mining, the flow characteristic law under different operating conditions can be extracted from a large amount of operating data, and the intelligent identification and optimization of the valve flow characteristic mode can be realized. In addition, by segmenting the flow characteristic curve and performing independent optimization, the linear error of each segment can be accurately identified and adjusted in a targeted manner, solving the nonlinear and abnormal area problems that are difficult to handle with traditional methods, and innovatively introducing local optimization and nonlinear compensation technology, while improving the valve performance, ensuring the linear relationship between the flow and the opening in the entire adjustment range, greatly improving the stability and reliability of the system.

[0051] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0053] Figure 1 is a flow chart of a method for optimizing the flow monitoring of valves of a unit based on data mining according to an exemplary embodiment;

[0054] Figure 2It is a principle block diagram of a unit valve flow monitoring and optimization system based on data mining according to an exemplary embodiment. DETAILED DESCRIPTION

[0055] As used herein, the term "plurality" means two or more than two, unless otherwise specified.

[0056] In this document, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B means: A or B.

[0057] In this article, the term "and / or" is a description of the association relationship between objects, indicating that three relationships may exist. For example, A and / or B means: A or B, or, A and B.

[0058] It should be understood that, although the various steps in the flow chart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the figure may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0059] Each module in the device or system of the present application can be implemented in whole or in part by software, hardware, or a combination thereof. The above modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above modules.

[0060] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0061] Figure 1 An embodiment of a method for optimizing flow monitoring of valves of a unit based on data mining according to the present invention is shown.

[0062] In this optional embodiment, the unit valve flow monitoring optimization method based on data mining includes:

[0063] Step S101, obtaining the operation data of the valve of the unit during operation, constructing a flow model based on the valve structure characteristics and the fluid flow state, and analyzing the linearity of the flow in the valve based on the characteristic flow area identification method.

[0064] Step S102: Based on the data mining clustering algorithm, data is extracted from the linearity analysis results, the valve flow characteristics are clustered according to different modes, and the flow characteristic rules under different operating conditions are identified.

[0065] Step S103, divide the valve flow characteristic curve into multiple segments, perform independent linearity optimization on each segment, analyze the linear error of each segment, identify the key area for valve performance optimization, and optimize and adjust the flow characteristic part with abnormal areas based on the piecewise linear analysis results.

[0066] In this optional embodiment, the operation data of the valve of the unit is obtained during operation, a flow model based on the valve structure characteristics and the fluid flow state is constructed, and the linearity of the flow in the valve is analyzed based on the characteristic flow area identification method, including: collecting the operation data of the valve during operation through the sensor, denoising, removing abnormal values ​​and standardizing the operation data, and recording the timestamp information of the acquisition time of the operation data. The operation data includes valve opening, flow, pressure, temperature and fluid type; obtaining the structural characteristics and fluid flow state of the valve of the unit, and constructing a flow model formed by the valve body and the internal fluid based on the valve structure characteristics and the fluid flow state, and using the flow model to show the characteristic flow area and flow behavior of the valve under different openings and working conditions; based on the characteristic flow area of ​​the valve, calculating the flow at different openings, analyzing the linear relationship between different flow values ​​and the opening, obtaining the linearity analysis result of the valve flow, and identifying the existing linearity error.

[0067] It should be noted that the characteristic flow area (A) of a valve usually depends on the valve opening (x) and its geometric characteristics. Different types of valves (such as ball valves, butterfly valves, gate valves, etc.) have different flow characteristics, so the relationship between opening and flow is also different. The flow state of a fluid is usually described by a flow equation.

[0068] Assuming that the fluid is incompressible, the relationship between the flow rate Q, the characteristic flow area A and the flow velocity v is: Q = A·v. In addition, the flow velocity v is related to factors such as the pressure difference and density of the fluid. Common flow equations include the Bernoulli equation and the modified Darcy-Weisbach equation. For simplicity, if the effects of pressure drop and fluid friction are ignored, it can be assumed that there is a linear relationship between the flow velocity and the opening.

[0069] In this optional embodiment, based on the characteristic flow area of ​​the valve, the flow rate at different openings is calculated, the linear relationship between different flow values ​​and the opening is analyzed, and the linearity analysis result of the valve flow is obtained. The identification of the existing linearity error includes: obtaining the relationship between different flow rates and valve openings through a flow model, drawing flow characteristic curves at different openings, and establishing a flow relationship expression including the characteristic flow area and the fluid pressure difference; by measuring the corresponding data of the actual flow rate and the opening, minimizing the error between the actual flow rate and the theoretical flow rate as the correction target, using the characteristic flow area identification method to correct the actual flow characteristics, and by fitting the relationship between the valve opening and the actual flow rate, the characteristic flow area and flow characteristics of the valve are obtained; the goodness of fit between the valve opening and the flow rate is calculated, and the linearity of the flow curve is judged by the goodness of fit. If the goodness of fit value is closer to 1, it indicates that the relationship between the flow rate and the opening is close to linear. If the goodness of fit value is smaller, it is determined that the relationship between the flow rate and the opening is close to linearity, and there is a linearity error.

[0070] It should be noted that for the characteristic flow area identification, the actual flow characteristics of the valve are corrected by measuring the corresponding data of the actual flow rate and the opening, and comparing them with the theoretical value. Usually, the characteristic flow area identification method can fit the relationship between the valve opening and the actual flow rate through an optimization algorithm (such as the least squares method), and then derive the flow area and flow characteristics of the valve at different openings, and the flow area identification goal is to minimize the error between the actual flow rate and the theoretical flow rate.

[0071] For flow linearity analysis, flow linearity analysis is used to determine whether the relationship between flow and opening is close to the ideal linear relationship. Ideally, there should be a linear relationship between flow and opening, but in reality, nonlinear phenomena often occur due to factors such as valve design and fluid flow.

[0072] Specifically, if R 2 The value is close to 1, indicating that the relationship between flow rate and opening is relatively linear; if R 2 The lower the value, the less obvious the linear relationship is and the larger the linearity error is. 2 Areas with large linearity errors can be identified as areas with large linearity errors. For opening intervals with obvious linearity errors, the control mechanism of the valve can be adjusted or the flow characteristics can be corrected through optimization strategies. For example, the design parameters of the valve can be adjusted according to the identification results, or a compensation algorithm can be introduced to adjust the flow characteristics. Using nonlinear control strategies or piecewise linear optimization methods, the relationship between flow and opening in different opening intervals can be gradually adjusted to improve the flow linearity of the valve.

[0073] In this optional embodiment, the flow relationship expression is:

[0074]

[0075] In the formula, Q represents the flow rate; C represents the flow coefficient; A represents the characteristic flow area under the opening x; ΔP represents the fluid pressure difference.

[0076] The formula for calculating goodness of fit is:

[0077]

[0078] In the formula, R 2 Indicates goodness of fit; Q i Indicates the actual flow rate; represents the fitted value; represents the average value of the flow rate; n represents the number of data points.

[0079] In this optional embodiment, based on the data mining clustering algorithm, data extraction is performed on the linearity analysis results, the valve flow characteristics are clustered according to different modes, and the flow characteristic laws under different operating conditions are identified, including: performing linearity analysis on each operating condition of the valve, evaluating whether the relationship between different openings and flows is close to linear, and combining the linearity analysis results with the original operating data to form a feature data set; based on a clustering algorithm optimized by a gap statistics algorithm, the gap statistics of different cluster numbers are calculated, and the cluster number that maximizes the gap statistics is selected as the optimal number of clusters; hierarchical clustering is used to cluster the feature data in the feature data set, and the data is divided according to the optimal number of clusters, and each data point is assigned to a cluster, different clusters represent different flow characteristic modes, and the cluster center represents the typical flow characteristic of the cluster, and the flow characteristic laws of different clusters are identified.

[0080] It should be noted that the improved data mining clustering method based on the gap statistics algorithm (Gap Statistic Algorithm, GSA) is mainly used to determine the optimal number of clusters, and cluster the valve flow characteristics according to different modes by extracting data from the linearity analysis results, so as to identify the flow characteristics under different operating conditions. The gap statistics method evaluates the quality of the clustering results by comparing the clustering effect of the data set with the effect of the reference distribution.

[0081] In this optional embodiment, a clustering algorithm optimized based on a gap statistics algorithm calculates gap statistics for different numbers of clusters, and selects the number of clusters that maximizes the gap statistics as the optimal number of clusters, including: selecting a range of cluster numbers, randomly selecting different numbers of clusters within the range, and for each number of clusters, clustering the feature data set using a K-means clustering algorithm, and calculating the total clustering error; randomly generating a reference data set that has the same features as the feature data set, and clustering the reference data set, repeatedly generating multiple reference data sets, and obtaining clustering errors of multiple reference data sets; calculating the gap statistics for each reference data set, and selecting the number of clusters that maximizes the gap statistics as the optimal number of clusters.

[0082] In this optional embodiment, the calculation formula of the gap statistic is:

[0083]

[0084] Where Gap(k) represents the gap statistic; B represents the number of reference data sets; SSE k Represents the clustering error of the feature data set; SSE b,k represents the clustering error of the b-th reference dataset.

[0085] In this optional embodiment, the valve flow characteristic curve is divided into multiple segments, and independent linearity optimization is performed on each segment. The linear error of each segment is analyzed, the key area for valve performance optimization is identified, and the flow characteristic part with abnormal areas is optimized and adjusted according to the segmented linear analysis results, including: based on the clustering results of the valve flow characteristics, the valve flow characteristic curve is divided into multiple segments using the opening range and the flow change law, each segment contains a group of points with similar flow characteristics, and the corresponding division point is selected according to the coordinates of the cluster center point; the relationship between the flow and the opening of each segment is subjected to linear regression analysis, The linear regression model is used to fit the data points of each segment to obtain the fitting equation, and the goodness of fit and linear error of each segment are calculated; the segments whose goodness of fit and linear error both meet the preset thresholds are marked as key areas, and a feedback adjustment mechanism is introduced to adjust the valve opening control signal according to the linear error size of each segment, and the valve opening is controlled using a PID controller; any segment whose goodness of fit and linear error do not meet the preset thresholds is marked as an abnormal area, and local optimization and nonlinear compensation methods are used to compensate for the nonlinear relationship between flow and opening in the abnormal area, and the linearity of each segment is gradually optimized through multiple iterations.

[0086] It should be noted that the nonlinear error of each segment can be feedback-adjusted through the control strategy. For example, an adaptive control algorithm can be designed to adjust the valve opening control signal according to the error size of each segment. The control algorithm can be used for adjustment. If it is found that the error of a certain segment is large, the valve adjustment mechanism or control algorithm can be adjusted. Common adjustment methods include: 1. PID controller adjustment: optimize the parameters of the PID controller so that it can more accurately adjust the valve opening within the segment. 2. Fuzzy control: In the nonlinear area of ​​flow and opening, fuzzy control methods can be used to generate adaptive rules according to different operating conditions.

[0087] In some cases, the flow characteristics of the valve can be described by some mathematical models, such as polynomial fitting or exponential model. If it is found that the linearity of a certain segment is poor, you can try to introduce polynomial regression or other nonlinear regression models to describe the flow characteristics of the segment. In this way, each segment can be optimized to make its flow characteristics closer to the ideal linearity.

[0088] For some sections that are difficult to adjust with simple control algorithms, you can consider changing the mechanical structure of the valve, such as improving the rigidity of the regulating mechanism or adjusting the sensitivity of the valve opening. Through these adjustments, the valve response within a specific opening range is made more linear.

[0089] In this optional embodiment, any segment that does not meet the preset threshold value in the goodness of fit and the linear error is marked as an abnormal area, and the nonlinear relationship between the flow and the opening in the abnormal area is compensated by local optimization and nonlinear compensation. The linearity of each segment is gradually optimized through multiple iterations, including: local optimization of the abnormal area, replacing the linear model with a nonlinear regression model to fit the relationship between the flow and the opening of the corresponding segment in the abnormal area, and reducing the fitting error in the abnormal area; for nonlinear problems that cannot be solved by linear or polynomial regression, a nonlinear compensation method is introduced to correct the nonlinear error caused by external factors by adding a dynamic compensation mechanism; the fitting error and the nonlinear error are optimized through iteration, and the goodness of fit and the residual are recalculated after each iteration until the preset value is reached, and the iteration is stopped.

[0090] It should be noted that based on the linearity analysis, areas with poor linearity and large residuals are identified. Usually, these areas may be manifestations of unstable valve performance or nonlinear problems caused by operating conditions (such as fluid pressure, temperature, etc.).

[0091] For these abnormal areas, more sophisticated optimization methods can be used, such as: 1. Local optimization: For the abnormal segments, local optimization processing is performed, which may involve changing the valve adjustment strategy, updating the control algorithm parameters or adjusting the valve working mode. 2. Nonlinear compensation: Compensation for the nonlinear relationship between flow and opening can be performed by adding nonlinear correction factors in the control system or using dynamic compensation technology.

[0092] During the optimization process, multiple iterations may be required to gradually improve the linearity of each segment. Through the feedback mechanism, the optimization scheme of each segment is continuously adjusted until a relatively ideal flow-opening relationship is achieved.

[0093] Figure 2 An embodiment of a unit valve flow monitoring and optimization system based on data mining of the present invention is shown.

[0094] In this optional embodiment, the unit valve flow monitoring and optimization system based on data mining includes:

[0095] The data acquisition and processing module 201 is used to obtain the operation data of the unit valve during operation, construct a flow model based on the valve structure characteristics and the fluid flow state, and analyze the linearity of the flow in the valve based on the characteristic flow area identification method.

[0096] The cluster analysis identification module 202 is used to extract data from the linearity analysis results based on the data mining clustering algorithm, cluster the valve flow characteristics according to different modes, and identify the flow characteristic rules under different operating conditions.

[0097] The segmented optimization and adjustment module 203 is used to divide the valve flow characteristic curve into multiple segments, perform independent linear optimization on each segment, analyze the linear error of each segment, identify the key area for valve performance optimization, and optimize and adjust the flow characteristic part with abnormal areas based on the segmented linear analysis results.

[0098] The present invention is not limited to the structures which have been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for optimizing the flow monitoring of unit valves based on data mining, characterized in that: include: Obtain the operating data of the unit valve during operation, build a flow model based on the valve structure characteristics and fluid flow state, and analyze the linearity of the flow in the valve based on the characteristic flow area identification method; Based on the data mining clustering algorithm, the linearity analysis results are extracted, the valve flow characteristics are clustered according to different modes, and the flow characteristics under different operating conditions are identified; The valve flow characteristic curve is divided into multiple segments, and independent linearity optimization is performed on each segment. The linear error of each segment is analyzed to identify the key areas for valve performance optimization. Based on the results of the piecewise linear analysis, the flow characteristic parts with abnormal areas are optimized and adjusted.

2. The method for optimizing the flow rate monitoring of valves of a unit based on data mining according to claim 1 is characterized in that: The acquisition of the operating data of the unit valve during operation, the construction of a flow model based on the valve structure characteristics and the fluid flow state, and the analysis of the linearity of the flow in the valve based on the characteristic flow area identification method include: Collecting the operating data of the valve in operation through the sensor, denoising, removing outliers and standardizing the operating data, and recording the timestamp information of the collection time of the operating data; Obtain the structural characteristics and fluid flow state of the unit valve, build a flow model formed by the valve body and internal fluid based on the valve structural characteristics and fluid flow state, and use the flow model to display the characteristic flow area and flow behavior of the valve under different opening degrees and working conditions; Based on the characteristic flow area of ​​the valve, the flow rate at different openings is calculated, the linear relationship between different flow values ​​and openings is analyzed, the linearity analysis results of the valve flow are obtained, and the existing linearity errors are identified.

3. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 2 is characterized in that: The characteristic flow area of ​​the valve is used to calculate the flow rate at different openings, analyze the linear relationship between different flow values ​​and openings, obtain the linearity analysis result of the valve flow, and identify the existing linearity errors, including: The relationship between different flow rates and valve openings is obtained through the flow model, the flow characteristic curves under different openings are drawn, and the flow relationship expression including the characteristic flow area and the fluid pressure difference is established; By measuring the corresponding data of actual flow rate and opening, minimizing the error between actual flow rate and theoretical flow rate is taken as the correction target, and the characteristic flow area identification method is used to correct the actual flow characteristics. By fitting the relationship between valve opening and actual flow rate, the characteristic flow area and flow characteristics of the valve are obtained. The goodness of fit between the valve opening and the flow rate is calculated, and the linearity of the flow curve is judged by the goodness of fit. If the goodness of fit value is closer to 1, it means that the relationship between the flow rate and the opening is close to linear. If the goodness of fit value is smaller, it means that the relationship between the flow rate and the opening is close to linearity, and there is a linearity error.

4. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 3 is characterized in that: The calculation formula for the goodness of fit is: In the formula, R 2 Indicates goodness of fit; Q i Indicates the actual flow rate; represents the fitted value; represents the average value of the flow rate; n represents the number of data points.

5. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 1 is characterized in that: The data mining clustering algorithm is used to extract data from the linearity analysis results, cluster the valve flow characteristics according to different modes, and identify the flow characteristics under different operating conditions, including: Perform linearity analysis on each operating condition of the valve to evaluate whether the relationship between different openings and flow rates is close to linearity, and combine the linearity analysis results with the original operating data to form a characteristic data set; Based on the clustering algorithm optimized by the gap statistics algorithm, the gap statistics of different cluster numbers are calculated, and the cluster number that maximizes the gap statistics is selected as the optimal cluster number; Hierarchical clustering is used to cluster the feature data in the feature data set, and the data is divided according to the optimal number of clusters. Each data point is assigned to a cluster. Different clusters represent different traffic characteristic patterns. The cluster center represents the typical traffic characteristics of the cluster, and the traffic characteristic rules of different clusters are identified.

6. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 5 is characterized in that: The clustering algorithm optimized based on the gap statistics algorithm calculates the gap statistics of different cluster numbers, and selects the cluster number that maximizes the gap statistics as the optimal cluster number, including: Select a range of cluster numbers, randomly select different cluster numbers within the range, and for each cluster number, use the K-means clustering algorithm to cluster the feature data set and calculate the total clustering error; Randomly generate a reference data set with the same features as the feature data set, cluster the reference data set, repeatedly generate multiple reference data sets, and obtain the clustering errors of the multiple reference data sets; The gap statistics of each of the reference data sets are calculated, and the number of clusters that maximizes the gap statistics is selected as the optimal number of clusters.

7. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 6 is characterized in that: The calculation formula of the gap statistics is: Where Gap(k) represents the gap statistic; B represents the number of reference data sets; SSE k Represents the clustering error of the feature data set; SSE b,k represents the clustering error of the b-th reference dataset.

8. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 1, characterized in that: The method of dividing the valve flow characteristic curve into multiple segments, performing independent linearity optimization on each segment, analyzing the linear error of each segment, identifying the key area for valve performance optimization, and optimizing and adjusting the flow characteristic part in the abnormal area according to the segmented linear analysis results includes: Based on the clustering results of valve flow characteristics, the valve flow characteristic curve is divided into multiple segments using the opening range and flow change law. Each segment contains a group of points with similar flow characteristics, and the corresponding division point is selected according to the coordinates of the cluster center point. Perform linear regression analysis on the relationship between the flow rate and the opening of each segment, use the linear regression model to fit the data points of each segment, obtain the fitting equation, and calculate the fitting goodness of fit and linear error of each segment; The segments where both the goodness of fit and the linear error meet the preset thresholds are marked as key areas, and a feedback adjustment mechanism is introduced to adjust the valve opening control signal according to the linear error size of each segment. The valve opening is controlled using a PID controller. Any segment that does not meet the preset threshold in the goodness of fit and linear error is marked as an abnormal area. The nonlinear relationship between flow and opening in the abnormal area is compensated by local optimization and nonlinear compensation. The linearity of each segment is gradually optimized through multiple iterations.

9. The method for optimizing the flow rate monitoring of unit valves based on data mining according to claim 8 is characterized in that: The segment that does not meet the preset threshold value in the goodness of fit and linear error is marked as an abnormal area, and the nonlinear relationship between the flow and the opening in the abnormal area is compensated by local optimization and nonlinear compensation, and the linearity of each segment is gradually optimized through multiple iterations, including: Perform local optimization on the abnormal area and replace the linear model with a nonlinear regression model to fit the flow and opening relationship of the corresponding section in the abnormal area, thus reducing the fitting error in the abnormal area; For nonlinear problems that cannot be solved by linear or polynomial regression, nonlinear compensation method is introduced to correct nonlinear errors caused by external factors by adding dynamic compensation mechanism; The fitting error and nonlinear error are optimized through iteration. After each iteration, the goodness of fit and residual are recalculated until the preset value is reached and the iteration is stopped.

10. A unit valve flow monitoring and optimization system based on data mining, characterized in that: include: The data acquisition and processing module is used to obtain the operating data of the unit valve during operation, build a flow model based on the valve structure characteristics and fluid flow state, and analyze the linearity of the flow in the valve based on the characteristic flow area identification method; Cluster analysis and identification module, which is used to extract data from linearity analysis results based on data mining clustering algorithm, cluster valve flow characteristics according to different modes, and identify flow characteristics under different operating conditions; The segmented optimization and adjustment module is used to divide the valve flow characteristic curve into multiple segments, perform independent linear optimization on each segment, analyze the linear error of each segment, identify the key areas for valve performance optimization, and optimize and adjust the flow characteristic parts in abnormal areas based on the segmented linear analysis results.

Images