Oil depot whole-process comprehensive management monitoring method and system

Abstract

The invention discloses an oil depot full-process comprehensive management monitoring method and system, particularly relates to the field of oil depot full-process comprehensive management monitoring, and aims to solve the problems that data correlation analysis among different links is insufficient, and potential systematic risks are difficult to comprehensively identify and predict. A transfer entropy method is introduced into upstream real-time data and historical association mapping, information transmission strength between upstream and downstream key parameters is quantitatively analyzed, and an effective causal link is established; in combination with dynamic time warping and synchronous entropy correlation degree quantity, trace deviation characteristics are matched with downstream key parameter cluster information, and a deviation propagation path is accurately drawn; strictly judging a non-stationary conduction trend from the aspects of time and structure by using a multi-scale complexity increment index and cluster structure change measurement; on the basis, a mixed integer nonlinear programming optimization model is adopted, a downstream link pre-intervention strategy is quantitatively formulated, and efficient solving of parameter adjustment and early warning signal level decision is achieved.

Description

Technical Field

[0001] The present invention relates to the field of comprehensive management and monitoring of the entire process of an oil depot, and more specifically, to a method and system for comprehensive management and monitoring of the entire process of an oil depot. Background Art

[0002] With the continuous growth of energy demand and the increasing diversification of oil types, the operation and management of modern oil depots are becoming increasingly complex, covering multiple key links such as the reception, storage, transportation, blending, scheduling and delivery of oil products. In order to ensure the continuity, safety and efficiency of oil supply, the oil depot management system needs to achieve real-time monitoring and comprehensive management of the entire process. This not only requires the system to have a high degree of automation and intelligence capabilities, but also to be able to process massive amounts of multi-source data to ensure coordination and optimized operation between various links. At present, many oil depots rely on decentralized monitoring equipment and traditional data processing methods, making it difficult to achieve efficient coordination and risk warning of the overall process.

[0003] However, the existing oil depot full-process management and monitoring technology faces many challenges in practical applications. First, the insufficient data correlation analysis between different links makes it difficult to fully identify and predict potential systemic risks; second, the monitoring system responds slowly when dealing with complex working conditions, making it difficult to detect and handle abnormal deviations in a timely manner; in addition, traditional data processing methods are inefficient when processing high-dimensional, multi-variable real-time data, which limits the implementation of intelligent decision support. These technical problems not only affect the safety and efficiency of oil depot operations, but also increase operating costs and management difficulties. Therefore, it is urgent to develop a comprehensive management and monitoring method and system that can fully integrate data from various processes, realize deep correlation analysis, and have real-time early warning capabilities, so as to improve the intelligence level of the entire process of the oil depot and the scientific nature of operational management. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method and system for comprehensive management and monitoring of the entire process of an oil depot. By introducing the transfer entropy method in the upstream real-time data and historical correlation mapping, the information transmission intensity between the upstream and downstream key parameters is quantitatively analyzed, and an effective causal link is established; by combining dynamic time warping and synchronous entropy correlation measurement, the trace deviation characteristics are matched with the downstream key parameter cluster information, and the deviation propagation path is accurately outlined; through the multi-scale complexity increment index and cluster structure change measurement, the identified deviation characteristics and cluster contours are multi-dimensionally examined, and the non-stationary conduction trend is strictly determined from the time and structure level; on this basis, a mixed integer nonlinear programming optimization model is used to quantify the pre-intervention strategy for the downstream link, and efficient solution of parameter adjustment and warning signal level decision under multiple constraints is achieved; thereby ensuring that potential risks are identified, quantitatively predicted and effectively suppressed as soon as possible, significantly improving the warning and control capabilities of the entire process monitoring of the oil depot, comprehensively solving the problem of insufficient data correlation analysis between different links, and improving the safety and operational efficiency of the system, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for comprehensive management and monitoring of the entire process of an oil depot, comprising the steps of:

[0007] Obtain the real-time data of upstream process parameters and the historical correlation mapping matrix to obtain the upstream parameter feature data set to be monitored;

[0008] Perform deviation feature identification on the upstream parameter feature data set to generate a deviation feature profile set and a downstream key parameter cluster profile corresponding to the deviation feature profile;

[0009] According to the deviation characteristic profile and the downstream key parameter cluster profile, it is judged whether the trace deviation presents a non-stationary conduction trend;

[0010] If the slight deviation shows a non-stationary transmission trend, the pre-correction plan for the downstream link will be determined based on the pre-intervention strategy. The pre-correction plan includes the quantitative value of key parameter adjustment and the quantitative level of the early warning signal.

[0011] In a preferred embodiment, the real-time data of the upstream process parameters and the historical correlation mapping matrix are obtained to obtain the upstream parameter feature data set to be monitored. The specific processing logic is described as follows:

[0012] A1. First, at the data level, the upstream oil pipeline data is used as the basic input variable set, and the downstream data is used as the target response variable set to construct a multidimensional data structure; the multidimensional data structure is time-aligned and the parameters are normalized to make each parameter dimension in a comparable space with the same dimension; then, the causal direction measurement method based on conditional probability density distribution is used to construct the association mapping matrix based on transfer entropy: Where X represents a single upstream parameter or parameter combination, Y represents a downstream parameter, p(·) is the joint probability density function, and TE X→Y Indicates the amount of information transferred from the upstream parameter X to the downstream parameter Y.

[0013] A2. Then, the association channels with high transfer entropy values ​​are selected in the association mapping matrix to identify the key potential influencing transmission paths; specifically, if the amount of information transmitted by a certain upstream parameter to the downstream parameter exceeds the corresponding threshold, the corresponding upstream parameter and its association path will be marked as key monitoring objects.

[0014] A3. After the processing is completed, an association mapping matrix with quantitative causal orientation and path strength calibration is output, from which the upstream parameter feature data set to be monitored is extracted; each upstream parameter in the upstream parameter feature data set is marked as an upstream key parameter.

[0015] In a preferred embodiment, based on the determined upstream parameter feature data set, deviation feature recognition is carried out to generate a deviation feature profile set and its corresponding downstream key parameter cluster profile; the specific processing logic is as follows:

[0016] B1. First, for each upstream key parameter in the upstream parameter feature data set, the dynamic time warping algorithm is applied to match and analyze its real-time data sequence with the preset benchmark operation mode; by calculating the dynamic time warping distance between the real-time data sequence and the benchmark mode, if the dynamic time warping distance exceeds the preset threshold, it is marked that the upstream key parameter in the corresponding time period has deviation characteristics.

[0017] B2. Secondly, for each marked deviation time period, extract the data changes of the downstream parameters in the corresponding time window; adopt the synchronization correlation method based on synchronization entropy to evaluate the response intensity and change trend of the downstream parameters in the deviation period; specifically, calculate the synchronization entropy value of the downstream parameters in the deviation time window, and compare it according to the preset correlation threshold. If the synchronization entropy value of the downstream parameters in the deviation time window does not exceed the correlation threshold, it is considered that the downstream parameters have a significant response to the deviation of the upstream key parameters in the corresponding time window and are marked as downstream key parameters.

[0018] B3. After identifying the downstream key parameters, cluster analysis is performed to form a cluster profile of the downstream key parameters; the specific steps are as follows:

[0019] Feature vector construction: For each affected downstream key parameter, construct its response feature vector in multiple deviation time windows. The response feature vector contains the synchronous entropy value in each time window.

[0020] Similarity measurement: A similarity measurement method based on mutual information is used to calculate the similarity between each downstream key parameter and other downstream key parameters;

[0021] Cluster analysis: Based on the calculated mutual information matrix, the spectral clustering algorithm is used to cluster the downstream key parameters;

[0022] According to the clustering results, each cluster is defined to contain a set of downstream key parameters with similar response patterns; each cluster profile describes the overall response characteristics of the downstream key parameters when the upstream key parameters deviate.

[0023] B4. Finally, the identified deviation features are paired with their corresponding downstream key parameter cluster profiles to form a deviation feature profile set C S and its corresponding downstream key parameter cluster profile C Y Each pair Contains the deviation characteristic profile of upstream key parameters within a specific deviation time period and its downstream key parameter clustering Response profile.

[0024] In a preferred embodiment, in the deviation feature profile set and the downstream key parameter cluster profile, each pair The deviation performance of upstream key parameters corresponding to a specific period of time and its response pattern on the cluster of downstream key parameters; in order to determine whether the trace deviation presents a non-stationary transmission trend, it is necessary to proceed from the following steps:

[0025] C1. Temporal embedding and feature stability quantification:

[0026] Will Nested and arranged in time series order to form a multi-period profile sequence: For each pair Combining the association mapping matrix with the downstream key parameter cluster structure, the following metric term Γ is defined k : In the formula, is the upstream key parameter X i For downstream key parameter Y j The transfer entropy value of The structural weight is obtained by weighting the internal similarity of the downstream key parameter cluster and the synchronous entropy characteristics.

[0027] C2. Construction of multi-period non-stationarity determination criteria:

[0028] In order to determine the non-stationary characteristics, the multi-period sequence analysis method is used to analyze the sequence {Γ 1 ,Γ 2 ,…,Γ n}Measure non-stationarity; use the complexity increment discrimination index based on multi-scale time-frequency embedding; define the multi-scale feature complexity increment index ΔΛ to capture the change direction and amplitude of the measurement items at different time levels: Among them, f(·,·) is a function with monotonicity analysis ability, which is used to measure the deviation direction and degree of information structure change of measurement item values ​​at different scales l in multiple time periods, and L is the number of analysis scale layers; if the multi-scale feature complexity increment index continues to grow positively at multiple consecutive time levels, it indicates that there is a gradual structural deviation in the conduction trend in the time dimension.

[0029] C3. Result feedback and decision support:

[0030] If the multi-scale feature complexity increment index exceeds the critical threshold, it means that the conduction of slight deviations has a non-stationary conduction trend.

[0031] In a preferred embodiment, under the premise that non-stationarity already exists, it is necessary to determine the pre-correction scheme of the downstream link based on the pre-intervention strategy, so as to quantitatively adjust the downstream key parameters and determine the quantitative level of the early warning signal, thereby forming effective intervention and risk suppression measures for the subsequent links, which is carried out from the following steps:

[0032] D1. Construction of pre-intervention strategy decision model:

[0033] Define the downstream key parameters to be adjusted and the corresponding warning signal levels as decision variables, solve them through multi-objective optimization strategies, and make the decision output meet the requirements of reducing the impact of non-stationary conduction trends, controlling adjustment costs, and improving system safety margins; define the decision objective function as follows: Where: Z(Y j ) is the downstream key parameter Y j The quantified value of the modified gain indicates the contribution of the parameter adjustment to the reduction of the non-stationary conduction intensity; is the safety margin weighting factor; U(X i ) is the quantitative value of the resource cost required to implement indirect regulation of the upstream key parameter associated channel; α and β are used to regulate the weight ratio of gain and cost in the objective function.

[0034] D2. Solving multiple constraints and determining the quantification level of early warning signals:

[0035] In the solution process, the mixed integer nonlinear programming algorithm and distributed iterative optimization method are used to ensure that the decision variables, including the adjustment value of the downstream key parameters and the quantization level of the early warning signal, are within the feasible domain; the quantization level of the early warning signal is a discrete variable, corresponding to different levels of operation intervention instructions and response time requirements; the quantization adjustment value of the downstream key parameters is a continuous variable, which must meet the process safety limit and the upper and lower bounds defined;

[0036] As the iterative solution proceeds, the algorithm continuously checks the improvement of the decision objective function and the satisfaction of the constraints; when the decision objective function value can no longer be optimized, the optimal solution is output as a pre-correction plan.

[0037] D3. Solution output:

[0038] After the optimization solution is completed, the quantified parameter adjustment value and warning signal level are output to the downstream control system.

[0039] An oil depot full-process integrated management and monitoring system, comprising: a data acquisition module, a feature recognition module, a trend analysis module and a correction decision module;

[0040] Data acquisition module: obtains the real-time data of upstream process parameters and the historical association mapping matrix, obtains the upstream parameter feature data set to be monitored, and passes the acquired association mapping matrix and upstream parameter feature data set to the feature recognition module;

[0041] Feature recognition module: performs deviation feature recognition on the upstream parameter feature data set, generates a deviation feature profile set and a downstream key parameter cluster profile corresponding to the deviation feature profile, and transmits the downstream key parameter cluster profile to the trend analysis module;

[0042] Trend analysis module: Based on the deviation characteristic profile and the downstream key parameter cluster profile, it determines whether the trace deviation presents a non-stationary conduction trend, and transmits the judgment result to the correction decision module;

[0043] Correction decision module: If the slight deviation shows a non-stationary transmission trend, the pre-correction plan for the downstream link will be determined based on the pre-intervention strategy. The pre-correction plan includes the quantitative value of parameter adjustment and the quantitative level of the early warning signal.

[0044] The technical effects and advantages of the oil depot full-process comprehensive management and monitoring method and system of the present invention are as follows:

[0045] By introducing the transfer entropy method in the upstream real-time data and historical correlation mapping, the information transmission intensity between the upstream and downstream key parameters is quantitatively analyzed, and an effective causal link is established; by combining dynamic time warping and synchronous entropy correlation measurement, the trace deviation characteristics are matched with the downstream key parameter cluster information, and the deviation propagation path is accurately outlined; through the multi-scale complexity increment index and cluster structure change measurement, the identified deviation characteristics and cluster contours are examined in multiple dimensions, and the non-stationary conduction trend is strictly determined from the time and structure levels; on this basis, the optimization model is used to quantify the pre-intervention strategy for the downstream links, and efficient solution of parameter adjustment and warning signal level decision-making is achieved; thereby making up for the problem of insufficient data correlation analysis between links, ensuring the early identification, quantitative prediction and effective suppression of potential risks, and improving the warning and control capabilities of the entire process monitoring of the oil depot. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flow chart of a method for comprehensive management and monitoring of the entire process of an oil depot according to the present invention;

[0047] Figure 2 The present invention is a structural schematic diagram of an oil depot full-process comprehensive management and monitoring system. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Embodiment 1: Figure 1 The present invention provides a comprehensive management and monitoring method for the entire process of an oil depot, comprising:

[0050] Obtain the real-time data of upstream process parameters and the historical correlation mapping matrix to obtain the upstream parameter feature data set to be monitored;

[0051] Perform deviation feature identification on the upstream parameter feature data set to generate a deviation feature profile set and a downstream key parameter cluster profile corresponding to the deviation feature profile;

[0052] According to the deviation characteristic profile and the downstream key parameter cluster profile, it is judged whether the trace deviation presents a non-stationary conduction trend;

[0053] If the slight deviation shows a non-stationary transmission trend, the pre-correction plan for the downstream link will be determined based on the pre-intervention strategy. The pre-correction plan includes the quantitative value of key parameter adjustment and the quantitative level of the early warning signal.

[0054] In the comprehensive management of the whole process, the potential impact of the micro-deviation of the upstream oil pipeline parameters can easily spread to the key links downstream through the association path. In order to achieve effective early warning and intervention, a set of precise association mapping systems must be established in the data acquisition and initial analysis stages. In this regard, the real-time data of the upstream parameters and the historical association mapping matrix are organically integrated to form the upstream parameter feature data set to be monitored, and the potential transmission path of the upstream micro-deviation to the key parameter cluster of the subsequent process is clarified through specific data processing logic and quantitative indicators.

[0055] Obtain the real-time data of upstream process parameters and the historical correlation mapping matrix to obtain the upstream parameter feature data set to be monitored. The specific processing logic is as follows:

[0056] A1. First, at the data level, the upstream oil pipeline data, such as pressure, flow, and temperature parameters, are used as the basic input variable set, and the downstream data, such as tank area volume changes, feed valve switching frequency, mixing density and viscosity parameters, and outbound scheduling timing parameters, are used as the target response variable set to construct a multidimensional data structure. The multidimensional data structure is time-aligned and parameters are normalized so that each parameter dimension is in a comparable space with the same dimension. Then, the causal pointing measurement method based on conditional probability density distribution is used to construct the association mapping matrix based on transfer entropy: Where X represents a single upstream parameter or parameter combination, Y represents a downstream parameter, p(·) is the joint probability density function, and TE X→Y It represents the amount of information transferred from the upstream parameter X to the downstream parameter Y. By calculating the conditional information gain of the upstream parameter to the downstream parameter, the direction and strength of the causal relationship are screened out. For the three core parameters of upstream pressure, flow rate, and temperature, the transfer entropy is calculated and summarized one by one with the target response variable set, and the results are included in the association mapping matrix, so that each element of the association mapping matrix corresponds to the information transmission intensity and directionality between a pair of upstream and downstream parameters.

[0057] A2. Then, high transfer entropy value associated channels are selected in the association mapping matrix to identify key potential influencing conduction paths. Specifically, if the amount of information transmitted by a certain upstream parameter to the downstream parameter exceeds the corresponding threshold, the corresponding upstream parameter and its associated path are marked as key monitoring objects. On this basis, a multivariate regression analysis of the composite parameter set is performed to ensure that the sensitivity and accuracy of trace deviation detection can be improved by combining feature dimensions without relying on a single parameter. This regression analysis is based on the transfer entropy value and the related information gain level, and the parameter weights are solved by iteratively optimizing the directional constraints that minimize the fitting error.

[0058] A3. After completing the above processing, an association mapping matrix with quantitative causal orientation and path strength calibration is output, and a set of upstream parameter feature data sets to be monitored are extracted from it. Each upstream parameter in the upstream parameter feature data set is marked as an upstream key parameter; the upstream parameter feature data set explicitly includes the upstream key parameter feature values ​​and their causal association information that have the greatest potential impact on the downstream links. In the subsequent steps, when the deviation characteristics of the key parameters contained in the upstream parameter feature data set are identified, the upstream and downstream association path information provided by the association mapping matrix will be directly referenced to ensure that when a slight deviation occurs, the judgment can be based on a clear causal link rather than an isolated data point, thereby laying the foundation for the implementation of the overall early warning mechanism.

[0059] Through the above processing logic, upstream parameters and downstream key parameters are established in a rigorous framework of information transmission and causal association, and transfer entropy is used as the core metric to quantitatively characterize the potential conduction path. On this basis, the output correlation mapping matrix and the upstream parameter feature data set to be monitored provide an accurate and efficient data basis for the subsequent steps of deviation feature identification, non-stationary conduction judgment and pre-correction plan formulation.

[0060] In the oil depot's full-process integrated management and monitoring system, ensuring that the slight deviation of upstream oil pipeline parameters can be timely identified and effectively transmitted to downstream key links is the core of preventing the spread of systemic risks. After step one constructs the correlation mapping matrix between upstream parameters and downstream key parameters through the transfer entropy method, and screens out upstream parameters with significant information transmission tendencies, step two aims to conduct in-depth analysis of these upstream parameter feature data sets to be monitored. Specifically, it is necessary to identify the deviation characteristics that occur in actual operation and correspond them to the change profiles of downstream key parameters, forming a clear set of deviation feature profiles and their associated downstream key parameter cluster profiles, which provides a basis for subsequent non-stationary transmission trend judgment.

[0061] Based on the determined upstream parameter feature data set, deviation feature identification is carried out to generate a deviation feature profile set and its corresponding downstream key parameter cluster profile. The specific processing logic is as follows:

[0062] B1. First, for each upstream key parameter in the upstream parameter feature data set, the dynamic time warping algorithm is applied to match and analyze its real-time data sequence with the preset benchmark operation mode. The dynamic time warping algorithm can effectively handle the nonlinear alignment problem in the time series and identify small but continuous deviation patterns. By calculating the dynamic time warping distance between the real-time data sequence and the benchmark mode, if the dynamic time warping distance exceeds the preset threshold, it is marked that the upstream key parameter has deviation characteristics in the corresponding time period.

[0063] B2. Secondly, for each marked deviation time period, extract the data changes of the downstream parameters in the corresponding time window. Use the synchronous correlation method based on synchronous entropy to evaluate the response intensity and change trend of the downstream parameters during the deviation period. Specifically, calculate the synchronous entropy value of the downstream parameter in the deviation time window, and compare it according to the preset correlation threshold. If the synchronous entropy value of the downstream parameter in the deviation time window does not exceed the correlation threshold, it is considered that the downstream parameter has a significant response to the deviation of the upstream key parameter in the corresponding time window and is marked as a downstream key parameter.

[0064] B3. After identifying the downstream key parameters, cluster analysis is required to form a cluster profile of the downstream key parameters. The specific steps are as follows:

[0065] Feature vector construction: For each affected downstream key parameter, construct its response feature vector in multiple deviation time windows. The response feature vector contains the synchronous entropy value in each time window.

[0066] Similarity measurement: A similarity measurement method based on mutual information is used to calculate the similarity between each downstream key parameter and other downstream key parameters;

[0067] Cluster analysis: Based on the calculated mutual information matrix, the spectral clustering algorithm is used to cluster the downstream key parameters. Spectral clustering can effectively handle the nonlinear structure of high-dimensional data and is suitable for complex downstream key parameter response patterns. The specific steps include:

[0068] A similarity graph is constructed, where nodes represent downstream key parameters and edge weights are the mutual information between the corresponding downstream key parameters and other downstream key parameters.

[0069] Compute the Laplacian matrix of the graph.

[0070] Solve the first rated eigenvectors of the Laplace matrix to form the feature space.

[0071] The K-means algorithm is applied in the feature space to assign downstream key parameters into different clusters.

[0072] According to the clustering results, each cluster is defined to contain a set of downstream key parameters with similar response patterns. Each cluster profile describes the overall response characteristics of the downstream key parameters when the upstream key parameters deviate.

[0073] B4. Finally, the identified deviation features are paired with their corresponding downstream key parameter cluster profiles to form a deviation feature profile set C S and its corresponding downstream key parameter cluster profile C Y Each pair Contains the deviation characteristic profile of upstream key parameters within a specific deviation time period and its downstream key parameter clustering By constructing these profile pairs, we can systematically capture the specific impact patterns of upstream micro-deviations on downstream key parameter clusters, providing detailed data support for subsequent non-stationary conduction trend judgments.

[0074] The dynamic time warping algorithm is used to accurately identify the deviation characteristics of upstream key parameters, and the synchronous entropy synchronous correlation measurement method is combined to evaluate the response strength of downstream key parameters, and the spectral clustering algorithm is further used to generate the cluster profile of downstream key parameters. This step ensures that when there is a slight deviation in the upstream key parameters, its specific manifestation in time and space can be accurately identified, and its impact intensity and pattern on the downstream key links can be quantified. The generated deviation feature profile set and downstream key parameter cluster profile provide the necessary quantitative basis for the judgment of non-stationary conduction trends in subsequent steps, further enhancing the early warning capability and risk prevention and control level of the comprehensive management and monitoring system of the whole process.

[0075] In the comprehensive management and monitoring system of the whole process, ensuring that the slight deviation of the upstream key parameters can be timely identified whether it will be transmitted to the downstream links and present non-stationary characteristics is the key to achieving accurate early warning and precise intervention. In the previous step, the upstream parameter feature data set to be monitored has been obtained and the association mapping matrix has been constructed. Based on this data set, a clear set of deviation feature profiles and their corresponding downstream key parameter cluster profiles have been identified. Now it is necessary to judge from these profile pairs whether the conduction of the slight deviation is out of the stable mode, that is, to judge whether it produces a quantifiable non-stationary conduction trend in time and correlation structure, so as to lay a quantitative foundation for subsequent pre-intervention and decision correction.

[0076] In the set of deviation feature profiles and the downstream key parameter cluster profiles, each pair The deviation performance of upstream key parameters in a specific period and its response pattern on the cluster of downstream key parameters. In order to determine whether the slight deviation presents a non-stationary transmission trend, it is necessary to proceed from the following steps:

[0077] C1. Temporal embedding and feature stability quantification:

[0078] Will Nested and arranged in time series order to form a multi-period profile sequence: For each pair Extract multiple causal indicator sets that can reflect the characteristics of deviation conduction. For example, combining the association mapping matrix with the downstream key parameter cluster structure, the following metric term Γ can be defined k : In the formula, is the upstream key parameter X iFor downstream key parameter Y j The transfer entropy value of It is the structural weight obtained by weighting the internal similarity of the downstream key parameter cluster and the synchronous entropy characteristics. The metric is used to characterize the unified performance of the overall information intensity of the deviation transmission and the downstream cluster response intensity during this period.

[0079] C2. Construction of multi-period non-stationarity determination criteria:

[0080] In order to determine the non-stationary characteristics, the multi-period sequence analysis method is used to analyze the sequence {Γ 1 ,Γ 2 ,…,Γ n} to measure non-stationarity. Here, we can use a complexity increment discrimination index based on multi-scale time-frequency embedding. For example, we define a multi-scale feature complexity increment index ΔΛ to capture the change direction and amplitude of the measurement item at different time levels: Among them, f(·,·) is a function with monotonic analysis ability, which is used to measure the deviation direction and information structure change degree of the measurement item value at different scales l between multiple time periods, and L is the number of analysis scales. If the multi-scale feature complexity increment index continues to grow positively at multiple consecutive time levels, it indicates that there is a gradual structural deviation in the conduction trend in the time dimension.

[0081] C3. Result feedback and decision support:

[0082] Once it is determined that the conduction of the micro-deviation has a non-stationary conduction trend, that is, the multi-scale feature complexity increment index exceeds the critical threshold, the judgment result is output to the subsequent steps. The subsequent process will provide a pre-correction plan for the downstream link based on the non-stationary judgment result through the pre-intervention strategy decision model, including the quantitative value of the key parameter adjustment and the quantitative level of the early warning signal.

[0083] By using multi-scale complexity incremental analysis and correlation structure topology change measurement methods, the deviation characteristic profile and downstream key parameter cluster profile sequence can be strictly examined at the quantitative and structural levels. When the judgment results show that the slight deviation presents a non-stationary transmission trend, it can ensure that the subsequent downstream link pre-correction plan based on the pre-intervention strategy is provided with accurate and traceable decision-making basis. With this rigorous logic and innovative quantitative means, the prediction ability and intervention efficiency of the full-process comprehensive management and monitoring system are greatly enhanced.

[0084] In the comprehensive management and monitoring of the entire process of the oil depot, when the upstream key parameters show a non-stationary transmission trend, a pre-correction plan needs to be quickly formulated and implemented to prevent the potential risk from spreading further. In the previous step: establish an association map through transfer entropy, and select the upstream key parameter feature data set to be monitored; identify the deviation feature profile and the downstream key parameter cluster profile; based on these profiles, combine the time series multi-scale complexity increment index and the cluster structure diversity change measurement to rigorously determine the non-stationary characteristics of the deviation transmission.

[0085] At present, under the premise of clarifying that non-stationarity already exists, it is necessary to determine the pre-correction plan for the downstream links based on the pre-intervention strategy, so as to quantitatively adjust the downstream key parameters and determine the quantitative level of the early warning signal, and then form effective intervention and risk suppression measures for the subsequent links. It needs to be carried out from the following steps:

[0086] D1. Construction of pre-intervention strategy decision model:

[0087] The downstream key parameters to be adjusted and the corresponding warning signal levels are defined as decision variables, and solved through multi-objective optimization strategies to ensure that the decision output can simultaneously meet the requirements of reducing the impact of non-stationary conduction trends, controlling adjustment costs, and improving system safety margins. Define the decision objective function as follows:

[0088] Where:

[0089] Z(Y j ) is the downstream key parameter Y j The quantitative value of the correction gain indicates the contribution of the parameter adjustment to reducing the intensity of non-stationary transmission. This quantitative value is calculated by evaluating the impact of the downstream key parameters on the overall stability of the system before and after the parameter adjustment, combined with historical data and real-time monitoring results. The quantitative value of the correction gain reflects the effectiveness of the adjustment measures in reducing the risk transmission process, ensuring that the adjustment operation can improve the system's protection capabilities in a targeted manner, thereby optimizing the safety and reliability of the entire process.

[0090] It is a safety margin weighting coefficient, which is used to balance the importance and safety sensitivity of different parameters; this weighting coefficient is determined based on a comprehensive assessment of the key role of each downstream key parameter in the overall operation of the system and its impact on safety performance. By introducing the safety margin weighting coefficient, resources and adjustment efforts can be reasonably allocated in the optimization objective function to ensure that key parameters with high importance and high sensitivity are given higher priority, thereby achieving the best balance between safety and efficiency in the multi-parameter control process.

[0091] U(X i) is the quantitative value of the resource cost required to implement indirect regulation of the upstream key parameter-related channels; it is specifically used to measure the various types of resources consumed in the process of parameter adjustment, including but not limited to energy consumption, operating costs, equipment wear and maintenance costs, etc. This quantitative value is obtained through a detailed analysis and calculation of the actual implementation cost of the control measures, aiming to take resource utilization efficiency into consideration during the optimization process. Through the quantitative value of resource cost, the decision-making model can minimize resource consumption and operating costs while achieving system stability, thereby improving the economy and sustainability of the overall management system.

[0092] α and β are used to adjust the weight ratio of gain and cost in the objective function.

[0093] D2. Solving multiple constraints and determining the quantification level of early warning signals:

[0094] In the solution process, the mixed integer nonlinear programming (MINLP) algorithm and distributed iterative optimization methods are used to ensure that the decision variables, including the adjustment values ​​of downstream key parameters and the quantization level of early warning signals, are within the feasible domain. The quantization level of the early warning signal is a discrete variable, corresponding to different levels of operation intervention instructions and response time requirements; the quantization adjustment value of the downstream key parameter is a continuous variable, which must meet the process safety limit and the upper and lower bounds defined.

[0095] As the iterative solution proceeds, the algorithm continuously checks the improvement of the decision objective function and the satisfaction of the constraint conditions. When the system can no longer optimize the decision objective function value, it outputs the optimal solution as the pre-correction solution.

[0096] Mixed integer nonlinear programming (MINLP) is an advanced mathematical optimization method that combines the characteristics of integer programming and nonlinear programming. In MINLP, decision variables include discrete integer variables and continuous real variables, while the objective function and its constraints are nonlinear. This method can effectively solve problems with complex decision structures and nonlinear interactions, and is widely used in engineering design, energy system optimization, supply chain management and other fields. Through MINLP, the optimization model can simultaneously handle the discrete selection and nonlinear relationship of variables, so as to find the optimal solution of the objective function while satisfying all constraints. However, since MINLP problems are usually highly computationally complex and difficult to solve, it is usually necessary to use heuristic algorithms, branch and bound methods or other advanced optimization techniques to seek approximate optimal solutions in order to achieve efficient decision support.

[0097] D3. Solution output:

[0098] After the optimization solution is completed, the quantified parameter adjustment value and warning signal level are output to the downstream control system. After obtaining the solution, the downstream link performs corresponding control operations and monitors the execution results in real time during the subsequent operation cycle. Through this closed-loop feedback mechanism, the new data is looped back to the initial processing flow to dynamically correct the decision model parameters (α, β) and associated path weights of the subsequent cycle, thereby continuously improving the flexibility and accuracy of decision-making in long-term operation.

[0099] Through multi-objective optimization strategies and mixed integer nonlinear solutions, a quantitative and operational correction solution is provided for the deviation of upstream key parameters that have been determined to be non-stationary transmission trends. With the support of rigorous data foundation and clear technical logic, this solution can adapt to different degrees and types of non-stationary risks, flexibly adjust the adjustment values ​​and early warning signal levels of downstream key parameters, and realize efficient and forward-looking intervention in the comprehensive management and monitoring system of the whole process. This kind of technical path for quantitative decision-making and dynamic adaptation has significantly enhanced the system's prevention and control capabilities and the robustness of long-term operation.

[0100] Embodiment 2: Figure 2 The present invention provides an oil depot full-process comprehensive management and monitoring system, including: a data acquisition module, a feature recognition module, a trend analysis module and a correction decision module;

[0101] Data acquisition module: obtains the real-time data of upstream process parameters and the historical association mapping matrix, obtains the upstream parameter feature data set to be monitored, and passes the acquired association mapping matrix and upstream parameter feature data set to the feature recognition module;

[0102] Feature recognition module: performs deviation feature recognition on the upstream parameter feature data set, generates a deviation feature profile set and a downstream key parameter cluster profile corresponding to the deviation feature profile, and transmits the downstream key parameter cluster profile to the trend analysis module;

[0103] Trend analysis module: Based on the deviation characteristic profile and the downstream key parameter cluster profile, it determines whether the trace deviation presents a non-stationary conduction trend, and transmits the judgment result to the correction decision module;

[0104] Correction decision module: If the slight deviation shows a non-stationary transmission trend, the pre-correction plan for the downstream link will be determined based on the pre-intervention strategy. The pre-correction plan includes the quantitative value of parameter adjustment and the quantitative level of the early warning signal.

[0105] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0106] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0107] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0108] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A comprehensive management and monitoring method for the entire process of an oil depot, characterized in that: Includes steps: Obtain the real-time data of upstream process parameters and the historical correlation mapping matrix to obtain the upstream parameter feature data set to be monitored; Perform deviation feature identification on the upstream parameter feature data set to generate a deviation feature profile set and a downstream key parameter cluster profile corresponding to the deviation feature profile; According to the deviation characteristic profile and the downstream key parameter cluster profile, it is judged whether the trace deviation presents a non-stationary conduction trend; If the slight deviation shows a non-stationary transmission trend, the pre-correction plan for the downstream link will be determined based on the pre-intervention strategy. The pre-correction plan includes the quantitative value of key parameter adjustment and the quantitative level of the early warning signal.

2. The method for comprehensive management and monitoring of the entire process of an oil depot according to claim 1 is characterized in that: Obtain the real-time data of upstream process parameters and the historical correlation mapping matrix to obtain the upstream parameter feature data set to be monitored. The specific processing logic is as follows: A1. First, at the data level, the upstream oil pipeline data is used as the basic input variable set, and the downstream data is used as the target response variable set to construct a multidimensional data structure; The multidimensional data structure is time-aligned and parameters are normalized so that each parameter dimension is in a comparable space with the same dimension. Then, the causal direction measurement method based on conditional probability density distribution is used to construct the association mapping matrix based on the transfer entropy: Where X represents a single upstream parameter or parameter combination, Y represents a downstream parameter, p(·) is the joint probability density function, and TE X→Y Indicates the amount of information transferred from the upstream parameter X to the downstream parameter Y.

3. The method for comprehensive management and monitoring of the entire process of an oil depot according to claim 2 is characterized in that: A2. Then, high transfer entropy value correlation channels are selected in the correlation mapping matrix to identify key potential impact transmission paths; specifically, if the amount of information transmission from a certain upstream parameter to a downstream parameter exceeds the corresponding threshold, the corresponding upstream parameter and its correlation path are marked as key monitoring objects; A3. After the processing is completed, an association mapping matrix with quantitative causal orientation and path strength calibration is output, from which the upstream parameter feature data set to be monitored is extracted; Each upstream parameter in the upstream parameter feature dataset is marked as an upstream key parameter.

4. The method for comprehensive management and monitoring of the entire process of an oil depot according to claim 3 is characterized in that: Based on the determined upstream parameter feature data set, deviation feature identification is carried out to generate a deviation feature profile set and its corresponding downstream key parameter cluster profile; the specific processing logic is as follows: B1. First, for each upstream key parameter in the upstream parameter feature data set, the dynamic time warping algorithm is applied to match and analyze its real-time data sequence with the preset benchmark operation mode; by calculating the dynamic time warping distance between the real-time data sequence and the benchmark mode, if the dynamic time warping distance exceeds the preset threshold, it is marked that the upstream key parameter in the corresponding time period has deviation characteristics; B2. Secondly, for each marked deviation time period, extract the data changes of the downstream parameters in the corresponding time window; adopt the synchronization correlation method based on synchronization entropy to evaluate the response intensity and change trend of the downstream parameters in the deviation period; specifically, calculate the synchronization entropy value of the downstream parameters in the deviation time window, and compare it according to the preset correlation threshold. If the synchronization entropy value of the downstream parameters in the deviation time window does not exceed the correlation threshold, it is considered that the downstream parameters have a significant response to the deviation of the upstream key parameters in the corresponding time window and are marked as downstream key parameters.

5. The method for comprehensive management and monitoring of the entire process of an oil depot according to claim 4 is characterized in that: B3. After identifying the downstream key parameters, cluster analysis is performed to form a cluster profile of the downstream key parameters; the specific steps are as follows: Feature vector construction: For each affected downstream key parameter, construct its response feature vector in multiple deviation time windows. The response feature vector contains the synchronous entropy value in each time window. Similarity measurement: A similarity measurement method based on mutual information is used to calculate the similarity between each downstream key parameter and other downstream key parameters; Cluster analysis: Based on the calculated mutual information matrix, the spectral clustering algorithm is used to cluster the downstream key parameters; Based on the clustering results, each cluster is defined to contain a set of downstream key parameters with similar response patterns; Each cluster profile describes the overall response characteristics of the downstream key parameter when the upstream key parameter deviates; B4. Finally, the identified deviation features are paired with their corresponding downstream key parameter cluster profiles to form a deviation feature profile set C S and its corresponding downstream key parameter cluster profile C Y Each pair Contains the deviation characteristic profile of upstream key parameters within a specific deviation time period and its downstream key parameter clustering Response profile.

6. A method for comprehensive management and monitoring of the entire process of an oil depot according to claim 5, characterized in that: In the set of deviation feature profiles and the downstream key parameter cluster profiles, each pair The deviation performance of upstream key parameters corresponding to a specific period of time and its response pattern on the cluster of downstream key parameters; in order to determine whether the trace deviation presents a non-stationary transmission trend, it is necessary to proceed from the following steps: C1. Temporal embedding and feature stability quantification: Will Nested and arranged in time series order to form a multi-period profile sequence: For each pair Combining the association mapping matrix with the downstream key parameter cluster structure, the following metric term Γ is defined k : In the formula, is the upstream key parameter X i For downstream key parameter Y j The transfer entropy value of The structural weight is obtained by weighting the internal similarity of the downstream key parameter cluster and the synchronous entropy characteristics.

7. A method for comprehensive management and monitoring of the entire process of an oil depot according to claim 6, characterized in that: C2. Construction of multi-period non-stationarity determination criteria: In order to determine the non-stationary characteristics, the multi-period sequence analysis method is used to analyze the sequence {Γ 1 ,Γ 2 ,…,Γ n }Measure non-stationarity; use the complexity increment discrimination index based on multi-scale time-frequency embedding; define the multi-scale feature complexity increment index ΔΛ to capture the change direction and amplitude of the measurement items at different time levels: Among them, f(·,·) is a function with monotonic analysis ability, which is used to measure the deviation direction and information structure change degree of the measurement item value at different scales l between multiple time periods, and L is the number of analysis scales. If the multi-scale feature complexity increment index continues to grow positively at multiple consecutive time levels, it indicates that there is a gradual structural deviation in the conduction trend in the time dimension. C3. Result feedback and decision support: If the multi-scale feature complexity increment index exceeds the critical threshold, it means that the conduction of slight deviations has a non-stationary conduction trend.

8. The method for comprehensive management and monitoring of the entire process of an oil depot according to claim 7 is characterized in that: At present, under the premise of clarifying that non-stationarity already exists, it is necessary to determine the pre-correction plan for the downstream links based on the pre-intervention strategy, so as to quantitatively adjust the downstream key parameters and determine the quantitative level of the early warning signal, and then form effective intervention and risk suppression measures for the subsequent links, which can be carried out from the following steps: D1. Construction of pre-intervention strategy decision model: Define the downstream key parameters to be adjusted and the corresponding warning signal levels as decision variables, solve them through multi-objective optimization strategies, and make the decision output meet the requirements of reducing the impact of non-stationary conduction trends, controlling adjustment costs, and improving system safety margins; define the decision objective function as follows: Where: Z(Y j ) is the downstream key parameter Y j The quantitative value of the modified gain indicates the contribution of the parameter adjustment to reducing the non-stationary conduction intensity; G Yj is the safety margin weighting factor; U(X i ) is the quantitative value of the resource cost required to implement indirect regulation of the upstream key parameter associated channel; α and β are used to regulate the weight ratio of gain and cost in the objective function.

9. The method for comprehensive management and monitoring of the entire process of an oil depot according to claim 8 is characterized in that: D2. Solving multiple constraints and determining the quantification level of early warning signals: In the solution process, the mixed integer nonlinear programming algorithm and distributed iterative optimization method are used to ensure that the decision variables, including the adjustment value of the downstream key parameters and the quantization level of the early warning signal, are within the feasible domain; the quantization level of the early warning signal is a discrete variable corresponding to different levels of operation intervention instructions and response time requirements; The quantitative adjustment value of the downstream key parameter is a continuous variable, which must meet the process safety limit and the defined upper and lower bound constraints; As the iterative solution proceeds, the algorithm continuously checks the improvement of the decision objective function and the satisfaction of the constraint conditions; when the decision objective function value can no longer be optimized, the optimal solution is output as a pre-correction solution; D3. Solution output: After the optimization solution is completed, the quantified parameter adjustment value and warning signal level are output to the downstream control system.

10. An oil depot full-process integrated management and monitoring system, used to implement an oil depot full-process integrated management and monitoring method as claimed in any one of claims 1 to 9, characterized in that: include: Data acquisition module, feature recognition module, trend analysis module and correction decision module; Data acquisition module: obtains the real-time data of upstream process parameters and the historical association mapping matrix, obtains the upstream parameter feature data set to be monitored, and passes the acquired association mapping matrix and upstream parameter feature data set to the feature recognition module; Feature recognition module: performs deviation feature recognition on the upstream parameter feature data set, generates a deviation feature profile set and a downstream key parameter cluster profile corresponding to the deviation feature profile, and transmits the downstream key parameter cluster profile to the trend analysis module; Trend analysis module: Based on the deviation characteristic profile and the downstream key parameter cluster profile, it determines whether the trace deviation presents a non-stationary conduction trend, and transmits the judgment result to the correction decision module; Correction decision module: If the slight deviation shows a non-stationary transmission trend, the pre-correction plan for the downstream link will be determined based on the pre-intervention strategy. The pre-correction plan includes the quantitative value of parameter adjustment and the quantitative level of the early warning signal.

Images