Thermal power plant whole-process intelligent diagnosis and self-optimization control method based on big data

Abstract

The invention relates to the crossing field of thermal power generation technology and industrial internet, in particular to a thermal power plant whole-process intelligent diagnosis and self-optimization control method based on big data. Comprising the following steps: S1, acquiring multi-source data of a thermal power plant; s2, acquiring a chemical heat independent link, a heat engine independent link and an electromechanical independent link, and constructing an energy path network by taking equipment ranges required by the respective independent links as original points; s3, acquiring a cross device set and a cross flow table in the three independent links, calculating respective cross device coupling degrees and cross flow coupling degrees, and screening out a first cross abnormal point and a second cross abnormal point of the thermal power plant; and S4, the efficiency deviation is compared with the specified deviation of the first cross abnormal point and the second cross abnormal point, and a control optimization instruction set is generated. According to the invention, the thermal power plant is diagnosed, controlled and optimized through the influence of cross devices and cross processes on different processes among the energy conversion links of the thermal power plant.

Description

Technical Field

[0001] This invention relates to the intersection of thermal power generation technology and the industrial internet, and in particular to a method for intelligent diagnosis and self-optimization control of the entire process of thermal power plants based on big data. Background Technology

[0002] As the energy industry transforms towards high efficiency, cleanliness, and intelligence, thermal power plants, as a crucial source of electricity supply, urgently need to improve their operational efficiency and intelligence level. Traditional thermal power plant operation and management mainly rely on decentralized control systems and offline analysis based on fixed thresholds or experience, resulting in the following technical bottlenecks: First, data from each subsystem (such as boiler, turbine, and generator) is isolated, lacking a unified and dynamic correlation model that reflects the continuous conversion of the entire energy process, making it difficult to diagnose the root causes of energy efficiency losses at the system level. Second, there is a lack of quantitative analysis methods for the complex physical and functional coupling relationships between links, often only able to handle single equipment or local problems, unable to identify cross-link collaborative constraints, leading to optimization measures being locally effective but globally suboptimal. Third, existing control strategies are mostly statically set or based on simple feedback, unable to adaptively adjust according to dynamic operating conditions and internal system contradictions, restricting the continuous optimization of the overall energy efficiency of the power plant and the limits of safe and economical operation.

[0003] Meanwhile, due to errors and adjustments in the previous stage of energy conversion, the cross-connected devices in thermal power plants may fail to guarantee the efficiency and stability of energy conversion in the next stage, and the probability of abnormal energy conversion cannot be reliably predicted. Therefore, there is an urgent need to develop a big data-based intelligent diagnosis and self-optimization control method for the entire process of thermal power plants. Summary of the Invention

[0004] To overcome the shortcomings of errors in cross-functional devices and processes in different stages of a thermal power plant, which lead to inefficient operation of the next stage and the inability to reliably predict the probability of anomalies, this invention provides a method for intelligent diagnosis and self-optimization control of the entire thermal power plant process based on big data.

[0005] A big data-based intelligent diagnosis and self-optimization control method for the entire process of thermal power plants includes the following steps: S1: Acquire multi-source data from the thermal power plant, including a set of conversion devices, an energy conversion sequence, and actual energy conversion efficiency; S2: Divide the thermal power plant area into independent segments according to the energy conversion sequence, obtain the independent segments of chemical heat, heat engine and electromechanical systems, and construct the energy path network of the thermal power plant with the equipment range required by each independent segment as the origin; S3: Based on the set of conversion devices and mapping relationships, obtain the set of cross-devices and cross-process tables in the three independent links, calculate the coupling degree of each cross-device and the coupling degree of each cross-process, and diagnose the first and second cross-abnormal points of the thermal power plant. S4: Based on the average efficiency deviation of the energy path network, and compared with the abnormal efficiency deviation of the first and second cross anomalies, generate a control optimization instruction set and set early warning for observation points.

[0006] Preferably, the acquisition of multi-source data from the thermal power plant includes a set of conversion devices, an energy conversion sequence, and actual energy conversion efficiency. Specifically, the set of conversion devices includes a complete set of physical devices used for energy conversion during the operation of the thermal power plant; the energy conversion sequence includes the current energy conversion process from input to output, requiring the collection of data on fuel type, feed rate, and power generation output; the energy conversion sequence includes the conversion from chemical energy to thermal energy, from thermal energy to mechanical energy, and from mechanical energy to electrical energy; and the actual energy conversion efficiency includes the actual conversion efficiency between two different energy sources, calculated using real-time measurement data.

[0007] Preferably, the step of dividing the thermal power plant area into independent segments according to the energy conversion sequence to obtain independent chemical-thermal segments, independent heat engine segments, and independent electromechanical segments, and constructing the energy path network of the thermal power plant with the equipment range required for each independent segment as the origin, includes: dividing the independent segment of chemical energy to heat energy conversion into chemical-thermal segments, the independent segment of heat energy to mechanical energy conversion into heat engine segments, and the independent segment of mechanical energy to electrical energy conversion into electromechanical segments according to the energy conversion sequence of the thermal power plant; constructing the energy path network of the thermal power plant with the three independent segments of the thermal power plant as the origin and the input parameters with different energy sources as the origin; and using the output of the three independent segments as the input of the next segment, using the actual energy conversion efficiency as the first-level edge weight and the endpoint energy conversion efficiency as the second-level edge weight; the path of the energy path network includes the actual energy transmission path of energy conversion, and the mapping relationship of the devices in the conversion process is marked on the actual energy conversion path.

[0008] Preferably, the input parameters with different energy sources as origins include: the energy conversion paths of the thermal power plant include a first energy conversion path, a second energy conversion path, and a third energy conversion path; the first energy conversion path includes the thermal power plant converting energy from chemical energy to electrical energy under normal conditions; the second energy conversion path includes the thermal power plant converting energy from thermal energy to electrical energy; the third energy conversion path includes the thermal power plant converting energy from mechanical energy to electrical energy; the different energy sources of the thermal power plant, including chemical energy, thermal energy, and mechanical energy, are obtained according to the three energy conversion paths; the input parameters are the outputs of the previous independent links; the first-level edge is the edge obtained in the energy path network from chemical energy to the electrical energy output position of the thermal power plant through the first energy conversion path; the second-level edge is the edge obtained in the energy conversion network through the second and third energy conversion paths; the endpoint energy conversion efficiency is the energy conversion efficiency in the direct conversion process from chemical energy and thermal energy to electrical energy.

[0009] Preferably, the step of marking the mapping relationship of devices during the conversion process on the actual energy conversion path includes: constructing a mapping table of conversion devices based on the set of conversion devices, the mapping table including all conversion devices and operation processes required for energy conversion on the current energy conversion path; performing topology mapping on the current energy conversion path using the mapping table; obtaining the static and dynamic attributes of the conversion devices in the mapping table using the set of conversion devices, the static attributes including the physical form of the current conversion device; the dynamic attributes including the real-time status of the current conversion device; calculating the device contribution of the current conversion device on the energy conversion path using the static and dynamic attributes of the conversion device, specifically calculating the device contribution by the ratio of the number of operation processes using the conversion device in the current operation process to the total number of operation processes; dynamically adjusting the mapping table based on the device contribution; and deleting or silencing conversion devices whose device contribution is lower than the device contribution threshold.

[0010] Preferably, the step of obtaining the cross-device sets and cross-process tables in three independent stages based on the conversion device set and mapping relationship, calculating the cross-device coupling degree and cross-process coupling degree of each, and diagnosing the first and second cross-abnormal points of the thermal power plant includes: based on the mapping relationship after deleting silent processes, the mapping relationship of the energy path network includes the various physical devices and process information required for energy conversion in the current independent stage on the current path; obtaining the cross-device set and cross-process table of the entire process in the thermal power plant based on the mapping relationship and conversion device set, wherein the cross-device set includes a set of physical devices that simultaneously serve multiple independent stages, and the cross-process table includes a stage process table that simultaneously serves multiple independent stages; and calculating the cross-device set and cross-process table using the cross-coupling degree formula. The cross-device coupling degree and cross-process coupling degree of the cross-process table include the degree of mutual confusion and influence between cross-devices and cross-processes between two independent links. A first coupling degree and a second coupling degree are preset for the cross-device set and the cross-process table. The cross-device coupling degree and cross-process coupling degree are compared with the first and second coupling degrees to diagnose the first and second cross-abnormal points in the energy path network. The first cross-abnormal point includes independent links where either the cross-device or cross-process is abnormal. The second cross-abnormal point includes independent links where both the cross-device and cross-process are abnormal. If the current energy conversion path contains both the first and second cross-abnormal points, they are marked on the current energy conversion path. The cross-coupling degree formula is: ; in, The cross-coupling degree of the cross-connect device; The degree of cross-coupling in the cross-process; This represents the theoretical maximum efficiency of the current cross-connect device; This represents the theoretical maximum efficiency of the current cross-process; The contribution of the current device to the device's performance; For cross-connect devices or cross-connect processes in independent stages and independent links Mutual information; For cross-connect devices or cross-connect processes in independent stages and independent links Information entropy; For cross-connect devices or cross-connect processes in independent stages and independent links The minimum information entropy; This is the ratio of the first-level edge weight between two processes in an energy path network to the second-level edge weight of the starting process.

[0011] Preferably, the first crossover anomaly point is defined as a situation where the coupling degree of the crossover device is higher than a first coupling degree and the coupling degree of the crossover process is lower than a second coupling degree. This includes situations where the first crossover anomaly point is a case where the coupling degree of the crossover device is higher than the first coupling degree and the coupling degree of the crossover process is lower than the second coupling degree, and a situation where the coupling degree of the crossover process is higher than the second coupling degree and the coupling degree of the crossover device is lower than the first coupling degree. The second crossover anomaly point is a case where the coupling degree of the crossover device is higher than the first coupling degree and the coupling degree of the crossover process is higher than the second coupling degree. The maximum efficiency of the crossover device and the crossover process is calculated using the efficiency formula, where the efficiency formula is: ; For cross-connect devices, The theoretical maximum efficiency is the maximum usable energy contained in a unit mass of substance. The specific enthalpy of the substance in its current state is obtained by consulting the Morrill vapor diagram based on the actual temperature and pressure of the current device. is the specific entropy of matter in its current state; For matter in environmental state Temperature and Specific enthalpy under pressure; For matter in environmental state Temperature and The specific entropy under pressure is obtained by querying the Morrill vapor diagram based on the actual temperature and pressure of the current device. For cross-processes, the theoretical maximum efficiency includes the sum of the theoretical maximum efficiency of all devices required in the current process.

[0012] Preferably, the step of generating a control optimization instruction set and setting observation point early warning based on the average efficiency deviation of the energy path network and comparing it with the abnormal efficiency deviations of the first and second cross anomalies includes: presetting a first deviation threshold and a second deviation threshold for the local and global efficiencies of the thermal power plant based on the first-level and second-level edge weights of the energy path network; calculating the average local efficiency deviation and the average global efficiency deviation of each independent link of the thermal power plant using the first and second deviation thresholds; wherein the average local efficiency deviation includes local efficiencies in each independent link where the difference from historical local efficiency exceeds the first deviation threshold, and the difference is the average local efficiency deviation; and the average global efficiency deviation includes local efficiencies in the overall energy conversion process where the difference from historical global efficiency exceeds the second deviation threshold. The global efficiency is calculated, where the difference is the average global efficiency deviation. The abnormal efficiency deviation includes abnormal local efficiency deviation and abnormal global efficiency deviation. Based on the average local efficiency deviation and average global efficiency deviation of the thermal power plant, the abnormal local efficiency deviation and abnormal global efficiency deviation of the two intersecting abnormal points are dynamically compensated, calibrated, and avoided. Dynamic calibration instructions for the thermal power plant are generated, and warnings are issued for the independent links that are avoided. The local efficiency of the thermal power plant includes the actual energy conversion efficiency between the two independent links in the thermal power plant where the intersecting abnormal point occurs. The global efficiency of the thermal power plant includes the final energy conversion efficiency obtained by the thermal power plant according to three different energy conversion paths. At the same time, observation points are set according to the positions of the first intersecting abnormal point, the second intersecting abnormal point, and historical abnormal points to make real-time predictions of the energy path network.

[0013] Preferably, the step of setting observation points based on the locations of the first crossover anomaly point, the second crossover anomaly point, and historical anomalies to perform real-time prediction of the energy path network includes: calculating the probability of anomalies occurring in the energy path network based on the locations of the first crossover anomaly point, the second crossover anomaly point, and historical anomalies; simultaneously, pre-setting observation points for the energy path network; and updating the energy path network in real time. The observation points include actual energy conversion paths of independent chemical-thermal links, independent heat-engine links, and independent electromechanical links. Real-time observation of the actual energy conversion paths is performed, and early warnings and avoidance are issued for energy path networks with high anomaly probability. The location of the observation points and the avoidance strategy are selected by comparing the magnitude of the anomaly probability, wherein the formula for calculating the anomaly probability is: ; in, This represents the probability of an anomaly occurring in the current energy conversion path. A set of nodes, including ; This is the normalization constant; Let be the set of outliers, if If it is 1, it means If an anomaly occurs at this location, If it is 0, it means No abnormalities were found at this location; For nodes The nodal potential function, For the edge The edge potential function is used to construct a stack of contradiction points. Based on the probability of occurrence of contradiction points in the energy path network, the sequence of contradiction points in the energy path network is obtained. The data in the contradiction point sequence are pushed onto the stack in ascending order. The time interval between the start of chemical energy conversion at the thermal power plant is used as the refresh cycle of the contradiction point sequence. After each refresh cycle, the contradiction point sequence stack is popped. The node where the popped contradiction point data is located is the specific location of the observation point. The thermal power plant is given an early warning based on the location of the observation point and the probability of occurrence of anomalies.

[0014] Preferably, the step of dynamically compensating for and controlling the abnormal local and global efficiencies at two intersection anomaly points based on the average local efficiency deviation and the average global efficiency deviation of the thermal power plant further includes: adjusting the values ​​of the abnormal local efficiency deviation and the abnormal global efficiency deviation to within the deviation range specified by the average local efficiency deviation and the average global efficiency deviation, while observing the change in the coupling degree formula to generate an instruction set for the energy path network and physical devices, and controlling the thermal power plant through the dynamic calibration instructions.

[0015] The beneficial effects of this invention are: 1. This invention constructs an energy path network with independent links as nodes and dynamic efficiency as edges, mapping the originally isolated equipment operation data into a unified system energy efficiency status diagram, thus laying a dynamic data foundation for global analysis and diagnosis; 2. This invention analyzes the mapping relationship of the energy path network, extracts cross-connection devices and cross-connection processes, and innovatively defines two quantitative indicators: cross-connection device coupling degree and cross-connection process coupling degree. This enables the objective and automatic screening of the first and second cross-connection anomalies that restrict the overall performance of the system, transforming the traditional empirical fault judgment into precise contradiction location based on data. 3. This invention links and compares the real-time efficiency deviation of the energy path network with the specified deviation of the identified anomalies, so that the generation of control commands directly targets the fundamental contradictions that cause system efficiency losses. It realizes the leap from parameter adjustment to intelligent decision-making to contradiction coordination, thereby driving the thermal power plant's operating state to continuously and adaptively approach the global optimum, significantly improving the energy efficiency level and the degree of intelligent operation of the entire process. Attached Figure Description

[0016] Figure 1 A schematic diagram illustrating a method for intelligent diagnosis and self-optimization control of the entire process of thermal power plants based on big data; Figure 2 This is a schematic diagram of the energy path network for a big data-based intelligent diagnosis and self-optimization control method for the entire process of thermal power plants. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example: A big data-based intelligent diagnosis and self-optimization control method for the entire process of thermal power plants, such as... Figure 1-2 As shown, it includes the following steps: S1: Acquire multi-source data from the thermal power plant, including a set of conversion devices, an energy conversion sequence, and actual energy conversion efficiency; S2: Divide the thermal power plant area into independent segments according to the energy conversion sequence, obtain the independent segments of chemical heat, heat engine and electromechanical systems, and construct the energy path network of the thermal power plant with the equipment range required by each independent segment as the origin; S3: Based on the set of conversion devices and mapping relationships, obtain the set of cross-devices and cross-process tables in the three independent links, calculate the coupling degree of each cross-device and the coupling degree of each cross-process, and diagnose the first and second cross-abnormal points of the thermal power plant. S4: Based on the average efficiency deviation of the energy path network, and compared with the abnormal efficiency deviation of the first and second cross anomalies, generate a control optimization instruction set and set early warning for observation points.

[0019] Acquiring multi-source data from a thermal power plant, including a set of conversion devices, an energy conversion sequence, and actual energy conversion efficiency, comprising: the set of conversion devices including a complete set of physical devices used for energy conversion during the operation of the thermal power plant; the energy conversion sequence including the current energy conversion process from input to output, requiring the collection of data on fuel type, feed rate, and power generation output, including conversion from chemical energy to thermal energy, from thermal energy to mechanical energy, and from mechanical energy to electrical energy; and the actual energy conversion efficiency including the actual conversion efficiency between two different energy sources, calculated through real-time measurement data.

[0020] It should be noted that the output energy parameters and the input energy parameters of each independent link are recorded according to specific time periods, constructing a directed line segment in the order of the thermal power plant's time. By setting anchor points on the current directed line segment, the thermal power plant is divided into time series. The energy conversion sequence includes the thermal power plant from source energy to terminal energy, as well as the input energy of various energy sources. The input energy includes various fuels based on chemical energy, various high-temperature solids, gases and liquids based on thermal energy, and various mechanical devices based on mechanical energy. The conversion device set is obtained by statistically analyzing all physical devices in the thermal power plant to obtain the physical devices used for energy conversion.

[0021] Based on the energy conversion sequence, the thermal power plant area is divided into independent segments, namely, chemical-thermal independent segments, heat engine independent segments, and electromechanical independent segments. Using the equipment range required for each independent segment as the origin, an energy path network for the thermal power plant is constructed. This includes: dividing the independent segments converting chemical energy to heat energy into chemical-thermal independent segments, the independent segments converting heat energy to mechanical energy into heat engine independent segments, and the independent segments converting mechanical energy to electrical energy into electromechanical independent segments. Using these three independent segments as the origin and different energy sources as the input parameters, the energy path network for the thermal power plant is constructed. The outputs of the three independent segments are used as the inputs of the next segment. The actual energy conversion efficiency is used as the first-level edge weight, and the final energy conversion efficiency is used as the second-level edge weight. The paths in the energy path network include the actual energy transmission paths of the energy conversion process, and the mapping relationships of the devices during the conversion process are marked on the actual energy conversion paths.

[0022] It should be noted that the aforementioned independent chemical-thermal link includes the specific process and period of converting chemical energy into thermal energy, as well as the change in energy state. The specific construction method of the energy path network is as follows: the independent chemical-thermal link, the independent heat engine link, the independent electromechanical link, and the final energy output of the thermal power plant are respectively taken as the four origins of the energy path network. The energy conversion sequence of the thermal power plant is chemical energy to thermal energy, thermal energy to mechanical energy, and mechanical energy to electrical energy. The first-level edge weights include the independent chemical-thermal link, the independent heat engine link, and the independent electromechanical link, and the second-level edge weights include the independent chemical-thermal link and the independent heat engine link, wherein the first-level edge weights and second-level edge weights of the independent electromechanical link are the same. The mapping relationship of the energy path network includes the functional boundary of the form conversion and the physical boundary of the main equipment. The functional boundary includes the energy state conversion process in the current energy conversion link, and the physical boundary includes the main conversion equipment required for energy conversion in the current energy conversion link.

[0023] The input parameters, with different energy sources as origins, include: the energy conversion paths of a thermal power plant, comprising a first energy conversion path, a second energy conversion path, and a third energy conversion path. The first energy conversion path includes the thermal power plant converting energy from chemical energy to electrical energy under normal conditions; the second energy conversion path includes the thermal power plant converting energy from thermal energy to electrical energy; and the third energy conversion path includes the thermal power plant converting energy from mechanical energy to electrical energy. The different energy sources of the thermal power plant, including chemical energy, thermal energy, and mechanical energy, are obtained based on these three energy conversion paths. The input parameters are the outputs of the previous independent stage. The first-level edge is the edge obtained in the energy path network through the first energy conversion path, converting from chemical energy to the electrical energy output position of the thermal power plant. The second-level edge is the edge obtained in the energy conversion network through the second and third energy conversion paths. The endpoint energy conversion efficiency is the energy conversion efficiency in the direct conversion process from chemical energy and thermal energy to electrical energy.

[0024] It should be noted that the input sources of the independent chemical-thermal link include various combustible chemical solids in the boiler; the independent heat engine link includes frictional heat energy, chemical heat energy, and other wind-powered heat energy; the independent electromechanical link includes gas mechanical work, gravity mechanical work, and wind-powered mechanical work; the input parameters of the independent chemical-thermal link are obtained through the theoretical chemical energy of various chemical fuels at the boiler inlet; the output parameters of the independent chemical-thermal link are obtained through the steam flow meter and pressure transmitter at the boiler outlet; and the output parameters of the independent heat engine link are obtained through the dynamometer and speed sensor at the turbine-generator coupling.

[0025] The mapping relationship of devices during the conversion process is marked on the actual energy conversion path, including: constructing a mapping table of conversion devices based on the set of conversion devices, the mapping table including all conversion devices and operation processes required for energy conversion on the current energy conversion path; performing topology mapping on the current energy conversion path using the mapping table; obtaining the static and dynamic attributes of the conversion devices in the mapping table through the set of conversion devices, the static attributes including the physical form of the current conversion device; the dynamic attributes including the real-time status of the current conversion device; calculating the device contribution of the current conversion device on the energy conversion path using the static and dynamic attributes, specifically calculating the device contribution by the ratio of the number of operation processes using the conversion device in the current operation process to the total number of operation processes; dynamically adjusting the mapping table based on the device contribution; and deleting or silencing conversion devices whose device contribution is lower than the device contribution threshold.

[0026] It should be noted that the device contribution threshold is obtained by averaging the contribution of all devices throughout the entire process of a thermal power plant. The formula for device contribution is: ; in, The contribution of the device to the current stage. The frequency of use of the device in the current stage. This represents the total frequency of all devices in the current stage. For devices The remaining functional capabilities of the component after failure. The overall design functionality of the current stage, and For fixed weight parameters, the default values ​​are 0.5 and 1.5. The specific implementation method is to calculate the average device contribution of all devices in the entire process of the thermal power plant, set the device contribution threshold as the average device contribution, and if the device contribution of the conversion device in the current independent link is less than the average device contribution, then the current device is deleted from the mapping table of the energy path network.

[0027] Based on the conversion device set and mapping relationship, obtain the cross-device set and cross-process table for three independent links, calculate the cross-device coupling degree and cross-process coupling degree for each, and diagnose the first and second cross-abnormal points of the thermal power plant, including: based on the mapping relationship after deleting silent processes, the mapping relationship of the energy path network includes the various physical devices and process information required for energy conversion in the current independent link on the current path; obtain the cross-device set and cross-process table for the entire process in the thermal power plant based on the mapping relationship and conversion device set, the cross-device set includes the physical device set that serves multiple independent links simultaneously, and the cross-process table includes the process table that serves multiple independent links simultaneously; calculate the cross-device set and cross-process table using the cross-coupling degree formula. The cross-device coupling degree and cross-process coupling degree of the process table include the degree of mutual confusion and influence between cross-devices and cross-processes between two independent links. A first coupling degree and a second coupling degree are preset for the cross-device set and the cross-process table. The cross-device coupling degree and cross-process coupling degree are compared with the first and second coupling degrees to diagnose the first and second cross-abnormal points in the energy path network. The first cross-abnormal point includes independent links where either the cross-device or cross-process is abnormal. The second cross-abnormal point includes independent links where both the cross-device and cross-process are abnormal. If the current energy conversion path contains both the first and second cross-abnormal points, they are marked on the current energy conversion path. The cross-coupling degree formula is: ; in, The cross-coupling degree of the cross-connect device; The degree of cross-coupling in the cross-process; This represents the theoretical maximum efficiency of the current cross-connect device; This represents the theoretical maximum efficiency of the current cross-process; The contribution of the current device to the device's performance; For cross-connect devices or cross-connect processes in independent stages and independent links Mutual information; For cross-connect devices or cross-connect processes in independent stages and independent links Information entropy; For cross-connect devices or cross-connect processes in independent stages and independent links The minimum information entropy; This is the ratio of the first-level edge weight between two processes in an energy path network to the second-level edge weight of the starting process.

[0028] It should be noted that the first and second coupling degrees are calculated based on the historical average values ​​of the coupling degrees of cross-connected devices and cross-process coupling degrees throughout the entire thermal power plant process. For the efficiency of cross-connected devices to theoretically convert all of their energy into useful work, As an independent segment and independent links The degree of interdependence between cross-connect devices and cross-connect processes is specifically addressed by implementing a method where the cross-connect devices are known to be in independent stages. In this state, independent steps can be reduced. The magnitude of uncertainty is determined by the following formulas for calculating mutual information and information entropy: in, for Information entropy in the current process; For a single random variable The probability of taking a specific value; For random variables and random variables Simultaneously take specific value pairs The possibility; for The specific implementation method is as follows: collect the current independent link From the time series data, obtain the time series of the current stage, based on specific values. Calculate the probability of the ratio of data in a time series; for The specific implementation method is as follows: simultaneously collect independent data. and independent links The time-series data is used to obtain a first sequence and a second sequence. The first and second sequences are then discretely divided. The first and second sequences are further divided hierarchically according to their sequence attributes. The number of data in each level is then statistically analyzed, and the probability is calculated. The theoretical maximum efficiency includes the limit efficiency that the current device or process can achieve under the current hardware and thermodynamic boundaries, excluding the influence of equipment design, material properties, and environmental conditions.

[0029] The first crossover anomaly is defined as a situation where the coupling degree of the crossover device is higher than a first coupling degree and the coupling degree of the crossover process is lower than a second coupling degree. This includes situations where the first crossover anomaly is defined as either a crossover device coupling degree higher than the first coupling degree and a crossover process coupling degree lower than the second coupling degree, or a crossover process coupling degree higher than the second coupling degree and a crossover device coupling degree lower than the first coupling degree. The second crossover anomaly is defined as a situation where the coupling degree of the crossover device is higher than the first coupling degree and the crossover process coupling degree is higher than the second coupling degree. The maximum efficiency of the crossover device and crossover process is calculated using the efficiency formula, where the efficiency formula is: ; For cross-connect devices, The theoretical maximum efficiency is the maximum usable energy contained in a unit mass of substance. The specific enthalpy of the substance in its current state is obtained by consulting the Morrill vapor diagram based on the actual temperature and pressure of the current device. is the specific entropy of matter in its current state; For matter in environmental state Temperature and Specific enthalpy under pressure; For matter in environmental state Temperature and The specific entropy under pressure is obtained by querying the Morrill vapor diagram based on the actual temperature and pressure of the current device. For cross-processes, the theoretical maximum efficiency includes the sum of the theoretical maximum efficiency of all devices required in the current process.

[0030] It should be noted that the first cross-connection anomaly includes specific equipment health status, selection matching, and problems with local control loops, which cause huge errors in the use of various devices in the physical world. The second cross-connection anomaly includes design defects, control strategy mismatch, and unreasonable operation mode based on the first cross-connection anomaly. If the coupling degree of cross-connection devices and the coupling degree of cross-connection processes are both lower than the first coupling degree and the second coupling degree, then the current energy conversion path is normal.

[0031] Based on the average efficiency deviation of the energy path network, and compared with the abnormal efficiency deviations of the first and second crossover anomalies, a control optimization instruction set is generated and observation point early warnings are set. This includes: based on the first-level and second-level edge weights of the energy path network, pre-setting first and second deviation thresholds for the local and global efficiencies of the thermal power plant; calculating the average local efficiency deviation and average global efficiency deviation of each independent component of the current thermal power plant using the first and second deviation thresholds; the average local efficiency deviation includes local efficiencies in each independent component where the difference from historical local efficiency exceeds the first deviation threshold, where the difference is the average local efficiency deviation; and the average global efficiency deviation includes global efficiencies in the overall energy conversion process where the difference from historical global efficiency exceeds the second deviation threshold. The system calculates the local efficiency, where the difference is the average global efficiency deviation. The abnormal efficiency deviation includes abnormal local efficiency deviation and abnormal global efficiency deviation. Based on the average local efficiency deviation and average global efficiency deviation of the thermal power plant, the abnormal local efficiency deviation and abnormal global efficiency deviation of two intersecting abnormal points are dynamically compensated, calibrated, and avoided. Dynamic calibration instructions for the thermal power plant are generated, and warnings are issued for the independent links that are avoided. The local efficiency of the thermal power plant includes the actual energy conversion efficiency between two independent links in the thermal power plant where the intersecting abnormal point occurs. The global efficiency of the thermal power plant includes the final energy conversion efficiency obtained by the thermal power plant according to three different energy conversion paths. At the same time, observation points are set according to the positions of the first intersecting abnormal point, the second intersecting abnormal point, and historical abnormal points to make real-time predictions of the energy path network.

[0032] It should be noted that the local efficiency, global efficiency, first efficiency deviation threshold, second efficiency deviation threshold, abnormal local efficiency deviation, and abnormal global efficiency deviation of a thermal power plant are all confirmed based on two independent links in the thermal power plant where a cross-abnormality point occurs. The first efficiency deviation threshold is determined based on the difference between the maximum and minimum local efficiency of the thermal power plant under normal conditions. The second efficiency deviation threshold is determined based on the difference between the maximum and minimum global efficiency of the thermal power plant under normal conditions. The abnormal local efficiency deviation is the difference between the actual energy conversion efficiency and the historical local efficiency between the two independent links where a cross-abnormality point occurs. The abnormal global efficiency is the difference between the global efficiency of the thermal power plant and the historical global efficiency after a cross-abnormality point occurs. If multiple first cross-abnormalities and second cross-abnormalities occur in the entire process, the abnormal global efficiency deviation is the difference between the current global efficiency of the thermal power plant and the historical global efficiency when multiple first cross-abnormalities and second cross-abnormalities exist simultaneously.

[0033] Based on the locations of the first crossover anomaly point, the second crossover anomaly point, and historical anomalies, observation points are set up to perform real-time prediction of the energy path network. This includes: calculating the probability of anomalies occurring in the energy path network based on the locations of the first crossover anomaly point, the second crossover anomaly point, and historical anomalies; simultaneously, pre-setting observation points for the energy path network; and updating the energy path network in real time. The observation points include actual energy conversion paths of independent chemical-thermal links, independent heat-engine links, and independent electromechanical links. Real-time observation of actual energy conversion paths is conducted, and early warnings and avoidance measures are implemented for energy path networks with high anomaly probabilities. The location of the observation points and the avoidance measures are selected by comparing the magnitude of the anomaly probabilities, where the anomaly probability calculation formula is: ; in, This represents the probability of an anomaly occurring in the current energy conversion path. A set of nodes, including ; This is the normalization constant; Let be the set of outliers, if If it is 1, it means If an anomaly occurs at this location, If it is 0, it means No abnormalities were found at this location; For nodes The nodal potential function, For the edge The edge potential function is used to construct a stack of contradiction points. Based on the probability of occurrence of contradiction points in the energy path network, the sequence of contradiction points in the energy path network is obtained. The data in the contradiction point sequence are pushed onto the stack in ascending order. The time interval between the start of chemical energy conversion at the thermal power plant is used as the refresh cycle of the contradiction point sequence. After each refresh cycle, the contradiction point sequence stack is popped. The node where the popped contradiction point data is located is the specific location of the observation point. The thermal power plant is given an early warning based on the location of the observation point and the probability of occurrence of anomalies.

[0034] It should be noted that if the sequence of contradictions is pushed onto the stack in ascending order, the top element of the stack will be the largest data in the sequence of contradictions. After each refresh cycle, the top element of the stack will be popped and the sequence of contradictions will be refreshed. In the current refresh cycle, the top element of the stack will be used as the observation point. The next cycle will perform the same operation, and the position of the observation point will be refreshed according to the refresh cycle. and The specific implementation method is to calculate and obtain the potential function through the node potential function and the edge potential function, wherein the formula for the node potential function is: ; in, and The adjustable weighting parameters are obtained through cross-coupling degree and local efficiency deviation. The historical average cross-coupling degree The standard deviation of cross-coupling. This represents the local efficiency deviation of the node. This represents the historical average of local efficiency deviations. Let be the standard deviation of the local efficiency deviation; where the boundary potential function is: ; in, For the edge The weights between them For nodes and The interaction coefficient between them is obtained by preset using historical data. Specifically, the value of the interaction coefficient is preset to 0.5. Two types of anomalies are marked in the energy path network. A preset time range is used to obtain the number of times the two types of anomalies appear on each energy conversion path within the current time range. Specifically, the preset time range is 30 days. Based on the observation points of the energy path network, the number of anomalies appearing in each independent link of the energy path network within 30 days is observed in real time, including the number of first cross anomalies and the number of second cross anomalies. At the same time, the historical average number of first cross anomalies and the historical average number of second cross anomalies of each energy conversion path are compared. The probability of anomalies in the energy path network is obtained by calculating the ratio of the number of first cross anomalies to the historical average number of first cross anomalies.

[0035] Based on the average local efficiency deviation and average global efficiency deviation of the thermal power plant, dynamic compensation calibration and control avoidance are performed on the abnormal local efficiency and abnormal global efficiency at two cross-abnormal points. This also includes: adjusting the values ​​of the abnormal local efficiency deviation and abnormal global efficiency deviation to the deviation range specified by the average local efficiency deviation and average global efficiency deviation, respectively, while observing the change in the coupling degree formula to generate an instruction set for the energy path network and physical devices, and controlling the thermal power plant through the dynamic calibration instructions.

[0036] It should be noted that, based on historical data, the initial values ​​of the first and second deviation thresholds are determined by the statistical characteristics of efficiency fluctuations under stable operating conditions. These values ​​are then adjusted according to the equipment's aging and maintenance cycle to obtain the final values ​​of the first and second deviation thresholds. Adjustments are made based on changes in other independent variables in the coupling formula, such as the actual temperature and pressure inside the boiler in the independent thermal link. The specific method for generating the instruction set involves receiving efficiency deviation signals and anomaly information as input, simulating different control actions to change the independent variables in the formula, such as regulating valve opening, pump speed, and set temperature, and assessing their impact on network edge weights. By continuously optimizing the control actions, the optimal combination of control variables is found that allows the deviation value to decrease to the threshold range most quickly and stably. After the instructions are executed, the system monitors new efficiency and coupling data in real time through preset observation points to verify the calibration effect.

[0037] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for intelligent diagnosis and self-optimization control of the whole process of a thermal power plant based on big data, characterized in that, Includes the following steps: S1: Acquire multi-source data from the thermal power plant, including a set of conversion devices, an energy conversion sequence, and actual energy conversion efficiency; S2: Divide the thermal power plant area into independent segments according to the energy conversion sequence, obtain the independent segments of chemical heat, heat engine and electromechanical systems, and construct the energy path network of the thermal power plant with the equipment range required by each independent segment as the origin; S3: Based on the set of conversion devices and mapping relationships, obtain the set of cross-devices and cross-process tables in the three independent links, calculate the coupling degree of each cross-device and the coupling degree of each cross-process, and diagnose the first and second cross-abnormal points of the thermal power plant. S4: Based on the average efficiency deviation of the energy path network, and compared with the abnormal efficiency deviation of the first and second cross anomalies, generate a control optimization instruction set and set early warning for observation points.

2. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 1, characterized in that, The acquisition of multi-source data from the thermal power plant includes a set of conversion devices, an energy conversion sequence, and actual energy conversion efficiency. Specifically: the set of conversion devices comprises a complete set of physical devices used for energy conversion throughout the entire process of energy conversion during the operation of the thermal power plant; the energy conversion sequence includes the current energy conversion process from input to output, requiring the collection of data on fuel type, feed rate, and power generation output; the energy conversion sequence includes the conversion from chemical energy to thermal energy, from thermal energy to mechanical energy, and from mechanical energy to electrical energy; and the actual energy conversion efficiency includes the actual conversion efficiency between two different energy sources, calculated through real-time measurement data.

3. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 1, characterized in that, The process involves dividing the thermal power plant area into independent segments based on the energy conversion sequence, identifying independent chemical-thermal segments, independent heat engine segments, and independent electromechanical segments. Using the equipment range required for each independent segment as the origin, an energy path network for the thermal power plant is constructed. This includes: dividing the independent segments converting chemical energy to thermal energy into chemical-thermal segments, the independent segments converting thermal energy to mechanical energy into heat engine segments, and the independent segments converting mechanical energy to electrical energy into electromechanical segments. Using these three independent segments as origins and different energy sources as input parameters, the energy path network for the thermal power plant is constructed. The outputs of the three independent segments are used as inputs for the next segment. The actual energy conversion efficiency is used as the first-level edge weight, and the final energy conversion efficiency is used as the second-level edge weight. The paths in the energy path network include the actual energy transmission paths of the energy conversion process, and the mapping relationships of the devices during the conversion process are marked on the actual energy conversion paths.

4. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 3, characterized in that, The input parameters, with different energy sources as origins, include: the energy conversion paths of a thermal power plant include a first energy conversion path, a second energy conversion path, and a third energy conversion path. The first energy conversion path includes the thermal power plant converting energy from chemical energy to electrical energy under normal conditions; the second energy conversion path includes the thermal power plant converting energy from thermal energy to electrical energy; and the third energy conversion path includes the thermal power plant converting energy from mechanical energy to electrical energy. The different energy sources of the thermal power plant, including chemical energy, thermal energy, and mechanical energy, are obtained based on the three energy conversion paths. The input parameters are the outputs of the previous independent stage. The first-level edge is the edge obtained in the energy path network through the first energy conversion path, converting from chemical energy to the electrical energy output position of the thermal power plant. The second-level edge is the edge obtained in the energy conversion network through the second and third energy conversion paths. The endpoint energy conversion efficiency is the energy conversion efficiency in the direct conversion process from chemical energy and thermal energy to electrical energy.

5. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 3, characterized in that, The step of mapping the device relationships during the conversion process onto the actual energy conversion path includes: constructing a mapping table of conversion devices based on a set of conversion devices, the mapping table including all conversion devices and operation processes required for energy conversion on the current energy conversion path; performing topology mapping on the current energy conversion path using the mapping table; obtaining the static and dynamic attributes of the conversion devices in the mapping table using the set of conversion devices, the static attributes including the physical form of the current conversion device; the dynamic attributes including the real-time status of the current conversion device; calculating the device contribution of the current conversion device on the energy conversion path using the static and dynamic attributes, the device contribution including the number of times the current conversion device is used in the current independent stage and its functional importance, specifically calculated as the ratio of the number of operation processes using the conversion device in the current operation process to the total number of operation processes; dynamically adjusting the mapping table based on the device contribution; and deleting or silencing conversion devices whose device contribution is lower than a device contribution threshold.

6. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 1, characterized in that, The process involves obtaining the sets of cross-devices and cross-process tables for three independent stages based on the conversion device set and mapping relationship, calculating the cross-device coupling degree and cross-process coupling degree for each, and diagnosing the first and second cross-abnormalities in the thermal power plant. This includes: based on the mapping relationship after deleting silent processes, the mapping relationship of the energy path network includes various physical devices and process information required for energy conversion in the current independent stage on the current path; obtaining the set of cross-devices and cross-process tables for the entire process in the thermal power plant based on the mapping relationship and conversion device set, wherein the cross-device set includes a set of physical devices that simultaneously serve multiple independent stages, and the cross-process table includes a stage process table that simultaneously serves multiple independent stages; and calculating the cross-device set and cross-process coupling degree using the cross-coupling degree formula. The process table includes the coupling degree of cross-devices and cross-processes. The coupling degree of cross-devices and cross-processes includes the degree of mutual interference between cross-devices and cross-processes between two independent links. A first coupling degree and a second coupling degree are preset for the cross-device set and the cross-process table. The coupling degree of cross-devices and cross-processes are compared with the first and second coupling degrees to diagnose the first and second cross-abnormalities in the energy path network. The first cross-abnormality includes independent links where either the cross-device or cross-process is abnormal. The second cross-abnormality includes independent links where both the cross-device and cross-process are abnormal. If the current energy conversion path contains both the first and second cross-abnormalities, they are marked on the current energy conversion path. The formula for the cross-coupling degree is: ； in, The cross-coupling degree of the cross-connect device; The degree of cross-coupling of cross-processes; This represents the theoretical maximum efficiency of the current cross-connect device; This represents the theoretical maximum efficiency of the current cross-process; The contribution of the current device to the device's performance; For cross-connect devices or cross-connect processes in independent stages and independent links Mutual information; For cross-connect devices or cross-connect processes in independent stages and independent links Information entropy; For cross-connect devices or cross-connect processes in independent stages and independent links The minimum information entropy; This is the ratio of the first-level edge weight between two processes in an energy path network to the second-level edge weight of the starting process.

7. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 6, characterized in that, The first crossover anomaly is defined as a situation where the coupling degree of the crossover device is higher than a first coupling degree and the coupling degree of the crossover process is lower than a second coupling degree. This includes situations where the first crossover anomaly is defined as either a crossover device coupling degree higher than the first coupling degree and a crossover process coupling degree lower than the second coupling degree, or a crossover process coupling degree higher than the second coupling degree and a crossover device coupling degree lower than the first coupling degree. The second crossover anomaly is defined as a situation where the coupling degree of the crossover device is higher than the first coupling degree and the crossover process coupling degree is higher than the second coupling degree. The maximum efficiency of the crossover device and crossover process is calculated using the efficiency formula, where the efficiency formula is: ; For cross-connect devices, The theoretical maximum efficiency is the maximum usable energy contained in a unit mass of substance. The specific enthalpy of the substance in its current state is obtained by consulting the Morrill vapor diagram based on the actual temperature and pressure of the current device. is the specific entropy of matter in its current state; For matter in environmental state Temperature and Specific enthalpy under pressure; For matter in environmental state Temperature and The specific entropy under pressure is obtained by querying the Morrill vapor diagram based on the actual temperature and pressure of the current device. For cross-processes, the theoretical maximum efficiency includes the sum of the theoretical maximum efficiency of all devices required in the current process.

8. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 1, characterized in that, The step of generating a control optimization instruction set and setting early warning observation points based on the average efficiency deviation of the energy path network and comparing it with the abnormal efficiency deviations of the first and second crossover anomalies includes: pre-setting a first deviation threshold and a second deviation threshold for the local and global efficiencies of the thermal power plant based on the first-level and second-level edge weights of the energy path network; calculating the average local efficiency deviation and average global efficiency deviation of each independent link of the thermal power plant using the first and second deviation thresholds; the average local efficiency deviation includes local efficiencies in each independent link where the difference from historical local efficiency exceeds the first deviation threshold, where the difference is the average local efficiency deviation; and the average global efficiency deviation includes local efficiencies in the overall energy conversion process where the difference from historical global efficiency exceeds the second deviation threshold. Global efficiency, where the difference is the average global efficiency deviation, and the abnormal efficiency deviation includes abnormal local efficiency deviation and abnormal global efficiency deviation. Based on the average local efficiency deviation and average global efficiency deviation of the thermal power plant, the abnormal local efficiency deviation and abnormal global efficiency deviation of two intersecting anomalies are dynamically compensated, calibrated and avoided, and dynamic calibration instructions for the thermal power plant are generated, as well as warnings are issued for the independent links that are avoided. The local efficiency of the thermal power plant includes the actual energy conversion efficiency between two independent links in the thermal power plant where the intersecting anomaly occurs. The global efficiency of the thermal power plant includes the final energy conversion efficiency obtained by the thermal power plant according to three different energy conversion paths. At the same time, observation points are set according to the positions of the first intersecting anomaly, the second intersecting anomaly, and historical anomalies to make real-time predictions of the energy path network.

9. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 8, characterized in that, The step of setting observation points based on the locations of the first, second, and historical anomalies to perform real-time prediction of the energy path network includes: calculating the probability of anomalies in the energy path network based on the locations of the first, second, and historical anomalies; simultaneously, pre-setting observation points for the energy path network and updating the energy path network in real time. The observation points include actual energy conversion paths of independent chemical-thermal links, independent heat-engine links, and independent electromechanical links. Real-time observation of actual energy conversion paths is conducted, and early warnings and avoidance measures are implemented for energy path networks with high anomaly probabilities. The location of the observation points and the avoidance measures are selected by comparing the magnitude of the anomaly probabilities, where the anomaly probability calculation formula is: ; in, This represents the probability of an anomaly occurring in the current energy conversion path. A set of nodes, including ; This is the normalization constant; Let be the set of outliers, if If it is 1, it means If an anomaly occurs at this location, If it is 0, it means No abnormalities were found at this location; For nodes The nodal potential function, For the edge The edge potential function is used to construct a stack of contradiction points. Based on the probability of occurrence of contradiction points in the energy path network, the sequence of contradiction points in the energy path network is obtained. The data in the contradiction point sequence are pushed onto the stack in ascending order. The time interval between the start of chemical energy conversion at the thermal power plant is used as the refresh cycle of the contradiction point sequence. After each refresh cycle, the contradiction point sequence stack is popped. The node where the popped contradiction point data is located is the specific location of the observation point. The thermal power plant is given an early warning based on the location of the observation point and the probability of occurrence of anomalies.

10. The intelligent diagnosis and self-optimization control method for the entire process of thermal power plants based on big data as described in claim 8, characterized in that, The method of dynamically compensating for and controlling the abnormal local and global efficiencies at two intersection anomaly points based on the average local and average global efficiency deviations of the thermal power plant also includes: adjusting the values ​​of the abnormal local and global efficiency deviations to within the deviation ranges specified by the average local and average global efficiency deviations, while observing the changes in the coupling formula to generate an instruction set for the energy path network and physical devices, and controlling the thermal power plant through the dynamic calibration instructions.

Images