Automatic turning method and system thereof

Abstract

The invention provides an automatic barring method and system, and relates to the technical field of power station maintenance, and the method comprises the steps: obtaining first information and second information, the first information being a sensor deployment scheme of historical barring operation, sensor data and a corresponding barring effect evaluation index, and the second information being real-time monitoring data; performing quantum annealing processing according to the first information to obtain third information; performing Kalman filtering processing according to the first information and the second information to obtain fourth information; dynamically planning and processing fifth information according to the third information and the fourth information; and performing fuzzy logic reasoning processing according to the first information and the fifth information to obtain sixth information. According to the invention, based on quantum annealing processing and dynamic planning processing, the intelligent design of the optimal control strategy of the barring operation is realized by combining the graph construction, the random forest algorithm, the shortest path algorithm and other technologies, the barring operation can be automatically adjusted according to the real-time monitoring data and the barring effect evaluation index, and the barring efficiency and precision are improved.

Description

Technical Field

[0001] The invention relates to the field of power station maintenance, and in particular to an automatic turning method and a system thereof. Background Art

[0002] With the rapid development of the power industry and the continuous advancement of power unit technology, power units play a vital role in the power generation process. In the operation and maintenance process of the unit, turning the gear as a part of the maintenance is crucial because it directly affects the stability and performance of the unit. Turning the gear is a precise adjustment process that requires accurate measurement and adjustment of parameters such as the swing, level, and air gap of various parts of the unit. The existing turning method mainly relies on manual operation, and the turning process is highly dependent on the experience and skill level of the operator. There are problems such as cumbersome operation, low efficiency, and low precision, which makes it difficult to meet the needs of the modern power industry for efficient and accurate maintenance.

[0003] Based on the above problems, there is an urgent need for an automatic cranking method and system. Summary of the invention

[0004] The purpose of the present invention is to provide an automatic cranking method, device, equipment and readable storage medium to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows: In a first aspect, the present application provides an automatic cranking method, comprising: Acquire first information and second information, wherein the first information is a sensor deployment scheme, sensor data, and corresponding cranking effect evaluation index of historical cranking operations, and the second information is real-time monitoring data; Performing quantum annealing processing on the first information to obtain third information, wherein the third information is a key feature and pattern recognition result of the turning operation; Performing Kalman filtering on the first information and the second information to obtain fourth information, wherein the fourth information is an estimated value of the current cranking state; Performing dynamic programming to process fifth information according to the third information and the fourth information, wherein the fifth information is an optimal control strategy for the turning operation; The sixth information is obtained by performing fuzzy logic reasoning processing on the first information and the fifth information, and the sixth information is a control parameter suitable for actual operation.

[0005] In a second aspect, the present application also provides an automatic turning system, comprising: An acquisition module, used to acquire first information and second information, wherein the first information is a sensor deployment scheme, sensor data and corresponding evaluation index of a historical cranking operation, and the second information is real-time monitoring data; An extraction module, configured to perform quantum annealing processing on the first information to obtain third information, wherein the third information is a key feature and pattern recognition result of the turning operation; A prediction module, configured to perform Kalman filtering on the first information and the second information to obtain fourth information, wherein the fourth information is an estimated value of a current cranking state; A planning module, used for dynamically planning and processing fifth information according to the third information and the fourth information, wherein the fifth information is an optimal control strategy for the turning operation; The reasoning module is used to perform fuzzy logic reasoning processing on the first information and the fifth information to obtain sixth information, where the sixth information is a control parameter suitable for actual operation.

[0006] The beneficial effects of the present invention are: The present invention is based on quantum annealing processing and dynamic programming processing, combined with graph construction, random forest algorithm, shortest path algorithm and other technologies, to achieve intelligent design of the optimal control strategy for the turning operation. It can automatically adjust the turning operation according to real-time monitoring data and turning effect evaluation indicators, thereby improving the turning efficiency and accuracy.

[0007] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0009] Figure 1 A schematic diagram of the automatic cranking method according to an embodiment of the present invention; Figure 2 It is a topological diagram of the turning gear system described in an embodiment of the present invention. DETAILED DESCRIPTION

[0010] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0011] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0012] Embodiment 1: This embodiment provides an automatic cranking method.

[0013] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400 and step S500.

[0014] Step S100, obtaining first information and second information, the first information being the sensor deployment scheme, sensor data and corresponding cranking effect evaluation index of historical cranking operations, and the second information being real-time monitoring data; It can be understood that the sensor deployment plan, sensor data and corresponding turning effect evaluation indicators of the historical turning operation cover information such as the center deviation of the axis of each section, the swing value and the orientation. The sensor deployment plan involves the installation position of each key part of the unit and the type of sensor. The sensor data refers to various parameters and status data obtained by these sensors, such as the swing value, offset, etc. The corresponding turning effect evaluation indicators are calculated based on these data, such as the swing, swing orientation, etc. The real-time monitoring data includes data obtained in real time by sensors during the turning process, which is mainly used to monitor the status and parameter changes of various parts of the unit, as well as environmental conditions in real time. These data include parameters such as the swing, level, roundness and air gap of various parts of the current unit. Furthermore, if Figure 2 As shown, Figure 2The topology diagram of the turning system is shown, through which we can intuitively understand how these data are integrated and applied in the turning system. Specifically, the system is divided into three parts: data acquisition, data collection, and data analysis and control. Data acquisition includes acquiring various sensor signals, such as the air supply pipe turning signal, the slip ring turning signal, the upper guide turning signal, etc., which are responsible for monitoring the status of specific components. For example, the air supply pipe turning signal is related to the air supply system of the turbine, and is used to monitor and control the status of the air supply pipe during the turning process; the slip ring turning signal is used to monitor the position and status of the slip ring during the turning process; and the upper guide turning signal focuses on the status of the upper guide bearing. Data acquisition devices, such as the upper guide turning data acquisition device, are responsible for collecting key parameters such as the roundness signals of the upper and lower end faces of the rotor and stator. These devices ensure the accuracy and real-time nature of the data and provide a basis for subsequent data analysis. The roundness signal of the upper end face of the rotor and the roundness signal of the upper end face of the stator are used to monitor the roundness of the upper end faces of the rotor and stator to ensure the balance and efficiency of the machinery. The rotor lower end face roundness signal and the stator lower end face roundness signal monitor the roundness of the rotor and stator lower end faces respectively. For data analysis and control, the cranking data storage and analysis host is the core of this part, which processes and analyzes the data collected from the data acquisition device. Through these data analysis, the cranking control signals, such as cranking control signals, including start, stop, speed control, etc., can be optimized to ensure the accuracy of operation and mechanical balance. The gear position signal is used to monitor the precise position of the gear, while the thrust head cranking signal and the lower guide cranking signal monitor the status of the thrust head and the lower guide bearing respectively. The mirror plate cranking signal is used to monitor the status of the mirror plate component. The lower wind tunnel cranking data acquisition device and the water wheel room cranking data acquisition device collect data from the lower wind tunnel area and the water wheel room during the cranking process, respectively, while the water guide cranking signal is related to the water guide bearing and is used to monitor the status of the water guide bearing.

[0015] Step S200, performing quantum annealing processing on the first information to obtain third information, where the third information is a key feature and pattern recognition result of the turning operation; It should be noted that quantum annealing is an optimization algorithm that finds the global optimal solution by simulating the evolution of quantum states in energy functions. In this embodiment, the quantum annealing algorithm is applied to the processing of historical cranking operation data to discover key features and patterns. These key features include the change rules of parameters such as swing and offset at different parts of the unit, as well as data patterns and features under different cranking modes.

[0016] Step S300, performing Kalman filtering processing on the first information and the second information to obtain fourth information, where the fourth information is an estimated value of the current cranking state; It can be understood that the Kalman filter estimates the state of the system by continuously updating and fusing the measurement data from the sensor and the predicted value of the system model. In this step, the fourth information will provide an estimate of the current cranking state through Kalman filter processing, including parameters such as the swing and offset of each part of the unit. These estimates will have high accuracy and stability, which will help to monitor changes and abnormal conditions in the cranking process in real time and provide important reference for subsequent decision-making and adjustments.

[0017] Step S400, performing dynamic planning to process fifth information according to the third information and the fourth information, wherein the fifth information is an optimal control strategy for the turning operation; It is understandable that dynamic programming is an optimization method used to solve multi-stage decision-making problems, selecting the current optimal decision at each stage, and ultimately reaching the global optimal solution. Through dynamic programming processing, the fifth information will provide the optimal control strategy for the cranking operation, including the adjustment plan of each part of the unit, the operation sequence, etc. These strategies will be based on historical cranking operation data and real-time monitoring data, combined with system models and prediction results, to maximize the cranking effect, optimize the cranking process, and meet specific cranking requirements and constraints.

[0018] Step S500: Perform fuzzy logic reasoning on the first information and the fifth information to obtain sixth information, where the sixth information is a control parameter suitable for actual operation.

[0019] It can be understood that fuzzy logic reasoning is a reasoning method based on fuzzy set theory, which is suitable for dealing with problems with strong uncertainty and ambiguity. Specifically, in this step, fuzzy logic reasoning will infer the control parameters suitable for the current turning operation based on the sensor deployment scheme and turning effect evaluation index in the historical data, as well as the adjustment scheme and operation sequence in the optimal control strategy. These control parameters involve aspects such as sensor position adjustment, operation sequence optimization, and turning effect evaluation index adjustment. Through fuzzy logic reasoning processing, the sixth information will provide specific control parameters required for actual operation, which can be directly applied to the turning operation to achieve a more accurate, efficient and controllable turning process.

[0020] It should be noted that step S200 includes step S210, step S220, step S230 and step S240.

[0021] Step S210, performing graph construction processing according to the first information to obtain seventh information, the seventh information includes a vector graph set, and the structure of each vector graph includes sensor nodes, sensor data attribute nodes, turning effect evaluation index attribute nodes, and association relationships and spatial position relationships between sensors; It can be understood that the sensor nodes can be represented as vertices or nodes in the graph, and each sensor corresponds to a node. The sensor data attribute node includes the data type and data accuracy collected by the sensor. The turning effect evaluation index attribute node includes swing, offset, vibration, etc. The association relationship and spatial position relationship between sensors indicates that there is a certain association or spatial position relationship between sensors, such as distance, angle, etc.

[0022] Step S220: taking the seventh information as input, applying the random forest algorithm to perform feature selection and weight calculation, and obtaining the eighth information by determining the importance of each feature and the degree of influence on the turning effect, wherein the eighth information includes a feature weight set and feature importance information; It should be noted that random forest is an integrated learning method that selects and predicts features by constructing multiple decision trees, and has high accuracy and robustness. The specific steps include feature selection and feature weight calculation. In the feature selection stage, the algorithm will construct a decision tree through multiple iterations, and determine which features have a greater impact on the turning effect based on the best feature splitting at each node. In the feature weight calculation stage, the random forest algorithm will calculate the weight of the feature based on the usage and influence of each feature in multiple decision trees, reflecting the importance of each feature in the turning operation. The feature weight set and feature importance information obtained through this step provide a key reference for subsequent quantum annealing processing and fuzzy logic reasoning.

[0023] Step S230: Perform data embedding processing according to the seventh information and the eighth information, and use the feature weight set and the feature importance information as attributes of the node in the corresponding vector map to obtain ninth information; This step aims to embed feature weights and importance information into the node attributes of the vector diagram to enrich the information content of the vector diagram. Through such data embedding processing, feature weights and importance attributes related to the turning operation can be given to each node, making the vector diagram more representative and information-rich.

[0024] Step S240: taking the ninth information as input, applying a quantum annealing algorithm to optimize the graph, and obtaining the third information by simulating the evolution of the quantum state in the energy function to find the global optimal solution.

[0025] It is understandable that the quantum annealing algorithm uses the structure and node attributes of the vector graph as the problem space, and adjusts the state of the nodes to find the state configuration that minimizes the energy function, thereby obtaining the optimal turning operation strategy. During the optimization process, the algorithm takes into account the feature weights and importance information embedded in the node attributes to ensure that the selected operation strategy can achieve the best turning effect after comprehensively considering the weights and importance of each feature. By simulating the evolution of the quantum state, the quantum annealing algorithm can search in the solution space of the problem and gradually optimize the state configuration of the vector graph until the global optimal solution is found where the energy function reaches the minimum value.

[0026] It should be noted that step S240 includes step S241, step S242, step S243 and step S244.

[0027] Step S241: According to the sensor deployment scheme and the corresponding turning effect evaluation index, the sensor deployment efficiency index is calculated by using the graph-based shortest path algorithm; This step calculates the sensor deployment efficiency index by applying a graph-based shortest path algorithm. This index can help evaluate the effectiveness and efficiency of sensor deployment, thereby guiding the optimization of the turning operation. By calculating the sensor deployment efficiency index, the contribution and efficiency of each sensor deployment scheme in the turning operation can be evaluated, thereby providing guidance for optimizing the deployment of sensors. This will help improve the accuracy and efficiency of the turning operation, thereby achieving better turning results. The calculation formula for the sensor deployment efficiency index is: ; in, represents the sensor deployment efficiency indicator, , Indicates the position number of the sensor. Indicates important positions related to turning operations. represents the set of all key parts, express Sensors to key parts The path length, represents the weight of the key part k, express Sensors and The distance between sensors, express Sensors and The association weights between sensors.

[0028] Step S242, constructing a function according to the sensor deployment scheme, the turning effect evaluation index and the sensor deployment efficiency index to obtain an energy function; The energy function is: ; in, represents the energy function, represents the number of sensors, Indicates the number of key parts, represents the number of evaluation indicators, Indicates the serial number of the sensor. Indicates the serial number of the key part, Indicates the serial number of the evaluation index, Indicates Sensor to The path length of the key parts, Indicates The deployment efficiency index of sensors is Indicates The sensor for the The score of the evaluation index.

[0029] Step S243, performing quantum state simulation evolution processing according to the energy function to obtain the evolution process; It can be understood that quantum state simulation evolution processing is a method of optimizing search using the idea of ​​quantum computing. Its main idea is to represent the state of the sensor deployment scheme as a quantum state and search for the optimal solution by simulating the evolution of the quantum state under the influence of the energy function. During the evolution process, the state of the sensor deployment scheme is adjusted through quantum gate operations in the hope of achieving the goal of minimizing the energy function.

[0030] Step S244: performing a solution processing according to the evolution process to obtain a global optimal solution of the energy function, and outputting the global optimal solution as the third information.

[0031] Preferably, in this step, an optimization algorithm is used to search for the global optimal solution, such as a genetic algorithm, a particle swarm algorithm, etc. These algorithms can continuously search for better solutions in the search space until a stop condition is reached or a global optimal solution is found. Specifically, during the operation of the algorithm, the energy function value corresponding to each sensor deployment scheme is evaluated and compared to determine which scheme has the smallest energy function value. This process requires multiple iterations until the convergence condition is reached. Finally, the sensor deployment scheme with the smallest energy function value is selected as the global optimal solution.

[0032] It should be noted that step S300 includes step S310, step S320, step S330 and step S340.

[0033] Step S310: construct a state space model of the turning gear system according to the first information, wherein the state space model includes state variables, state transfer equations and observation equations of the system, and the state variables include swing, level and air gap parameters of various parts of the unit; First, it is necessary to define the state variables of the system. In this embodiment, the state variables include the swing, level and air gap parameters of various parts of the unit. These state variables describe the internal state of the system, such as the swing of the unit rotor and stator and the gap between the various parts of the unit. Next, it is necessary to establish a state transfer equation. The state transfer equation describes how the system state changes over time. In the cranking system, the state transfer equation is established based on the laws of physics and the principles of system dynamics, taking into account the interaction between the various parts of the unit during the cranking process and the influence of external disturbances. Finally, the observation equation is determined, which describes the relationship between the system state and the observable data. In this step, the observation equation is established using the real-time monitoring data obtained by the sensor, and the estimated value of the cranking system state is compared with the actual observation value to verify the accuracy and reliability of the state space model. The establishment of this model can help the system to accurately grasp and effectively control its own state, and improve the safety and efficiency of the cranking operation.

[0034] Step S320, performing prediction processing according to the state transition equation of the state space model, predicting the state of the system at the next moment based on the current state and the dynamic properties of the system to obtain a prediction result; It is understandable that the state transition equation describes how the system state evolves over time, so the future state can be predicted based on the current state and the dynamic characteristics of the system. The prediction result is an estimate or prediction of the system state, which provides important information about the future behavior of the system. These prediction results can be used to formulate control strategies for turning operations, optimize system performance, or conduct risk assessments. For example, based on the prediction results, the parameters of the turning operation can be adjusted or appropriate measures can be taken to deal with possible abnormal situations, thereby improving the stability and safety of the system. Through the prediction processing of this step, the behavior of the system can be better understood, potential problems can be discovered in a timely manner, and appropriate measures can be taken to adjust them, thereby effectively improving the operating efficiency and reliability of the turning system.

[0035] Step S330: update processing is performed according to the prediction results and the real-time monitoring data, the real-time monitoring data is linked to the state space using the observation equation, and the system state is updated by the Kalman filter algorithm to obtain an update result; It can be understood that the real-time monitoring data is mapped into the state space through the observation equation, so as to be compared and adjusted with the predicted system state. Then, the predicted system state is fused with the real-time monitoring data using the Kalman filter algorithm to obtain a more accurate estimate of the system state. The Kalman filter algorithm can dynamically adjust the weights of the prediction results and the real-time monitoring data according to the dynamic properties of the system and the characteristics of the measurement noise, thereby achieving an optimized estimate of the system state. Through the update process, the prediction error can be corrected in time, making the estimate of the system state more accurate and reliable.

[0036] Step S340: Perform data extraction processing according to the update result to obtain fourth information.

[0037] It is understood that the updated result can obtain the latest estimate of the system state, which contains information about the swing, level and air gap parameters of each part of the turning system. This information is essential for understanding the current state of the turning system and the operating status of the system. In the data extraction process, the parts related to the required information are selected and extracted from the updated system state, which involves screening, converting or combining the state variables to obtain a specific data form.

[0038] It should be noted that step S400 includes step S410, step S420, step S430 and step S440.

[0039] Step S410, constructing a state space and an action space according to the third information and the fourth information, wherein the state space is used to represent the cranking state, and the action space is used to represent the cranking operation; It can be understood that the state space is a set describing the state of the turning system. In the state space, each state is a specific state that the system may be in. These states include various performance parameters of the turning system, environmental conditions, and other factors that affect the behavior of the system. Using the update results extracted from the fourth information, the system state can be represented as a vector, which contains various important parameters of the turning system, such as swing, level, and air gap. The action space is a set describing the turning operation. In the action space, each action is an operation or strategy that the system can perform. These actions include adjusting the turning parameters, changing the working mode, or other operations that affect the system state. Using the optimal control strategy in the third information, the turning operation can be represented as an action set, which contains various operation modes that the system can take. By formalizing the state space and action space, a basis is provided for subsequent dynamic programming processing.

[0040] Step S420: Establish a state transition model and a reward function based on the state space and the action space, and in combination with the motion law of the rotating parts during the cranking process, the force conditions, and the degree of influence of different cranking operations on the system performance. The state transition model is used to describe the probability distribution of the system transferring to the next state under a given state and action, and the reward function is used to evaluate the quality of each state action. It is understandable that establishing a state transition model requires understanding the motion characteristics and force conditions of the turning system, and considering the impact of different operations on the system in different states. The actions in the turning system include adjusting the axis, changing the swing, adjusting the level, and other operations. Different operations may have different degrees of impact on the performance of the system. For example, adjusting the axis may change the stability of the system, while changing the swing may affect the efficiency of the system. Therefore, when establishing a state transition model, it is necessary to consider the impact of different operations on the system in different states and incorporate these impacts into the state transition probability. The reward function is defined according to the performance indicators required in the turning process. For example, the turning efficiency can be measured by minimizing energy loss or maximizing output power, and stability can be evaluated by reducing the swing change. Further, by comprehensively considering the state transition model and the reward function, a reinforcement learning model can be established to select the optimal operation strategy under a given turning state. This model will be continuously optimized by learning historical data and interacting with the environment to achieve the optimization of the turning operation and performance improvement.

[0041] Step S430, performing iterative optimization processing of the state value function based on the state transition model and the reward function, and obtaining an optimized value function by calculating the value function of each state in each iteration until converging to the optimal solution; It should be noted that step S430 includes step S431, step S432, step S433, step S434 and step S435.

[0042] Step S431: Initialize the state value function to a suitable initial estimated value.

[0043] Taking into account the characteristics and data of the cranking system in this scenario, the initial estimate is determined based on the statistical characteristics of historical cranking data and real-time monitoring data to reflect the current state of the system as accurately as possible.

[0044] Step S432: In each iteration, all possible states are traversed, and the value function of each state is calculated using the state transition model and the reward function.

[0045] This step combines the state transition model and the reward function, combining the expected return of each action with the probability of state transition to derive the value function of each state. This value function represents the expected return of performing each action in the current state, thus helping to evaluate the pros and cons of each state.

[0046] Step S433: According to the update rule of dynamic programming, the value function of each state is updated according to the Bellman equation.

[0047] This step takes into account the interrelationships and transition probabilities between states, as well as the expected rewards from performing different actions. A discount factor is also taken into account during the update process to balance the trade-off between current rewards and future rewards.

[0048] Step S434: After each iteration, a convergence check is performed to check whether the difference between the current state value function and the value function of the previous iteration is less than a preset threshold. If the convergence condition is met, the iteration is stopped; otherwise, the iteration continues.

[0049] This step aims to ensure that the state-value function gradually converges to the optimal solution, that is, the maximum expected reward that can be obtained by taking the optimal action in each state.

[0050] Step S435: In the iterative process, the above steps are repeatedly performed until the state value function converges to the optimal solution. In this way, the optimal strategy of the system can be obtained to guide the turning operation, thereby maximizing the system performance.

[0051] Step S440: According to the optimized value function, an action that maximizes the value function is selected as the optimal control strategy.

[0052] Through this process, the optimal turning operation strategy can be determined for the system. This strategy is based on the comprehensive search and evaluation of the system state space by the dynamic programming algorithm, as well as the analysis and optimization of the possible actions in each state.

[0053] It should be noted that step S500 includes step S510, step S520, step S530 and step S540.

[0054] Step S510, performing fuzzy processing according to the first information and the unit characteristics to construct a fuzzy logic rule, wherein the fuzzy logic rule includes a relationship between a control parameter and a cranking operation feature; Specifically, fuzzy logic rules are based on expert experience, and can also be learned from historical data using data mining and machine learning techniques. These rules will describe the degree of influence of different control parameters on the characteristics of the cranking operation, and how to adjust these parameters in different situations to optimize the cranking effect. Through this step, a fuzzy knowledge base can be established, which contains a rich relationship between control parameters and cranking operation characteristics. This will provide a basis for subsequent fuzzy logic reasoning, enabling the system to automatically derive control parameters suitable for actual operations based on real-time monitoring data and unit characteristics, thereby realizing intelligent optimization of cranking operations.

[0055] Step S520, constructing a fuzzy reasoning mathematical model according to fuzzy logic rules; It can be understood that the fuzzy reasoning mathematical model is based on the fuzzy logic system, which includes fuzzy sets, fuzzy relations and fuzzy reasoning rules. In this model, fuzzy sets are used to describe the fuzzy information of real-time monitoring data and unit characteristics, fuzzy relations are used to represent the relationship between different control parameters and the characteristics of the turning operation, and fuzzy reasoning rules are used to guide the reasoning process. Specifically, fuzzy sets are used to represent the fuzzy concepts of various monitoring indicators, such as "high", "low", "normal", etc. Then, through fuzzy relations, these fuzzy concepts are mapped to the fuzzy values ​​of the control parameters to describe the association between them. Finally, fuzzy reasoning rules are used to reason based on the conditions and conclusions defined in the fuzzy logic rules to determine the optimal control parameters. Through this step, a fuzzy reasoning mathematical model is established that can derive control parameters suitable for actual operations based on real-time monitoring data and unit characteristics.

[0056] Step S530, inputting the fifth information into the fuzzy reasoning mathematical model to perform fuzzy reasoning processing to obtain a reasoning result; It can be understood that in the fuzzy reasoning process, the fuzzy reasoning system will perform fuzzy reasoning on the fuzzy input values ​​according to the conditions and conclusions defined in the fuzzy logic rules, thereby generating fuzzy output values. These output values ​​describe the degree of influence of different control parameters on the turning operation characteristics, as well as the recommended operating measures.

[0057] Step S540: Defuzzify the inference result, and convert the fuzzy control parameter into an actual operation parameter to obtain sixth information.

[0058] It is understood that defuzzification is the process of converting the fuzzified output results into clear and specific values ​​or instructions. In this step, the fuzzy control parameters are mapped to actual operating parameters using the results of fuzzy reasoning and pre-defined defuzzification rules. This process involves the conversion of fuzzy sets, the calculation of the center of gravity of fuzzy sets or other defuzzification methods to obtain definite and accurate control parameters. The results of the defuzzification process will be provided to the control module of the system to guide the actual turning operation. These operating parameters can be specific control instructions, adjustment parameters or other operations that need to be performed to achieve an optimized turning effect.

[0059] It should be noted that step S520 includes step S521, step S522 and step S523.

[0060] Step S521, converting the fuzzy logic rules, by converting the language description of each fuzzy rule into a mathematical expression of fuzzy logic to obtain a mathematical form rule base of fuzzy logic; It is understandable that in the process of converting fuzzy rules described in natural language into mathematical expressions of fuzzy logic, it is necessary to define input and output variables and select corresponding membership functions; then, keywords will be converted into logical operators and membership functions; each rule needs to determine the weight, which reflects the degree of its influence on the output result; at the same time, determine the logical connectives to connect different conditions; finally, the converted rules are recorded in the form of mathematical expressions in the rule base, which includes input and output variables, logical expressions and weight information. This rule base provides the basis for subsequent fuzzy reasoning.

[0061] Step S522: construct and process the mathematical form rule base, and establish a fuzzy reasoning mechanism by defining the membership functions of the input variables and the output variables and the logical rules between the fuzzy rules; It can be understood that according to the fuzzy rules in the rule base, the logical relationship between the fuzzy rules is determined, and fuzzy logic operators (such as AND, OR, NOT, etc.) are usually used to connect different conditions. These logical rules determine how the fuzzy reasoning system derives the fuzzy value of the output variable based on the fuzzy value of the input variable. By establishing a fuzzy reasoning mechanism, the fuzzy value of the output variable can be derived based on the fuzzy value of the input variable and the rules defined in the fuzzy rule base, thereby realizing the fuzzy reasoning process. This process allows the system to make decisions and inferences in the face of uncertainty and ambiguity, improving the adaptability and robustness of the system.

[0062] Step S523: Verify and optimize the mathematical form rule base and the fuzzy reasoning mechanism, and integrate them to obtain a fuzzy reasoning mathematical model.

[0063] It is understandable that the verification process includes checking the logical relationship of fuzzy rules, and adjusting and optimizing the parameters of membership functions, so that the fuzzy reasoning system can more accurately reflect the actual situation. The optimization methods include adjusting the weights and logical relationships of fuzzy rules, optimizing the shape and parameters of membership functions, and increasing or decreasing the number of input variables and output variables. Through verification and optimization processing, a more accurate and reliable fuzzy reasoning mathematical model can be obtained, which can effectively handle the fuzziness and uncertainty of input data, and make inferences and decisions based on pre-set rules and logical relationships, thereby providing more intelligent and reliable control parameters for turning operations.

[0064] Embodiment 2: This embodiment provides an automatic cranking system, including: An acquisition module, used to acquire first information and second information, the first information being a sensor deployment scheme, sensor data and corresponding evaluation index of a historical cranking operation, and the second information being real-time monitoring data; An extraction module, used for performing quantum annealing processing on the first information to obtain third information, where the third information is a key feature and pattern recognition result of the turning operation; A prediction module, configured to perform Kalman filtering on the first information and the second information to obtain fourth information, where the fourth information is an estimated value of a current cranking state; A planning module, used for dynamically planning and processing fifth information according to the third information and the fourth information, wherein the fifth information is an optimal control strategy for the turning operation; The reasoning module is used to perform fuzzy logic reasoning on the first information and the fifth information to obtain the sixth information, where the sixth information is a control parameter suitable for actual operation.

[0065] In a specific embodiment of the present disclosure, the extraction module includes: A first construction unit is used to perform graph construction processing according to the first information to obtain seventh information, wherein the seventh information includes a vector graph set, and the structure of each vector graph includes a sensor node, a sensor data attribute node, a turning effect evaluation index attribute node, and an association relationship and a spatial position relationship between sensors; a first calculation unit, configured to use the seventh information as input, apply a random forest algorithm to perform feature selection and weight calculation, and obtain eighth information by determining the importance of each feature and the degree of influence on the turning effect, wherein the eighth information includes a feature weight set and feature importance information; a first embedding unit, configured to perform data embedding processing according to the seventh information and the eighth information, and use the feature weight set and the feature importance information as attributes of the node in the corresponding vector graph to obtain ninth information; The first optimization unit is used to take the ninth information as input, apply the quantum annealing algorithm to optimize the graph, and obtain the third information by simulating the evolution of the quantum state in the energy function to find the global optimal solution.

[0066] In a specific embodiment of the present disclosure, the first optimization unit includes: The second calculation unit is used to calculate the sensor deployment efficiency index by using the graph-based shortest path algorithm according to the sensor deployment scheme and the corresponding turning effect evaluation index; The second construction unit is used to perform function construction processing according to the sensor deployment plan, the turning effect evaluation index and the sensor deployment efficiency index to obtain an energy function; A first simulation unit is used to perform quantum state simulation evolution processing according to an energy function to obtain an evolution process; The first processing unit is used to perform a solution processing according to the evolution process to obtain a global optimal solution of the energy function, and output the global optimal solution as the third information.

[0067] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An automatic turning method, characterized in that: include: Acquire first information and second information, wherein the first information is a sensor deployment scheme, sensor data, and corresponding cranking effect evaluation index of historical cranking operations, and the second information is real-time monitoring data; Performing quantum annealing processing on the first information to obtain third information, wherein the third information is a key feature and pattern recognition result of the turning operation; Performing Kalman filtering on the first information and the second information to obtain fourth information, wherein the fourth information is an estimated value of the current cranking state; Performing dynamic programming to process fifth information according to the third information and the fourth information, wherein the fifth information is an optimal control strategy for the turning operation; The sixth information is obtained by performing fuzzy logic reasoning processing on the first information and the fifth information, and the sixth information is a control parameter suitable for actual operation.

2. The automatic turning method according to claim 1, characterized in that: Performing quantum annealing processing on the first information to obtain third information includes: Performing graph construction processing according to the first information to obtain seventh information, the seventh information includes a vector graph set, and the structure of each vector graph includes a sensor node, a sensor data attribute node, a turning effect evaluation index attribute node, and an association relationship and a spatial position relationship between sensors; The seventh information is used as input, and a random forest algorithm is applied to perform feature selection and weight calculation, and the eighth information is obtained by determining the importance of each feature and the degree of influence on the turning effect, wherein the eighth information includes a feature weight set and feature importance information; Performing data embedding processing according to the seventh information and the eighth information, and using the feature weight set and the feature importance information as attributes corresponding to the nodes in the vector graph to obtain ninth information; The ninth information is used as input, and the graph is optimized using a quantum annealing algorithm. The third information is obtained by simulating the evolution of the quantum state in the energy function to find the global optimal solution.

3. The automatic turning method according to claim 2, characterized in that: The ninth information is used as input, and the graph is optimized by applying a quantum annealing algorithm, and the third information is obtained by finding a global optimal solution by simulating the evolution of the quantum state in the energy function, including: According to the sensor deployment scheme and the corresponding turning effect evaluation index, a sensor deployment efficiency index is calculated using a graph-based shortest path algorithm; Performing function construction processing according to the sensor deployment scheme, the turning effect evaluation index and the sensor deployment efficiency index to obtain an energy function; Performing quantum state simulation evolution processing according to the energy function to obtain an evolution process; A global optimal solution of the energy function is obtained by performing a solution processing according to the evolution process, and the global optimal solution is output as the third information.

4. The automatic cranking method according to claim 1, characterized in that: Performing Kalman filtering on the first information and the second information to obtain fourth information includes: constructing a state space model of the turning gear system according to the first information, wherein the state space model includes state variables, state transfer equations and observation equations of the system, and the state variables include swing, level and air gap parameters of various parts of the unit; Perform prediction processing according to the state transfer equation of the state space model, predict the state of the system at the next moment based on the current state and the dynamic properties of the system to obtain a prediction result; Performing update processing according to the prediction result and the real-time monitoring data, using the observation equation to link the real-time monitoring data with the state space, and updating the system state through the Kalman filter algorithm to obtain an update result; Data extraction processing is performed according to the update result to obtain fourth information.

5. The automatic cranking method according to claim 1, characterized in that: Performing dynamic programming to process fifth information according to the third information and the fourth information includes: A state space and an action space are constructed according to the third information and the fourth information, wherein the state space is used to represent the cranking state, and the action space is used to represent the cranking operation; According to the state space and the action space, and in combination with the motion law of the rotating parts during the cranking process, the force conditions, and the degree of influence of different cranking operations on the system performance, a state transition model and a reward function are established. The state transition model is used to describe the probability distribution of the system transferring to the next state under a given state and action, and the reward function is used to evaluate the quality of each state action; Iteratively optimizing the state value function based on the state transition model and the reward function, and obtaining the optimized value function by calculating the value function of each state in each iteration until converging to the optimal solution; According to the optimized value function, the action that maximizes the value function is selected as the optimal control strategy.

6. The automatic cranking method according to claim 1, characterized in that: Performing fuzzy logic reasoning on the fifth information to obtain sixth information includes: Performing fuzzy processing according to the first information and the unit characteristics to construct a fuzzy logic rule, wherein the fuzzy logic rule includes a relationship between a control parameter and a cranking operation feature; A fuzzy reasoning mathematical model is constructed according to the fuzzy logic rules; Inputting the fifth information into the fuzzy reasoning mathematical model to perform fuzzy reasoning processing to obtain a reasoning result; The inference result is defuzzified, and the fuzzified control parameter is converted into an actual operation parameter to obtain the sixth information.

7. The automatic cranking method according to claim 8, characterized in that: The fuzzy reasoning mathematical model is constructed according to the fuzzy logic rules, including: The fuzzy logic rules are converted to obtain a mathematical form rule base of fuzzy logic by converting the language description of each fuzzy rule into a mathematical expression of fuzzy logic; According to the mathematical form rule base, a fuzzy reasoning mechanism is established by defining the membership functions of input variables and output variables and the logical rules between fuzzy rules; Verification and optimization processing are performed according to the mathematical form rule base and fuzzy reasoning mechanism, and a fuzzy reasoning mathematical model is obtained through integration.

8. An automatic turning system, characterized in that: include: An acquisition module, used to acquire first information and second information, wherein the first information is a sensor deployment scheme, sensor data and corresponding evaluation index of a historical cranking operation, and the second information is real-time monitoring data; An extraction module, configured to perform quantum annealing processing on the first information to obtain third information, wherein the third information is a key feature and pattern recognition result of the turning operation; A prediction module, configured to perform Kalman filtering on the first information and the second information to obtain fourth information, wherein the fourth information is an estimated value of a current cranking state; A planning module, used for dynamically planning and processing fifth information according to the third information and the fourth information, wherein the fifth information is an optimal control strategy for the turning operation; The reasoning module is used to perform fuzzy logic reasoning processing on the first information and the fifth information to obtain sixth information, where the sixth information is a control parameter suitable for actual operation.

9. The automatic turning system according to claim 8, characterized in that: The extraction module comprises: A first construction unit is configured to perform graph construction processing according to the first information to obtain seventh information, wherein the seventh information includes a set of vector graphs, and the structure of each of the vector graphs includes a sensor node, a sensor data attribute node, a turning effect evaluation index attribute node, and an association relationship and a spatial position relationship between sensors; a first calculation unit, configured to use the seventh information as input, apply a random forest algorithm to perform feature selection and weight calculation, and obtain eighth information by determining the importance of each feature and the degree of influence on the turning effect, wherein the eighth information includes a feature weight set and feature importance information; a first embedding unit, configured to perform data embedding processing according to the seventh information and the eighth information, and use the feature weight set and the feature importance information as attributes corresponding to a node in the vector graph to obtain ninth information; The first optimization unit is used to take the ninth information as input, apply the quantum annealing algorithm to optimize the graph, and obtain the third information by simulating the evolution of the quantum state in the energy function to find the global optimal solution.

10. The automatic turning system according to claim 9, characterized in that: The first optimization unit comprises: A second calculation unit is used to calculate the sensor deployment efficiency index by using a graph-based shortest path algorithm according to the sensor deployment scheme and the corresponding turning effect evaluation index; A second construction unit is used to perform function construction processing according to the sensor deployment scheme, the turning effect evaluation index and the sensor deployment efficiency index to obtain an energy function; A first simulation unit, used for performing quantum state simulation evolution processing according to the energy function to obtain an evolution process; The first processing unit is used to perform a solution processing according to the evolution process to obtain a global optimal solution of the energy function, and output the global optimal solution as the third information.

Images