Parameter sharing system and method for multi-unit energy efficiency diagnosis model of thermal power plant

Abstract

The invention relates to the technical field of control systems, in particular to a thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system and method, and the system comprises an operation condition feature extraction module which receives the real-time operation data of a target unit, extracts a boiler model, a steam turbine model and a current load value, and carries out the calculation of the current load value according to the boiler model; and screening matched units in the whole plant unit list, and obtaining a reference set of the same type of units. According to the method, by collecting operation data of a target unit in real time, extracting a boiler model and a steam turbine model and automatically screening units of the same type, a reference unit group is obtained for parameter sharing, and an energy efficiency benchmark with actual representativeness is established; meanwhile, non-linear weighted calculation is carried out by using power supply coal consumption, station service power consumption rate and NOx emission indexes, so that the diagnosis reference can sensitively and accurately reflect the real operation performance.

Description

Technical Field

[0001] The present invention relates to the technical field of control systems, and in particular to a system and method for sharing parameters of energy efficiency diagnosis models of multiple units in a thermal power plant. Background Art

[0002] The field of control system technology mainly involves the real-time collection and processing of various operating data and status parameters of equipment or process through automation and information technology, and the implementation of intelligent monitoring, optimization and regulation to achieve stable, safe and efficient operation of the system.

[0003] Existing technologies rely on real-time monitoring of equipment operating data and triggering of conventional alarm thresholds. They lack differentiation between unit types and real-time operating conditions, making it difficult to adapt to the characteristics of different units. The resulting diagnostic results are not universal enough, making it difficult to effectively guide equipment control and regulation, thereby reducing the economic benefits of operation. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a system and method for sharing parameters of energy efficiency diagnosis models of multiple units in a thermal power plant.

[0005] To achieve the above-mentioned object, the present invention adopts the following technical solution: A multi-unit energy efficiency diagnosis model parameter sharing system for a thermal power plant includes: The operating condition feature extraction module receives the real-time operating data of the target unit, extracts the boiler model, turbine model, and current load value, and selects matching units from the plant-wide unit list based on the boiler model to obtain a reference set of similar units. An energy efficiency diagnostic benchmark generation module queries the power supply coal consumption, plant power consumption rate, and NOx emission indicators of each unit based on the reference set of similar units, calculates the operating parameters of each unit in the reference set of similar units, and generates aggregated energy efficiency diagnostic benchmark parameters; An operation deviation quantification diagnosis module obtains the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit based on the aggregated energy efficiency diagnosis benchmark parameters, calculates the difference between the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit and the corresponding parameter items in the aggregated energy efficiency diagnosis benchmark parameters, obtains a multi-dimensional operation parameter deviation value, calculates each item of the multi-dimensional operation parameter deviation value to obtain a comprehensive deviation index, compares the comprehensive deviation index with a set alarm threshold, and outputs a unit energy efficiency status alarm signal; The diagnostic parameter feedforward preset module monitors the load change trend of the target unit and calculates the load change rate based on the energy efficiency status alarm signal of the unit, calculates the target load range of the next time period according to the load change rate, generates a predicted operating condition range identifier, calls the predicted operating condition range identifier to trigger the diagnosis of the new operating condition, and sends the diagnostic result to the monitoring node of the unit to establish a feedforward diagnostic parameter set.

[0006] Preferably, the steps for obtaining the reference set of similar units are: Receive real-time operating data of the target unit, extract the boiler model identifier, the turbine model identifier, and the current load value respectively, aggregate the extracted boiler model identifier, the turbine model identifier, and the current load value parameters into an operating condition identifier set, and generate an operating condition identifier set; Based on the operating condition identification set, locate the corresponding item of the boiler model identification, traverse all unit information in the plant unit list, read the content of the boiler model identification field of each unit one by one, and perform string matching and comparison between the boiler model identification field and the boiler model identification in the operating condition identification set to obtain a boiler model matching unit set; Based on the boiler model matching unit set, the host number of each boiler model matching unit is called, and the units that are currently operating normally and are not under maintenance or start-up and shutdown are screened. The units that meet the conditions are recorded as similar reference units to generate a similar unit reference set.

[0007] Preferably, the steps for obtaining the aggregate energy efficiency diagnostic benchmark parameters are: Based on the reference set of similar units, the operation database interface of each reference unit is called to extract the complete power supply coal consumption record series, the plant power consumption rate record series, and the nitrogen oxide emission concentration record series within the target time period. All sequence data are synchronously segmented, and missing time points are filled by linear interpolation to obtain the power supply coal consumption time series set, the plant power consumption rate time series set, and the nitrogen oxide emission concentration time series set; Calculate the comprehensive performance score of each reference unit based on the time series of power supply coal consumption, plant power consumption rate, and nitrogen oxide emission concentration. Based on the comprehensive performance score, an aggregate energy efficiency diagnostic benchmark parameter is calculated.

[0008] Preferably, the step of obtaining the multi-dimensional operating parameter deviation value is: Call the real-time data acquisition interface of the target unit to collect the real-time values ​​of the main steam pressure, main steam temperature and reheat steam temperature under the current operating conditions, and summarize the real-time values ​​according to the acquisition timestamp to form the real-time parameter sequence of the target unit; Based on the real-time parameter sequence of the target unit, the corresponding main steam pressure benchmark value, main steam temperature benchmark value and reheat steam temperature benchmark value at the same acquisition timestamp in the aggregated energy efficiency diagnosis benchmark parameters are called one by one, and the difference between the real-time value and the benchmark value is calculated item by item at the corresponding time point to obtain the main steam pressure difference sequence, the main steam temperature difference sequence and the reheat steam temperature difference sequence.

[0009] Preferably, the step of obtaining the multidimensional operating parameter deviation value further includes: combining the difference at each time point based on the main steam pressure difference sequence, the main steam temperature difference sequence and the reheat steam temperature difference sequence to form the multidimensional operating parameter deviation value.

[0010] Preferably, the steps of obtaining the energy efficiency status alarm signal of the unit are: Based on the multidimensional operating parameter deviation value, the main steam pressure deviation value, the main steam temperature deviation value, and the reheat steam temperature deviation value corresponding to each time stamp are read one by one, and the absolute values ​​of the main steam pressure deviation value, the main steam temperature deviation value, and the reheat steam temperature deviation value at each time stamp are calculated respectively to form a multidimensional operating parameter absolute deviation value set; Based on the multi-dimensional operating parameter absolute deviation value set, the main steam pressure absolute deviation value, the main steam temperature absolute deviation value, and the reheat steam temperature absolute deviation value are squared and summed in sequence for each time stamp, and then a square root operation is performed to form a comprehensive deviation index sequence; Based on the comprehensive deviation index sequence, the comprehensive deviation index at each timestamp is compared with the preset unit energy efficiency status alarm threshold one by one. If the comprehensive deviation index is greater than or equal to the alarm threshold, it is marked as an alarm state; otherwise, it is marked as a normal state, forming a unit energy efficiency status alarm signal.

[0011] Preferably, the steps of obtaining the predicted operating condition interval identifier are: Based on the energy efficiency status alarm signal of the unit, the power output value and the corresponding time stamp of the target unit in a continuous time period are read, the power value and the time stamp of the adjacent time period are sequentially calculated to generate the unit time load change rate, and all the unit time load change rates are arranged in chronological order as a load change rate sequence to obtain the load change rate sequence; Calculating upper and lower boundary values ​​of the predicted operating condition based on the load change rate sequence; Based on the predicted power upper boundary and the predicted power lower boundary, the upper and lower boundary values ​​are combined with the current timestamp and the predicted time interval length to form a predicted operating condition interval identifier.

[0012] Preferably, the steps of acquiring the feedforward diagnostic parameter set are: Calling the predicted operating condition interval identifier, parsing the predicted power upper limit and predicted power lower limit recorded therein, performing real-time monitoring of the predicted operating condition interval identifier, and retrieving the numerical difference between the real-time power output data and the upper and lower limits of the predicted interval in the real-time monitoring, to generate a real-time monitoring record of the operating condition deviation; Based on the real-time monitoring record of the operating condition deviation, it is determined whether the real-time power output data of the target unit continues to exceed the upper or lower limit of the prediction interval. If the real-time power output data exceeds either boundary, real-time diagnosis is triggered, and the real-time status database is called to perform operating condition diagnosis and generate real-time diagnosis results.

[0013] Preferably, the step of acquiring the feedforward diagnostic parameter set further comprises: sending a diagnostic message to a monitoring node of the target unit based on the real-time diagnostic result to establish a feedforward diagnostic parameter set.

[0014] The present invention also provides a method for sharing parameters of a multi-unit energy efficiency diagnosis model in a thermal power plant, comprising the following steps: Receive the real-time operating data of the target unit, extract the boiler model, turbine model, and current load value, and filter matching units from the plant-wide unit list based on the boiler model to obtain a reference set of similar units. Based on the reference set of similar units, query the power supply coal consumption, plant power consumption rate and NOx emission indicators of each unit, calculate the operating parameters of each unit in the reference set of similar units, and generate aggregated energy efficiency diagnosis benchmark parameters; Based on the aggregated energy efficiency diagnostic benchmark parameters, the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit are obtained, and the current parameters are subtracted one by one from the corresponding parameter items in the aggregated energy efficiency diagnostic benchmark parameters to obtain a multi-dimensional operating parameter deviation value. Each item of the multi-dimensional operating parameter deviation value is calculated to obtain a comprehensive deviation index, and the comprehensive deviation index is compared with a set alarm threshold to output a unit energy efficiency status alarm signal; Based on the energy efficiency status alarm signal of the unit, the load change trend of the target unit is monitored and the load change rate is calculated. The target load range of the next time period is calculated according to the load change rate, and a predicted operating condition range identifier is generated. The predicted operating condition range identifier is called to trigger the diagnosis of the new operating condition, and the diagnosis result is sent to the monitoring node of the unit to establish a feedforward diagnostic parameter set.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In this invention, by collecting target unit operating data in real time, extracting boiler and turbine models and automatically screening similar units, a reference unit group is obtained for parameter sharing, establishing a representative energy efficiency benchmark. Simultaneously, nonlinear weighted calculations are performed using power supply coal consumption, plant power consumption rate, and NOx emission indicators, enabling the diagnostic benchmark to sensitively and accurately reflect actual operating performance. The aggregated parameters formed based on this can be compared one by one with the real-time main steam pressure, main steam temperature, and reheat steam temperature to obtain a precisely quantified multi-dimensional parameter deviation. The degree of operating state deviation is then determined through calculation, enabling rapid identification of abnormal operating conditions and outputting alarm status, assisting in timely proactive intervention. Furthermore, by real-time monitoring of load change trends and rates, the operating range of the next cycle is predicted and asymmetric expansion is performed, improving the accuracy and sensitivity of load forecasting. Simultaneously, new operating condition diagnosis is performed in advance, enabling the unit to proactively respond to operating condition changes, improving operational stability and the foresight of energy efficiency management, and reducing operating energy consumption and emissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] See also Figure 1 The present invention provides a technical solution: a multi-unit energy efficiency diagnosis model parameter sharing system for a thermal power plant, comprising: The operating condition feature extraction module receives the real-time operating data of the target unit, extracts the boiler model, turbine model, and current load value, and selects matching units from the plant-wide unit list based on the boiler model to obtain a reference set of similar units. The energy efficiency diagnostic benchmark generation module queries the power supply coal consumption, plant power consumption rate, and NOx emission indicators of each unit based on a reference set of similar units. It calculates the operating parameters of each unit in the reference set of similar units and generates aggregated energy efficiency diagnostic benchmark parameters. The operation deviation quantification diagnosis module obtains the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit based on the aggregated energy efficiency diagnosis benchmark parameters, and calculates the difference between the corresponding parameter items in the aggregated energy efficiency diagnosis benchmark parameters one by one to obtain the multi-dimensional operation parameter deviation value. The various items of the multi-dimensional operation parameter deviation value are calculated to obtain a comprehensive deviation index, which is then compared with the set alarm threshold to output the unit energy efficiency status alarm signal; The diagnostic parameter feedforward preset module monitors the load change trend of the target unit and calculates the load change rate based on the unit energy efficiency status alarm signal. It then deduces the target load range of the next time period based on the load change rate, generates a predicted operating condition range identifier, calls the predicted operating condition range identifier to trigger the diagnosis of the new operating condition, and sends the diagnostic results to the unit's monitoring node to establish a feedforward diagnostic parameter set.

[0019] The steps to obtain the reference set of similar units are as follows: Receive real-time operating data of the target unit, extract the boiler model identifier, the turbine model identifier, and the current load value respectively, aggregate the extracted boiler model identifier, the turbine model identifier, and the current load value parameters into an operating condition identifier set, and generate an operating condition identifier set; Based on the operating condition identification set, locate the corresponding item of the boiler model identification, traverse all unit information in the plant unit list, read the content of the boiler model identification field of each unit one by one, and perform string matching and comparison between the boiler model identification field and the boiler model identification in the operating condition identification set to obtain the boiler model matching unit set; Based on the boiler model matching unit set, the host number of each boiler model matching unit is called, and the units that are currently operating normally and are not under maintenance or start-up and shutdown are screened. The units that meet the conditions are recorded as similar reference units to generate a reference set of similar units.

[0020] Specifically, the real-time operating data of the target unit is received. This data is a structured data stream containing multiple measurement point information that is continuously pushed from the power plant distributed control system (DCS) or real-time database (PISystem) at a frequency of once per second. Each record in the data stream contains a timestamp, a unique unit number, and a key-value pair of various operating parameters, such as "boiler_model: DG1025 / 18.2-II4", "turbine_model: N300-16.7 / 538 / 538", and "current_load_MW: 250.5". First, according to the target The unique number of the unit locks the corresponding data stream. Then, for each received real-time data record, the boiler model identifier with the key name "boiler_model", the turbine model identifier with the key name "turbine_model", and the current load value with the key name "current_load_MW" are directly accessed and extracted by parsing the data structure. The extracted current load value is then verified for validity to confirm that its value is between 0 and the rated capacity of the unit. Subsequently, the verified boiler model identifier, turbine model identifier, and current load value are combined as an independent structured object to generate an operating condition identifier set.

[0021] Based on the operating condition identification set, the boiler model identification is first read out, such as "DG1025 / 18.2-II4", and then the list of all units in the plant stored in the local or server database is accessed. The list is a pre-configured static data table that contains the basic information of all units in the plant. Its data structure includes fields such as "unit number", "boiler model identification", "turbine model identification", "rated capacity", and "design coal type". The system starts the traversal program, starting from the first row of records in the list, and reads the string content of the "boiler model identification" field line by line. Before comparison, the read list is checked. The boiler model identifier and the boiler model identifier obtained from the operating condition identifier set are standardized. This process includes removing blank characters at both ends of the string and converting them to uppercase to eliminate matching failures caused by inconsistent data entry formats. Then, the two standardized strings are compared character by character for exact equality. If the two strings are exactly the same, the "unit number" in the row record is added to a temporary set. This process continues until all unit records in the list are traversed. Finally, all successfully matched unit numbers in the temporary set are summarized to obtain a boiler model matching unit set.

[0022] Based on the boiler model matching unit set, the system iteratively processes each unit number in the set. For each unit number, it queries the real-time database for a set of key performance indicators of its current operating status, including three data tags: "unit operating status code", "maintenance flag" and "real-time load". Subsequently, the system executes a three-step screening logic to determine whether the unit is in a stable operating state. The first step is to check the "unit operating status code". The status code must be the preset "grid-connected operation" status value, such as the value 5. Any other status code such as "starting" or "shutdown" is considered to be unsatisfied. The second step is to check the "maintenance flag". The Boolean value must be "false", indicating that the unit is not currently in maintenance or test mode. The third step is to determine whether the unit is in the start-stop state. The process calculates the absolute value of its load change rate and compares it with a preset stable operation threshold. The threshold is set based on historical operation data analysis. For example, for a 300MW unit, the load fluctuation within 5 minutes during stable operation usually does not exceed 3MW. Therefore, the stable operation threshold is set to 0.6MW / minute. The calculation method is to take the absolute value of the difference between the real-time load at the current moment and the real-time load 5 minutes ago, and then divide it by the time interval of 5 minutes. If the calculated load change rate is less than 0.6MW / minute, it is considered that the unit is not in the start-up and shutdown process. Only units that meet the above three conditions at the same time will have their unit numbers recorded. After traversing all units in the boiler model matching unit set, all unit numbers that meet the conditions are summarized to generate a reference set of similar units.

[0023] The steps for obtaining the aggregate energy efficiency diagnostic benchmark parameters are as follows: Based on a reference set of similar units, the operating database interface of each reference unit is called to extract the complete power supply coal consumption record series, plant power consumption rate record series, and nitrogen oxide emission concentration record series within the target time period. All sequence data are synchronously segmented, and missing time points are filled in through linear interpolation to obtain the power supply coal consumption time series set, plant power consumption rate time series set, and nitrogen oxide emission concentration time series set; Based on the time series set of power supply coal consumption, plant power consumption rate and nitrogen oxide emission concentration, the comprehensive performance score of each reference unit is calculated. The calculation expression is: ; in, For the The comprehensive performance score of the reference unit, 、 、 Respectively The average power supply coal consumption, average plant power consumption rate and average nitrogen oxide emission concentration of each reference unit are 、 、 and 、 、 are the minimum and maximum values ​​of each indicator in all reference units, 、 、 is the importance weight corresponding to each indicator, satisfying ; Based on the comprehensive performance score, the aggregate energy efficiency diagnostic benchmark parameters are calculated using the following formula: ; in, For the moment Aggregate energy efficiency diagnostic benchmark parameters, Representative Reference unit at time Real-time values ​​of operating parameters, is the corresponding comprehensive performance score, is the total number of reference units.

[0024] Specifically, based on a reference set of similar units, for each reference unit number in the set, a data request is initiated through its corresponding operation database interface, which is a standard OPC (open platform communication) or proprietary database API, to extract the operation data in the past 24 hours. The specific requested measurement point labels are "power supply coal consumption", "plant power consumption rate" and "nitrogen oxide emission concentration". The acquired data is a set of discrete data points with high-precision timestamps. To ensure that the data of all units are aligned in time, the system establishes a unified time grid with a fixed interval of 1 minute, dividing 24 hours into 1440 time points. Then, the system traverses the original data sequence of each unit, and aggregates the data points that fall within the same 1-minute time window into a single time point at that minute by calculating the arithmetic mean. The system calculates the value of the asynchronous original sequence and converts it into a synchronous time series. In this process, if no data is collected at a certain time point, the point is marked as missing. The system will check the number of consecutive missing time points. If the consecutive missing time points exceed 15 minutes, the data of the unit during this period is considered unreliable and the unit will be eliminated from this calculation. For short-term data missing of less than 15 minutes, the system uses linear interpolation to fill the missing value, that is, linearly calculate and fill the missing value based on the values ​​of the two most recent valid data points before and after the missing point and their timestamps. Finally, three time-synchronized and data-complete sequences with a length of 1440 data points are generated for each reference unit, and the similar sequences of all units are aggregated to obtain the power supply coal consumption time series set, the plant power consumption rate time series set and the nitrogen oxide emission concentration time series set.

[0025] formula: The benefit of the formula lies in that it integrates multiple key energy efficiency and environmental protection indicators (power supply coal consumption, plant power consumption rate, and nitrogen oxide emissions) into a single comprehensive performance score through nonlinear transformation. Among them, the form of (1-normalized value) is used to convert indicators with lower values ​​into an evaluation system with higher scores. The application of the exponential function can significantly widen the score gap between units with different performance, making the weight advantage of excellent units more prominent. Finally, the hyperbolic tangent function smoothly maps the score to the interval (0, 1), forming a stable and standardized weight value. This comprehensive scoring mechanism avoids the one-sidedness of single indicator evaluation and can more comprehensively and objectively quantify the comprehensive operating level of each reference unit. 、 、 The steps to obtain the parameters are as follows: these three parameters represent the The average power supply coal consumption, average plant power consumption rate and average nitrogen oxide emission concentration of each reference unit in the target time period (for example, the past 24 hours) are obtained by taking the arithmetic mean of the power supply coal consumption time series, plant power consumption rate time series and nitrogen oxide emission concentration time series obtained in the previous step. For example, for reference unit 1, its power supply coal consumption time series contains 1440 data points in the past 24 hours, then It is the sum of the 1440 power supply coal consumption values ​​divided by 1440. For example, if the average power supply coal consumption of reference unit 1 is 295g / kWh, the average plant power rate is 5.5%, and the average nitrogen oxide emission concentration is 45mg / m³, then , , .

[0026] 、 、 and 、 、 The steps for obtaining parameters are as follows: these parameters represent the minimum and maximum values ​​of the corresponding indicators in all reference units, and the acquisition method is to traverse the average indicator value of all reference units ( 、 、 ), and find the global maximum and minimum values. For example, if the reference set of similar units contains 3 units, their average power supply coal consumption is 295g / kWh, 292g / kWh, and 298g / kWh respectively, then , .

[0027] 、 、 The steps for obtaining the parameters are as follows: these three parameters are the importance weights of each indicator, which are determined by the analytic hierarchy process (AHP). First, a review team consisting of the power plant chief engineer, operation experts, etc. compares the three indicators of power supply coal consumption (H), plant power consumption rate (E), and nitrogen oxide emissions (N) in pairs to construct a judgment matrix. For example, if the experts believe that "H is obviously more important than E", the value is assigned to 3; "H is extremely important than N", the value is assigned to 5; "E is obviously more important than N", the value is assigned to 3, and the judgment matrix is ​​obtained. , by calculating the maximum eigenvalue of the matrix and its corresponding eigenvector, and normalizing the eigenvector to obtain the weight, the calculation process is: first normalize each column of the matrix, then sum the normalized matrix by row, and finally normalize the row and vector again to obtain the weight vector. After calculation, we can get , , , and satisfies .

[0028] Calculation process: Taking a set containing three reference units as an example, its average index and weight are as follows: Reference Unit 1: g / kWh, %, mg / m³; Refer to Unit 2: g / kWh, %, mg / m³; Refer to Unit 3: g / kWh, %, mg / m³; Global maximum value: , ; , ; , ; Importance Weight: , , ; Calculate the overall performance score for reference unit 2 : Normalized energy consumption item: ; Normalized electrical rate term: ; Normalized emissions: ; Compute the weighted sum: ; Calculate the final score: ; The results show that the comprehensive performance score of reference unit 2 is 0.954, which is very close to 1, indicating that this unit has the best operating performance in the current reference set after comprehensively considering the three aspects of power supply coal consumption, plant power consumption rate and nitrogen oxide emissions.

[0029] formula: ,The benefit of the formula is that it creates a dynamic ,performance-oriented diagnostic benchmark that utilizes the ,comprehensive performance score calculated in the previous step. As a weight, the real-time operating parameters of similar units (such as main steam pressure, temperature, etc.) are weighted averaged to make the units that are currently operating more efficiently and more environmentally friendly (i.e. Units with higher values ​​have a greater say in the formation of the benchmark, so that the generated aggregate energy efficiency diagnostic benchmark parameters are not only achievable but also represent the best operating practices among similar units, providing a more challenging and instructive benchmark for the target units. The steps to obtain the parameter are as follows: Reference unit at time The real-time value of a specific operating parameter (such as main steam pressure) is obtained directly from the synchronized time series data set processed in the first step, according to the unit number and timestamp Perform query extraction. For example, to calculate the main steam pressure benchmark at 10:30, query the main steam pressure measurement point values ​​of reference units 1, 2, and 3 at 10:30 respectively. If the measured values ​​are 16.80MPa, 16.85MPa, and 16.75MPa respectively, then , , .

[0030] The steps to obtain the parameter are as follows: the parameter is the first The overall performance score of the reference unit.

[0031] The parameter is obtained by counting the number of elements in the reference set of similar units. For example, if there are 3 reference units that meet the conditions in the set, then .

[0032] Calculation process: Continuing to use the three reference units in the previous example, their comprehensive performance scores are Already calculated, now need to calculate at time Main steam pressure reference value : The comprehensive performance scores of each unit are known: , , ; At the moment Main steam pressure of each unit for: MPa; MPa; MPa; Compute the numerator of a weighted sum: ; ; ; Calculate the denominator of the weight: ; Calculate the baseline value: MPa; This result shows that at time The main steam pressure benchmark value in the aggregate energy efficiency diagnosis benchmark parameters is 16.799 MPa. This value is not the average value of the pressure of each reference unit (16.80 MPa), but is offset to the pressure value of reference unit 2 with the highest comprehensive performance score (16.85 MPa). This shows that the benchmark value is dynamically generated based on the current optimal operating practice. It will serve as the "gold standard" for whether the main steam pressure of the target unit deviates at this moment.

[0033] The steps for obtaining the multi-dimensional operating parameter deviation value are as follows: Call the real-time data acquisition interface of the target unit to collect the real-time values ​​of the main steam pressure, main steam temperature and reheat steam temperature under the current operating conditions, and summarize the real-time values ​​according to the acquisition timestamp to form the real-time parameter sequence of the target unit; Based on the real-time parameter sequence of the target unit, the corresponding main steam pressure benchmark value, main steam temperature benchmark value, and reheat steam temperature benchmark value at the same acquisition time stamp in the aggregated energy efficiency diagnosis benchmark parameters are called one by one. The difference between the real-time value and the benchmark value is calculated item by item at the corresponding time point to obtain the main steam pressure difference sequence, main steam temperature difference sequence, and reheat steam temperature difference sequence; Based on the main steam pressure difference sequence, the main steam temperature difference sequence and the reheat steam temperature difference sequence, the difference at each time point is combined to form a multi-dimensional operating parameter deviation value.

[0034] Specifically, the real-time data acquisition interface of the target unit is called. This interface continuously obtains data streams from the unit's distributed control system (DCS) at a frequency of once per second through the open platform communication protocol (OPC). The system subscribes to specific measurement point tags based on the unique identifier of the target unit, such as the measurement point tag "21MPA101AA001" for the main steam pressure, the measurement point tag "21MTA101AA001" for the main steam temperature, and the measurement point tag "21MTA201AA001" for the reheat steam temperature. For each set of real-time values ​​collected with a timestamp, the system will first verify the validity and compare it with the safe operating limits specified in the unit design manual. For example, the collected main steam pressure value is compared with the range of 15MPa to 18MPa, and the main steam temperature and reheat steam temperature are compared with the range of 535℃ to 545℃. Only data points whose values ​​are within their respective valid ranges will be accepted. After verification, the real-time values ​​of the main steam pressure, main steam temperature and reheat steam temperature at the same timestamp are combined into a structured data record. For example, the data record at "10:30:01" is {pressure: 16.82MPa, main temperature: 539.5℃, reheat temperature: 540.2℃}. These data records are arranged in timestamp order to form a real-time parameter sequence of the target unit.

[0035] Based on the real-time parameter sequence of the target unit, the system traverses each data record in the sequence in timestamp order. For each timestamp, such as "10:30:01", the system reads the real-time parameters at that moment from the real-time parameter sequence of the target unit, namely {pressure: 16.82MPa, main temperature: 539.5℃, retemperature: 540.2℃}. At the same time, the system uses the same timestamp "10:30:01" as an index to query the corresponding benchmark value in the aggregated energy efficiency diagnosis benchmark parameters generated in the previous step. The query result is {pressure benchmark: 16.799MPa, main temperature benchmark: 540.0℃, retemperature benchmark: 541.0 ℃}, then the system performs difference calculations on these three parameters one by one, specifically by subtracting the corresponding benchmark values ​​from the real-time values, that is, the main steam pressure difference is calculated as 16.82 minus 16.799 to obtain +0.021MPa, the main steam temperature difference is calculated as 539.5 minus 540.0 to obtain -0.5℃, and the reheat steam temperature difference is calculated as 540.2 minus 541.0 to obtain -0.8℃. The three calculated differences are recorded together with their corresponding timestamps and stored in three independent sequences respectively. After traversing all time points, the main steam pressure difference sequence, the main steam temperature difference sequence, and the reheat steam temperature difference sequence are finally obtained.

[0036] Based on the main steam pressure difference sequence, main steam temperature difference sequence and reheat steam temperature difference sequence, since these three sequences are generated based on the same set of timestamps, the system performs synchronous iteration based on the timestamps. For each time point, for example, "10:30:01", the system takes out the pressure difference +0.021 at that moment from the main steam pressure difference sequence, takes out the temperature difference -0.5 at that moment from the main steam temperature difference sequence, and takes out the temperature difference -0.8 at that moment from the reheat steam temperature difference sequence. Then, the system combines these three independent timestamps. The quantity differences are combined into a three-dimensional vector in a pre-set order (for example, pressure, main temperature, and retemperature), namely [+0.021, -0.5, -0.8]. This vector is an instance of a multidimensional operating parameter deviation value at that moment. It represents the direction and magnitude of the deviation of the current operating state of the target unit from the optimal benchmark in three-dimensional space. The system repeats this combination operation for all time points within the target time period, and arranges the three-dimensional vectors generated at each time point in chronological order, ultimately forming a time series composed of multidimensional vectors, namely, the multidimensional operating parameter deviation value.

[0037] The steps for obtaining the unit energy efficiency status alarm signal are as follows: Based on the multi-dimensional operating parameter deviation values, the main steam pressure deviation value, main steam temperature deviation value, and reheat steam temperature deviation value corresponding to each time stamp are read one by one, and the absolute values ​​of the main steam pressure deviation value, main steam temperature deviation value, and reheat steam temperature deviation value at each time stamp are calculated respectively to form a multi-dimensional operating parameter absolute deviation value set; Based on the absolute deviation value set of multi-dimensional operating parameters, the main steam pressure absolute deviation value, the main steam temperature absolute deviation value, and the reheat steam temperature absolute deviation value are squared and summed in sequence at each time stamp, and then the square root operation is performed to form a comprehensive deviation index sequence; Based on the comprehensive deviation index sequence, the comprehensive deviation index at each timestamp is compared with the preset unit energy efficiency status alarm threshold one by one. If the comprehensive deviation index is greater than or equal to the alarm threshold, it is marked as an alarm state, otherwise it is marked as a normal state, forming a unit energy efficiency status alarm signal.

[0038] Specifically, based on the multidimensional operating parameter deviation value, the system traverses and processes each three-dimensional vector in the sequence in timestamp order. For any timestamp, such as "10:30:01", the system first reads its corresponding multidimensional operating parameter deviation value, that is, the vector [+0.021, -0.5, -0.8]. The three components of this vector are the main steam pressure deviation value, the main steam temperature deviation value, and the reheat steam temperature deviation value, respectively. Then, the system independently performs absolute value calculation on each component in the vector, and calculates the main steam pressure deviation value +0.021MPa as 0.021MPa. , the main steam temperature deviation value of -0.5℃ is calculated as 0.5℃, and the reheat steam temperature deviation value of -0.8℃ is calculated as 0.8℃. Through this calculation, whether the original deviation is positive or negative, it is converted into a positive value indicating the degree of deviation. The three values ​​after absolute value calculation, namely 0.021, 0.5 and 0.8, are recombined with the original timestamp "10:30:01" into a new record. The system repeats this process for all time points in the multidimensional operating parameter deviation value sequence, and all newly generated records are collected to form a multidimensional operating parameter absolute deviation value set.

[0039] Based on the set of absolute deviation values ​​of multi-dimensional operating parameters, the system processes the records in the set by timestamp. Before performing the calculation, in order to eliminate the influence of different physical dimensions (MPa and ℃) and the difference in numerical range, the system first normalizes the absolute deviation value of main steam pressure, main steam temperature and reheat steam temperature. The divisor used for normalization is the pre-set maximum allowable deviation value of each parameter, which is determined by the unit design specifications and long-term operating experience. For example, the maximum allowable deviation value of main steam pressure is set to 0.5MPa, and the maximum allowable deviation values ​​of main and reheat steam temperatures are both set to 5℃. For the absolute deviation value record {0.021Mpa} with the timestamp "10:30:01", the system calculates the absolute deviation value of the main steam pressure, the main steam temperature and the reheat steam temperature. Pa, 0.5°C, 0.8°C}, the normalized calculations are 0.021 divided by 0.5 to get 0.042, 0.5 divided by 5 to get 0.1, and 0.8 divided by 5 to get 0.16, respectively. This results in a dimensionless normalized absolute deviation vector [0.042, 0.1, 0.16]. The system then squares and sums the three components of this vector, that is, it calculates the square of 0.042 plus the square of 0.1 plus the square of 0.16, to get 0.037364. Finally, it takes the square root of this sum to obtain the comprehensive deviation index at that time point, which is approximately 0.193. Each calculated comprehensive deviation index is associated with its corresponding timestamp to form a comprehensive deviation index sequence.

[0040] Based on the comprehensive deviation index sequence, the system processes each comprehensive deviation index in the sequence one by one and compares it with the preset unit energy efficiency status alarm threshold. The setting of the alarm threshold is based on the statistical analysis of the unit's historical operating data. The specific method is to first screen out all the periods in the past year when the load of the target unit is stable in the range of 90% to 100% of the rated load and all energy efficiency indicators are rated as excellent, extract the comprehensive deviation index data corresponding to these periods, form a sample set representing the "healthy state", and then calculate the arithmetic mean and standard deviation of the sample set. For example, the average comprehensive deviation index in the healthy state is calculated to be 0.08, and the standard deviation is 0.09. The standard deviation is 0.04, and the alarm threshold is set as the mean value plus three times the standard deviation based on the "3-sigma" principle in statistics, that is, 0.08 plus 3 multiplied by 0.04 equals 0.20. For the comprehensive deviation index at any time stamp, for example, 0.193, the system compares it with the alarm threshold of 0.20. Since 0.193 is less than 0.20, the time point is marked as normal. If the comprehensive deviation index at another time point is 0.21, then because it is greater than or equal to 0.20, the time point will be marked as an alarm state. The system arranges the status marks (normal or alarm) of all time points in chronological order to form the unit energy efficiency status alarm signal.

[0041] The steps to obtain the predicted operating condition interval identifier are: Based on the unit energy efficiency status alarm signal, the power output value and corresponding timestamp of the target unit in a continuous time period are read, the power value and timestamp of the adjacent time period are calculated in sequence to generate the unit time load change rate, and all the unit time load change rates are arranged in chronological order as a load change rate sequence to obtain the load change rate sequence; According to the load change rate sequence, the upper and lower boundary values ​​of the predicted working condition are calculated using the following formula: ; ; in, To predict the upper power limit, To predict the lower power boundary, is the power output value at the current moment, is the exponentially weighted moving average of the load change rate, is the standard deviation of the load change rate series, is the length of the predicted time interval, To expand the scale of the confidence coefficient control interval, is the asymmetric adjustment coefficient, is the trend direction function, and the value selection rule is , indicating the current load change direction; Based on the predicted power upper boundary and the predicted power lower boundary, the upper and lower boundary values ​​are combined with the current timestamp and the predicted time interval length to form a predicted operating condition interval identifier.

[0042] Specifically, based on the unit energy efficiency status alarm signal, once the alarm status mark is received, the system immediately starts the reading program of the historical power data of the target unit, and retrieves the power output value and corresponding timestamp with a sampling interval of 1 minute in the past 60 minutes through the real-time database interface to form a power time series containing 60 data points. For example, the data point sequence may be [(10:00, 250.0MW), (10:01, 250.2MW), (10:02, 250.1MW), ..., (11:00, 255 .0MW)], the system then traverses the sequence, performs operations on every two adjacent data points, subtracts the power value of the previous time point from the power value of the latter time point, and divides it by the fixed time interval of 1 minute, thereby calculating the load change rate per unit time within that minute. For example, the load change rate at 10:01 is (250.2 minus 250.0) divided by 1, which is +0.2MW / minute. This calculation will generate a new sequence containing 59 rate values, and these rate values ​​are arranged in chronological order to obtain the load change rate sequence.

[0043] formula: , The benefit of the formula is that it constructs an asymmetric dynamic prediction interval and introduces an asymmetric adjustment coefficient and trend direction function , which allows the width of the prediction interval to be adaptively adjusted based on the current trend of load changes (increasing or decreasing). When the load is on an upward trend, the upper boundary will have a wider margin than the lower boundary, and vice versa. This design abandons the rigidity of traditional symmetrical prediction intervals and better adapts to the inertia and nonlinear characteristics of the thermal power unit load regulation process. This provides a more accurate and reliable range of future operating conditions, providing high-quality input for downstream diagnostic parameter feedforward presets. The parameter acquisition steps are as follows: the parameter represents the unit power output value at the current moment when the prediction calculation is triggered. It is directly obtained through the target unit real-time data acquisition interface without any transformation. Its value is the starting point of the calculation. For example, when the calculation is triggered at 11:00:00, the power output value at that moment read from the real-time database is 255.0MW, then .

[0044] The parameter is obtained as follows: This parameter is the exponentially weighted moving average of the load change rate, which is used to smoothly reflect the core trend of the load change. Its calculation is based on the load change rate sequence obtained in the previous step. The calculation formula is: ,in is the latest load change rate, is the smoothing coefficient, Sure, For the review period, according to the load dispatch response characteristics of the power plant, set Minutes, then , if the latest value of the load change rate sequence is +0.9MW / min, and the last minute's is +0.75MW / min, then the current MW / min.

[0045] The steps for obtaining the parameter are as follows: the parameter represents the uncertainty or fluctuation of the load change rate. Its value is obtained by calculating the standard deviation of the load change rate sequence obtained in the previous step (for example, 59 rate values ​​in the past 60 minutes). For example, after calculating the rate sequence in the past hour, we get MW / min.

[0046] The parameter acquisition steps are as follows: This parameter is the length of the predicted time interval. Its setting basis is the typical lead time required for the unit control system to adjust the parameters and take effect. For large thermal power units, it usually takes 10 to 20 minutes of preparation time to cope with a large range of load changes. Therefore, a fixed prediction time is set based on experience, for example minute.

[0047] The steps to obtain the parameter are as follows: the parameter is the confidence coefficient, which is used to control the expansion scale of the prediction interval. Its value is related to the expected prediction confidence. Referring to statistical theory, if a confidence level of about 95% is required, The value is usually taken as 1.96. In order to leave more sufficient margin in engineering applications, combined with the requirements of the grid safety operation regulations, Set to 2.0.

[0048] The steps for obtaining the parameter are as follows: the parameter is an asymmetric adjustment coefficient with a value range of [0, 1]. It is used to adjust the asymmetry of the upward or downward trend of the prediction interval. Its setting is based on the analysis of the historical variable load process data of the unit. By statistically analyzing the asymmetry of the amplitude of the actual load exceeding the linear prediction under different variable load rates, for example, the analysis found that during the load increase process of the unit, the amplitude of the actual load exceeding the predicted mean is 40% larger than the amplitude below the predicted mean, then it can be set .

[0049] The steps to obtain the parameter are as follows: This parameter is a trend direction function, which is used to determine the main trend direction of the current load change. Its value is The symbol of If it is positive, the value is +1; if it is negative, the value is -1; if it is 0, the value is 0, for example, if ,but .

[0050] Calculation process: Use the above parameter values ​​to deduce the example: MW, MW / min, MW / min, min, , , .

[0051] First calculate the predicted central value: MW.

[0052] Calculate the upper boundary extension: MW.

[0053] Calculate the lower boundary extension: MW.

[0054] Calculate the final upper and lower bounds: MW.

[0055] MW.

[0056] The results show that within the next 15 minutes, the power output of the target unit is expected to fluctuate between 261.255MW and 279.255MW. This range represents the range of possible operating conditions in the future. Due to the current upward trend in load, the upper boundary of this range has a much larger expansion (12.6MW) than the predicted central value, which is much larger than the contraction (5.4MW) of the lower boundary, accurately reflecting the asymmetric impact of the trend.

[0057] Based on the predicted power upper boundary of 279.255MW and the predicted power lower boundary of 261.255MW, the system structures these two values ​​with the current timestamp that triggers this calculation, such as "11:00:00", and the set prediction time interval length, i.e., 15 minutes. Specifically, the system creates a data object containing four fields, namely "prediction start time", "prediction duration", "prediction power lower limit" and "prediction power upper limit", and fills in the corresponding values ​​to generate a record: {Prediction start time: "11:00:00", prediction duration: 15 minutes, prediction power lower limit: 261.255MW, prediction power upper limit: 279.255MW}. This structured data object is the prediction operating condition interval identifier.

[0058] The steps for obtaining the feedforward diagnostic parameter set are: Call the predicted operating condition interval identifier, parse the predicted power upper limit and predicted power lower limit recorded therein, perform real-time monitoring of the predicted operating condition interval identifier, and retrieve the numerical difference between the real-time power output data and the upper and lower limits of the predicted interval in the real-time monitoring to generate a real-time monitoring record of the operating condition deviation; Based on the real-time monitoring records of operating condition deviations, it is determined whether the real-time power output data of the target unit continues to exceed the upper or lower limit of the prediction interval. If the real-time power output data exceeds either boundary, a real-time diagnosis is triggered, and the real-time status database is called to perform operating condition diagnosis and generate real-time diagnosis results; Based on the real-time diagnosis results, the diagnosis message is sent to the monitoring node of the target unit to establish a feedforward diagnosis parameter set.

[0059] Specifically, the system calls the predicted operating condition interval identifier, and first parses out the recorded predicted power upper limit, such as 279.255MW, and predicted power lower limit, such as 261.255MW, as well as the predicted start time "11:00:00" and the predicted duration of 15 minutes. Starting from this start time, a real-time monitoring cycle is started, and the real-time power output data of the target unit is obtained through the real-time data acquisition interface at a frequency of once every 10 seconds. For each collected power value, the system calculates the difference between its value and the upper and lower limits of the prediction interval. If the real-time power output data is within the interval, the difference value is If it is 0, if it is higher than the upper limit, the difference value is the real-time value minus the upper limit; if it is lower than the lower limit, the difference value is the real-time value minus the lower limit. The difference value, real-time power value, timestamp and current status (within the interval, exceeding the upper limit, or below the lower limit) calculated each time are packaged into a record. For example, at "11:05:30", if the real-time power is 280.5MW, a record {timestamp: "11:05:30", real-time power: 280.5MW, status: "exceeding the upper limit", difference value: +1.245MW} is generated. These records are stored in chronological order to form a real-time monitoring record of operating condition deviation.

[0060] Based on the real-time monitoring and recording of operating condition deviations, the system starts a continuous judgment logic. The core of this logic is to judge whether the unit power "continuously" exceeds the prediction interval. The "continuous" here is defined by a sliding time window and a counter mechanism. The size of the sliding window is set according to the response time of the unit load adjustment, for example, it is set to 120 seconds. The counter records the number of times the prediction interval is exceeded within the window. The system sets a trigger threshold. The basis for determining the threshold is to avoid false alarms caused by transient disturbances or measurement noise. By analyzing the short-term fluctuations under normal operating conditions in historical data, it is set that when more than 80% of the data points in the 120-second window (that is, at least 10 times, corresponding to 12 When all 12 sampling points within 10 seconds show that they exceed the same boundary, it is judged as "continuously exceeding". For example, starting from "11:05:30", the system finds that the status of 10 consecutive sampling points is "exceeding the upper limit". At this time, the trigger condition is met, and the system immediately triggers a real-time diagnosis process. The process first calls the real-time status database to extract the real-time values ​​of a series of key parameters related to the current operating conditions (power of approximately 280MW), such as coal feed rate, air supply volume, steam temperature and pressure at all levels, etc., and then re-executes a complete energy efficiency diagnosis calculation, including generating new aggregated energy efficiency diagnosis benchmark parameters, calculating multi-dimensional operating parameter deviation values ​​and comprehensive deviation indicators, and generating real-time diagnosis results.

[0061] Based on the real-time diagnostic results, which are structured information packets, they contain comprehensive deviation indicators for the current operating conditions, deviation values ​​for various parameters, and possible causes preliminarily determined based on a preset fault diagnosis rule base. For example, "excessive reheater water spray leads to low reheat steam temperature." The system formats this information packet and converts it into a message format that conforms to a communication protocol (such as Modbus or a proprietary protocol) that can be received by the target unit monitoring node (usually the operator or engineer station). The message content clearly includes the alarm level, the recommended operating parameter range under the new operating conditions (for example, a 5% reduction in the opening of the reheater water spray control valve), and the corresponding diagnostic baseline value. The system then sends this diagnostic message directly to the designated monitoring node IP address and port via the power plant's internal LAN. An alarm window pops up on the operator station's monitoring interface, displaying detailed diagnostic and recommendation information. This process establishes a dynamically updated, forward-looking parameter guidance set, known as a feedforward diagnostic parameter set.

Claims

1. A parameter sharing system for energy efficiency diagnosis model of multiple units in thermal power plants, characterized in that: The system comprises: The operating condition feature extraction module receives the real-time operating data of the target unit, extracts the boiler model, turbine model, and current load value, and selects matching units from the plant-wide unit list based on the boiler model to obtain a reference set of similar units. An energy efficiency diagnostic benchmark generation module queries the power supply coal consumption, plant power consumption rate, and NOx emission indicators of each unit based on the reference set of similar units, calculates the operating parameters of each unit in the reference set of similar units, and generates aggregated energy efficiency diagnostic benchmark parameters; An operation deviation quantification diagnosis module obtains the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit based on the aggregated energy efficiency diagnosis benchmark parameters, calculates the difference between the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit and the corresponding parameter items in the aggregated energy efficiency diagnosis benchmark parameters, obtains a multi-dimensional operation parameter deviation value, calculates each item of the multi-dimensional operation parameter deviation value to obtain a comprehensive deviation index, compares the comprehensive deviation index with a set alarm threshold, and outputs a unit energy efficiency status alarm signal; The diagnostic parameter feedforward preset module monitors the load change trend of the target unit and calculates the load change rate based on the energy efficiency status alarm signal of the unit, calculates the target load range of the next time period according to the load change rate, generates a predicted operating condition range identifier, calls the predicted operating condition range identifier to trigger the diagnosis of the new operating condition, and sends the diagnostic result to the monitoring node of the unit to establish a feedforward diagnostic parameter set.

2. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 1 is characterized in that: The steps for obtaining the reference set of similar units are as follows: Receive real-time operating data of the target unit, extract the boiler model identifier, the turbine model identifier, and the current load value respectively, aggregate the extracted boiler model identifier, the turbine model identifier, and the current load value parameters into an operating condition identifier set, and generate an operating condition identifier set; Based on the operating condition identification set, locate the corresponding item of the boiler model identification, traverse all unit information in the plant unit list, read the content of the boiler model identification field of each unit one by one, and perform string matching and comparison between the boiler model identification field and the boiler model identification in the operating condition identification set to obtain a boiler model matching unit set; Based on the boiler model matching unit set, the host number of each boiler model matching unit is called, and the units that are currently operating normally and are not under maintenance or start-up and shutdown are screened. The units that meet the conditions are recorded as similar reference units to generate a similar unit reference set.

3. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 1, characterized in that: The steps for obtaining the aggregate energy efficiency diagnostic benchmark parameters are as follows: Based on the reference set of similar units, the operation database interface of each reference unit is called to extract the complete power supply coal consumption record series, the plant power consumption rate record series, and the nitrogen oxide emission concentration record series within the target time period. All sequence data are synchronously segmented, and missing time points are filled by linear interpolation to obtain the power supply coal consumption time series set, the plant power consumption rate time series set, and the nitrogen oxide emission concentration time series set; Calculate the comprehensive performance score of each reference unit based on the time series of power supply coal consumption, plant power consumption rate, and nitrogen oxide emission concentration. Based on the comprehensive performance score, an aggregate energy efficiency diagnostic benchmark parameter is calculated.

4. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 1, characterized in that: The steps for obtaining the multi-dimensional operating parameter deviation value are as follows: Call the real-time data acquisition interface of the target unit to collect the real-time values ​​of the main steam pressure, main steam temperature and reheat steam temperature under the current operating conditions, and summarize the real-time values ​​according to the acquisition timestamp to form the real-time parameter sequence of the target unit; Based on the real-time parameter sequence of the target unit, the corresponding main steam pressure benchmark value, main steam temperature benchmark value and reheat steam temperature benchmark value at the same acquisition timestamp in the aggregated energy efficiency diagnosis benchmark parameters are called one by one, and the difference between the real-time value and the benchmark value is calculated item by item at the corresponding time point to obtain the main steam pressure difference sequence, the main steam temperature difference sequence and the reheat steam temperature difference sequence.

5. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 4 is characterized in that: The step of obtaining the multidimensional operating parameter deviation value also includes: combining the difference at each time point based on the main steam pressure difference sequence, the main steam temperature difference sequence and the reheat steam temperature difference sequence to form the multidimensional operating parameter deviation value.

6. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 1, characterized in that: The steps for obtaining the energy efficiency status alarm signal of the unit are as follows: Based on the multidimensional operating parameter deviation value, the main steam pressure deviation value, the main steam temperature deviation value, and the reheat steam temperature deviation value corresponding to each time stamp are read one by one, and the absolute values ​​of the main steam pressure deviation value, the main steam temperature deviation value, and the reheat steam temperature deviation value at each time stamp are calculated respectively to form a multidimensional operating parameter absolute deviation value set; Based on the multi-dimensional operating parameter absolute deviation value set, the main steam pressure absolute deviation value, the main steam temperature absolute deviation value, and the reheat steam temperature absolute deviation value are squared and summed in sequence for each time stamp, and then a square root operation is performed to form a comprehensive deviation index sequence; Based on the comprehensive deviation index sequence, the comprehensive deviation index at each timestamp is compared with the preset unit energy efficiency status alarm threshold one by one. If the comprehensive deviation index is greater than or equal to the alarm threshold, it is marked as an alarm state; otherwise, it is marked as a normal state, forming a unit energy efficiency status alarm signal.

7. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 1, characterized in that: The steps for obtaining the predicted operating condition interval identifier are: Based on the energy efficiency status alarm signal of the unit, the power output value and the corresponding time stamp of the target unit in a continuous time period are read, the power value and the time stamp of the adjacent time period are sequentially calculated to generate the unit time load change rate, and all the unit time load change rates are arranged in chronological order as a load change rate sequence to obtain the load change rate sequence; Calculating upper and lower boundary values ​​of the predicted operating condition based on the load change rate sequence; Based on the predicted power upper boundary and the predicted power lower boundary, the upper and lower boundary values ​​are combined with the current timestamp and the predicted time interval length to form a predicted operating condition interval identifier.

8. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 1, characterized in that: The steps for obtaining the feedforward diagnostic parameter set are: Calling the predicted operating condition interval identifier, parsing the predicted power upper limit and predicted power lower limit recorded therein, performing real-time monitoring of the predicted operating condition interval identifier, and retrieving the numerical difference between the real-time power output data and the upper and lower limits of the predicted interval in the real-time monitoring, to generate a real-time monitoring record of the operating condition deviation; Based on the real-time monitoring record of the operating condition deviation, it is determined whether the real-time power output data of the target unit continues to exceed the upper or lower limit of the prediction interval. If the real-time power output data exceeds either boundary, real-time diagnosis is triggered, and the real-time status database is called to perform operating condition diagnosis and generate real-time diagnosis results.

9. The thermal power plant multi-unit energy efficiency diagnosis model parameter sharing system according to claim 8, characterized in that: The step of acquiring the feedforward diagnostic parameter set further includes: sending a diagnostic message to a monitoring node of the target unit based on the real-time diagnostic result to establish a feedforward diagnostic parameter set.

10. The method for sharing parameters of a multi-unit energy efficiency diagnosis model in a thermal power plant according to any one of claims 1 to 9, characterized in that: The following steps are involved: Receive the real-time operating data of the target unit, extract the boiler model, turbine model, and current load value, and filter matching units from the plant-wide unit list based on the boiler model to obtain a reference set of similar units. Based on the reference set of similar units, query the power supply coal consumption, plant power consumption rate and NOx emission indicators of each unit, calculate the operating parameters of each unit in the reference set of similar units, and generate aggregated energy efficiency diagnosis benchmark parameters; Based on the aggregated energy efficiency diagnostic benchmark parameters, the current main steam pressure, main steam temperature, and reheat steam temperature of the target unit are obtained, and the current parameters are subtracted one by one from the corresponding parameter items in the aggregated energy efficiency diagnostic benchmark parameters to obtain a multi-dimensional operating parameter deviation value. Each item of the multi-dimensional operating parameter deviation value is calculated to obtain a comprehensive deviation index, and the comprehensive deviation index is compared with a set alarm threshold to output a unit energy efficiency status alarm signal; Based on the energy efficiency status alarm signal of the unit, the load change trend of the target unit is monitored and the load change rate is calculated. The target load range of the next time period is calculated according to the load change rate, and a predicted operating condition range identifier is generated. The predicted operating condition range identifier is called to trigger the diagnosis of the new operating condition, and the diagnosis result is sent to the monitoring node of the unit to establish a feedforward diagnostic parameter set.

Images