A Deep Peak Shaving System for Thermal Power Units Based on Waste Heat Recovery

Abstract

This invention relates to the field of energy-saving technology and discloses a deep peak-shaving system for thermal power units based on waste heat recovery. The system includes: a monitoring unit for real-time monitoring of the unit's waste heat resource operation data and working data; a central control unit for adjusting deep peak-shaving by matching different waste heat resources according to the demand characteristics of deep peak-shaving; and a first control module for generating deep peak-shaving data and extracting demand characteristics based on the data; a second control module for classifying each waste heat resource according to its operation data; a third control module for matching corresponding waste heat resources according to the demand characteristics; a fourth control module for setting the utilization status of each waste heat resource based on its operation data; and a storage unit for storing the operation data and working data acquired by the monitoring unit. This application effectively reduces resource waste and improves resource utilization.

Description

Technical Field

[0001] This application relates to the field of energy-saving technology, and in particular to a deep peak-shaving system for thermal power units based on waste heat recovery. Background Technology

[0002] With the rapid development of new power systems based on new energy sources, the proportion of intermittent power sources such as wind power and photovoltaic power is constantly increasing, and the demand for flexible resource regulation in the power grid is becoming increasingly urgent. As the ballast stone of the current power supply, coal-fired power units are shifting from baseload power sources to peak-shaving power sources, and undertaking deep peak-shaving tasks has become the norm. Deep peak-shaving requires units to operate stably for a long time under conditions far below their design rated load.

[0003] When thermal power units are carrying out deep peak shaving, there will be waste heat resources with different contents and energy densities in different locations. These waste heat resources usually cannot be effectively utilized. There will be different needs and different focuses at different times during the deep peak shaving process. Conventional peak shaving methods cannot flexibly regulate waste heat resources to meet the deep peak shaving needs while ensuring the safe operation of thermal power units. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a deep peak-shaving system for thermal power units based on waste heat recovery, aiming to effectively reduce resource waste and improve resource utilization. Through detailed analysis of the deep peak-shaving process, it effectively regulates the peak-shaving process at different stages.

[0005] In some embodiments of this application, a deep peak-shaving system for thermal power units based on waste heat recovery is provided, characterized in that it includes: The monitoring unit is used to monitor the real-time operation data of the waste heat resources of the unit and the working data of the thermal power unit. The central control unit is used to adjust the deep peak shaving according to the demand characteristics of different waste heat resources; The central control unit includes: The first control module is used to generate deep peak-shaving data and extract required features based on the deep peak-shaving data. The second control module is used to classify each waste heat resource into different levels based on the operating data of the waste heat resources. The third control module is used to match corresponding waste heat resources according to the aforementioned demand characteristics; The fourth control module is used to set the utilization status of each waste heat resource based on the operating data of the waste heat resources. The storage unit is used to store the operational and working data acquired by the monitoring unit.

[0006] In some embodiments of this application, the first control module is further configured to: Receive grid dispatch instructions and generate deep peak-shaving data for the forecast period; The deep peak tuning is divided into multiple time periods; Extract real-time deep peak-shaving data from the deep peak-shaving data. Determine the real-time deep peak shaving period and generate a first demand feature based on the deep peak shaving stage; Extract real-time deep peak-shaving data and define the required characteristics.

[0007] In some embodiments of this application, dividing the deep peak modulation into multiple time periods includes: Extract load decrease data for the forecast period; Set a first descent threshold and a second descent threshold; Obtain the volatility of the load decrease data, and filter the load decrease data based on the volatility to obtain the second load decrease data; Load decrease data exceeding the first decrease threshold is set as the first stage, load decrease data below the second decrease threshold is set as the third stage, and load decrease data exceeding the second decrease threshold but below the first decrease threshold is set as the second stage.

[0008] In some embodiments of this application, the step of filtering the load drop data to obtain the second load drop data includes: Calculate the volatility of the load decrease data and set a volatility threshold; The time interval corresponding to the load decrease data with volatility less than the volatility threshold is set as the flat value interval; Replace the load decrease data in the flat range with the average load decrease data in the flat range, and set it as the second load decrease data.

[0009] In some embodiments of this application, calculating the volatility of the load decline data includes: Obtain load decrease data for the forecast period; Generate a load decrease curve based on the load decrease data; When the load decrease data corresponding to different time points on the load decrease curve are the same, the time period corresponding to the different time points is set as the first fluctuation range; When the load decrease data in the first fluctuation range includes either the first decrease threshold or the second decrease threshold, the first fluctuation range is set as the second fluctuation range. Extract the range of load decrease data corresponding to each second fluctuation interval, and generate the volatility based on the range.

[0010] In some embodiments of this application, generating the first requirement feature includes: Obtain real-time deep peak shaving periods and operating data of thermal power units; Based on the domain knowledge graph, a type of parameter is selected from the working data corresponding to the deep peak-shaving period; Obtain the rate of change of each type of parameter during the real-time deep peak-shaving period; Among the parameters, the one with the largest rate of change is set as the first requirement feature.

[0011] In some embodiments of this application, the extraction of real-time deep peak-shaving data and the setting of required characteristics include: Obtain the operating data of the thermal power unit; Based on the work data, determine whether the requirement corresponding to the first requirement feature is valid; When the condition is met, the first demand feature is set as the demand feature; when the condition is not met, the parameter with the largest rate of change among the other parameters is selected as the demand feature. Based on the aforementioned demand characteristics, a matching waste heat resource level is generated.

[0012] In some embodiments of this application, the second control module is further configured to: Obtain operational and historical operational data of waste heat resources; Extract the response time and adjustment accuracy of each waste heat resource; Based on the response time and adjustment accuracy, waste heat resources are divided into multiple levels; Labels are assigned to waste heat resources of various levels, and the labels contain the level data and response time data of the corresponding waste heat resources.

[0013] In some embodiments of this application, the fourth control module is further configured to: Obtain operational data for the waste heat resources at the specified level; Based on operational data, the waste heat resources at each of the aforementioned levels are sorted. Obtain the dispatchable range of each waste heat resource; According to the aforementioned order, each waste heat resource is called up until the power grid dispatch instructions are met.

[0014] In some embodiments of this application, the sorting of waste heat resources at each level based on operational data includes: Obtain the response time data and adjustment accuracy data of the waste heat resource; Set time-based and precision-based sorting metrics; Among waste heat resources with the same time ranking index, they are ranked according to the aforementioned precision ranking index; Generate a sorted list of all schedulable waste heat resources.

[0015] Compared with existing technologies, the deep peak-shaving system for thermal power units based on waste heat recovery in this application embodiment has the following advantages: This system enables intelligent segmentation and precise demand identification during deep peak shaving. Through the parsing of power grid dispatch instructions and analysis of real-time operational data by the first control module, the peak shaving process can be dynamically divided into multiple distinct time periods. Based on domain knowledge graphs and real-time data changes, the system accurately extracts the most critical and urgent demand characteristics for each time period. This dynamic demand identification based on process characteristics changes the past control model, making control strategies more targeted and forward-looking.

[0016] Furthermore, a dynamic assessment and capacity grading system for waste heat resources was established. The second control module scientifically classifies waste heat resources into multiple levels and assigns them dynamic labels based on dynamic operational data such as temperature, response time, and adjustment accuracy, rather than solely on fixed design parameters. This enables the system to clearly understand the current availability and adjustment potential of various resources, laying a data foundation for precise scheduling. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a deep peak-shaving system for thermal power units based on waste heat recovery, as described in an embodiment of this application. Detailed Implementation

[0018] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but are not intended to limit the scope of this application.

[0019] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0020] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0021] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0022] like Figure 1 As shown in the figure, an embodiment of this application provides a deep peak-shaving system for thermal power units based on waste heat recovery, comprising: The monitoring unit is used to monitor the real-time operation data of the waste heat resources of the unit and the working data of the thermal power unit. The central control unit is used to adjust the deep peak shaving according to the demand characteristics of different waste heat resources; The central control unit includes: The first control module is used to generate deep peak-shaving data and extract required features based on the deep peak-shaving data. The second control module is used to classify each waste heat resource into different levels based on the operating data of the waste heat resources. The third control module is used to match corresponding waste heat resources according to demand characteristics; The fourth control module is used to set the utilization status of each waste heat resource based on the operation data of the waste heat resources; The storage unit is used to store the operational and working data acquired by the monitoring unit.

[0023] Specifically, the monitoring unit collects two types of core data streams. The first type is "operational data on waste heat resources," which refers to the energy flow information released into the environment from various stages of the thermal cycle of the thermal power unit that is not utilized. Specific monitoring targets include: flue gas waste heat (such as the temperature and flow rate of flue gas at the boiler economizer outlet and air preheater outlet), turbine exhaust waste heat (such as the enthalpy of low-pressure cylinder exhaust steam, and the inlet and outlet water temperatures and flow rates of the associated circulating cooling water system or air-cooled island), and low-grade extraction steam waste heat (such as the pressure, temperature, and available margin of the turbine's low-pressure connecting pipe or the fifth and sixth stage extraction steam). The second type is "operational data of the thermal power unit," which refers to key parameters reflecting the operating status, safety boundaries, and environmental indicators of the main unit. Specific monitoring targets include: generator active power (i.e., the unit's real-time load), pressure and temperature of main steam and reheat steam, boiler feedwater flow rate and temperature, flue gas temperature at the inlet of the denitrification system, steam flow rate of the turbine's low-pressure cylinder, steam drum water level, and furnace negative pressure. This unit connects to the power plant's existing sensor network and data bus via a standard industrial communication protocol. After filtering, range conversion, and preliminary verification of the raw signals, it generates real-time data packets in a unified format and sends them to the central control unit.

[0024] Specifically, the central control unit is the decision-making and control hub of the system, consisting of four logically connected and operationally coordinated control modules. It receives data streams from the monitoring unit and external scheduling commands, performs a series of calculations, analyses, and decisions, and ultimately outputs control commands for the waste heat recovery device and related systems.

[0025] Specifically, the storage unit is used to persistently save the raw data stream collected by the monitoring unit, the intermediate data generated by each module of the central control unit during processing, the final generated control command sequence, and the system's own operation log.

[0026] Specifically, demand characteristics refer to the specific objectives that urgently require intervention or compensation through external heat at certain moments during the deep peak-shaving process to ensure the safe, environmental, and stable operation of the generating units. These characteristics are not fixed but dynamically evolve based on the stage of the peak-shaving process and the real-time status of the generating units. The demand characteristics are extracted by the first control module. For example, upon receiving a grid dispatch instruction that the load will decrease from 90% to 30% within 30 minutes, the first control module will combine the generating unit model to make predictions and divide this process into three stages. In each stage, it will identify the most prominent contradiction based on real-time monitored thermal power unit operating data.

[0027] Specifically, classifying waste heat resources refers to the process by which the second control module categorizes and classifies waste heat sources based on their dynamic performance parameters, rather than their physical location or fixed design values. The purpose of this classification is to establish an evaluation system for internal heat sources to accurately match dynamically changing demands. The classification is primarily based on two core dynamic parameters: response time and adjustment accuracy.

[0028] Specifically, based on the state and matching decisions, a set of specific, executable operation instructions is set for each called waste heat resource, along with monitoring and feedback on the execution effect of these instructions. This final execution stage is handled by the fourth control module, which transforms the abstract matching scheme into actions that the field equipment can understand and execute.

[0029] In some embodiments of this application, the first control module is further configured to: Receive grid dispatch instructions and generate deep peak-shaving data for the forecast period; The deep peak-shaving process is divided into multiple time periods; Extracting real-time deep peak-shaving data from deep peak-shaving data Determine the real-time deep peak shaving period and generate the first demand feature based on the deep peak shaving stage; Extract real-time deep peak-shaving data and define the required characteristics.

[0030] Specifically, the first control module is responsible for analyzing external demand and identifying the main internal contradictions. Its workflow is as follows: First, it receives grid dispatch instructions, which are typically structured data messages containing elements such as target load value and expected arrival time. After parsing the instructions, the module generates deep peak-shaving data for the forecast period. That is, based on the current unit status and instruction requirements, it simulates and generates a reference load change curve from the current load to the target load and the corresponding time axis. Next, the deep peak-shaving is divided into multiple time periods. This division is not a simple equal division of time, but a dynamic process based on the load change rate and the safety margin of key parameters. The implementation steps are as follows: Extract the load reduction data for the forecast period, that is, the first derivative of the load with respect to time in the above reference curve. Set a first reduction threshold and a second reduction threshold. These two thresholds are preset according to the unit characteristics. For example, the first reduction threshold is set to 2% rated load / minute, and the second reduction threshold is set to 0.5% rated load / minute. Obtain the volatility of the load reduction data, that is, calculate the standard deviation or range of the load reduction rate over a period of time to characterize the smoothness of the load reduction process. The second load decline data is obtained by filtering the load decline data based on volatility. The purpose is to eliminate signal noise and short-term fluctuations that could interfere with the stage determination. Specifically, this involves: calculating the volatility and setting a volatility threshold; marking the time interval corresponding to relatively stable load decline data segments with volatility below this threshold as the "average value interval"; replacing the original data with the average load decline rate within this average value interval, thus obtaining a smoother "second load decline data" curve that better reflects the true intention. Finally, the second load decline data is used to divide the system into stages: the period exceeding the first decline threshold is divided into the first stage, the rapid decline stage, characterized by a rapid load decline and the thermal system facing a significant inertial impact; the period below the second decline threshold is divided into the third stage, the low-load stable operation stage, characterized by the load reaching or approaching the target low value and requiring long-term stable operation; the period between the two thresholds is divided into the second stage, the smooth transition stage, characterized by a steady load decline and the need for fine-tuning the system to cope with the approaching safety and environmental boundaries.

[0031] In some embodiments of this application, deep peak modulation is divided into multiple time periods, including: Extract load decrease data for the forecast period; Set a first descent threshold and a second descent threshold; Obtain the volatility of the load decline data, and filter the load decline data based on the volatility to obtain the second load decline data; Load decrease data exceeding the first decrease threshold is set as the first stage, load decrease data below the second decrease threshold is set as the third stage, and load decrease data exceeding the second decrease threshold but below the first decrease threshold is set as the second stage.

[0032] Specifically, extracting load decline data for the forecast period means separating the data sequence that represents the decrease in active power of the unit over time from the deep peak-shaving data for the forecast period. This is the load decline data, which is usually represented as a series of time point-load value data pairs.

[0033] Specifically, the time point is collected every 10 seconds.

[0034] Specifically, setting the first and second load reduction thresholds means that, based on the unit's operating characteristics and safety constraints, two critical values ​​for the load reduction rate are preset. The first threshold is a reduction of 3% of the rated load per minute, used to identify the boundary of "rapid change"; the second threshold is set at 1% of the rated load per minute, used to identify the boundary of "smooth change". These two thresholds divide the dynamic range of load reduction into three intervals.

[0035] Specifically, obtaining the volatility of load decline data and filtering it based on that volatility to obtain the second load decline data is a data smoothing preprocessing step. Since actual load commands or unit responses may exhibit slight fluctuations, directly using the raw data to divide time periods could lead to misjudgments. Therefore, it is necessary to calculate the volatility of the original load decline data sequence—that is, a measure of instability between adjacent data points—and based on this, filter the data to remove meaningless minor fluctuations, resulting in a smoother curve that better reflects the overall trend, i.e., the second load decline data.

[0036] Specifically, the final operation of time period division is as follows: the time periods corresponding to data points with values ​​greater than the first decrease threshold in the second load decrease data are classified as the first stage requiring rapid response; the time periods corresponding to data points with values ​​less than the second decrease threshold are classified as the third stage requiring stable maintenance; and the time periods corresponding to data points with values ​​between the two thresholds are classified as the second stage of transitional nature.

[0037] In some embodiments of this application, filtering the load drop data to obtain second load drop data includes: Calculate the volatility of load decline data and set a volatility threshold; Set the time interval corresponding to the load decrease data with volatility less than the volatility threshold as the flat value interval; Replace the load decrease data in the flat range with the average load decrease data in the flat range, and set it as the second load decrease data.

[0038] Specifically, calculating the volatility of load decline data and setting a volatility threshold involves using the standard deviation of the data within a sliding window to quantify the severity of fluctuations in the original load decline data within each local time window, thus obtaining a volatility sequence. Simultaneously, based on historical operating experience or statistical analysis, an acceptable upper limit for volatility is set as the volatility threshold to determine which fluctuations are significant.

[0039] Specifically, setting the time interval corresponding to load decline data with volatility less than the volatility threshold as the flat value interval means that during the load decline process, those continuous time periods with volatility below the threshold, i.e., relatively stable load changes, are identified and marked as the flat value interval.

[0040] Specifically, replacing the load decrease data in the flat period with the average load decrease data of the flat period, designated as the second load decrease data, means: for each marked flat period, calculate the arithmetic mean of all the original load decrease data within that period. Then, replace the original load decrease data value corresponding to each time point within that period with this average value. After performing this operation on all flat periods, the data in the non-flat periods remain unchanged, and the final new data sequence is the smoothed second load decrease data.

[0041] In some embodiments of this application, the calculation of the volatility of load decline data includes: Obtain load decrease data for the forecast period; Generate a load decline curve based on load decline data; When the load decrease data corresponding to different time points on the load decrease curve are the same, the time period corresponding to the different time points is set as the first fluctuation range. When the load decrease data in the first fluctuation range includes either the first decrease threshold or the second decrease threshold, the first fluctuation range is set as the second fluctuation range. Extract the range of load decrease data corresponding to each second fluctuation interval, and generate volatility based on the range.

[0042] Specifically, obtaining load decline data for the forecast period and generating a load decline curve based on the load decline data means: first, obtaining raw time-load data pairs, and then connecting these points in a coordinate system to form a load decline curve that reflects the load decline trend over time.

[0043] Specifically, when the load decline data at different time points on the load decline curve is the same, defining the time period corresponding to each different time point as the first fluctuation range means: scanning along the load decline curve to identify continuous time periods where the load value on the curve remains constant. These load plateau periods, regardless of their length, are initially defined as the first fluctuation range.

[0044] Specifically, when the load decline data in the first fluctuation range includes either a first decline threshold or a second decline threshold, setting the first fluctuation range as the second fluctuation range means that not all load plateau periods are worth monitoring. The system will check whether the constant load value in each first fluctuation range is exactly equal to the pre-set first decline threshold or second decline threshold.

[0045] Specifically, extracting the range of load decline data corresponding to each second fluctuation interval and generating volatility based on the range means that for each second fluctuation interval, the range refers to the constant value of the load value in that interval. The system can analyze the distribution of load values ​​in multiple such intervals, for example, calculating the variance or coefficient of variation of these constant values, or directly using the degree of deviation of these constant values ​​from the overall trend line to comprehensively generate an indicator characterizing the degree of data clustering or dispersion near key points. This indicator can then serve as a component or correction basis for volatility used in data filtering.

[0046] In some embodiments of this application, generating the first requirement feature includes: Obtain real-time deep peak shaving periods and operating data of thermal power units; Based on the domain knowledge graph, select one type of parameter from the working data corresponding to the deep peak shaving period; Obtain the rate of change of each type of parameter during the real-time deep peak-shaving period; Among a class of parameters, the parameter with the largest rate of change is set as the first requirement feature.

[0047] Specifically, obtaining real-time deep peak shaving periods and thermal power unit operating data means: first, determining that the current unit is in a specific time period, and then simultaneously obtaining the latest thermal power unit operating data uploaded by the monitoring unit within that time period.

[0048] Specifically, based on the domain knowledge graph, selecting one type of parameter from the working data corresponding to the deep peak-shaving period refers to: the system has a built-in domain knowledge graph, which is a structured database storing professional knowledge in the thermal power field. It defines which categories of operating parameters typically require the most attention during different peak-shaving periods (such as rapid load reduction, transition, and low-pressure steady-state). For example, during a rapid load reduction period, the knowledge graph might associate parameters such as "main steam temperature" and "main steam pressure" as key monitoring parameters; during a low-load steady-state period, it might associate parameters such as "SCR inlet flue gas temperature" and "low-pressure cylinder steam flow." The system queries and retrieves the corresponding parameter category list from the knowledge graph based on the current time period.

[0049] Specifically, obtaining the rate of change of each type of parameter during the real-time deep peak-shaving period means that for each type of parameter selected from the knowledge graph, the system calculates its rate of change in the current real-time period (e.g., how many degrees Celsius / minute the temperature has dropped, or how many megapascals / minute the pressure has dropped in the past minute).

[0050] Specifically, setting the parameter with the largest rate of change among a class of parameters as the primary demand feature means: comparing the absolute values ​​of the rates of change of all parameters of interest, identifying the problem represented by the parameter that changes most drastically, deviates from the normal or set value the fastest, as the most pressing contradiction at present, and setting it as the primary demand feature.

[0051] In some embodiments of this application, real-time deep peak-shaving data is extracted, and demand characteristics are set, including: Obtain operating data from thermal power units; Based on the work data, determine whether the requirement corresponding to the first requirement feature is valid; When the condition is met, the first demand characteristic is set as the demand characteristic; when the condition is not met, among the other types of parameters, the parameter with the largest rate of change is selected as the demand characteristic. Based on the demand characteristics, a matching waste heat resource level is generated.

[0052] Specifically, acquiring the operating data of the thermal power unit and determining whether the requirement corresponding to the first demand characteristic is valid based on the operating data means that the system not only looks at the rate of change, but also makes logical judgments based on the absolute value of the parameters. For example, if the first demand characteristic is "to prevent the SCR inlet flue gas temperature from being too low," the system will check whether the current actual SCR inlet flue gas temperature is lower than or close to the minimum allowable operating temperature of the catalyst. If the actual value is very close to the lower limit, the requirement is still valid even if its rate of change is not the maximum.

[0053] Specifically, when the condition is met, the first requirement feature is set as the requirement feature. When the condition is not met, among the other parameters, the parameter with the highest rate of change is selected as the requirement feature. This is a priority verification mechanism. If real-time data verification shows that the problem pointed to by the first requirement feature is more important, it is ultimately confirmed as the requirement feature to be solved. If the verification fails, it reverts to the alternative solutions. From the same parameter list defined by the knowledge graph, the problem corresponding to the parameter with the second highest rate of change is selected as a new candidate requirement feature, and similar verification may be performed again until a pressing requirement that has been verified by data is found.

[0054] Specifically, generating a matching waste heat resource level based on demand characteristics means that after determining the specific demand characteristics, the system needs to translate them into capability requirements for waste heat resources. For example, if the demand requires fast response speed and accurate heat replenishment, the system will generate a resource level request, such as "requiring the use of fast-response waste heat resources with a certain temperature range".

[0055] In some embodiments of this application, the second control module is further configured to: Obtain operational and historical operational data of waste heat resources; Extract the response time and adjustment accuracy of each waste heat resource; Based on response time and adjustment accuracy, waste heat resources are divided into multiple levels; Labels are assigned to waste heat resources of each level. The labels contain the level data and response time data of the corresponding waste heat resources.

[0056] Specifically, acquiring operational and historical operational data of waste heat resources means that the second control module not only receives waste heat data reported by the monitoring unit in real time, but also retrieves recent historical data from the storage unit to analyze the performance trends and reliability of waste heat resources.

[0057] Specifically, the response time and adjustment accuracy of extracting each waste heat resource are two key dynamic performance indicators. Response time refers to the time elapsed from when the system issues a command to utilize the waste heat resource until the actual heat output of the resource reaches 90% of the commanded value; it reflects the resource's agility. Adjustment accuracy refers to the deviation range (e.g., ±15%) between the stable value of the waste heat output and the commanded set value during actual operation; it reflects the resource's controllability.

[0058] Specifically, classifying waste heat resources into multiple levels based on response time and adjustment accuracy means that the system sets classification rules. For example, waste heat with a response time in the second range and high adjustment accuracy is classified as fast response type; waste heat with a response time in the minute range and medium adjustment accuracy is classified as medium-speed buffer type; and waste heat with a slow response but large capacity is classified as slow-speed reserve type.

[0059] Specifically, assigning tags to waste heat resources at different levels, with each tag containing the corresponding level data and response time data, means that the system creates a dynamic tag for each specific waste heat recovery device or path. This tag not only records its level but also updates its current response time data and other key attributes in real time, such as the current available heat capacity and geographical location.

[0060] In some embodiments of this application, the fourth control module is further configured to: Obtain operational data on waste heat resources at different levels; Based on operational data, waste heat resources at each level are sorted. Obtain the dispatchable range of each waste heat resource; According to the order, each waste heat resource is called up until the power grid dispatch instructions are met.

[0061] Specifically, acquiring operational data on waste heat resources by grade means that the fourth control module obtains graded and tagged waste heat resource information from the second control module, especially its real-time operating status, such as whether a certain heat pump is currently available and how much output it has.

[0062] Specifically, sorting waste heat resources at different levels based on operational data means that within the same level, resources are ranked according to their real-time performance indicators (response time, available capacity, and adjustment accuracy) to determine the priority order for their use.

[0063] Specifically, obtaining the dispatchable range of each waste heat resource means determining the minimum and maximum values ​​of its output heat that can be safely and continuously adjusted under the current operating conditions; this is the dispatchable range. This is crucial to ensuring the feasibility of control commands.

[0064] Specifically, the process of prioritizing and calling upon various waste heat resources until the grid dispatch instructions are met means that the fourth control module receives the matching results from the third control module. It then sends control instructions to the resource ranked first within that priority level, causing it to output heat. If a single resource cannot meet all the demand, the next highest-ranked resource is called, and this cycle continues until the total output of all called resources meets the requirements of the current deep peak-shaving phase.

[0065] In some embodiments of this application, waste heat resources at different levels are sorted based on operational data, including: Acquire response time data and regulation accuracy data of waste heat resources; Set time-based and precision-based sorting metrics; Among waste heat resources with the same time ranking index, they are ranked according to the accuracy ranking index; Generate a sorted list of all schedulable waste heat resources.

[0066] Specifically, among waste heat resources with the same time ranking index, sorting by precision ranking index means that when the response times of two or more waste heat resources are very close (or at the same level), it is impossible to distinguish their quality simply by time. In this case, the system introduces adjustment precision as a secondary sorting key, ranking resources with higher adjustment precision first among those with the same response time.

[0067] Specifically, generating a sorted list of all schedulable waste heat resources means comprehensively applying the aforementioned primary and secondary sorting rules to calculate and sort all currently available waste heat resources, ultimately generating an ordered list. This list clearly indicates which specific device or path should be prioritized when a resource of a certain level needs to be accessed, providing a clear action guide for the execution of the fourth control module.

[0068] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of this application, and these improvements and substitutions should also be considered within the scope of protection of this application.

Claims

1. A deep peak shaving system for a thermal power generating unit based on waste heat recovery, characterized in that, include: The monitoring unit is used to monitor the real-time operation data of the waste heat resources of the unit and the working data of the thermal power unit. The central control unit is used to adjust the deep peak shaving according to the demand characteristics of different waste heat resources; The central control unit includes: The first control module is used to generate deep peak-shaving data and extract required features based on the deep peak-shaving data. The second control module is used to classify each waste heat resource into different levels based on the operating data of the waste heat resources. The third control module is used to match corresponding waste heat resources according to the aforementioned demand characteristics; The fourth control module is used to set the utilization status of each waste heat resource based on the operating data of the waste heat resources. The storage unit is used to store the operational and working data acquired by the monitoring unit.

2. The system according to claim 1, wherein, The first control module is also used for: Receive grid dispatch instructions and generate deep peak-shaving data for the forecast period; The deep peak modulation is divided into multiple time periods; Extract real-time deep peak-shaving data from the deep peak-shaving data. Determine the real-time deep peak shaving period and generate a first demand feature based on the deep peak shaving stage; Extract real-time deep peak-shaving data and define the required characteristics.

3. The system for deep load following of a thermal power generating unit based on waste heat recovery according to claim 2, characterized in that, The process of dividing the deep peak tuning into multiple time periods includes: Extract load decrease data for the forecast period; Set a first descent threshold and a second descent threshold; Obtain the volatility of the load decrease data, and filter the load decrease data based on the volatility to obtain the second load decrease data; Load decrease data exceeding the first decrease threshold is set as the first stage, load decrease data below the second decrease threshold is set as the third stage, and load decrease data exceeding the second decrease threshold but below the first decrease threshold is set as the second stage.

4. The system according to claim 3, wherein, The process of filtering the load decrease data to obtain the second load decrease data includes: Calculate the volatility of the load decrease data and set a volatility threshold; The time interval corresponding to the load decrease data with volatility less than the volatility threshold is set as the flat value interval; Replace the load decrease data in the flat range with the average load decrease data in the flat range, and set it as the second load decrease data.

5. A system for deep load following of a thermal power plant based on waste heat recovery according to claim 4, characterized in that, The calculation of the volatility of the load decline data includes: Obtain load decrease data for the forecast period; Generate a load decrease curve based on the load decrease data; When the load decrease data corresponding to different time points on the load decrease curve are the same, the time period corresponding to the different time points is set as the first fluctuation range; When the load decrease data in the first fluctuation range includes either the first decrease threshold or the second decrease threshold, the first fluctuation range is set as the second fluctuation range. Extract the range of load decrease data corresponding to each second fluctuation interval, and generate the volatility based on the range.

6. A system for deep load following of a thermal power plant based on waste heat recovery according to claim 5, characterized in that, The generation of the first requirement feature includes: Obtain real-time deep peak shaving periods and operating data of thermal power units; Based on the domain knowledge graph, a type of parameter is selected from the working data corresponding to the deep peak-shaving period; Obtain the rate of change of each type of parameter during the real-time deep peak-shaving period; Among the parameters, the one with the largest rate of change is set as the first requirement feature.

7. The system for deep load following of a thermal power generating unit based on waste heat recovery according to claim 6, characterized in that, The extraction of real-time deep peak-shaving data and the setting of required characteristics include: Obtain the operating data of the thermal power unit; Based on the work data, determine whether the requirement corresponding to the first requirement feature is valid; When the condition is met, the first demand feature is set as the demand feature; when the condition is not met, the parameter with the largest rate of change among the other parameters is selected as the demand feature. Based on the aforementioned demand characteristics, a matching waste heat resource level is generated.

8. The system for deep load following of a thermal power generating unit based on waste heat recovery according to claim 1, characterized in that, The second control module is also used for: Obtain operational and historical operational data of waste heat resources; Extract the response time and adjustment accuracy of each waste heat resource; Based on the response time and adjustment accuracy, waste heat resources are divided into multiple levels; Labels are assigned to waste heat resources of various levels, and the labels contain the level data and response time data of the corresponding waste heat resources.

9. The system for deep load following of a thermal power generating unit based on waste heat recovery according to claim 1, characterized in that, The fourth control module is also used for: Obtain operational data for the waste heat resources at the specified level; Based on operational data, the waste heat resources at each of the aforementioned levels are sorted. Obtain the dispatchable range of each waste heat resource; According to the aforementioned order, each waste heat resource is called up until the power grid dispatch instructions are met.

10. The system for deep load following of a thermal power generating unit based on waste heat recovery according to claim 9, characterized in that, The sorting of waste heat resources at each level based on operational data includes: Obtain the response time data and adjustment accuracy data of the waste heat resource; Set time-based and precision-based sorting metrics; Among waste heat resources with the same time sorting index, they are sorted according to the aforementioned precision sorting index to generate a sorted list of all schedulable waste heat resources.

Images