A control method and system for zero discharge of desulfurization wastewater in a thermal power plant

Abstract

The present application belongs to the field of zero emission treatment and intelligent control technology of desulfurization wastewater in thermal power plant, and discloses a control method and system for zero emission of desulfurization wastewater in thermal power plant, comprising: a data fusion unit collects desulfurization wastewater operation data, and obtains a credible state parameter after fusing the low-credibility data replaced by historical alternative data through credibility discrimination; a trend risk unit obtains evolution parameters, scaling and zero emission load index accordingly; a double-dimensional decision unit outputs operation parameters according to the double index; a time lag control unit divides fast and slow control sequences according to response time; a double path unit selects one of the control or reconstruction paths to generate deviation to feed back to the two units in the front end. In the present application, the fusion state parameter is obtained by relying on data credibility processing, the time sequence control strategy is constructed by combining multiple indexes, the parameter adjustment or process reconstruction path is selected, and the closed loop is formed by relying on the double channel feedback of operation deviation to complete the collaborative control of zero emission of desulfurization wastewater.

Description

Technical Field

[0001] This invention belongs to the field of zero-discharge treatment and intelligent control technology of desulfurization wastewater from thermal power plants, specifically relating to a control method and system for zero discharge of desulfurization wastewater from thermal power plants. Background Technology

[0002] Coal-fired power plants typically employ wet flue gas desulfurization (FGD) processes to control sulfur dioxide emissions during coal-fired power generation. This process generates a certain amount of desulfurization wastewater. This wastewater is usually characterized by high salinity, high hardness, and complex composition. Its treatment system generally operates in conjunction with processes such as evaporation and concentration, softening pretreatment, and crystallization desalination to achieve water resource recovery and pollutant reduction. In engineering applications, zero-discharge desulfurization wastewater systems typically maintain stable operation by monitoring water quality parameters online and adjusting reagent dosage, evaporation temperature, and circulation flow rate in conjunction with evaporation system operating parameters.

[0003] Most existing zero-discharge control systems for desulfurization wastewater adopt single real-time parameter feedback control. Because the operating status data is easily affected by high-salt, high-hardness and high-suspended-solids environments, the status information is distorted and lacks the ability to judge trends and loads in coordination, resulting in an imbalance between scaling risk and treatment load. Summary of the Invention

[0004] To overcome the above shortcomings, the present invention aims to provide a control method and system for zero discharge of desulfurization wastewater from thermal power plants, which improves the problem that most existing zero discharge control systems for desulfurization wastewater adopt single real-time parameter feedback control, which is prone to causing an imbalance between scaling risk and treatment load.

[0005] To achieve the above objectives, the present invention provides the following technical solution: On the one hand, the present invention provides an intelligent control method for zero discharge of desulfurization wastewater from thermal power plants, comprising the following steps: The system acquires desulfurization wastewater operation status data, calculates the credibility of the operation status data, and generates alternative status data based on historical operation data when the credibility of the operation status data is lower than the preset conditions. The system then merges the operation status data or the alternative status data to generate a credible fused status parameter. State evolution parameters are generated within multiple consecutive time windows, and a scaling risk index is generated based on the reliable fusion of state parameters and state evolution parameters, while a zero-emission load index is also generated. A two-dimensional state space is constructed based on the scaling risk index and the zero-emission load index, and two-dimensional operating state parameters are generated based on the two-dimensional state space. Acquire two-dimensional operating state parameters and response time data of each controlled object, and generate a timing control strategy containing fast control sequences and slow control sequences based on the response time data; The parameter adjustment path or process reconfiguration path is executed according to the time-series control strategy. The parameter adjustment path is used to adjust the operating status of the desulfurization wastewater treatment unit, and the process reconfiguration path is used to switch the zero-emission treatment process. Based on the actual operating status and the target operating status, the operating deviation parameters are generated and then reused in the generation process of the reliable fusion state parameters, scaling risk index and zero-emission load index.

[0006] Preferably, the process of acquiring the desulfurization wastewater operating status data includes: Acquire influent flow rate data, conductivity data, pH data, temperature data, and turbidity data; Obtain calcium ion concentration data, magnesium ion concentration data, chloride ion concentration data, and sulfate ion concentration data; Timestamp the influent flow rate data, conductivity data, pH data, temperature data, turbidity data, calcium ion concentration data, magnesium ion concentration data, chloride ion concentration data, and sulfate ion concentration data. Based on a unified sampling period, time-aligned data with timestamps are generated to produce desulfurization wastewater operation status data.

[0007] Preferably, the reliability calculation of the operational status data includes the following steps: Calculate the variation range of ion concentration data within a continuous time window in the operational status data; The degree of data loss in the time series of ion concentration data is calculated to generate continuous calculation results; Consistency calculations were performed to verify the correspondence between ion concentration data and conductivity data; The deviation between the ion concentration data and the corresponding data distribution in historical operation data is calculated. The reliability of the operational status data is generated based on the calculation results of the change range, consistency, and deviation.

[0008] Preferably, generating alternative status data based on historical operational data includes the following steps: Extract conductivity data, pH data, turbidity data, and influent flow rate data from historical operating data; Construct a mapping relationship between operation status variables based on historical operation data; The operating condition range is determined based on conductivity data, pH data, turbidity data, and influent flow rate data. Alternative ion concentration data are generated based on the mapping relationship corresponding to the operating condition range.

[0009] Preferably, generating state evolution parameters within multiple consecutive time windows includes the following steps: The trusted fusion state parameters are arranged in chronological order to generate a trusted fusion state parameter time series. The trusted fusion state parameter time series is divided into multiple adjacent time windows by a sliding window according to a preset time span. Differential calculations are performed on the trusted fusion state parameters within each time window to generate a sequence of state changes. The state change sequence is normalized across a time window to generate state evolution parameters.

[0010] Preferably, the step of generating the scaling risk index based on trusted fusion state parameters and state evolution parameters includes the following steps: Obtain online concentration ratio and evaporation temperature; A scaling risk index is generated by combining the reliable fusion state parameters, the state evolution parameters, the online concentration ratio, and the evaporation temperature.

[0011] Preferably, the construction of the two-dimensional state space based on the scaling risk index and the zero-emission load index includes the following steps: The scaling risk index is normalized to generate standardized scaling risk coordinate values; The zero-emission load index is normalized to generate standardized zero-emission load coordinate values; A two-dimensional coordinate mapping relationship is constructed using standardized scaling risk coordinates as the first dimension and standardized zero-emission load coordinates as the second dimension. The state region is divided based on the two-dimensional coordinate mapping relationship to generate a two-dimensional state space.

[0012] Preferably, the step of generating a timing control strategy containing fast control sequences and slow control sequences based on response time data includes the following steps: collecting response time data and constructing a response time sequence; sorting the response time sequence according to the response time value; dividing the controlled object into a fast control sequence and a slow control sequence based on the response time sorting result; and generating a timing control strategy based on the execution order of the fast control sequence and the slow control sequence. The process reconfiguration path for switching the zero-emission treatment process includes the following steps: switching the operating mode between a normal operating mode, an enhanced hardening removal mode, a high-load operating mode, and a risk protection mode; wherein, the normal operating mode is softening followed by evaporation and then crystallization; the enhanced hardening removal mode is increasing the intensity of the softening reaction followed by evaporation and then crystallization; the high-load operating mode is pre-concentration followed by evaporation and then crystallization; and the risk protection mode is reducing the online concentration ratio, reducing the evaporation load, and enhancing hardening removal.

[0013] On the other hand, the present invention also provides an intelligent control system for zero discharge of desulfurization wastewater from thermal power plants. This system is used to implement the intelligent control method for zero discharge of desulfurization wastewater from thermal power plants as described above, comprising: The data fusion processing unit is used to acquire desulfurization wastewater operation status data, calculate the credibility of the operation status data, generate alternative status data based on historical operation data when the credibility of the operation status data is lower than the preset conditions, and fuse the operation status data or alternative status data to generate a credible fusion status parameter. The trend risk calculation unit is used to receive trusted fusion state parameters, generate state evolution parameters within multiple consecutive time windows, and generate a scaling risk index based on the trusted fusion state parameters and state evolution parameters, while also generating a zero-emission load index. A two-dimensional state decision unit is used to construct a two-dimensional state space based on the scaling risk index and the zero-emission load index, and to generate two-dimensional operating state parameters based on the two-dimensional state space. The time-delay cooperative control unit is used to acquire two-dimensional operating state parameters and response time data of each controlled object, and generate a timing control strategy containing fast control sequence and slow control sequence based on the response time data; The dual-path closed-loop control unit is used to execute parameter adjustment path or process reconfiguration path according to the timing control strategy. The parameter adjustment path is used to adjust the operating state of the processing unit, and the process reconfiguration path is used to switch the zero-emission treatment process. It generates operating deviation parameters based on the actual operating state and the target operating state, and feeds the operating deviation parameters back to the data fusion processing unit and the trend risk calculation unit.

[0014] Preferably, in the dual-path closed-loop control unit, feeding back the operating deviation parameters to the data fusion processing unit and the trend risk calculation unit includes the following steps: The operating deviation parameters are decomposed according to the type of operating state variable to generate a set of state deviation components; Match and map the set of state deviation components with the corresponding variables in the running state data; The matched state deviation components are written into the state input terminal of the data fusion processing unit. Write the running deviation parameters into the time window update terminal of the trend risk calculation unit.

[0015] The present invention has the following beneficial effects: This invention provides an intelligent control method for zero discharge of desulfurization wastewater from thermal power plants. First, it collects operational status data of the desulfurization wastewater and verifies its reliability. Due to the high salt, high hardness, and high suspended solids characteristics of desulfurization wastewater, online monitoring data is prone to distortion and loss. Directly using raw data can lead to biased control decisions. Therefore, this invention uses historical operational data to generate alternative data for data with insufficient reliability, and then integrates this data to obtain reliable fusion state parameters. This ensures the accuracy of all subsequent analysis and control actions from the data source. Next, it calculates state evolution parameters using multiple continuous time windows, and combines these with the reliable fusion state parameters to calculate the scaling risk index and the zero-discharge load index. This overcomes the limitations of existing technologies that rely solely on instantaneous parameters to determine operating conditions, enabling the tracking of dynamic changes in the system's operating status. It can predict scaling risks in a timely manner and accurately reflect the actual processing load of the system. Finally, it constructs a two-dimensional state space using these two indices and generates two-dimensional operational status parameters, combining the two key indicators of scaling risk and processing load to determine the overall operating condition, making the control basis more comprehensive. By dividing the response time of each controlled object into fast and slow control sequences and formulating a timing control strategy, the system can match the response speed of different control links and avoid operational fluctuations caused by unreasonable timing of control actions. Based on the timing control strategy, the system selects parameter adjustment paths or process reconfiguration paths as needed to complete the adjustment of the desulfurization wastewater treatment unit's operating conditions and the switching of the zero-discharge treatment process. Simultaneously, it compares the actual operating state with the target state to obtain operating deviation parameters, which are then fed back to the front end to participate in the generation of reliable fusion state parameters and two types of indices, forming a complete closed-loop control. This entire solution changes the existing single real-time parameter feedback control mode, progressively realizing data correction, trend analysis, time-sequenced regulation, and dynamic closed-loop optimization. It effectively solves the problems of scaling risk and imbalance in treatment load control, making the desulfurization wastewater zero-discharge system operate more stably and reliably. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the architecture of an intelligent control system for zero discharge of desulfurization wastewater from a thermal power plant, as proposed in this invention. Figure 2 This is a schematic diagram of the workflow of the data fusion processing unit proposed in an embodiment of the present invention; Figure 3 This is a schematic diagram of the workflow of the trend risk calculation unit and the two-dimensional state decision unit proposed in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the workflow of the time-delay collaborative control unit generating the timing control strategy proposed in an embodiment of the present invention; Figure 5 This is a schematic diagram of the process reconfiguration path and feedback update workflow of the dual-path closed-loop control unit proposed in this embodiment of the invention. Detailed Implementation

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

[0018] This invention provides an intelligent control method for zero discharge of desulfurization wastewater from thermal power plants, comprising the following steps: The system acquires desulfurization wastewater operation status data, calculates the credibility of the operation status data, and generates alternative status data based on historical operation data when the credibility of the operation status data is lower than the preset conditions. The system then merges the operation status data or the alternative status data to generate a credible fused status parameter. State evolution parameters are generated within multiple consecutive time windows, and a scaling risk index is generated based on the reliable fusion of state parameters and state evolution parameters, while a zero-emission load index is also generated. A two-dimensional state space is constructed based on the scaling risk index and the zero-emission load index, and two-dimensional operating state parameters are generated based on the two-dimensional state space. Acquire two-dimensional operating state parameters and response time data of each controlled object, and generate a timing control strategy containing fast control sequences and slow control sequences based on the response time data; The parameter adjustment path or process reconfiguration path is executed according to the time-series control strategy. The parameter adjustment path is used to adjust the operating status of the desulfurization wastewater treatment unit, and the process reconfiguration path is used to switch the zero-emission treatment process. Based on the actual operating status and the target operating status, the operating deviation parameters are generated and then reused in the generation process of the reliable fusion state parameters, scaling risk index and zero-emission load index.

[0019] By adopting the above technical solution, reliable fusion state parameters are generated based on the reliability of operating status data. Combined with state evolution parameters, scaling risk index, zero-discharge load index, and dual-dimensional operating status parameters, a time-series control strategy is constructed and a parameter adjustment path or process reconfiguration path is executed to achieve dual-channel feedback closed-loop control of operating deviation parameters. This improves the problem that most existing desulfurization wastewater zero-discharge control systems use single real-time parameter feedback control. Because operating status data is easily affected by high-salt, high-hardness, and high-suspended-solids environments, the state information is distorted and lacks the ability to judge trends and loads, resulting in an imbalance between scaling risk and treatment load.

[0020] To achieve the above method, in embodiments of the present invention, the present invention provides an intelligent control system for achieving zero discharge of desulfurization wastewater from thermal power plants, such as... Figure 1 As shown, it includes: The data fusion processing unit is used to acquire desulfurization wastewater operation status data, calculate the credibility of the operation status data, generate alternative status data based on historical operation data when the credibility of the operation status data is lower than the preset conditions, and fuse the operation status data or alternative status data to generate a credible fusion status parameter. like Figure 2 As shown, further, in the data fusion processing unit, acquiring the desulfurization wastewater operating status data includes the following steps: Acquire influent flow rate data, conductivity data, pH data, temperature data, and turbidity data; Obtain calcium ion concentration data, magnesium ion concentration data, chloride ion concentration data, and sulfate ion concentration data; Timestamp the influent flow rate data, conductivity data, pH data, temperature data, turbidity data, calcium ion concentration data, magnesium ion concentration data, chloride ion concentration data, and sulfate ion concentration data. Based on a unified sampling period, time-aligned data with timestamps are generated to produce desulfurization wastewater operation status data.

[0021] Specifically, this data fusion processing unit achieves a unified representation of operational status information by acquiring, determining the reliability of, generating alternative data, and fusing the operational status data of desulfurization wastewater. The acquisition of operational status data is based on multi-source online monitoring and a time alignment mechanism, with the influent flow rate data denoted as... The conductivity data is denoted as , which characterizes the flow rate intensity of wastewater entering the system. Used to characterize the overall level of dissolved salts, pH data are denoted as Used to characterize the acid-base environment of the system, temperature data is denoted as The ambient temperature conditions used in evaporation processes and chemical reactions; turbidity data is denoted as... Used to characterize the content of suspended particulate matter, calcium ion concentration is denoted as Magnesium ion concentration is denoted as The chloride ion concentration is denoted as The sulfate ion concentration is denoted as The aforementioned monitoring data are acquired through online sensors and timestamped to form a time-series dataset. This dataset is then time-aligned based on a unified sampling period to eliminate data offsets caused by asynchronous sampling. The time alignment process can be represented as follows: at the sampling time... Construct the state vector below satisfy:

[0022] This state vector serves as the input to the operational state data for subsequent reliability calculations. The reliability calculation is constructed based on the magnitude of the operational state data's variation within a continuous time window and its coupling consistency with the conductivity variable. The magnitude of the variation can be expressed in a difference form. , ; and extend it to all ion variables to form a change vector. Simultaneously, a consistency relationship of conductivity was established. ; where the function This indicates that the salinity contribution mapping relationship obtained by fitting historical operating data is not used directly when the confidence level is lower than the set conditions. Instead, alternative values ​​are generated based on the state mapping relationship established by historical operating data. The historical operating data contains a historical state sequence that is consistent with the variable structure of the current operating condition. Through the working condition matching function Determine the set of historical operating conditions to which the current operating status belongs. And calculate the expected value of ion variables within this set of operating conditions. , , , As an alternative state data output, it is ultimately passed through the fusion function. The trusted fusion state parameters are obtained by weighted fusion of trusted online data and alternative data. ;in The credible fusion state parameters are used as the basic input for the subsequent trend risk calculation unit as the scaling risk index and zero emission load index, thereby providing a unified state representation basis for subsequent state evolution analysis and two-dimensional control decision-making.

[0023] Furthermore, in the data fusion processing unit, the credibility calculation of the operating status data includes the following steps: Calculate the variation range of ion concentration data within a continuous time window in the operational status data; The degree of data loss in the time series of ion concentration data is calculated to generate continuous calculation results; Consistency calculations were performed to verify the correspondence between ion concentration data and conductivity data; The deviation between the ion concentration data and the corresponding data distribution in historical operation data is calculated. The reliability of the operational status data is generated based on the calculation results of the change range, consistency, and deviation.

[0024] Specifically, this credibility calculation process corresponds to the credibility status screening stage in the data fusion processing unit. Its inputs are ion concentration data, conductivity data, and historical operating data from the operating status data, where the ion concentration data includes calcium ion concentration. magnesium ion concentration Chloride ion concentration and sulfate ion concentration These data originate from the online ion detection unit and are processed through time alignment to form a unified time-series state vector. Conductivity data is denoted as This is used to reflect the overall changes in dissolved salt content; historical operating data is recorded as... This is used to provide a reference benchmark for operating condition distribution; the reliability calculation first calculates the variation amplitude of ion concentration data within a continuous time window, and constructs the variation using a difference form. , , , ; and combine them into a vector of change magnitude: ;in This indicates the time interval between adjacent sampling times; secondly, it calculates the degree of data gaps in the ion concentration data over the time series and sets an indicator function. Indicates at the sampling time If valid ion concentration data exists, the continuous calculation result is expressed as follows: ;in This indicates the number of sampling points within the time window; subsequently, a consistency calculation is performed on the correspondence between ion concentration data and conductivity data to construct a functional relationship. Consistency deviation is expressed in the form of error. Further, the deviation between the ion concentration data and the corresponding data distribution in historical operating data was calculated using a distribution distance function. This indicates the difference between the current state and the historical operating condition distribution, where The multivariate Euclidean distance is represented; finally, a confidence function is constructed based on the calculation results of the magnitude of change, continuity, consistency, and distribution bias. ;in The weighting coefficients are and satisfy the following conditions: In this scheme, the weighting coefficients are preferably evenly distributed to avoid the dominance of a single indicator, and the credibility calculation results are... Used to determine whether runtime status data enters the trusted fusion path, when When the value falls below a set threshold, an alternative state generation process based on historical operational data is triggered. The alternative result and the original operational state data are then input into a fusion function to generate reliable fusion state parameters. This parameter serves as the unified input basis for state evolution analysis and risk index construction in the subsequent trend risk calculation unit.

[0025] Furthermore, in the data fusion processing unit, the process of generating alternative state data based on historical operational data includes the following steps: Extract conductivity data, pH data, turbidity data, and influent flow rate data from historical operating data; Construct a mapping relationship between operation status variables based on historical operation data; The operating condition range is determined based on conductivity data, pH data, turbidity data, and influent flow rate data. Alternative ion concentration data are generated based on the mapping relationship corresponding to the operating condition range.

[0026] Specifically, this alternative state data generation process corresponds to the missing measurement compensation and state reconstruction stage in the data fusion processing unit. Its inputs are historical operating data and a set of variables with insufficient reliability in the current operating state data. The historical operating data includes conductivity data. pH data Turbidity data and inflow rate data The above data comes from the long-term operation records of the desulfurization wastewater system and has been synchronized with time and archived with operating conditions to form a historical status set. When the reliability of the current operating status data is insufficient, the alternative generation process is triggered; firstly, conductivity data, pH data, turbidity data, and influent flow rate data are extracted from historical operating data to construct a basic set of operating condition variables. This is used to characterize the system's operational boundary conditions; a mapping relationship between operational state variables is constructed based on historical operational data, and this mapping relationship is expressed as a function. This is used to establish the mapping relationship between basic operating condition variables and ion concentration states, i.e. ;in This represents the set of historical ion concentration states. This mapping relationship is constructed by pairing operating condition variables and ion concentration variables synchronously sampled from historical operating data. Subsequently, based on conductivity data, pH data, turbidity data, and influent flow rate data, the operating condition interval is determined, and the current operating condition variables are constructed as a vector. And through the working condition distance function Matching is performed in the historical operating condition set, where The Euclidean distance metric between the current operating condition and historical operating conditions is used to select the subset of historical operating conditions with the smallest distance as the current operating condition interval. ;in Representing multidimensional Euclidean distance; finally, generating alternative ion concentration data based on the mapping relationship corresponding to the operating condition interval, and incorporating the current operating condition variables. Input the corresponding working condition range mapping function Obtain the output of the alternative ion concentration ;in This substitute ion concentration data is used to replace the original ion concentration data in the subsequent fusion calculation process when the reliability of the operating status data is lower than a set condition. It is also input into the fusion function together with the reliable operating status data to generate reliable fusion status parameters. It serves as the input basis for the state evolution analysis and scaling risk assessment of the trend risk calculation unit.

[0027] like Figure 3 As shown, the trend risk calculation unit is used to receive trusted fusion state parameters, generate state evolution parameters within multiple consecutive time windows, and generate a scaling risk index based on the trusted fusion state parameters and state evolution parameters, while also generating a zero-emission load index. Furthermore, in the trend risk calculation unit, generating state evolution parameters within multiple consecutive time windows includes the following steps: The trusted fusion state parameters are arranged in chronological order to generate a trusted fusion state parameter time series. The trusted fusion state parameter time series is divided into multiple adjacent time windows by a sliding window according to a preset time span. Differential calculations are performed on the trusted fusion state parameters within each time window to generate a sequence of state changes. The state change sequence is normalized across a time window to generate state evolution parameters.

[0028] Specifically, the state evolution parameter generation process corresponds to the time dynamic modeling stage in the trend risk calculation unit, and its input is the reliable fusion state parameters. This parameter, output by the data fusion processing unit, includes a unified set of state variables such as influent flow rate, conductivity, pH, temperature, turbidity, and calcium, magnesium, chloride, and sulfate ions. It is used to characterize the comprehensive operating status of the desulfurization wastewater system at each sampling time. First, the reliable fused state parameters are arranged in chronological order to form a time series. ;in Indicates the first The sampling time series is derived from the online monitoring and fusion output results within a continuous time window; based on this, a sliding window is formed according to the preset time span time series to create multiple sets of adjacent time windows. Each time window is represented as ;in The window length represents the system-defined time coverage range, used to describe state changes within a local time period. Differential calculations are performed on the trusted fusion state parameters within each time window to generate a sequence of state changes; this process is represented as follows: ;in The time interval between adjacent sampling moments is denoted as . This is used to characterize the rate of state change within a local time window; subsequently, the state change sequence is normalized across time windows to eliminate differences between different window scales. The processing method is as follows: The denominator represents the maximum change within the corresponding window, used to standardize the scale of change between different time windows, ultimately yielding the state evolution parameters. The state evolution parameter serves as the input basis for the trend risk calculation unit. Together with the reliable fusion state parameter, it participates in the construction of the scaling risk index and the zero emission load index, which are used for the subsequent two-dimensional state space mapping and control strategy generation of the two-dimensional state decision unit.

[0029] Furthermore, in the trend risk calculation unit, the generation of the scaling risk index based on the trusted fusion state parameters and state evolution parameters includes the following steps: Obtain online concentration ratio and evaporation temperature; A scaling risk index is generated by combining the reliable fusion state parameters, the state evolution parameters, the online concentration ratio, and the evaporation temperature.

[0030] Specifically, the scaling risk index generation process corresponds to the multi-source coupled assessment stage in the trend risk calculation unit, and its inputs include reliable fusion state parameters. State evolution parameters The online concentration ratio and evaporation temperature are used to characterize the enrichment of dissolved salts in the system. The online concentration ratio is derived from the ratio of the conductivity of the circulating liquid to the conductivity of the feed liquid in the evaporation system. It is expressed as follows: ;in Indicates the conductivity of the circulating fluid. The conductivity of the feed liquid is represented by , and the evaporation temperature is denoted as . The parameters are derived from the online temperature sensor output of the evaporation unit and are used to characterize the phase change enhancement conditions; reliable fusion state parameters The data originates from the output of the data fusion processing unit and includes comprehensive state variables such as calcium, magnesium, chloride, and sulfate ion concentrations, as well as flow rate and conductivity, and state evolution parameters. The scaling risk index, derived from the sliding window difference and normalization processing results in the trend risk calculation unit, is used to characterize the direction and intensity of system state changes over time. The scaling risk index generation process is achieved by constructing a multivariate coupling relationship function, the basic form of which is... ; where the function This represents the scaling sensitivity mapping relationship obtained based on historical operating data. In engineering implementation, it can be represented in a weighted structure form as follows: ;in This represents the hardness and salt content contribution function obtained from the state mapping of ion concentration. This represents the cumulative growth function obtained by mapping the trend of state changes. To unify the influence of different physical quantities by using weighting coefficients that satisfy normalization constraints, this scheme preferably uses offline calibration based on historical scaling data to avoid a single variable dominating the risk assessment output. (Scaling risk index) The output is a dimensionless risk characterization quantity, used as the input to the two-dimensional state decision unit as the horizontal risk dimension, and together with the zero-emission load index, it constructs a two-dimensional state space, thus providing a state basis for the selection of subsequent parameter adjustment paths or process reconfiguration paths. Among these, hardness and salinity contribution functions are included. Constructed based on the ratio of the product of calcium and magnesium ion concentrations to the sulfate ion concentration, i.e. ; The reference constant; cumulative growth function Constructed based on a weighted combination of the growth rates of each component in the state evolution parameters, i.e. ;in , , , These represent the growth rates of salinity, hardness, chloride ions, and sulfate ions, respectively. to These are the weighting coefficients. This function structure uses the time of occurrence of scaling events from historical operating data... and The samples were subjected to multiple regression calibration.

[0031] Furthermore, in the trend risk calculation unit, the generation of the zero-emission load index includes the following steps: generating the flow load LQ(t) = Q(t) based on the influent flow rate Q(t); and generating the salinity load LS(t) = Q(t) based on the product of the influent flow rate Q(t) and the total dissolved solids TDS(t). TDS(t); The hardness load LH(t) is generated by multiplying the influent flow rate Q(t) by the sum of the calcium ion concentration Ca(t) and the magnesium ion concentration Mg(t). (Ca(t)+Mg(t)); Obtain the evaporation load LE(t) and crystallization load LC(t); Weight and combine the flow load, salinity load, hardness load, evaporation load and crystallization load to generate the zero-emission load index. ;in to This is the load weighting coefficient, and it satisfies the normalization constraint. Evaporation load. The value is the ratio of the actual evaporation rate of the evaporation unit to the rated evaporation rate, and the crystallization load. The value is the ratio of the current solid-liquid separation rate of the crystallization unit to its rated processing capacity. This reflects the overall processing pressure of zero-emission systems within the future time window.

[0032] A two-dimensional state decision unit is used to construct a two-dimensional state space based on the scaling risk index and the zero-emission load index, and to generate two-dimensional operating state parameters based on the two-dimensional state space. Furthermore, in the two-dimensional state decision unit, the construction of the two-dimensional state space based on the scaling risk index and the zero-emission load index includes the following steps: The scaling risk index is normalized to generate standardized scaling risk coordinate values; The zero-emission load index is normalized to generate standardized zero-emission load coordinate values; A two-dimensional coordinate mapping relationship is constructed using standardized scaling risk coordinates as the first dimension and standardized zero-emission load coordinates as the second dimension. The state region is divided based on the two-dimensional coordinate mapping relationship to generate a two-dimensional state space.

[0033] Specifically, this two-dimensional state space construction process corresponds to the spatial mapping and region division stage in the state decision-making unit, and its input is the scaling risk index. With zero-emission load index The scaling risk index, derived from the trend risk calculation unit, is a comprehensive assessment result characterizing the probability of scaling in the system. The zero-emission load index, derived from the load assessment model, characterizes the processing pressure level of the evaporation crystallization system per unit time. First, the scaling risk index is normalized to eliminate dimensional differences; its standardization is expressed in interval mapping form. ;in and These represent the minimum and maximum values ​​of the scaling risk index in historical operating data, respectively, and the normalized result. This is used to represent standardized scaling risk coordinate values; similarly, the zero-emission load index is normalized to obtain standardized zero-emission load coordinate values. ;in and Let $\frac{1}{2}$ represent the minimum and maximum boundary values ​​of the zero-emission load index in historical operating data, respectively. The normalization result is used to unify the load dimension scale. Subsequently, a two-dimensional coordinate mapping relationship is constructed using the standardized scaling risk coordinate value as the first dimension coordinate and the standardized zero-emission load coordinate value as the second dimension coordinate. Its expression form is as follows: The coordinate points are derived from the synchronous calculation results of risk and load under the same sampling period, and are used to describe the system operating status in a two-dimensional space. Based on this, the state region is divided according to the two-dimensional coordinate mapping relationship, and the two-dimensional space is divided into multiple sets of operating regions according to the preset division rules. The region division is based on the quadrant assignment and boundary threshold determination of the coordinate points in the two-dimensional plane, ultimately generating a two-dimensional state space representation. This two-dimensional state space is used for subsequent two-dimensional operating state parameter output and serves as the input basis for the time-delay collaborative control unit to select parameter adjustment path or process reconfiguration path, thereby realizing a unified mapping and control decision basis between the risk dimension and the load dimension.

[0034] like Figure 4 As shown, the time-delay cooperative control unit is used to acquire two-dimensional operating state parameters and response time data of each controlled object, and to generate a timing control strategy containing fast control sequence and slow control sequence based on the response time data. Furthermore, in the time-delay cooperative control unit, the generation of a timing control strategy based on response time data, which includes fast control sequences and slow control sequences, includes the following steps: Collect response time data and construct a response time series; Sort the response time series according to the response time values; Based on the response time sorting results, the controlled objects are divided into fast control sequences and slow control sequences; A timing control strategy is generated based on the execution order of fast and slow control sequences.

[0035] Specifically, the generation process of this timing control strategy corresponds to the hierarchical control object and execution sequence construction stage in the time-delay collaborative control unit. Its inputs are the response time data of each control object and two-dimensional operating state parameters. The response time data originates from the dynamic response characteristic measurement results of each execution unit, used to characterize the time scale required for different control elements to reach steady state under disturbance input. The control objects include the evaporation temperature regulation unit, the circulation flow regulation unit, and the reagent dosing unit, etc. First, the response time data is collected and a response time sequence is constructed. In this process, the response time corresponding to each control object is recorded as... ,in Indicates the first The response delay time of each controlled object is derived from the results of step disturbance response tests or online identification under historical operating conditions, forming a response time series. The response time series was then sorted according to the response time values ​​using a sorting function. Generate ordered sequences satisfy This sorting process is used to establish the time-scale mapping relationship of the controlled objects; based on this, the controlled objects are divided into fast control sequences and slow control sequences according to the response time sorting results, where the fast control sequence is defined as satisfying... The set of control objects, the slow control sequence is defined as satisfying The set of controlled objects, where This represents the response time threshold, which is derived from the system stability margin analysis results or the statistical quantile value of historical operating data. Finally, a timing control strategy is generated based on the execution order of the fast control sequence and the slow control sequence. The execution relationship is that the adjustment action of the controlled object corresponding to the fast control sequence is executed first, followed by the adjustment action of the controlled object corresponding to the slow control sequence, thus forming a time-layered control sequence. ;in Represents a set of rapid control actions. This represents a set of slow control actions. This timing control strategy is used to input the dual-path closed-loop control unit to achieve the sequential coordinated execution of control links with different time scales, such as evaporation, flow rate and reagent dosing, and to provide the execution timing basis for triggering subsequent parameter adjustment paths and process reconfiguration paths.

[0036] like Figure 5 As shown, the dual-path closed-loop control unit is used to execute the parameter adjustment path or the process reconfiguration path according to the timing control strategy. The parameter adjustment path is used to adjust the operating state of the processing unit, and the process reconfiguration path is used to switch the zero-emission treatment process. It also generates operating deviation parameters based on the actual operating state and the target operating state, and feeds the operating deviation parameters back to the data fusion processing unit and the trend risk calculation unit. Furthermore, in the dual-path closed-loop control unit, the process reconfiguration path for switching the zero-emission treatment process includes the following steps: Switch between normal operation mode, enhanced hard removal mode, high load operation mode and risk protection mode; The conventional operating mode is that after softening, the material is evaporated and then crystallized. The enhanced hardening mode involves increasing the intensity of the softening reaction followed by evaporation to enter the crystallization process. The high-load operation mode involves pre-concentration followed by evaporation and then crystallization. The risk protection mode is to reduce the online concentration ratio, reduce the evaporation load, and strengthen hardening removal.

[0037] Specifically, this process reconfiguration path corresponds to the operation mode switching and process reconfiguration execution stage in the dual-path closed-loop control unit. Its inputs are two-dimensional operation state parameters, operation deviation parameters, and timing control strategies. The two-dimensional operation state parameters originate from the two-dimensional state space output of the state decision unit, representing the combined state of the scaling risk dimension and the zero-emission load dimension. The operation deviation parameters are derived from the difference calculation result between the actual operation state and the target operation state, representing the degree of current control deviation of the system. The timing control strategy originates from the time-delay collaborative control unit, used to determine the triggering order of the parameter adjustment path and the process reconfiguration path. During the execution of the process reconfiguration path, the operation mode is first mapped to a mode set. ;in This indicates the normal operating mode, where the process involves the softening unit outputting the solution, which then enters the evaporation unit and finally the crystallization unit. This indicates an enhanced hardening process, where the softening reaction intensity is increased before the material enters the evaporation unit and then the crystallization unit. This indicates a high-load operation mode. The process involves pre-concentration followed by evaporation and then crystallization. This indicates a risk protection mode, the process of which involves reducing the online concentration ratio and evaporation load while simultaneously strengthening the hardening unit; the operation mode switching is achieved through a state mapping function. The implementation takes two-dimensional operating state parameters as input. With operating deviation parameters The output is the current running mode. ;in Originating from two-dimensional state space coordinates Operating deviation parameters Derived from the calculation of the difference between the actual state vector and the target state vector ;in Indicates the actual operating status. Indicates the target's running state; mode switching is determined through a partitioning function, when... If it falls into a low-risk, low-load area, then choose When the risk of scaling increases and the load is low, choose When the load increases and the risk of scaling is low, choose When both scaling risk and load increase and operating deviation parameters exceed the set range, select... The determination process maps the two-dimensional state space partitioning results into discrete control mode outputs. The switched operating mode directly drives the change in the process connectivity of the evaporation, softening, pre-concentration and crystallization units, thereby forming a process reconfiguration path output. The operating deviation parameters are synchronously fed back to the data fusion processing unit and the trend risk calculation unit to update the state fusion results and trend calculation sequence, so that the subsequent scaling risk index and zero emission load index re-participate in the state space reconstruction and control strategy update of the two-dimensional state decision unit as the operating conditions change.

[0038] Furthermore, in the dual-path closed-loop control unit, feeding back the operating deviation parameters to the data fusion processing unit and the trend risk calculation unit includes the following steps: The operating deviation parameters are decomposed according to the type of operating state variable to generate a set of state deviation components; Match and map the set of state deviation components with the corresponding variables in the running state data; The matched state deviation components are written into the state input terminal of the data fusion processing unit. Write the running deviation parameters into the time window update terminal of the trend risk calculation unit.

[0039] Specifically, this feedback update process corresponds to the state write-back and dual-channel update mechanism in the dual-path closed-loop control unit. Its inputs include operating deviation parameters, actual operating state data, and the state input interfaces of the trend risk calculation unit and the data fusion processing unit. The operating deviation parameters are derived from the difference calculation result between the actual operating state and the target operating state, used to characterize the degree of deviation of the system in the two-dimensional state space. The actual operating state is obtained by fusing online measurement data from each desulfurization wastewater treatment unit, and the target operating state is determined by the expected operating range output by the two-dimensional state decision unit. First, the operating deviation parameters are decomposed according to the type of operating state variable. The operating deviation parameters are denoted as... And by splitting the set of state variables, we obtain a set of components:

[0040] Each component represents the deviation corresponding to flow rate, conductivity, pH, temperature, turbidity, and ion concentration, respectively. This decomposition process is based on the actual operating state vector. With the target running state vector The difference between each variable is calculated, i.e. Subsequently, the set of state deviation components is matched and mapped with the corresponding variables in the operating state data, where the operating state data is denoted as... The matching mapping process uses functional relationships. This ensures that each deviation component forms an update pair with its corresponding state variable. This mapping relationship is used to determine the target of deviation correction; subsequently, the matched state deviation components are written to the state input terminal of the data fusion processing unit, and the writing process is represented as follows: ;in Represents the original trusted fusion state parameters. This represents the updated state fusion input, used to correct subsequent fusion calculations of ion concentration and operating condition variables; finally, the operating deviation parameters are written to the time window update terminal of the trend risk calculation unit, and its update form is expressed as follows: ;in It represents a reliable fusion state time series. By introducing operational deviation parameters, the state evolution sequence is dynamically corrected, so that the calculation of state evolution parameters, scaling risk index and zero emission load index is reconstructed based on the latest deviation state, and the two-dimensional state decision unit is driven to update the two-dimensional state space simultaneously, providing a consistent state basis for the selection of subsequent parameter adjustment path and process reconstruction path.

[0041] As can be seen from the above solutions, the technical solution of the present invention has the following characteristics: 1. In this invention, a reliable fusion state parameter is generated based on the reliability of the operating state data. A time-series control strategy is constructed by combining the state evolution parameter, scaling risk index, zero-discharge load index and dual-dimensional operating state parameter, and the parameter adjustment path or process reconfiguration path is executed to achieve dual-channel feedback closed-loop control of the operating deviation parameter. This improves the problem that most existing desulfurization wastewater zero-discharge control systems use single real-time parameter feedback control. Because the operating state data is easily affected by high salt, high hardness and high suspended solids environment, the state information is distorted and lacks the ability to judge trends and loads, resulting in the control imbalance between scaling risk and treatment load.

[0042] 2. In this invention, by synchronously collecting desulfurization wastewater operation status data over time, analyzing the magnitude of ion concentration changes, calculating the degree of data loss, analyzing conductivity consistency, and calculating the distribution deviation of historical operation data, the reliability of the operation status data is generated and a reliable fusion status parameter is constructed. This improves the problem that most existing desulfurization wastewater monitoring systems use single online ion detection data to directly participate in control. Due to the drift or loss of online detection under high salt, high hardness, and high turbidity conditions, the operation status data is discontinuous or unreliable, resulting in control decisions based on distorted data.

[0043] 3. In this invention, by generating state evolution parameters within multiple continuous time windows and combining them with the scaling risk index and zero-discharge load index to construct a two-dimensional state space and generate two-dimensional operating state parameters, thereby forming a zoned control decision, the existing zero-discharge desulfurization wastewater systems mostly adopt empirical control methods based on instantaneous parameters. Due to the lack of a unified characterization of the salt accumulation trend and the coordinated changes in the system treatment load, the scaling risk identification is delayed and the evaporation and crystallization treatment capacity is mismatched with the wastewater load.

[0044] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent control method for zero discharge of desulfurization wastewater from thermal power plants, characterized in that, Includes the following steps: The system acquires desulfurization wastewater operation status data, calculates the reliability of the operation status data, and generates alternative status data based on historical operation data when the reliability of the operation status data is lower than the preset conditions. The system then merges the operation status data or the alternative status data to generate a reliable fused status parameter. State evolution parameters are generated within multiple consecutive time windows, and a scaling risk index is generated based on the reliable fusion of state parameters and state evolution parameters, while a zero-emission load index is also generated. A two-dimensional state space is constructed based on the scaling risk index and the zero-emission load index, and two-dimensional operating state parameters are generated based on the two-dimensional state space. Acquire two-dimensional operating state parameters and response time data of each controlled object, and generate a timing control strategy containing fast control sequences and slow control sequences based on the response time data; The parameter adjustment path or process reconfiguration path is executed according to the time-series control strategy. The parameter adjustment path is used to adjust the operating status of the desulfurization wastewater treatment unit, and the process reconfiguration path is used to switch the zero-emission treatment process. Based on the actual operating status and the target operating status, the operating deviation parameters are generated and then reused in the generation process of the reliable fusion state parameters, scaling risk index and zero-emission load index.

2. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 1, characterized in that, The process of acquiring the desulfurization wastewater operating status data includes: Acquire influent flow rate data, conductivity data, pH data, temperature data, and turbidity data; Obtain calcium ion concentration data, magnesium ion concentration data, chloride ion concentration data, and sulfate ion concentration data; Timestamp the influent flow rate data, conductivity data, pH data, temperature data, turbidity data, calcium ion concentration data, magnesium ion concentration data, chloride ion concentration data, and sulfate ion concentration data. Based on a unified sampling period, time-aligned data with timestamps are generated to produce desulfurization wastewater operation status data.

3. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 2, characterized in that, The credibility calculation of operational status data includes the following steps: Calculate the variation range of ion concentration data within a continuous time window in the operational status data; The degree of data loss in the time series of ion concentration data is calculated to generate continuous calculation results; Consistency calculations were performed to verify the correspondence between ion concentration data and conductivity data; The deviation between the ion concentration data and the corresponding data distribution in historical operation data is calculated. The reliability of the operational status data is generated based on the calculation results of the change range, consistency, and deviation.

4. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 2, characterized in that, The process of generating alternative status data based on historical operational data includes the following steps: Extract conductivity data, pH data, turbidity data, and influent flow rate data from historical operating data; Construct a mapping relationship between operation status variables based on historical operation data; The operating condition range is determined based on conductivity data, pH data, turbidity data, and influent flow rate data. Alternative ion concentration data are generated based on the mapping relationship corresponding to the operating condition range.

5. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 1, characterized in that, The process of generating state evolution parameters within multiple consecutive time windows includes the following steps: The trusted fusion state parameters are arranged in chronological order to generate a trusted fusion state parameter time series. The trusted fusion state parameter time series is divided into multiple adjacent time windows by a sliding window according to a preset time span. Differential calculations are performed on the trusted fusion state parameters within each time window to generate a sequence of state changes. The state change sequence is normalized across a time window to generate state evolution parameters.

6. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 1, characterized in that, The process of generating a scaling risk index based on trusted fusion state parameters and state evolution parameters includes the following steps: Obtain online concentration ratio and evaporation temperature; A scaling risk index is generated by combining the reliable fusion state parameters, the state evolution parameters, the online concentration ratio, and the evaporation temperature.

7. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 1, characterized in that, The construction of the two-dimensional state space based on the scaling risk index and the zero-emission load index includes the following steps: The scaling risk index is normalized to generate standardized scaling risk coordinate values; The zero-emission load index is normalized to generate standardized zero-emission load coordinate values; A two-dimensional coordinate mapping relationship is constructed using standardized scaling risk coordinates as the first dimension and standardized zero-emission load coordinates as the second dimension. The state region is divided based on the two-dimensional coordinate mapping relationship to generate a two-dimensional state space.

8. The intelligent control method for zero discharge of desulfurization wastewater from thermal power plants according to claim 1, characterized in that, The generation of a timing control strategy containing fast and slow control sequences based on response time data includes the following steps: collecting response time data and constructing a response time sequence; sorting the response time sequence according to the response time value; dividing the controlled object into a fast control sequence and a slow control sequence based on the response time sorting result; and generating a timing control strategy based on the execution order of the fast and slow control sequences. The process reconfiguration path for switching the zero-emission treatment process includes the following steps: switching the operating mode between a normal operating mode, an enhanced hardening removal mode, a high-load operating mode, and a risk protection mode; wherein, the normal operating mode is softening followed by evaporation and then crystallization; the enhanced hardening removal mode is increasing the intensity of the softening reaction followed by evaporation and then crystallization; the high-load operating mode is pre-concentration followed by evaporation and then crystallization; and the risk protection mode is reducing the online concentration ratio, reducing the evaporation load, and enhancing hardening removal.

9. An intelligent control system for zero discharge of desulfurization wastewater from thermal power plants, characterized in that, A smart control method for achieving zero discharge of desulfurization wastewater from a thermal power plant as described in any one of claims 1-8 includes: The data fusion processing unit is used to acquire the operating status data of desulfurization wastewater, calculate the credibility of the operating status data, generate alternative status data based on historical operating data when the credibility of the operating status data is lower than the preset conditions, and fuse the operating status data or alternative status data to generate a credible fusion status parameter. The trend risk calculation unit is used to receive trusted fusion state parameters, generate state evolution parameters within multiple consecutive time windows, and generate a scaling risk index based on the trusted fusion state parameters and state evolution parameters, while also generating a zero-emission load index. A two-dimensional state decision unit is used to construct a two-dimensional state space based on the scaling risk index and the zero-emission load index, and to generate two-dimensional operating state parameters based on the two-dimensional state space. The time-delay cooperative control unit is used to acquire two-dimensional operating state parameters and response time data of each controlled object, and generate a timing control strategy containing fast control sequence and slow control sequence based on the response time data; The dual-path closed-loop control unit is used to execute parameter adjustment path or process reconfiguration path according to the timing control strategy. The parameter adjustment path is used to adjust the operating state of the processing unit, and the process reconfiguration path is used to switch the zero-emission treatment process. It generates operating deviation parameters based on the actual operating state and the target operating state, and feeds the operating deviation parameters back to the data fusion processing unit and the trend risk calculation unit.

10. The intelligent control system for zero discharge of desulfurization wastewater from thermal power plants according to claim 9, characterized in that, In the dual-path closed-loop control unit, feeding back the operating deviation parameters to the data fusion processing unit and the trend risk calculation unit includes the following steps: The operating deviation parameters are decomposed according to the type of operating state variable to generate a set of state deviation components; Match and map the set of state deviation components with the corresponding variables in the running state data; The matched state deviation components are written into the state input terminal of the data fusion processing unit. Write the running deviation parameters into the time window update terminal of the trend risk calculation unit.

Images