CN115146871A - Intelligent desulfurization system based on data - Google Patents
Intelligent desulfurization system based on data Download PDFInfo
- Publication number
- CN115146871A CN115146871A CN202210901420.1A CN202210901420A CN115146871A CN 115146871 A CN115146871 A CN 115146871A CN 202210901420 A CN202210901420 A CN 202210901420A CN 115146871 A CN115146871 A CN 115146871A
- Authority
- CN
- China
- Prior art keywords
- data
- desulfurization system
- power station
- learning
- steady state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000006477 desulfuration reaction Methods 0.000 title claims abstract description 54
- 230000023556 desulfurization Effects 0.000 title claims abstract description 54
- 238000005457 optimization Methods 0.000 claims abstract description 23
- 238000013486 operation strategy Methods 0.000 claims abstract description 5
- 238000004134 energy conservation Methods 0.000 claims abstract 3
- 230000007613 environmental effect Effects 0.000 claims abstract 3
- 238000000034 method Methods 0.000 claims description 13
- 238000010845 search algorithm Methods 0.000 claims description 13
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 claims description 9
- 239000003546 flue gas Substances 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000011160 research Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims 1
- 239000000463 material Substances 0.000 abstract description 4
- 230000026676 system process Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 8
- 238000010248 power generation Methods 0.000 description 7
- 239000002002 slurry Substances 0.000 description 7
- RAHZWNYVWXNFOC-UHFFFAOYSA-N Sulphur dioxide Chemical compound O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 239000000126 substance Substances 0.000 description 4
- 241000196324 Embryophyta Species 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 239000003245 coal Substances 0.000 description 3
- 230000003472 neutralizing effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000003912 environmental pollution Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000003916 acid precipitation Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003009 desulfurizing effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000003517 fume Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/46—Removing components of defined structure
- B01D53/48—Sulfur compounds
- B01D53/50—Sulfur oxides
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/74—General processes for purification of waste gases; Apparatus or devices specially adapted therefor
- B01D53/80—Semi-solid phase processes, i.e. by using slurries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Biomedical Technology (AREA)
- Environmental & Geological Engineering (AREA)
- Human Resources & Organizations (AREA)
- General Chemical & Material Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Oil, Petroleum & Natural Gas (AREA)
- Analytical Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Primary Health Care (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Fuzzy Systems (AREA)
Abstract
The invention discloses an intelligent desulfurization system based on data, which comprises: the off-line learning module is used for reading historical data of the power station which has run from the database, then checking the historical data and eliminating error data; further searching data of the power station during steady state operation after eliminating error data; learning from steady state operating data; and the online optimization module is used for searching an optimal operation strategy in an offline learning result according to the current operation data and instruction data in a period of time so as to achieve the aim of energy conservation or environmental protection. The invention reduces the power consumption and material input of the equipment as much as possible under the condition of maintaining the work load of the power station and the condition that the desulfurization system processes a certain amount of emissions.
Description
Technical Field
The invention relates to the field of data analysis, in particular to an intelligent desulfurization system based on data.
Background
The electric power industry is the largest energy production and consumption industry in China, the electric power supports the development of scientific and technological civilization to a great extent, and the electric power is in society and industry, and is indispensable just like oxygen for human survival. Unlike hydroelectric power generation and wind power generation, thermal power generation is not limited by weather and geography, and can stably provide energy, so that thermal power still occupies most of the market of electric power. However, compared with water conservancy power generation and wind power generation, thermal power generation consumes more natural resources and causes more damage to the environment, wherein flue gas discharged by coal-fired units of thermal power plants becomes a main source of sulfur dioxide emission. The emission of sulfur dioxide is the main reason causing atmospheric environmental pollution and continuous aggravation of acid rain, so that the monitoring on the operation of the coal-fired desulfurization and denitrification unit facilities is strengthened, the online monitoring and analysis of desulfurization and denitrification data and the qualification of desulfurization and denitrification power price check of a power grid are greatly necessary, and the method has a very important significance for guaranteeing the safety and economic operation of the power grid. During operation of the plant, SO 2 The emissions of (c) are related to a number of factors, the most closely related of which is the operating load of the plant. Higher plant operating loads mean more equipment is in operation, and therefore more coal and alkaline materials are required to neutralize the emissions. If the supply of the coal and the neutralizing substances is insufficient, the power station can not meet the user requirements or the discharge substances exceed the standard, and if the supply of the coal and the neutralizing substances is excessive, certain energy waste can be caused. Many researches have been made to solve the problems of energy consumption and environmental pollution of thermal power generation, including the realization of planned energy input by using electricity utilization characteristics in different regions, and the load prediction of electricity utilization requirements by using traditional or neural networks and other methods.
Disclosure of Invention
The invention aims to provide an intelligent desulfurization system based on data aiming at the defects of the prior art.
The invention is realized by adopting the following technical scheme:
the invention provides an intelligent desulfurization system based on data, which comprises two modules, namely an off-line learning module and an on-line optimization module.
Without loss of generality, the invention is directed to a power station desulfurization system comprising a primary desulfurization tower and a secondary desulfurization tower. In the actual operation of the power station, because the unit load is influenced by a plurality of factors, the value of the unit load is continuously changed, when the unit load is in the process of rising or falling, the desulfurization system is in an unsteady state operation state, and various measured data can not well reflect the actual operation state of the desulfurization system. In order to ensure the validity of the learning result, the invention only considers the condition that the desulfurization system is in a steady-state operation state. For the learning process of the off-line learning module, the invention provides a set of algorithm aiming at large data steady-state value search, namely a sliding window search algorithm. In addition, in order to enable the system of the invention to be used in different power stations, the module also provides some settable parameters which can be set according to the actual conditions of the different power stations. The method specifically comprises the following steps:
in the data of the first-stage tower and the second-stage tower, the PH value of the effective data is within a reasonable value range, if any one of the PH values is larger than a set upper limit or lower than a set lower limit, the reference value of the data can be judged to be smaller, and therefore the upper and lower limits of the PH values of the first-stage tower and the second-stage tower are used as two settable parameters;
in the data of the first-stage tower and the second-stage tower, the desulfurization efficiency of effective data should be in a reasonable value range, if any one of the desulfurization efficiency values is lower than a set value, the reference value of the data can be judged to be smaller, so the lower limit of the desulfurization efficiency values of the first-stage tower and the second-stage tower should be taken as two settable parameters, and the reason that the upper limit is not taken as the settable parameter is that the desulfurization efficiency cannot reach 100% generally in an actual working condition;
because of the system provided by the invention, the online optimization is realizedThe chemical module has an energy-saving mode and an environment-friendly mode, and the clean flue gas SO needing to be treated in the two modes 2 The contents should be different, SO two different thresholds are set respectively, when the clean flue gas SO is in two modes 2 When the contents are respectively larger than the respective threshold values, the reference value of the data can be judged to be smaller, so that the two threshold values are used as two settable parameters;
in the data of the first-stage tower and the second-stage tower, when the working current and the working voltage of the desulphurization pump of the first-stage tower and the second-stage tower are lower than the set threshold values, the pump is considered to be closed, otherwise, the pump is opened. The conditions under which the desulfurization pump is turned on or off should be specifically set by the conditions of different power stations. Therefore, the lower limits of the working current and the voltage of the desulfurization pumps of the first-stage tower and the second-stage tower are used as settable parameters.
After the settable parameters are set, the huge data of the power station can be preliminarily screened, and since the screened data not only comprises the data of the power station in the steady-state operation period, but also comprises the data of the power station in the unsteady-state operation period, and the data in the unsteady-state operation period has little reference significance, the steady-state value search is carried out on the data by using the sliding window search algorithm provided by the invention in the next step, and the data in the steady-state operation period of the power station is reserved and processed in the next step.
The contents of the sliding window search algorithm include:
1) Selecting a steady-state load bandwidth N (namely when the unit loads in a continuous time are all in a certain fixed load interval, the time interval is judged to be a steady-state operation time interval, and the load interval is the steady-state load bandwidth). The selection of the steady-state load bandwidth is mainly based on the installed capacity of the power station unit, generally selected to be 1% -10% of the installed capacity, and can be selected according to specific working conditions. For example, the installed capacity of a certain unit is 600MW, and the steady-state load bandwidth is 3% of the installed capacity, so the steady-state load bandwidth is N =18MW.
2) A minimum steady state operating time period T is determined. During the operation of the system, the sampling of data has a certain time interval, which is set as s, and the minimum steady-state operation duration is set as T =15 minutes, which means that in a certain time period not less than 15 minutes, if all normal data points fall within the steady-state load bandwidth, the time period can be considered as a steady-state phase, whereas if 46 data points within 15 minutes except for abnormal points do not all fall within the steady-state load bandwidth, the time period is an unsteady state, which means that the power station is in a load ascending phase or a load descending phase.
3) And calculating a steady state value of the generator set during operation. The load of the first piece of data in the historical normal operation data of the unit is taken as an initial central point, a first stable load band is established with the set bandwidth N in the step 1), by respectively moving the stable load band up and down by N/2 units, (1 + N) stable load bands can be obtained altogether, and the maximum number of data points continuously contained in the 1+ N stable load bands is respectively found out. If the maximum value of the number of data points continuously contained in the nineteen steady-state load bands is smaller than 46, deleting the first data point and reconstructing 1+ N steady-state load bands by taking the second data point as an initial central point; if the maximum value of the number of data points continuously contained in the 1+ N steady-state load bands is greater than or equal to 46, selecting the load bandwidth containing the number of data points greater than or equal to 46 from the 1+ N load bandwidths by using a statistical technique, then performing an averaging operation on the center of the selected load bandwidth, and finally taking the obtained value as the load steady-state value of the period of time. And then, by taking the next data load after the time period as an initial central point, reconstructing 1+ N steady-state load bands with N as the bandwidth, and repeating the operations until all steady-state loads and corresponding steady-state load time periods are found.
4) Considering the abnormal load condition, the maximum allowable abnormal data number can be introduced to improve the robustness of the unit steady-state data search algorithm, and the fact that when the data point is in a steady state, the data point cannot jump out of the steady state due to large disturbance is substantially ensured.
It needs to be provided that the parameters in the sliding window search algorithm can be set independently by matching with different hardware, so as to be popularized to other power stations.
The contents of the data learning algorithm comprise:
because the data in the online optimization moduleThe matching function requires the local data generated in the offline module, so the generation of the local data is introduced first. The premise still is that the power station desulfurization system comprises a primary tower and a secondary tower. In the off-line learning module, a learning algorithm respectively learns the SO treatment per second of the first-stage tower desulfurization pump under various working conditions (divided according to a load section and a slurry PH section) 2 Maximum value of content; the learning algorithm respectively learns the SO treatment per second of the secondary tower desulfurization pump under various working conditions (divided according to the load section and the slurry PH section) 2 The maximum value of the content and the basic principle of learning are classified learning according to the working conditions. Firstly, classifying and storing all steady-state data points according to the starting state of a power station pump, naming a stored file by using the started pump, taking the case that five ABCDE desulfurization pumps are shared in a first-stage tower and three FGH pumps are shared in a second-stage tower as an example, an ABC folder shows that data of starting of the pumps A, B and C and stopping of the pumps D and E in the first-stage tower are stored under the folder; FG folder indicates that data of F, G pump on, H pump off in the secondary tower are stored under this folder. Then, a neural network is constructed, and the maximum SO per second processing of the first-stage tower and the second-stage tower under various working conditions is learned one by one 2 And (4) content. The load value of the data class is taken as an x axis, the pH value of the slurry is taken as a y axis, and SO is processed per second 2 Establishing a three-dimensional coordinate system for the content of the slurry in the z-axis, and establishing a neural network to obtain SO processed per second corresponding to the load section and the slurry PH value section 2 The maximum value of the amount. The operation conditions of the pumps of the first-stage tower and the second-stage tower of the desulfurization system can establish corresponding learning models. And finally, storing the learned knowledge locally for use in an online optimization module.
The content of the online optimization module is as follows:
firstly, the real-time working condition data of the power station is preliminarily calculated, and then the calculated data is searched and matched in an off-line learning result. The method for searching and matching comprises the following steps: firstly, the content of the instruction load stored locally is searched, and under the condition that the working equipment is found in the local content and reaches the instruction load, the current SO capable of being processed is further searched 2 The content data, including the PH of the effluent, were finally recorded. And the number of the searched results is reduced according to the number of the opening of the desulfurization pumpThe order of the system discharge PH is given in turn, thereby achieving the purpose of energy saving.
The present invention has at least the following advantageous effects
The research angle of the invention is different from the macroscopic analysis, and the invention is more focused on the optimization of the power station in the actual operation and more focused on the implementation of details. Aiming at a power station desulfurization system, the optimization idea of the invention is as follows: if a rule can be formulated that specifies different SO's at different loads 2 In the case of contents, minimum number of desulphurization pumps to be started and minimum input of neutralizing material. When different working conditions are processed, the optimal mode can be selected according to the rule to process the working conditions, and therefore the purpose of reducing resource consumption is achieved.
Specifically, the invention reduces the power consumption and material input of the equipment as much as possible while maintaining the work load of the power station and the desulfurization system to process a certain amount of emissions. The system mainly comprises two modules, namely an off-line learning module and an on-line optimization module. The offline learning module is mainly used for reading related historical data from a database of the power station, searching data in the historical data during steady-state operation, then learning from the steady-state operation data to obtain the optimal combination mode of the desulfurization pump and the corresponding PH value of the emission under different working conditions, and finally storing the learned content locally according to a set frame. The operation mechanism of the online optimization module is to search the optimal operation strategy of the desulfurization system in the offline learning result according to the current operation data and instruction data of the power station in a period of time. In order to ensure the robustness of the online optimization module, the module generally provides several optimal combinations matched with the current working conditions for operators to refer to, so that the aim of intelligent desulfurization is fulfilled.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flowchart illustrating the operation of a system including two modules according to the present invention.
FIG. 2 is a schematic diagram of a steady state value search algorithm.
Fig. 3 is a diagram for allowing an outlier to occur when searching for a steady-state value.
FIG. 4 is a diagram of the processing of SO per second by the learning algorithm to obtain each operating condition 2 And (3) constructing a model schematic diagram of the maximum content.
FIG. 5 is a frame for storing learned knowledge in a local area by a learning algorithm, wherein the left side shows a folder named by an opened pump, and the right side shows SO processed per second corresponding to different load segments and slurry pH value segments in a certain pump combination mode 2 The maximum value of the amount. The file names are named in load intervals, and the size of the intervals is adjustable in algorithm implementation.
Fig. 6 is a graph showing an operation load from 11 months, 1 day 0 to 11 months, 4 days, 0 of a certain power plant 2020.
Fig. 7 is a schematic diagram of search results of the sliding window steady-state value search algorithm.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The terms "comprises" and "comprising," and any variations thereof, in the present description and claims and drawings are intended to cover a non-exclusive inclusion, such as a list of steps or elements.
In the following, the technical solution of the present invention is further described in detail with reference to the accompanying drawings and examples, and in order to avoid loss of generality, the case provided herein is still based on the premise that the desulfurization system is divided into a first-stage tower and a second-stage tower.
As shown in fig. 1, the operation flow of the system provided by the present invention is shown. The method comprises the following steps:
1. by analyzing the data characteristics of the power station, the clean flue gas SO to be processed is obtained by utilizing the settable parameters provided by the system, such as the desulfurization efficiency of the desulfurizing tower 2 Content, PH value of the emissions to be processed and the like, and screening information in a power station database to select data with greater significance for algorithm learning.
2. And searching the screened data for the data of the desulfurization system in steady state operation by using the sliding window steady state value searching algorithm provided by the invention, namely the data of the generator set in steady state operation. The steady-state value search algorithm has been described in detail in the summary of the invention, fig. 2 is a schematic diagram of the steady-state value search algorithm, and fig. 3 is a schematic diagram of allowing an outlier to appear when searching for a steady-state value.
3. After data of the power station in the stable operation period are obtained, the maximum SO processing per second corresponding to the tower desulfurization pump under various working conditions (divided according to a load section and a slurry PH section) is learned from the data according to the learning algorithm provided by the invention 2 Content, storing the learned content locally according to a set frame, wherein the specific frame is elaborated in the invention content, and FIG. 4 is a graph for processing SO per second for each working condition obtained by a learning algorithm 2 Schematic of the maximum of the content. FIG. 5 is a framework in which the learning algorithm saves learned knowledge locally.
4. After the learned knowledge is stored locally, the online optimization module can be entered, and a near-real-time optimization suggestion is provided for the working condition of the power station. When the online optimization module runs, the module can read load data of five minutes before the current moment (the time is a settable parameter and is set according to the specific conditions of different power stations) and the desulfurized raw flue gas SO of the current moment from the database 2 Content data, communicating flue gas SO 2 Content data, desulfurization inlet flue gas volume data and load instructions. Firstly, the load data five minutes before the current time is averaged, then the calculated average value is differed with the current obtained load instruction, and if the absolute value of the difference between the average value and the load instruction is in an acceptable rangeIn the enclosure, the system can think that the current power station is in a stable state, and search in the learned knowledge to find the information which can meet the current working condition. If the absolute value of the difference between the average value and the load command is outside the acceptable range, the system considers that the power station operation state is in a transient state at the current moment and displays that the transient process is currently performed.
The specific flow proposed by the optimization module is that firstly, the instruction load is used for searching in all locally stored data once, the searching is started from the starting and stopping states of the power station pump during the searching, and if all load data in a certain starting and stopping state cannot be matched with the instruction load, the program directly enters the next state until the state data which can be matched with the instruction load is found. Next, it is determined that SO can be processed for this state under the instruction load 2 Whether the content meets the current emission index, if not, the algorithm will reduce the PH requirement and preferentially meet the SO requirement 2 The discharge index, if the PH requirement is lowered to also not meet the current discharge index, the algorithm will increase the number of pumps that are turned on. According to the process, the algorithm can match the instruction load in all local data once, and after all local data are traversed, the algorithm can obtain a series of working conditions that the working load meets the load instruction and the emission index reaches the standard. In order to achieve the energy-saving effect, the algorithm will preferentially suggest the working condition that the number of the desulphurization pumps is small, for example, the suggestion that two pumps are started is preferred to the suggestion that three pumps are started, and if the number of the pumps is the same, the algorithm will preferentially suggest the working condition with lower PH value.
5. It should be noted that, in the case of a two-stage desulfurization tower, the optimization proposal of the invention is to give a proposal of a first-stage tower and then give a proposal of a second-stage tower according to the fume data of the first-stage tower to the second-stage tower. In order to ensure the reliability of the invention, when the optimization proposal of the secondary tower is calculated, the invention calculates the maximum value of the theoretical calculation value and the actual operation value of the power station, thereby ensuring that the flue gas discharged to the atmosphere after passing through the secondary desulfurization tower meets the standard.
6. Through two modules, the purpose of intelligent desulfurization is realized. It should be noted, however, that the longer the system is in use, the more significant the optimization results. Because the data which is learned at the beginning contains a large amount of data with smaller reference value, the system extracts information with more reference value through one off-line learning, and optimizes the combination mode of the desulfurization pump by using the information. And the optimized data is used as historical data for off-line learning, so that a better optimization scheme can be continuously searched from the optimized data, and the optimization is always towards a better direction through iterative steps.
The system provided by the invention is already put into use in a desulphurization system of a certain power station, and fig. 6 and 7 are schematic diagrams of the results of the sliding window steady-state value search algorithm.
The embodiments of the present invention have been described in detail. However, the present invention is not limited to the above-described embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention.
The above embodiments are provided to further explain the objects, technical solutions and advantages of the present invention in detail, it should be understood that the above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the scope of the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (8)
1. Intelligent desulfurization system based on data, its characterized in that includes:
the off-line learning module is used for reading the historical data of the power station which has run from the database, then checking the historical data and eliminating error data; further searching data of the power station during steady state operation after eliminating error data; learning from steady state operating data;
and the online optimization module is used for searching an optimal operation strategy in an offline learning result according to the current operation data and instruction data in a period of time so as to achieve the aim of energy conservation or environmental protection.
2. The intelligent desulfurization system based on data of claim 1, characterized in that the off-line learning module reads the historical data of the power station operation from the database and screens the data, the purpose of the screening is to eliminate the error data caused by hardware and to retain the data with value, and the basis of the screening is a series of parameters set after the research.
3. The data-based intelligent desulfurization system of claim 2, wherein the parameters include: PH value of effluent in desulfurization system, desulfurization efficiency of desulfurization tower in desulfurization system, and clean flue gas SO to be treated by desulfurization system 2 And (4) content.
4. The system of claim 1, wherein the offline learning module searches for data during steady state operation of the power plant, and specifically comprises: providing a steady state value search algorithm and a sliding window search algorithm aiming at the big data of the power station; the algorithm has settable parameters for abnormal data detection and adjustment of algorithm accuracy, including steady state load bandwidth for adjustment of data fluctuation range and minimum steady state operation duration for adjustment of steady state value accuracy over time.
5. The intelligent desulfurization system based on data of claim 4, characterized in that the parameters in the sliding window search algorithm can be set autonomously with different hardware.
6. The system of claim 1, wherein the offline learning module learns from data in steady state operation, and specifically comprises: the learning algorithm can learn the optimal operation strategy of the desulfurization system under different working conditions, realizes offline in the learning process, and finally stores the learned contents according to a set frame.
7. The intelligent desulfurization system based on data of claim 1, wherein the online optimization module finds the optimal operation strategy in the results obtained by offline learning according to the real-time operation data of the current period of time, and specifically comprises: firstly, the real-time working condition data of the power station are preliminarily calculated, then the calculated data are searched and matched in the result of off-line learning, and the searched results are sequentially given according to the sequence that the number of the opening of the desulphurization pumps is small and the PH of the system emission is low, so that the purpose of energy conservation is achieved.
8. The intelligent data-based desulfurization system of claim 7, wherein the online optimization module is provided with an energy-saving mode and an environmental protection mode.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210901420.1A CN115146871A (en) | 2022-07-28 | 2022-07-28 | Intelligent desulfurization system based on data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210901420.1A CN115146871A (en) | 2022-07-28 | 2022-07-28 | Intelligent desulfurization system based on data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115146871A true CN115146871A (en) | 2022-10-04 |
Family
ID=83414785
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210901420.1A Pending CN115146871A (en) | 2022-07-28 | 2022-07-28 | Intelligent desulfurization system based on data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115146871A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109636001A (en) * | 2018-11-13 | 2019-04-16 | 北京国电龙源环保工程有限公司 | Desulfurization pulp feeding system pH value adjusting method, system and computer-readable medium based on big data |
-
2022
- 2022-07-28 CN CN202210901420.1A patent/CN115146871A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109636001A (en) * | 2018-11-13 | 2019-04-16 | 北京国电龙源环保工程有限公司 | Desulfurization pulp feeding system pH value adjusting method, system and computer-readable medium based on big data |
CN109636001B (en) * | 2018-11-13 | 2023-05-23 | 国能龙源环保有限公司 | Method, system and computer readable medium for adjusting pH value of desulfurization slurry supply system based on big data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Xu et al. | Carbon emission reduction and reliable power supply equilibrium based daily scheduling towards hydro-thermal-wind generation system: A perspective from China | |
CN108876163B (en) | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning | |
Gibadullin et al. | The development strategy of the environmental safety of the electric power complex | |
CN111522290A (en) | Denitration control method and system based on deep learning method | |
CN115146871A (en) | Intelligent desulfurization system based on data | |
CN109711589A (en) | A kind of SCR denitration system running optimizatin method and system based on big data | |
CN112506162B (en) | Oxidation air system control method based on data model and mechanism operation | |
CN107203687A (en) | Absorption tower sweetening process multiple target cooperative intelligent optimal control method | |
CN110624696B (en) | Intelligent electric dust removal energy-saving method | |
CN115309117B (en) | WFGD export SO based on data drive2Concentration prediction and intelligent optimization method | |
CN111917111B (en) | Method, system, equipment and storage medium for online evaluation of distributed photovoltaic power supply acceptance capacity of power distribution network | |
CN115094481A (en) | Modular alkaline electrolyzed water hydrogen production scheduling switching method adapting to wide power fluctuation | |
CN112365065A (en) | WFGD self-adaptive online optimization scheduling method | |
CN117595488A (en) | Power dispatching monitoring method and system based on load dynamic matching | |
CN117933548A (en) | Method for applying carbon emission management and optimization technology to transformer substation | |
CN110829484B (en) | Space-time decomposition-based global energy interconnection power balance optimization method | |
CN117479326A (en) | Wireless network resource scheduling optimization method and device based on flow prediction and storage medium | |
Ding et al. | Low carbon economic dispatch of power system at multiple time scales considering GRU wind power forecasting and integrated carbon capture | |
CN110570124A (en) | Economic quantification model for improving degree of distribution network operation performance by distribution network project | |
Wang et al. | Research on Intelligent Dispatching Optimization of New Energy Grid Considering the Impact of Wind Power | |
CN114924611B (en) | Photovoltaic cell maximum power point tracking method, device and medium | |
CN117394444B (en) | Direct-current power distribution network distribution robust optimization scheduling method based on deep learning assistance | |
Zhou et al. | Design of urban sludge emission reduction optimisation strategy based on fuzzy neural network | |
CN117498457A (en) | New energy optimal permeability determination method and system for electric power system | |
Lu et al. | Research on SO 2 emission prediction model of desulfurization system based on GRU |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |