Energy prediction method and device based on working condition change
By collecting and analyzing data from multiple energy devices, a working condition identification and prediction system was constructed, which solved the problems of data dependence and error accumulation in energy forecasting for steel enterprises, and realized the accurate scheduling and balance prediction of multiple energy media.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-18
- Publication Date
- 2026-03-27
AI Technical Summary
Existing steel energy forecasting methods suffer from problems such as strong data dependence, severe error accumulation, inability to predict multiple energy media, and low forecasting accuracy.
By collecting energy data from multiple energy devices, an energy database is formed. Sets of routine and important operating conditions are selected, energy fluctuation constants are set, an energy change fluctuation function is constructed, operating condition types are identified and tracked, and an energy change prediction curve is built, thereby reducing data dependence and improving prediction accuracy.
It enables accurate and reliable prediction of multiple energy media, reduces error accumulation, provides accurate reference information on energy balance, and improves the rationality and accuracy of energy dispatch.
Smart Images
Figure CN121745352A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of steel energy prediction technology, and in particular to an energy prediction method and apparatus based on changes in operating conditions. Background Technology
[0002] The steel industry is energy-intensive, with its production process involving the coordinated operation of multiple stages and equipment, resulting in large energy consumption and a wide variety of energy media, such as electricity, gas, and steam. In actual production, how to efficiently and accurately predict energy levels has become a key issue for steel companies to improve energy efficiency and reduce energy consumption. Current steel energy prediction methods are mainly divided into data-driven prediction methods and prediction methods that combine process and data. Data-driven methods primarily focus on modeling using deep learning and other techniques; while methods that combine process and data combine energy unit structure with historical data to predict energy fluctuations.
[0003] However, most existing technologies remain at the theoretical level, and traditional methods focus energy forecasting on the production and consumption forecasts of each user. Due to the large number of energy users and the influence of multi-user forecasting errors, existing solutions cannot guarantee the accuracy of forecast results and are prone to data dependence. When current production conditions change, the reliability of their forecast results will also decrease. Summary of the Invention
[0004] In view of the above analysis, the present invention aims to provide an energy prediction method and apparatus based on changes in operating conditions, in order to solve the problems of strong data dependence, serious error accumulation, inability to predict multiple energy media, and low prediction accuracy in the prior art.
[0005] On one hand, embodiments of the present invention provide an energy prediction method based on changes in operating conditions, including:
[0006] Collect energy data from multiple energy devices and create an energy database;
[0007] Based on the energy database, a set of normal operating conditions and a set of important operating conditions are obtained by filtering.
[0008] Set an energy fluctuation constant for the set of normal operating conditions; construct an energy change fluctuation function for the set of important operating conditions; obtain the current operating condition type identification result based on the real-time collected current energy data and the energy change fluctuation function;
[0009] Based on the energy database, the discriminant variables and discriminant conditions are determined; based on the real-time collected current energy data, discriminant variables, and discriminant conditions, the current operating condition type discrimination result is obtained; the consistency between the operating condition type discrimination result and the operating condition type identification result is compared, and the corrected operating condition distribution information is determined based on the comparison result;
[0010] Based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, an energy change prediction curve is constructed, and the flow prediction value of the energy medium in each energy device is obtained.
[0011] Furthermore, the process of collecting energy data from multiple energy devices and forming an energy database includes:
[0012] The energy data is cleaned and then categorized and stored using TDengine super tables to form the energy database.
[0013] The energy database includes at least: energy production and consumption data of each energy device, equipment operation data, and production and maintenance plan data.
[0014] Furthermore, based on the energy database, the set of normal operating conditions and the set of important operating conditions are obtained by filtering, including:
[0015] Based on the production and maintenance plan data and the energy production and consumption data, determine the changes in energy medium fluctuations under each operating condition:
[0016] When the fluctuation of the energy medium is unstable, the current operating condition is classified into the set of important operating conditions; otherwise, the current operating condition is classified into the set of normal operating conditions.
[0017] Further, the step involves setting an energy fluctuation constant for the set of normal operating conditions; constructing an energy change fluctuation function for the set of important operating conditions; and obtaining the current operating condition type identification result based on the real-time collected current energy data and the energy change fluctuation function, including:
[0018] Clustering algorithms are used to identify the characteristics of energy medium fluctuations under various operating conditions, and energy fluctuation curves under various operating conditions are extracted.
[0019] Based on the operating condition screening results, the energy change fluctuation curves corresponding to each normal operating condition with zero change are set as energy fluctuation constants; and an energy change fluctuation function is constructed based on the energy change fluctuation curves corresponding to each important operating condition.
[0020] The energy change fluctuation curve obtained from the real-time collected current energy data is compared with the energy change fluctuation curve under each operating condition, and the operating condition cluster with the closest target distance is selected as the current operating condition type identification result.
[0021] Further, the step of determining the discriminant variables and discriminant conditions based on the energy database; and obtaining the current operating condition type discrimination result based on the real-time collected current energy data, discriminant variables, and discriminant conditions, includes:
[0022] The discriminant variables and the discriminant conditions are determined based on the equipment operation data, and a set of discriminant variables for the corresponding working conditions is obtained based on the discriminant variables;
[0023] Based on the set of discriminant variables and the discriminant conditions, the real-time collected current energy data is discriminated to obtain the current operating condition type discrimination result.
[0024] Further, the step of comparing the consistency between the working condition type discrimination result and the working condition type identification result, and determining the corrected working condition distribution information based on the comparison result, includes:
[0025] If the comparison is consistent, the current working condition is corrected based on the working condition type identification result, and the corrected working condition distribution information is determined.
[0026] Otherwise, the corrected working condition distribution information is determined according to the pre-set logical rules.
[0027] Furthermore, the step of correcting the current working condition based on the working condition type identification result includes:
[0028] If it is identified that the current working condition has changed to the first working condition but the planned working condition is not the first working condition, then the time difference between the current working condition and the planned working condition is calculated, and the first working condition and its subsequent working conditions are brought forward.
[0029] If it is detected that the current working condition has not changed to the first working condition but the planned working condition is the first working condition, then the time difference between the current working condition and the planned working condition is calculated, and the first working condition and its subsequent working conditions are delayed.
[0030] Furthermore, based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, an energy change prediction curve is constructed, including:
[0031] Based on the operating condition distribution information, time nodes are defined, and different operating conditions are matched to the corresponding time periods according to the time sequence;
[0032] The energy change fluctuation function and energy fluctuation constant corresponding to each time period are integrated and calculated to construct the overall energy change prediction curve.
[0033] On the other hand, embodiments of the present invention provide an energy prediction device based on changes in operating conditions, including: a data acquisition module, a data processing module, and a model prediction module;
[0034] The data acquisition module is used to collect energy data from multiple energy devices;
[0035] The data processing module is used to process the energy data and form an energy database;
[0036] The model prediction module specifically includes:
[0037] The operating condition screening unit is used to screen and obtain a set of normal operating conditions and a set of important operating conditions based on the energy database.
[0038] The operating condition identification unit is used to set an energy fluctuation constant for the set of normal operating conditions; construct an energy change fluctuation function for the set of important operating conditions; and obtain the current operating condition type identification result based on the real-time collected current energy data and the energy change fluctuation function.
[0039] The operating condition tracking unit is used to determine the discrimination variables and discrimination conditions based on the energy database; obtain the current operating condition type discrimination result based on the real-time collected current energy data, discrimination variables and discrimination conditions; compare the consistency between the operating condition type discrimination result and the operating condition type identification result, and determine the corrected operating condition distribution information based on the comparison result;
[0040] The energy prediction unit is used to construct an energy change prediction curve based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, and to obtain the predicted flow rate of the energy medium in each energy device.
[0041] Furthermore, the energy prediction device also includes:
[0042] The terminal analysis module specifically includes:
[0043] A data processing setting unit is used to set the operation type and cleaning method of the energy data;
[0044] The operating condition definition unit is used to input newly created equipment operating condition information, discrimination variable information, and energy change fluctuation curve information, and to set the logical rules for operating condition identification.
[0045] The prediction execution unit is used to invoke the energy change prediction curve and calculate the predicted flow rate of the energy medium in real time based on user input.
[0046] The interface display unit is used to generate a Gantt chart based on the operating condition distribution information and to display the predicted flow rate of the energy medium after the prediction is completed.
[0047] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:
[0048] First, unlike related technologies which have limited energy medium types and cannot achieve multi-energy allocation, this invention integrates data resources of different energy mediums by collecting data from multiple energy devices and forming an energy database. This not only provides fluctuation prediction results for multiple energy mediums, but also provides dispatchers with accurate and reliable reference information on energy balance status.
[0049] Secondly, unlike related technologies which suffer from strong data dependence, severe error accumulation, and low prediction accuracy, this invention uses operating condition information as an important basis for energy prediction. Based on processes such as operating condition screening, energy curve identification, operating condition tracking, and prediction, an energy prediction system framework is built, which reduces data dependence, minimizes error accumulation, and improves prediction accuracy, thereby providing an important basis for the scheduling of different energy media.
[0050] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description
[0051] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.
[0052] Figure 1 This is a flowchart of an energy prediction method based on changes in operating conditions according to an embodiment of the present invention;
[0053] Figure 2 This is a schematic diagram illustrating the correction of operating condition delay in an embodiment of the present invention;
[0054] Figure 3 This is a schematic diagram illustrating the correction of the working condition in advance according to an embodiment of the present invention;
[0055] Figure 4 This is a schematic diagram illustrating energy curve prediction according to an embodiment of the present invention;
[0056] Figure 5 This is a schematic diagram of the main modules of the energy prediction device based on changes in operating conditions according to an embodiment of the present invention;
[0057] Figure 6 This is a schematic diagram showing the Gantt chart and prediction results of an embodiment of the present invention. Detailed Implementation
[0058] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.
[0059] A specific embodiment of the present invention discloses an energy prediction method based on changes in operating conditions, such as... Figure 1 As shown, the steps S1 to S5 are as follows:
[0060] Step S1: Collect energy data from multiple energy devices and form an energy database.
[0061] Multi-energy equipment is used for the conversion, supply, or consumption of energy media in the steel production process, including but not limited to gas equipment, steam equipment, power generation and transmission equipment, compressed air equipment, and technical gas equipment. Implementation requires first collecting energy data from multiple energy devices, then cleaning and processing the energy data, and finally classifying and storing the cleaned energy data using a TDengine super table to form the energy database. The energy data includes energy production and consumption data, equipment operation data, and production and maintenance plans. The data stored in the energy database has the same dimensions as the energy data, meaning the energy database includes at least: energy production and consumption data, equipment operation data, and production and maintenance plan data for each energy device. This provides a data foundation for subsequent collaborative optimization and precise allocation of multiple energy media.
[0062] Specifically, energy production and consumption data should include at least the flow data of media such as coal gas, steam, electricity, compressed air, and technical gases (such as oxygen, nitrogen, and argon); equipment operation data should include at least the operation data that can identify the working status of energy equipment, such as the hot air pressure and flow rate of blast furnace equipment, and the instantaneous flow rate of green pellets and the temperature of the roasting hood; production and maintenance plan data are business data, which should include at least production plans and maintenance plans, such as the blowing plan of converters.
[0063] Data can be collected through IoT gateways, temperature sensors, current sensors, PLC programmable logic controllers, and API interfaces of energy management systems. The data collection frequency may vary due to the influence of data characteristics, instruments, and prediction granularity.
[0064] Data processing can be divided into two parts: data cleaning and data storage. First, data cleaning includes, but is not limited to, filtering outliers, standardizing formats, using interpolation to fill in missing data, and using filtering algorithms to remove noise. Second, the processed data will be sent to an energy database for categorized storage. For example, a time-series database, TDengine (TSDB), will be used as the core storage engine, and deep optimizations will be performed to address the inherent characteristics of data in energy forecasting scenarios, such as massive volume, time series, and multi-source heterogeneity.
[0065] Compared with traditional relational databases, the embodiments of the present invention can provide an order of magnitude improvement in write and query performance and significantly reduce storage costs, thereby ensuring the real-time performance and economy of the energy forecasting system when processing large-scale, high-frequency operating data.
[0066] Preferably, to efficiently manage data generated by multiple industrial energy devices, this embodiment uses TDengine's SuperTable (STable) for database modeling. The SuperTable creates a standardized and unified data structure template (Schema) for a specific type of energy device, including collected metrics (such as pressure, flow rate, and temperature) and static tags (such as device ID and installation location). It automatically creates corresponding sub-tables for each device within that category, and these sub-tables inherit the SuperTable's data structure and can generate unique tag values. Thus, the cleaned energy production and consumption data, equipment operation data, and business data are stored in the energy database, enabling data retrieval in subsequent stages.
[0067] For example, to measure the production volume of coke oven gas, blast furnace gas, and converter gas, a supertable with timestamps and flow columns can be used as a template to quickly create three corresponding sub-tables and store the flow data for coke oven gas, blast furnace gas, and converter gas. A unique label can be defined for each energy medium to improve data query efficiency. It is evident that the database model built using supertables is suitable for scenarios requiring grouping, filtering, and aggregation analysis based on equipment attributes, and offers significant advantages in management and querying.
[0068] Step S2: Based on the energy database, a set of normal operating conditions and a set of important operating conditions are obtained by filtering.
[0069] Specifically, this includes: based on the production and maintenance plan data and the energy production and consumption data, determining the changes in energy medium fluctuations under each operating condition; when the changes in energy medium fluctuations are unstable, classifying the current operating condition into the set of important operating conditions; otherwise, classifying the current operating condition into the set of normal operating conditions.
[0070] Energy fluctuations exhibit different patterns under different operating conditions. Under normal operating conditions, energy fluctuations are relatively stable; however, the distribution of important operating conditions has a significant impact on energy fluctuations. Therefore, during the operating condition screening phase, operating conditions are screened according to the magnitude of the impact of energy fluctuations. The purpose is to identify non-routine operating conditions that require special attention and to establish energy change fluctuation curves for all operating conditions.
[0071] During implementation, the screening of important operating conditions needs to be based on key parameters. These key parameters are determined based on the production and maintenance plan data and energy production and consumption data in the aforementioned energy database. The production and maintenance plan data can be used to determine the time period in which the equipment operates under a certain condition. Then, by combining the energy production and consumption data to extract the energy medium fluctuation amount in that time period, different operating conditions can be screened and distinguished.
[0072] For example, for converter equipment, statistics can be collected based on two key parameters: the blowing plan and the converter gas recovery flow rate during the blowing period. Based on this, the fluctuation of the gas recovery flow rate during the period can be judged. If the fluctuation is stable, it is considered a normal operating condition; otherwise, it is considered an important operating condition. The standard for whether the fluctuation is stable is whether it is necessary to reduce the impact of the fluctuation on the energy system through scheduling means such as adjusting the generator set. If the fluctuation is small and no scheduling means are needed, it is considered a normal operating condition.
[0073] Furthermore, various operating conditions can be categorized to ultimately obtain a set of important operating conditions. Common operating conditions in This indicates the k-th critical operating condition of the j-th device for the i-th energy medium. This represents the k-th normal operating condition of the j-th device for the i-th energy medium.
[0074] Step S3: Set an energy fluctuation constant for the set of normal operating conditions; construct an energy change fluctuation function for the set of important operating conditions; and obtain the current operating condition type identification result based on the real-time collected current energy data and the energy change fluctuation function.
[0075] Specifically, a clustering algorithm is used to identify the characteristics of energy medium fluctuations under each operating condition, and the energy fluctuation curves under each operating condition are extracted. Based on the operating condition screening results, the energy fluctuation curves with zero fluctuation corresponding to each normal operating condition are set as energy fluctuation constants. An energy fluctuation function is constructed based on the energy fluctuation curves corresponding to each important operating condition. The energy fluctuation curves obtained from the real-time collected current energy data are compared with the energy fluctuation curves under each operating condition, and the operating condition cluster with the closest target distance is selected as the current operating condition type identification result.
[0076] During implementation, the operating condition identification stage mainly includes two parts: setting the energy medium change fluctuation curve under different operating conditions and identifying the energy change fluctuation curve.
[0077] For setting the energy change fluctuation curve, historical energy data from the energy database is read, and a clustering algorithm is used to identify the characteristics of energy medium changes under multiple operating conditions, resulting in the energy change fluctuation curve for each operating condition. It can be understood that the energy change fluctuation curve is based on the constant value of the conventional operating condition. Then, based on the screening results of the aforementioned conventional and important operating conditions, the energy change fluctuation function for the important operating condition is constructed, denoted as... At the same time, the energy fluctuations corresponding to normal operating conditions are set as constants, denoted as... That is, the change is 0; where, Let f(j) represent the energy change fluctuation function of the j-th device for the k-th key operating condition of the i-th energy medium; This represents the energy fluctuation constant of the j-th device under the k-th normal operating condition for the i-th energy medium; This indicates that the energy fluctuation of the j-th device for the i-th energy medium is stable at the same reference constant under normal operating conditions.
[0078] For the identification of operating condition energy curves, the identification results of the operating condition type can be used in the subsequent operating condition tracking stage. The current energy change fluctuation curve is obtained by collecting the current energy data in real time, and the curve is compared with the multi-operating condition energy change fluctuation curve obtained by the aforementioned clustering method and the similarity is calculated. The smaller the distance, the higher the similarity. Thus, the operating condition cluster with the closest target distance is selected as the identified operating condition type.
[0079] Step S4: Determine the discriminant variables and discriminant conditions based on the energy database; obtain the current operating condition type discrimination result based on the real-time collected current energy data, discriminant variables, and discriminant conditions; compare the consistency between the operating condition type discrimination result and the operating condition type identification result, and determine the corrected operating condition distribution information based on the comparison result.
[0080] Specifically, the discriminant variables and the discriminant conditions are determined based on the equipment operation data, and a set of discriminant variables for the corresponding operating condition is obtained based on the discriminant variables; based on the set of discriminant variables and the discriminant conditions, the real-time collected current energy data is discriminated to obtain the current operating condition type discrimination result.
[0081] Since the actual working conditions on the production site are dynamic, in order to obtain more accurate working condition information, it is necessary to identify changes in working conditions based on the current production situation and adjust the working condition information in the production plan in real time.
[0082] Therefore, during the operating condition tracking phase, the operating condition information generated based on production and maintenance plans can be adjusted and corrected in real time, and the energy change curve can be directly identified to determine the operating condition type. Specifically, based on the equipment operation data in the energy database, and according to process or expert knowledge, the discrimination variables and discrimination conditions of the energy equipment are determined. Then, the discrimination results and the energy curve identification results are combined to determine the actual operating condition of each piece of equipment at the current moment.
[0083] For example, blast furnace operating conditions include shutdown, reduced blast, furnace replacement, and normal operation; sintering operating conditions include shutdown for maintenance and normal operation. The discriminant variables for blast furnace can be set as hot blast pressure, hot / cold blast flow rate, etc., while the discriminant variables for sintering can be set as ignition gas, sintering machine speed, etc. For another example, for the reduced blast condition of a blast furnace, the discriminant variables can be set as blast furnace blast pressure and hot blast flow rate, with the discrimination conditions set as blast furnace blast pressure less than 380 kPa and greater than 230 kPa, and hot blast flow rate less than 310,000 cubic meters / hour and greater than 150,000 cubic meters / hour. Based on this, the discrimination method or conflict resolution logic of the discrimination method can be set when identifying the operating condition.
[0084] When using a discriminant variable for judgment, the discriminant variable can be set as κ, and the set of m corresponding working conditions can be denoted as {K1,...,K}. m Meanwhile, the set to which the discriminant variables for the corresponding working conditions belong is denoted as {Ω1,...,Ω}. m This determines the current operating condition K, and based on the real-time collected energy data, the discrimination variable, and the discrimination condition, the current operating condition type is determined. The specific judgment formula is as follows:
[0085]
[0086] Next, the consistency between the working condition type discrimination result and the working condition type identification result is compared: if they are consistent, the current working condition is corrected based on the working condition type identification result, and the corrected working condition distribution information is determined; otherwise, the corrected working condition distribution information is determined according to the pre-set logical rules. That is, the aforementioned energy change fluctuation function is called to judge the current working condition. If the judgment result is consistent with the judgment result obtained based on the discriminant variable method, the working condition is corrected; otherwise, the final working condition type needs to be determined according to the user's pre-set custom rule logic. For example, when a discriminant variable or condition for a certain working condition is missing, the identification result of the energy change fluctuation curve is used.
[0087] Preferably, correcting the current working condition based on the working condition type identification result includes:
[0088] If it is identified that the current working condition has changed to the first working condition but the planned working condition is not the first working condition, then the time difference Δt between the current working condition and the planned working condition is calculated, and the first working condition and its subsequent working conditions are advanced by Δt.
[0089] If it is detected that the current working condition has not changed to the first working condition but the planned working condition is the first working condition, then the time difference Δt between the current working condition and the planned working condition is calculated, and the first working condition and its subsequent working conditions are delayed by Δt.
[0090] Combination Figure 2 and Figure 3 As shown, when comparing the distribution of operating conditions, if the change in operating condition K is detected to be K... i Then search for the nearest plan K among all plans. i Calculate K under actual working conditions i The time difference Δt between the occurrence time and the planned occurrence time of the working condition is used to advance or delay the working condition and all subsequent plans by Δt, thereby realizing the automatic identification and correction of the working condition, and finally obtaining the working condition distribution that conforms to the actual situation. It can be seen that this embodiment not only utilizes real-time production information on site, but also reduces the energy prediction error introduced by production plan deviation.
[0091] For example, if it is detected that the current equipment has changed to the first operating condition but the operating condition was originally scheduled to start in 10 minutes, then the first operating condition is brought forward by 10 minutes, and all subsequent operating conditions are brought forward by 10 minutes in sequence; if it is detected that the first operating condition has not yet started, but the current operating condition was originally scheduled to start 10 minutes ago, then the operating condition and subsequent operating conditions are postponed by 10 minutes in sequence.
[0092] Step S5: Based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, construct the energy change prediction curve and obtain the predicted flow rate of the energy medium in each energy device.
[0093] Specifically, this includes: defining time nodes based on the operating condition distribution information, matching different operating conditions to corresponding time periods in chronological order; integrating and calculating the energy change fluctuation function and energy fluctuation constant corresponding to each time period to construct an overall energy change prediction curve, and calling the energy change prediction curve based on user input data and outputting the energy prediction value in real time.
[0094] For example, when predicting changes in the surplus of three types of coal gas in the steel industry, the energy prediction method of this invention can quickly determine the allocation of multiple energy devices based on data such as the gas consumption of the main process, the gas demand of energy units, and the status of gas holders. The surplus coal gas can be supplied to the generator set for power generation, while compensating for possible future changes, ensuring the balance between coal gas supply and demand, reducing coal gas emissions, and providing an important basis for energy dispatch.
[0095] Preferably, the energy forecasting stage forecasts the changes in energy production and consumption, rather than directly forecasting the actual energy production and consumption. The energy change forecast curve, as an important reference for energy dispatching methods, can be constructed according to the following process:
[0096] First, based on the operating condition distribution information, the multi-time period operating condition combination information within the production cycle is sorted in chronological order, with each two adjacent time points defined as a time period. This allows all operating conditions to be divided into multiple time nodes {t1, t2, ..., t...} in chronological order. m}, so that each time period {t l ,t l+1 Each item contains several key operating conditions and routine operating conditions for different energy equipment.
[0097] Secondly, for each time period {t} l ,t l+1 Calculate the energy change fluctuation function for different media i. The specific formula is as follows:
[0098]
[0099] Among them, {t l ,t l+1} represents multiple time periods obtained by sorting based on working condition distribution information; This represents the prediction function for the fluctuation curve of the change in the j-th device with respect to the i-th medium; Let X represent the energy change fluctuation function of the j-th device for the k-th key operating condition of the i-th energy medium; X represents the current operating condition type. This represents a set of important operating conditions; This represents the set of normal operating conditions.
[0100] Finally, by integrating and calculating the energy change fluctuation function and energy fluctuation constant corresponding to each time period, an overall energy change prediction curve can be constructed. The specific formula is as follows:
[0101]
[0102] F i (t), t=[t1,t m (5)
[0103] Among them, F i This represents the energy change fluctuation function of medium i obtained by summarizing m-1 time periods; This represents the energy change fluctuation function of medium i during the time period [t1,t2]. The function representing the fluctuation of energy change in medium i during the time interval [t2, t3]; representing [t] m-1 ,t m The energy change fluctuation function of medium i during time period; F i (t) represents the overall energy change prediction curve of medium i obtained by splicing and integrating; [t1,t] m ] represents t1 to t m The overall time period.
[0104] In some implementations, such as Figure 4 As shown, if we want to directly obtain the prediction results of the energy fluctuation curve, we can make adjustments while ensuring that the above ideas remain unchanged for the sake of easy calculation.
[0105] First, the energy fluctuation constants under normal operating conditions are selected and summed to obtain a fluctuation benchmark value for energy medium i under normal operating conditions, i.e. Secondly, based on the energy fluctuation baseline, the energy change fluctuation function corresponding to the key operating conditions of all equipment appearing in the Gantt chart is accumulated to the baseline value. The above yields the energy change prediction function F over the entire time period. i (t). That is, the aforementioned time period division and splicing process of the fluctuation curves of various energy changes are eliminated, thus improving the calculation efficiency. Figure 4 To facilitate description, all key operating conditions are aligned, and the curve for each condition is a schematic diagram of the set energy change fluctuation. When superimposing energy curves, the upper left corner of the operating condition block is used as the reference, and the curve is superimposed on the baseline curve of the energy prediction chart below. At the same time, the superimposed energy curves are smoothed to a certain extent during abrupt change periods.
[0106] Therefore, the embodiments of the present invention, on the one hand, integrate data resources of different energy media by collecting data from multiple energy devices and forming an energy database. This not only provides fluctuation prediction results for various energy media, but also provides dispatchers with accurate and reliable reference information on energy balance status. On the other hand, by using operating condition information as an important basis for energy prediction, and based on the processes of operating condition screening, identification, tracking, and prediction, an energy prediction system framework is built, which reduces data dependence, reduces error accumulation, and improves prediction accuracy, thereby providing an important basis for the dispatch of different energy media.
[0107] In another embodiment of the present invention, an energy prediction device based on changes in operating conditions is proposed, such as... Figure 5 As shown, it specifically includes a data acquisition module, a data processing module, and a model prediction module. The model prediction module further includes a working condition screening unit, a working condition identification unit, a working condition tracking unit, and an energy prediction unit, wherein:
[0108] The system includes: a data acquisition module for acquiring energy data from multiple energy devices; a data processing module for processing the energy data and forming an energy database; a working condition screening unit for filtering a set of regular working conditions and a set of important working conditions based on the energy database; a working condition identification unit for setting an energy fluctuation constant for the set of regular working conditions; constructing an energy change fluctuation function for the set of important working conditions; obtaining the current working condition type identification result based on the real-time acquired current energy data and the energy change fluctuation function; a working condition tracking unit for determining discriminant variables and discriminant conditions based on the energy database; obtaining the current working condition type discrimination result based on the real-time acquired current energy data, discriminant variables, and discrimination conditions; comparing the consistency between the working condition type discrimination result and the working condition type identification result, and determining the corrected working condition distribution information based on the comparison result; and an energy prediction unit for constructing an energy change prediction curve based on the working condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, and obtaining the predicted flow rate value of the energy medium in each energy device.
[0109] Preferably, the energy prediction device further includes: a terminal analysis module, which specifically includes: a data processing setting unit for setting the operation type and cleaning method of the energy data; a working condition definition unit for inputting newly created equipment working condition information, discrimination variable information, and energy change fluctuation curve information, and setting the logical rules for working condition identification; a prediction execution unit for calling the energy change prediction curve and calculating the predicted flow value of the energy medium in real time according to user input; and an interface display unit for generating a Gantt chart based on the working condition distribution information and displaying the predicted flow value of the energy medium after the prediction execution is completed.
[0110] Meanwhile, this energy forecasting device can integrate IoT protocols and gateways, time-series databases, data preprocessing, machine learning algorithms, and web-based visualization technologies, depending on the specific application scenario.
[0111] In some implementations, the terminal analysis module includes at least data processing operation options, operating condition definition, prediction execution, and interface display functions, so that users can perform relevant operations according to specific usage scenarios.
[0112] The data processing operation options allow users to select the type of data processing operation, including but not limited to filtering outliers, format standardization, interpolation, and filtering. The operating condition definition function allows users to manually input newly created equipment operating condition information, discrimination variable information, and energy fluctuation curve information; simultaneously, it can automatically call and obtain the aforementioned energy change fluctuation function. The prediction model call function allows users to select energy medium type, prediction time range, and whether to adjust operating conditions. Following the steps of organizing energy medium operating conditions, adjusting operating conditions, dividing the prediction range, and calling the specific algorithms in the aforementioned model prediction module, it ultimately calculates the required energy medium flow prediction value based on the energy change prediction curve.
[0113] Preferably, refer to Figure 6 As shown, this terminal analysis module can intuitively display the operating condition distribution information in the form of a Gantt chart. Figure 6 Different operating conditions are represented by rectangular blocks, with the start and end points of the blocks being the start and end points of the corresponding operating conditions. Rectangular blocks of different shades represent different operating condition types. After the user triggers the prediction execution function, the terminal analysis model can call the energy change prediction curve and the corrected operating condition distribution information from the model prediction module, thereby calculating and outputting the flow rate value of the energy medium in each energy device. Then, a Gantt chart is displayed on the terminal display interface, and the prediction results of the energy flow rate value can be intuitively displayed below the Gantt chart.
[0114] It is understood that the formulas and parameters, the framework and function settings of each module in the above embodiments are only for ease of understanding and simplification of description, and should not be construed as limitations on the present invention.
[0115] Therefore, the energy forecasting method and apparatus in this embodiment of the invention, as an upstream link in the energy dispatching of steel enterprises, can provide information on the fluctuation of various energy media for the energy dispatching of steel production, effectively improve the accuracy and robustness of energy forecasting, enhance the rationality and accuracy of dispatching operations, further ensure the balance of the energy system, reduce energy dissipation and lower production costs.
[0116] The above-described method and apparatus embodiments are based on the same principle, and their related aspects can be referenced from each other to achieve the same technical effect. For specific implementation processes, please refer to the foregoing embodiments, which will not be repeated here.
[0117] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.
[0118] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.
Claims
1. An energy forecasting method based on changes in operating conditions, characterized in that, include: Collect energy data from multiple energy devices and create an energy database; Based on the energy database, a set of normal operating conditions and a set of important operating conditions are obtained by filtering. Set an energy fluctuation constant for the set of normal operating conditions; construct an energy change fluctuation function for the set of important operating conditions; obtain the current operating condition type identification result based on the real-time collected current energy data and the energy change fluctuation function; Based on the energy database, determine the discriminant variables and discriminant conditions; Based on the real-time collected current energy data, discrimination variables, and discrimination conditions, the current operating condition type discrimination result is obtained; The consistency between the working condition type discrimination result and the working condition type identification result is compared, and the corrected working condition distribution information is determined based on the comparison result; Based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, an energy change prediction curve is constructed, and the flow prediction value of the energy medium in each energy device is obtained.
2. The energy forecasting method according to claim 1, characterized in that, The process of collecting energy data from multiple energy devices and forming an energy database includes: The energy data is cleaned and then categorized and stored using TDengine super tables to form the energy database. The energy database includes at least: energy production and consumption data of each energy device, equipment operation data, and production and maintenance plan data.
3. The energy forecasting method according to claim 2, characterized in that, Based on the energy database, the set of normal operating conditions and the set of important operating conditions are obtained by filtering, including: Based on the production and maintenance plan data and the energy production and consumption data, determine the changes in energy medium fluctuations under each operating condition: When the fluctuation of the energy medium is unstable, the current operating condition is classified into the set of important operating conditions; otherwise, the current operating condition is classified into the set of normal operating conditions.
4. The energy forecasting method according to claim 3, characterized in that, The first step is to set an energy fluctuation constant for the set of normal operating conditions; and to construct an energy change fluctuation function for the set of important operating conditions. Based on the real-time collected current energy data and the energy change fluctuation function, the current operating condition type identification result is obtained, including: Clustering algorithms are used to identify the characteristics of energy medium fluctuations under various operating conditions, and energy fluctuation curves under various operating conditions are extracted. Based on the operating condition screening results, the energy change fluctuation curves corresponding to each normal operating condition with zero change are set as energy fluctuation constants; and an energy change fluctuation function is constructed based on the energy change fluctuation curves corresponding to each important operating condition. The energy change fluctuation curve obtained from the real-time collected current energy data is compared with the energy change fluctuation curve under each operating condition, and the operating condition cluster with the closest target distance is selected as the current operating condition type identification result.
5. The energy forecasting method according to claim 4, characterized in that, The discriminant variables and discriminant conditions are determined based on the energy database; Based on the real-time collected current energy data, discrimination variables, and discrimination conditions, the current operating condition type discrimination result is obtained, including: The discriminant variables and the discriminant conditions are determined based on the equipment operation data, and a set of discriminant variables for the corresponding working conditions is obtained based on the discriminant variables; Based on the set of discriminant variables and the discriminant conditions, the real-time collected current energy data is discriminated to obtain the current operating condition type discrimination result.
6. The energy forecasting method according to claim 5, characterized in that, The step of comparing the consistency between the working condition type discrimination result and the working condition type identification result, and determining the corrected working condition distribution information based on the comparison result, includes: If the comparison is consistent, the current working condition is corrected based on the working condition type identification result, and the corrected working condition distribution information is determined. Otherwise, the corrected working condition distribution information is determined according to the pre-set logical rules.
7. The energy forecasting method according to claim 6, characterized in that, The step of correcting the current operating condition based on the operating condition type identification result includes: If it is identified that the current working condition has changed to the first working condition but the planned working condition is not the first working condition, then the time difference Δt between the current working condition and the planned working condition is calculated, and the first working condition and its subsequent working conditions are advanced by Δt. If it is detected that the current working condition has not changed to the first working condition but the planned working condition is the first working condition, then the time difference Δt between the current working condition and the planned working condition is calculated, and the first working condition and its subsequent working conditions are delayed by Δt.
8. The energy forecasting method according to claim 1, characterized in that, Based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, an energy change prediction curve is constructed, including: Based on the operating condition distribution information, time nodes are defined, and different operating conditions are matched to the corresponding time periods according to the time sequence; The energy change fluctuation function and energy fluctuation constant corresponding to each time period are integrated and calculated to construct the overall energy change prediction curve.
9. An energy prediction device based on changes in operating conditions, characterized in that, include: Data acquisition module, data processing module, and model prediction module; The data acquisition module is used to collect energy data from multiple energy devices; The data processing module is used to process the energy data and form an energy database; The model prediction module specifically includes: The operating condition screening unit is used to screen and obtain a set of normal operating conditions and a set of important operating conditions based on the energy database. The operating condition identification unit is used to set an energy fluctuation constant for the set of normal operating conditions; construct an energy change fluctuation function for the set of important operating conditions; and obtain the current operating condition type identification result based on the real-time collected current energy data and the energy change fluctuation function. The operating condition tracking unit is used to determine the discrimination variables and discrimination conditions based on the energy database; obtain the current operating condition type discrimination result based on the real-time collected current energy data, discrimination variables and discrimination conditions; compare the consistency between the operating condition type discrimination result and the operating condition type identification result, and determine the corrected operating condition distribution information based on the comparison result; The energy prediction unit is used to construct an energy change prediction curve based on the operating condition distribution information, the energy change fluctuation function, and the energy fluctuation constant, and to obtain the predicted flow rate of the energy medium in each energy device.
10. The energy prediction device according to claim 9, characterized in that, Also includes: The terminal analysis module specifically includes: A data processing setting unit is used to set the operation type and cleaning method of the energy data; The operating condition definition unit is used to input newly created equipment operating condition information, discrimination variable information, and energy change fluctuation curve information, and to set the logical rules for operating condition identification. The prediction execution unit is used to invoke the energy change prediction curve and calculate the predicted flow rate of the energy medium in real time based on user input. The interface display unit is used to generate a Gantt chart based on the operating condition distribution information and to display the predicted flow rate of the energy medium after the prediction is completed.