Dynamic liquid level control method based on water volume prediction and expert rule embedding

Abstract

The invention discloses a dynamic liquid level control method based on water volume prediction and expert rule embedding, and belongs to the technical field of water volume control. According to the method, the problems of extensive control of the liquid level of the clean water tank, conflict of man-machine control strategies and unnatural time interval connection in an existing method are solved, an appropriate liquid level control interval is matched according to daily water volume prediction grading, and it can be ensured that the liquid level of the clean water tank runs within a safe and energy-saving range in the water consumption peak period; the liquid level control interval can be dynamically corrected according to manual requirements, and through man-machine collaborative optimization, the manual interval is forced to cover and guarantee sudden requirements, equipment impact is avoided, and safe and stable water supply is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water quantity control, in particular to a dynamic liquid level control method based on water quantity prediction and expert rule embedding. BACKGROUND

[0002] In the urban water supply system, the clear water pool as the key buffer unit of water production and supply, its liquid level control directly determines the stability of the water supply system, energy consumption efficiency and equipment life. The traditional water supply scheduling focuses on pipe network pressure regulation or pump start-stop optimization, while the dynamic interval setting of the clear water pool liquid level is often simplified as fixed threshold control (such as setting high / low liquid level alarm limit). However, the clear water pool liquid level needs to respond to the fluctuation of water supply quantity and the change of user demand in real time. The fixed threshold cannot adapt to the instantaneous surge of water supply during peak period or the redundant water storage during low period, resulting in frequent triggering of limit alarm of liquid level. At the same time, the existing method lacks quantitative analysis of the correlation between historical liquid level and unit consumption, and it is difficult to achieve energy efficiency optimization through dynamic interval adjustment. In actual operation, when encountering emergencies, it is necessary to rely on experience to set the liquid level interval to ensure the safety of water supply, and the interval generated according to the data analysis algorithm may not be compatible with it, resulting in liquid level mutation at the time period junction. Such discontinuity easily causes frequent start-stop of equipment, and causes confusion of control logic.

[0003] Therefore, the existing needs are not met, and for this we propose a dynamic liquid level control method based on water quantity prediction and expert rule embedding. SUMMARY

[0004] The purpose of the present application is to provide a dynamic liquid level control method based on water quantity prediction and expert rule embedding, which dynamically generates a liquid level interval by combining the results of a water quantity prediction model, using historical daily water quantity, operating liquid level and unit consumption data, and embeds expert rules to achieve precise control of the clear water pool liquid level, which can realize unit consumption optimization under the premise of ensuring control continuity, and solves the problems raised in the above background art.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a dynamic liquid level control method based on water quantity prediction and expert rule embedding, comprising the following steps:

[0006] Water quantity grading and static interval construction, including data processing, energy-saving sample screening and static reference liquid level interval calculation;

[0007] Water quantity grading matches the reference liquid level interval, by obtaining the predicted daily water supply quantity provided by the external water quantity prediction system , matching the water quantity level , and loading the corresponding static interval;

[0008] Expert rule embedding, by presetting the artificial intervention period, and manually inputting the set liquid level interval within the time period, forming the control interval of time period-interval, and then embedding the control system as a man-machine collaborative decision;

[0009] Dynamic interval generation, by extracting all the transition zones 3 hours before and after the manual adjustment of the interval boundary, and adjusting the weight of the transition zone liquid level control instruction;

[0010] Output dynamic liquid level control instruction to the actuator to realize dynamic control of liquid level.

[0011] Further, data processing, including:

[0012] Statistical historical daily water supply , daily single consumption and daily hourly liquid level average value, and divide the historical daily water supply into grades, wherein the daily water supply is the collected data, and the daily single consumption is the calculated value.

[0013] Further, the calculation formula of daily single consumption is as follows:

[0014]

[0015] In the formula, is the daily single consumption, the unit is , that is, the average energy consumption per cubic meter of water produced, and the lower the energy consumption value represents the higher the energy efficiency.

[0016] Further, energy-saving sample screening, including:

[0017] For each grade , select the top 20% of daily single consumption as sample days, that is, in the same water volume grade, the top 20% of daily single consumption data represents the optimal energy-saving operation record of the water volume interval.

[0018] Further, the specific method of screening is:

[0019] Screen the data set in the same grade according to daily water supply;

[0020] Sort all sample days in the grade according to single consumption value from low to high;

[0021] Select the sample days ranked between 0~20% quantile points.

[0022] Further, static reference liquid level interval calculation, including:

[0023] Calculate the hourly liquid level average value of each energy-saving sample set The median and interquartile range of the liquid level of the energy-saving sample are calculated to generate the hourly static reference liquid level interval.

[0024] Further, the calculation formula of the hourly static reference liquid level interval is as follows:

[0025]

[0026] In the formula, are the ranks The hourly static liquid level interval lower limit and upper limit; is the liquid level interquartile range of the energy-saving sample; is the median of the liquid level of the energy-saving sample, that is, the sample daily data ranked between the 0~20% percentile points The hourly liquid level average value.

[0027] Further, in the step of generating a dynamic interval, the dynamic weight The Sigmoid function is adopted, and the specific formula is as follows:

[0028]

[0029] In the formula, is the boundary distance from the current time to the artificial interval, in hours; is the maximum transition distance, which is 3 hours; is the slope coefficient, The larger the value is, the steeper the transition of the function near the threshold value is; is the offset coefficient, which is used to adjust the center point of the function; when tends to 1, it indicates that the time period is far away from the artificial interval, and the static reference is completely adopted; is the dynamic weight, which is adjusted according to the boundary distance from the current time to the artificial interval, and the parameter is dynamically adjusted; when tends to 0, it indicates the artificial interval, and the artificial set value is forced to be matched.

[0030] Further, in the step of generating a dynamic interval, the dynamic center value of the transition zone is as follows:

[0031]

[0032] In the formula, is the dynamic center value of the transition zone; is the algorithm weight; are the ranks The hourly static liquid level interval lower limit and upper limit; , are the global minimum and global maximum of the transition zone respectively.

[0033] Further, the state interval is generated as shown in the following formula:

[0034]

[0035] In the formula, are the upper and lower limits of the dynamic interval; is the dynamic center value of the transition zone; is the original static liquid level interval half-width, .

[0036] Compared with the prior art, the beneficial effects of the present application are:

[0037] The present application matches the appropriate liquid level control interval according to the daily water quantity prediction, which can ensure that the liquid level of the clean water pool runs in a safe and energy-saving range during the water use peak period. In addition, according to the artificial needs, the liquid level control interval can be dynamically corrected, and through the man-machine collaborative optimization, the artificial interval is forced to cover the protection of the sudden demand, avoiding the impact of the equipment, and ensuring the safety and stability of water supply. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is the flowchart of the dynamic liquid level control method based on water quantity prediction and expert rule embedding of the present application. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0040] In order to solve the technical problems that the fixed threshold value in the existing liquid level control method cannot adapt to the fluctuation of water supply quantity, the man-machine control strategy conflicts, and the time period connection is unnatural, please refer to Figure 1 The technical solutions of the present embodiment are as follows:

[0041] The dynamic liquid level control method based on water quantity prediction and expert rule embedding comprises the following steps:

[0042] Water quantity grading and static interval construction, including data processing, energy-saving sample screening and static reference liquid level interval calculation;

[0043] Water quantity grading matches the reference liquid level interval, by acquiring the predicted daily water supply quantity provided by the external water quantity prediction system , matching the water quantity grade and load the corresponding static interval;

[0044] Expert rules are embedded by setting a manual intervention period and inputting the set liquid level interval during the period to form a control interval of period-interval, which is then embedded into the control system as a man-machine collaborative decision;

[0045] Dynamic interval generation is achieved by extracting 3 hours before and after the manual adjustment of interval boundaries as transition zones and adjusting the weight of the liquid level control instruction in the transition zone to achieve smooth transition of the liquid level. The liquid level interval input by the human being has absolute priority in a specific period to ensure water supply safety in emergency situations.

[0046] Output the dynamic liquid level control instruction to the actuator, wherein:

[0047] Reliable communication technology and protocols (such as industrial Ethernet, Profibus, etc.) are used to transmit dynamic liquid level control instructions. At the same time, communication redundancy mechanisms are established, such as dual-channel communication or communication backup devices, to prevent communication failures from causing control instructions to fail to transmit or transmit errors. In addition, the communication link is monitored and diagnosed in real time to discover and handle communication abnormalities in a timely manner, ensuring the reliability of control instruction transmission.

[0048] The actuator is a frequency converter, valve controller or motor driver. The dynamic liquid level control instruction is output to the actuator in the form of a digital signal or an analog signal. The actuator adjusts the operating parameters of the water supply equipment according to the received instruction to achieve dynamic control of the liquid level.

[0049] The technical effects of the above technical solutions are: first, through water quantity grading and static interval construction, the historical data can be accurately analyzed and processed, thereby screening out representative energy-saving samples, and thus providing a scientific and reasonable benchmark for subsequent liquid level control; second, by matching the benchmark liquid level interval with the water quantity grading, the predicted daily water supply amount provided by the external water quantity prediction system can be accurately matched with the corresponding water quantity grade, and the corresponding static interval is loaded, so that the water supply system can flexibly adjust the operating state according to the actual demand, thereby further optimizing the water supply effect and ensuring the stability and reliability of water supply; in addition, the embedded expert rules fully consider the importance of manual intervention, by presetting the manual intervention period and allowing manual input of the set liquid level interval to form a "period-interval" control interval, the embedded control system is used as a man-machine collaborative decision-making to ensure water supply safety and avoid water supply interruption or instability caused by system failure or abnormal conditions; at the same time, the dynamic interval generation extracts the 3 hours before and after the manual adjustment interval boundary as a transition zone, and adjusts the weight of the transition zone liquid level control instruction, which can realize the smooth transition of the liquid level and avoid the sharp fluctuation of the liquid level, further improving the stability and safety of the water supply system; finally, the dynamic liquid level control instruction is output to the actuator to realize accurate control of the water supply equipment, thereby ensuring the efficient, stable and safe operation of the entire water supply system.

[0050] In summary, the method has significant advantages in improving water supply efficiency, reducing energy consumption, enhancing water supply stability and safety, etc., and has important significance for improving the overall performance and management level of the water supply system.

[0051] Data processing includes:

[0052] Statistical historical daily water supply amount , daily specific energy consumption and hourly liquid level average value per day, and the historical daily water supply amount is divided into grades, wherein the daily water supply amount is the collected data, the daily specific energy consumption is a calculated value, and the calculation formula is as follows:

[0053]

[0054] In the formula, is the daily specific energy consumption, and the unit is , i.e. the average energy consumption per cubic meter of water produced, and the lower the energy consumption value represents the higher the energy efficiency.

[0055] As shown in Table 1, the grades are divided into six grades, and examples are as follows:

[0056] Rank Water supply (million tons) I II III IV V VI

[0057] The technical effects of the above technical solutions are: through the statistics and analysis of the historical daily water supply, daily unit consumption and daily hourly liquid level average value, the operation condition and energy efficiency level of the water supply system can be comprehensively mastered, the historical daily water supply is divided into different grades, which helps to more intuitively identify the change trend and abnormal situation of the water supply, and is convenient for targeted management and optimization, at the same time, the daily unit consumption as an important indicator of energy efficiency can clearly reflect the energy consumption in the water supply production process, provide data support for energy saving, and help to improve the overall energy efficiency of the water supply system.

[0058] Energy-saving sample screening, comprising:

[0059] For each grade , the daily unit consumption The top 20% of the sample day, that is, in the same water volume grade, the daily unit consumption The lowest 20% of the group data, which represents the optimal energy-saving operation record of the water volume interval;

[0060] The specific method of screening is:

[0061] Screening the data set of the same grade of daily water supply;

[0062] Sort all sample days in the grade by unit consumption value from low to high;

[0063] Select the sample day between the 0~20% quantile points.

[0064] The technical effects of the above technical solutions are: through the screening of energy-saving samples, the representative optimal energy-saving operation record can be accurately extracted from different water volume grades, ensuring the typicality and effectiveness of the samples, so that the selected sample day can truly reflect the best energy-saving state in each water volume interval, thereby helping to deeply tap the energy-saving potential and providing clear direction and reference basis for energy-saving optimization of the water supply system under different working conditions.

[0065] Static reference liquid level interval calculation, comprising:

[0066] Through calculating the median and interquartile range of the liquid level of each hour In the energy-saving sample set, the hourly static reference liquid level interval is generated, and the calculation formula is as follows:

[0067]

[0068] In the formula, The lower limit and upper limit of the hourly static liquid level interval are respectively The first Grade The quartile range of the liquid level of the energy-saving sample; The median of the liquid level of the energy-saving sample, i.e., the sample daily data ranked between the 0-20% percentile points The average value of the liquid level of the hour.

[0069] The technical effects of the above technical solutions are: by calculating the median and quartile range of the liquid level, the static reference liquid level interval generated thereby can effectively reflect the liquid level fluctuation range of each hour in the optimal energy-saving operation state, thereby helping to accurately control the liquid level and avoiding increased energy consumption or unstable operation caused by excessively high or low liquid level, and ensuring stable operation of the water supply system in an energy-saving and efficient state.

[0070] In summary, by accurately calculating the static reference liquid level interval of each hour based on the energy-saving sample set, a scientific and reasonable reference standard can be provided for the liquid level control of the water supply system.

[0071] In the step of water quantity grading matching reference liquid level interval, the prediction method used by the external water quantity prediction system is at least one of time series analysis method, machine learning algorithm or physical model simulation method, and before obtaining the predicted daily water supply quantity provided by the external water quantity prediction system, a plurality of different water quantity prediction systems (such as physical model-based prediction system, data-driven machine learning prediction system, etc.) should be evaluated and compared to select a water quantity prediction system with high prediction accuracy and strong reliability as the data source;

[0072] At the same time, a verification mechanism for the output data of the water quantity prediction system is established, the accuracy of the prediction result is evaluated by comparing with the actual monitoring data, and the prediction data is corrected if necessary, specifically:

[0073] Collect actual monitoring data within the corresponding time range of the output data of the water quantity prediction system, including the size of the water quantity, the change trend, etc.

[0074] Directly compare the prediction data output by the water quantity prediction system with the actual monitoring data;

[0075] Analyze whether the change trends of the prediction data and the actual monitoring data are consistent, for example, if the actual monitoring data shows that the water quantity is rising, but the prediction data shows that the water quantity is falling, then the prediction result may have a problem.

[0076] The technical effects of the above-mentioned technical solution are as follows: By comprehensively applying multiple water volume prediction methods and scientifically evaluating and comparing different prediction systems, a water volume prediction system with high prediction accuracy and high reliability can be selected as a data source, thereby providing an accurate and reliable basis for matching the benchmark liquid level range for water volume classification. At the same time, by establishing a verification mechanism, the predicted data can be compared with the actual monitoring data, which can not only effectively evaluate the accuracy of the prediction results, but also correct the prediction data when necessary, further improving the reliability of water volume prediction.

[0077] In the expert rule embedding step, a feedback mechanism for human-machine collaborative decision-making is established. The operational effects after each decision (such as level control accuracy, water supply stability, energy-saving effects, etc.) are fed back to the system and experts. The decision-making rules are continuously optimized by analyzing the feedback data. Specifically:

[0078] Data transmitted through various sensors (including but not limited to level sensors, water pressure sensors, and energy consumption monitoring equipment);

[0079] Statistical analysis was performed on the accuracy of liquid level control, the liquid level deviation within each decision cycle was analyzed, the rate of change of water supply pressure was analyzed, the standard deviation of water supply pressure within each decision cycle was analyzed, and the energy-saving effect was analyzed by calculating the ratio of actual energy consumption to theoretical energy consumption within each decision cycle.

[0080] Based on the above analysis results, the existing decision-making rules will be adjusted.

[0081] The technical effects of the above-mentioned technical solution are as follows: By establishing a feedback mechanism for human-machine collaborative decision-making, the actual operational effects (including key indicators such as liquid level control accuracy, water supply stability, and energy-saving effect) after each decision can be promptly fed back to the water supply system and experts. This allows for dynamic adjustment and optimization of decision-making rules based on actual operational data. At the same time, experts can further improve and adjust the rules based on the feedback information. This continuous optimization process not only improves the scientificity and accuracy of decision-making but also enhances the adaptability and flexibility of the water supply system, thereby enabling it to better cope with various complex situations in the operation of the water supply system.

[0082] In the dynamic interval generation step, dynamic weights The Sigmoid function is used, as shown in the following formula:

[0083]

[0084] In the formula, For the current moment Distance to the boundary of the artificial zone, in hours; The maximum transition distance is set at 3 hours. The slope coefficient, The greater, the steeper the transition of the function near the threshold; is an offset coefficient for adjusting the center point of the function; is expressed as a dynamic weight, according to the current time to the boundary distance of the manual interval is a dynamically adjusted parameter, when tends to 1, it means that the time period is far away from the manual interval, and the static reference is completely adopted; when tends to 0, it means the manual interval, and the manual set value is forced to match;

[0085] wherein the dynamic center value of the transition zone is specifically shown as follows:

[0086]

[0087] In the formula, is the dynamic center value of the transition zone; is the algorithm weight; are the grade The first hour static liquid level interval lower limit and upper limit; , are the global minimum value and the global maximum value of the transition zone, respectively;

[0088] wherein the dynamic interval is generated as follows:

[0089]

[0090] In the formula, is the upper limit and the lower limit of the dynamic interval; is the dynamic center value of the transition zone; is the original static liquid level interval half-width, .

[0091] The technical effect of the above technical scheme is that the dynamic weight adopts the Sigmoid function, which can flexibly adjust the upper limit and the lower limit of the dynamic interval according to the boundary distance of the current time and the manual interval, so as to realize the smooth transition of the liquid level control in different time periods. When the weight tends to 1, it means that the time period is far away from the manual interval, at this time the water supply system can completely rely on the static reference for control to ensure stable operation. When the weight tends to 0, it means that it is in the manual interval, at this time the water supply system can force to match the manual set value to meet the special needs under special working conditions. Through the above dynamic adjustment mechanism, not only the flexibility and adaptability of the liquid level control are improved, but also the relationship between manual intervention and automatic control is effectively balanced, further optimizing the operation efficiency and energy efficiency performance of the water supply system, and further enhancing the overall performance and reliability of the water supply system.

[0092] In order to more clearly illustrate the technical solutions of the present application, further description will be made below in combination with examples.

[0093] Example 1

[0094] Because there is external engineering needing water in the early morning period, water needs to be stored in advance, so the liquid level interval of 23:00-00:00 is artificially set as [2.2, 3] m.

[0095] According to the water supply prediction, it is determined that the daily water supply is 840,000 tons, and the level =V, and the corresponding static interval table is obtained according to historical data screening:

[0096] Time period 20:00-21:00 2.28 3.09 21:00-22:00 2.03 2.80 22:00-23:00 1.84 2.53 23:00-00:00 1.64 2.33 00:00-01:00 1.52 2.14 01:00-02:00 1.58 2.16 02:00-03:00 1.63 2.48

[0097] According to the formula, the weight and the dynamic reference center value are calculated, and the dynamic liquid level interval 1 is determined:

[0098] Time period Static reference center value weight liquid level interval half-width Dynamic center value Dynamic liquid level interval minimum value Dynamic liquid level interval maximum value 20:00-21:00 2.685 0.960 0.405 2.681 2.276 3.086 21:00-22:00 2.415 0.630 0.385 2.483 2.098 2.868 22:00-23:00 2.185 0.1058 0.345 2.556 2.211 2.901 23:00-00:00 1.985 - - - 2.2 3.0 00:00-01:00 1.83 0.1058 0.31 2.518 2.208 2.828 01:00-02:00 1.87 0.630 0.29 2.139 1.849 2.429 02:00-03:00 2.055 0.960 0.425 2.076 1.651 2.501

[0099] Working principle: through water quantity grading and static interval construction, data is processed, energy-saving samples are screened, and static reference liquid level interval is calculated. This process not only provides accurate basic data and reference interval for subsequent control, but also lays the foundation for energy saving for the entire control process, thereby helping to reduce energy consumption. Then, the water quantity grading matches the reference liquid level interval, so that the liquid level control can be dynamically adjusted according to the actual water supply demand, thereby ensuring that the water supply system can efficiently cope with different water quantity demand situations. In the expert rule embedding link, through the introduction of artificial intervention combined with expert experience, the shortcomings of automatic control in specific complex scenarios can be effectively made up, greatly enhancing the flexibility and adaptability of the water supply system. In the dynamic interval generation stage, by extracting the 3 hours before and after the artificial adjustment interval boundary as the transition zone, adjusting the transition zone liquid level control instruction weight, the liquid level can be smoothly transitioned, effectively reducing the influence of liquid level fluctuation on the water supply system, thereby helping to stabilize the water supply pressure, and further ensuring the continuity and stability of water supply. Finally, the dynamic liquid level control instruction is output to the actuator to realize dynamic liquid level control. The whole method through the close cooperation of each step not only improves the accuracy and adaptability of liquid level control, but also reduces energy consumption through energy-saving sample screening, enhances system flexibility through man-machine collaborative decision-making, guarantees water supply safety through smooth transition and priority setting, and finally realizes efficient, stable, energy-saving and safe water supply system operation.

[0100] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting, since the scope of the present application will be limited to the appended claims. It must be noted that, as used in the specification and the appended claims, the singular form "a," "an" and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to "a component" can include a plurality of components. Similarly, the words "comprise," "comprises," and "comprising," as well as the words "include," "includes," and "including," when used in this specification and in the following claims, are intended to specify the presence of stated features, regions, integers, steps, operations, elements, or components, but they do not preclude the presence or addition of one or more other features, regions, integers, steps, operations, elements, components, or groups thereof. Furthermore, these terms do not necessarily denote the presence of anything that can be claimed as new. The meaning of "a," "an," and "the" also includes plural references and plural forms, for example, "a" or "an" entity includes one or more entities.

[0101] While the embodiments of the application have been shown and described herein, it is understood that modifications, substitutions, changes, and alterations can be made by those skilled in the art without departing from the spirit and scope of the present application.

Claims

1. A dynamic liquid level control method based on water quantity prediction and expert rule embedding, characterized in that, The method comprises the following steps: Water volume grading and static interval construction, including data processing, energy-saving sample screening, and static reference liquid level interval calculation; The water amount is graded and matched with the reference liquid level interval, and the predicted daily water supply amount provided by an external water amount prediction system is acquired , the water amount is graded , and the corresponding static interval is loaded; Expert rule embedding, forming a time interval control interval by presetting a manual intervention period and manually inputting a set liquid level interval in the period, and embedding the control interval into a control system as a man-machine collaborative decision; Dynamic interval generation, extracting all 3 hours before and after the manual adjustment interval boundary as a transition zone, and adjusting the weight of the transition zone liquid level control instruction; Outputting a dynamic liquid level control instruction to an actuator to achieve dynamic control of the liquid level.

2. The dynamic liquid level control method based on water quantity prediction and expert rule embedding according to claim 1, characterized in that, Data processing includes: statistical historical daily water supply , daily specific consumption and hourly level average values, and dividing the historical daily water supply into grades, wherein the daily water supply is the collected data and the daily specific consumption is the calculated value.

3. The dynamic liquid level control method based on water quantity prediction and expert rule embedding according to claim 2, characterized in that, Daily consumption The calculation formula is as follows: ； In the formula, is the daily consumption, unit is That is, the average energy consumption per production of 1 cubic meter of water, the lower the energy consumption value represents the higher the energy efficiency.

4. The dynamic liquid level control method based on water quantity prediction and embedded expert rules of claim 1, wherein, Energy-saving sample screening includes: For each level Filter out daily unit consumption The first 20% are used as sample days, that is, within the same water volume level, the daily unit consumption The lowest 20% of the data represents the best energy-saving operating record for that water volume range.

5. The dynamic liquid level control method based on water quantity prediction and expert rule embedding according to claim 4, characterized in that, The specific method for screening is: Screening data sets with the same level of daily water supply; Sorting all sample days in the level from low to high according to the unit consumption value; Selecting sample days ranked between 0-20% quantile points.

6. The dynamic liquid level control method based on water quantity prediction and embedded expert rules of claim 1, wherein, Static reference liquid level interval calculation includes: By calculating each hour in the energy-saving sample set The median and interquartile range of the liquid level are used to generate the hourly static reference liquid level range.

7. The dynamic liquid level control method based on water quantity prediction and expert rule embedding according to claim 6, characterized in that, The calculation formula of the hourly static reference liquid level interval is as follows: ； wherein respectively the grade The lower and upper limits of the hourly static level interval; is the quartile range of the energy saving samples' liquid level; is the median of the energy saving samples' liquid level, i.e. the sample daily data ranked between the 0~20% percentile points is the hourly average of the liquid level.

8. The dynamic liquid level control method based on water quantity prediction and embedded expert rules of claim 1, wherein, In the step of dynamic interval generation, the dynamic weight The Sigmoid function is adopted, and is specifically as shown in the following formula: ； In the formula, is the current time is the distance to the boundary of the manual interval, in hours; is the maximum transition distance, taking 3 hours; is the slope coefficient, The larger the function, the steeper the transition near the threshold; is the offset coefficient, used to adjust the center point of the function; when tends to 1, it means that the time period is far away from the manual interval, and completely adopts the static reference; is expressed as a dynamic weight, according to the current time to the boundary of the manual interval is a dynamically adjusted parameter, when tends to 0, it means the manual interval, and forces to match the manual setting value.

9. The dynamic liquid level control method based on water quantity prediction and expert rule embedding according to claim 8, characterized in that, In the step of dynamic interval generation, the dynamic center value of the transition zone is as follows: ； wherein is the dynamic center value of the transition zone; is the algorithm weight; are the global minimum and global maximum values of the transition zone, respectively. The first are the lower and upper limits of the hourly static liquid level interval. , are the global minimum and global maximum values of the transition zone, respectively.

10. The dynamic liquid level control method based on water quantity prediction and expert rule embedding according to claim 9, characterized in that, The dynamic interval generation is as follows: ； wherein is the upper and lower limit of the dynamic interval; is the dynamic center value of the transition zone; is the original static liquid level interval half-width, .

Images