Industrial accumulation area water resource demand prediction and water-saving regulation and control method and system based on artificial intelligence

Abstract

The invention discloses an industrial accumulation area water resource demand prediction and water-saving regulation and control method and system based on artificial intelligence, and relates to the technical field of water resource regulation and control. Comprising the following steps: S1, establishing an evaluation index model according to preprocessed sensor data; s2, obtaining predicted water consumption from the preprocessed sensor data through a constructed water-saving model, and determining a preliminary water-saving scheme according to the predicted water consumption and the comprehensive evaluation score; and S3, adjusting the preliminary water-saving scheme, obtaining a risk assessment coefficient corresponding to each water-saving scheme, and determining a final water-saving scheme through the risk assessment coefficients. According to the method, the evaluation index model is established, so that the optimal water-saving path and implementation scheme can be provided according to the effect evaluation of different water-saving measures, the effective configuration of water resources is ensured, and the scientificity and accuracy of water resource management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of water resource regulation and control, and specifically to a method and system for water resource demand forecasting and water-saving regulation in industrial clusters based on artificial intelligence. Background Art

[0002] With the acceleration of industrialization, industrial clusters have become an important engine of economic development, but they are also facing severe challenges in water resource management and environmental protection. The large amount of industrial water consumption, complex and changeable water quality, and the difficulty of water pollution control make it difficult for traditional management models to meet the needs of modern industrial development.

[0003] At present, most industrial clusters rely on traditional statistical analysis or empirical rules to make water allocation and scheduling decisions. This method often has the following problems: first, insufficient data collection cannot accurately reflect the actual water use situation; second, the model construction is too simplified and fails to consider the impact of dynamic change factors; third, the emergency plan has a slow response speed and cannot respond to emergencies in a timely manner. In addition, the traditional management method is easily interfered by human factors, which increases operational risks.

[0004] The Chinese invention patent with publication number CN107808211A discloses a method for predicting regional water resource demand, including step one: determining the water demand prediction principle; step two: predicting water demand according to regional social and economic development goals and indicators, including industrial water demand prediction, tertiary industry water demand prediction, domestic water demand prediction, agricultural water demand prediction, and ecological environment water demand prediction; step three: determining the total regional water demand; step four: rationality analysis of water demand. The beneficial effects of the invention are: predicting water demand according to regional social and economic development goals and indicators, dividing water demand into five parts for prediction, determining the total regional water demand to be included in the urban water supply network leakage rate, and realizing the prediction of regional water resource demand trend by analyzing the dynamic process of the water demand system; analyzing the rationality of water demand, regulating regional water resources according to the three red lines and water price regulation mechanism, so as to realize the rational use of water resources at the regional and even basin level.

[0005] The above and similar water resource indicator systems tend to focus on the direct reduction of water consumption and large-scale overall application when regulating water resources, thus ignoring the impact of multiple dimensions such as water use efficiency and water quality protection. As a result, the water resource regulation plans obtained focus more on meeting short-term water demand, and cannot be customized according to the characteristics of different industries or enterprises, thus failing to solve the water resource management problem from the root. Summary of the invention

[0006] The purpose of the present invention is to provide an artificial intelligence-based method and system for predicting water resource demand and water-saving regulation in industrial clusters to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above purpose, the present invention provides the following technical solution: a method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence, comprising: S1: Establishing an evaluation index model: acquiring sensor data and preprocessing it, and establishing an evaluation index model based on the preprocessed sensor data; S2: Obtaining a preliminary water-saving plan: Obtaining the predicted water consumption by using the preprocessed sensor data through the constructed water-saving model, and determining a preliminary water-saving plan based on the predicted water consumption and the comprehensive evaluation score; S3: Obtaining the final water-saving plan: adjusting the preliminary water-saving plan and obtaining the risk assessment coefficient corresponding to each water-saving plan. The final water-saving plan is determined by the risk assessment coefficient, including: S3.1: Obtaining the risk assessment coefficient of the preliminary water-saving measures: Obtain the risk assessment coefficient of the preliminary water-saving measures through the risk assessment coefficient acquisition formula, specifically: ; in: is the risk assessment coefficient corresponding to the water-saving measures, is the severity of water conservation measures, is the occurrence frequency corresponding to the water-saving measures, The difficulty of detecting water-saving measures; S3.2: Obtaining risk assessment coefficients of different water-saving measures: adjusting the preliminary water-saving measures through the test module to obtain different combinations of water-saving measures and corresponding risk assessment coefficients; S3.3: Obtain the final water-saving plan: compare the risk assessment coefficients of the preliminary water-saving measures and the risk assessment coefficients corresponding to different combinations of water-saving measures to determine the minimum risk assessment coefficient. The water-saving measure corresponding to the minimum risk assessment coefficient is the final water-saving plan.

[0008] Furthermore, an evaluation index model is established, including: M1: Obtain water use efficiency scoring function: Obtain water use efficiency scoring function through water volume and temperature, specifically: ; in: is the water use efficiency scoring function, is the evaluation function of water quantity, is the evaluation function of temperature, is the weight of water, is the weight of temperature; M2: Obtain water quality protection scoring function: Obtain water quality protection scoring function through pH value, dissolved oxygen, chemical oxygen demand and biochemical oxygen demand, specifically: ; in: is the water quality protection scoring function, is the evaluation function of pH value, is the evaluation function of dissolved oxygen, is the evaluation function of chemical oxygen demand, is the evaluation function of biochemical oxygen demand, is the pH weight, is the weight of dissolved oxygen, is the weight of chemical oxygen demand, is the weight of biochemical oxygen demand; M3: Obtain water supply stability scoring function: Obtain water supply stability scoring function through water supply pressure and flow rate change rate, specifically: ; in: is the water supply stability scoring function, is the evaluation function of water supply pressure, is the evaluation function of the flow rate change rate, is the weight of water supply pressure, is the weight of the flow rate change rate; M4: Obtaining a comprehensive score: Obtain a comprehensive evaluation score based on the water use efficiency scoring function, water quality protection scoring function and water supply stability scoring function, specifically: ; in: is the comprehensive evaluation score, is the water use efficiency scoring function, is the water quality protection scoring function, is the water supply stability scoring function, is the weight of the water use efficiency scoring function, is the weight of the water quality protection scoring function, is the weight of the water supply stability scoring function.

[0009] Furthermore, the process of obtaining the weights of the water efficiency scoring function, the water quality protection scoring function and the water supply stability scoring function includes: N1: Set the weights for extreme weather conditions: adjust the weights for drought and flood periods according to the initial weights of each weight; N2: Obtain the comprehensive evaluation scores of each stage: According to the adjusted weights, obtain the comprehensive evaluation scores corresponding to each adjusted weight; N3: Determine the final weight size: Repeat steps N1 and N2 to adjust the weight size, and obtain the comprehensive evaluation score corresponding to each adjusted weight, and draw a time series graph corresponding to the weight based on all the comprehensive evaluation scores, and determine the optimal weight through the time series graph.

[0010] Furthermore, the weights of drought and flood periods are adjusted as follows: When in a drought period, the weight of the water use efficiency scoring function is increased, and the weight of the water quality protection scoring function and the weight of the water supply stability scoring function are decreased; When it is in the flood season, the weight of the water use efficiency scoring function is lowered, and the weight of the water quality protection scoring function and the weight of the water supply stability scoring function are increased.

[0011] Going a step further, get forecasted water usage, including: W1: Get the predicted water consumption corresponding to the first-layer model: Set multiple basic models and get the predicted water consumption corresponding to each basic model, specifically: ; in: is the predicted water consumption corresponding to the long short-term memory network, is the predicted water consumption corresponding to the gated circulation unit, is the predicted water consumption corresponding to the Transformer model, is the model function of the long short-term memory network, is the model function of the gated recurrent unit, is the model function of the Transformer model, is the feature vector input at time t, is the feature vector input at time t-1, is the feature vector input at time tn; W2: Obtain the predicted water consumption corresponding to the second-layer model: According to the predicted water consumption corresponding to each of the basic models and the preprocessed sensor data, a feature matrix is ​​constructed, and the feature matrix is ​​used as the input of the random forest model to obtain the final predicted water consumption, specifically: ; in: For the final predicted water consumption, is the model function of the random forest model, is the feature matrix.

[0012] Furthermore, a preliminary water conservation plan was determined, including: E1: Set water saving target: According to the final predicted water consumption, obtain the target water consumption, specifically: ; in: The target water consumption is To finally predict water consumption; E2: Divide time: Divide time according to the comprehensive evaluation score to determine the time of low water use efficiency, poor water quality protection and unstable water supply; E3: Develop preliminary water-saving measures: According to the divided time and the target water consumption, set water-saving measures, specifically: When the water use efficiency is low, avoid the concentrated time period of high water consumption operations; When the water quality protection is poor, reducing the operating load of the sewage treatment facilities; When the water supply is unstable, the water supply network is inspected and leaks are repaired.

[0013] An artificial intelligence-based water resource demand forecasting and water-saving regulation system for industrial clusters uses any of the artificial intelligence-based industrial cluster water resource demand forecasting and water-saving regulation methods described above.

[0014] Compared with the prior art, the present invention has the following beneficial effects: Firstly, the present invention establishes an evaluation index model through preprocessed sensor data, so that it can provide the optimal water-saving path and implementation plan according to the effect evaluation of different water-saving measures, thereby ensuring the effective allocation of water resources and improving the scientificity and accuracy of water resources management; Second: The present invention adjusts the preliminary customized water-saving plan through the test module, and obtains the risk assessment coefficient corresponding to the adjusted water-saving plan, so that users can update the strategy at any time according to the latest operating conditions and technological advances, thereby maintaining the timeliness and adaptability of water-saving measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the method for predicting water resource demand and water-saving regulation in industrial clusters in the present invention; Figure 2 It is a time series diagram corresponding to the comprehensive evaluation scores under different weight configurations in the present invention; Figure 3 A schematic diagram of a process for obtaining a comprehensive evaluation score in the present invention; Figure 4It is a schematic diagram of the process of obtaining the final predicted water consumption in the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] The existing water resource indicator system focuses more on the direct reduction of water consumption and the overall application of a large range when regulating water resources, thus ignoring the influence of multiple dimensions such as water efficiency and water quality protection, so that the water resource regulation scheme obtained focuses more on meeting the water demand in the short term, and cannot be customized according to the characteristics of different industries or enterprises, so that the management problem of water resources cannot be solved from the root. The technical solution of this application collects environmental and equipment operation data in real time through a sensor network, and constructs a water-saving model through a machine learning algorithm to obtain predicted water consumption. At the same time, the initial water-saving scheme is determined by predicting water consumption and comprehensive evaluation scores. Further, the user can adjust the specific measures in the preliminary water-saving scheme in the test environment, and obtain the risk assessment coefficients of different adjustment schemes, and compare them with the risk assessment coefficients of the preliminary water-saving scheme, and use the water-saving measures corresponding to the minimum risk assessment coefficient as the final water-saving scheme. Thus, a comprehensive and flexible water resource management framework is provided for industrial clusters, so that it can not only meet the needs of short-term emergency response, but also serve long-term planning and development goals.

[0018] refer to Figure 1-Figure 4 This embodiment provides a method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence, comprising the following steps: Step S1: Establish an evaluation index model, that is, by setting up a sensor network, obtaining sensor data, and performing corresponding preprocessing on the sensor data, and establishing an evaluation index model based on the preprocessed sensor data. The details are as follows: Step S1.1: Preprocessing sensor data. That is, according to the actual monitoring requirements (such as water volume, water quality, temperature, pressure, etc.), a sensor network is set up, and the sensor data obtained by the sensor network is preprocessed accordingly to obtain preprocessed sensor data.

[0019] In this embodiment, corresponding sensors are set, and the set sensors are set at positions that can accurately reflect the characteristics of the target area. In the specific implementation process, the water flow rate is measured by a water volume sensor, the pH sensor monitors the acidity and alkalinity of the water quality, the temperature sensor monitors the water temperature, and the pressure sensor records the water flow pressure. Furthermore, in the industrial cluster area, sensors are arranged at the main water inlet, water outlet and key water use points to ensure the representativeness and accuracy of the data.

[0020] Furthermore, according to actual usage requirements, the sampling frequency of the sensor can be set to every minute or every hour. At the same time, according to the set sampling frequency, the corresponding water volume, pH value, temperature and pressure are obtained regularly. And outliers and erroneous readings are removed through statistical methods, such as using the mean filter method to identify and exclude data that deviates from the mean by more than a certain multiple of the standard deviation, and using linear interpolation to fill in missing values. At the same time, sensor data of different units can also be converted to the same scale for subsequent analysis. For example, through minimum-maximum standardization or Z-score standardization, the data can be converted to the same scale.

[0021] Step S1.2: Construct an evaluation index model. That is, by obtaining three different evaluation dimensions of water use efficiency, water quality protection and water supply stability, and obtaining the evaluation function corresponding to each evaluation dimension, the overall total evaluation score is obtained. The details are as follows: Step M1: Obtain a water use efficiency scoring function. That is, water use efficiency is a measure of the amount of water required per unit time or per unit product output, which is used to obtain the relationship between water volume and temperature. Specifically, water volume directly affects resource consumption, while temperature may affect the evaporation rate of water and the cooling requirements in certain industrial processes. In this embodiment, the formula for obtaining the water use efficiency scoring function is specifically: ; in: is the water use efficiency scoring function, is the evaluation function of water quantity, is the evaluation function of temperature, is the weight of water, is the weight of temperature.

[0022] Furthermore, the evaluation function of the water volume in this embodiment is and temperature evaluation function The formula for obtaining is as follows: ; in: is the evaluation function of water quantity, is the evaluation function of temperature, is the water consumption of the ith sampling, For the ideal maximum water consumption, is the operating temperature of the ith sampling, is the total number of sampling times, is the ideal average operating temperature.

[0023] Step M2: Obtain a water quality protection scoring function. That is, water quality protection is to ensure that the water body is not polluted and maintain a good ecological health state. It is used to obtain the relationship between chemical components such as pH value, dissolved oxygen, chemical oxygen demand, and biochemical oxygen demand. In this embodiment, the formula for obtaining the water quality protection scoring function is specifically: ; in: is the water quality protection scoring function, is the evaluation function of pH value, is the evaluation function of dissolved oxygen, is the evaluation function of chemical oxygen demand, is the evaluation function of biochemical oxygen demand, is the pH weight, is the weight of dissolved oxygen, is the weight of chemical oxygen demand, is the weight of biochemical oxygen demand.

[0024] Furthermore, the pH evaluation function , evaluation function of dissolved oxygen , the evaluation function of chemical oxygen demand and the evaluation function of biochemical oxygen demand In the process of obtaining specific numerical values, the corresponding threshold range can be set according to actual needs, and the corresponding evaluation function can be obtained according to the set threshold range, that is, points will be deducted if the threshold range is exceeded (it can be specifically set according to actual needs, so it will not be specifically explained in this embodiment).

[0025] In the specific implementation process, for pH value, the corresponding threshold range is 6.5-8.5, that is, when the pH value obtained is within the range of 6.5-8.5, the evaluation function of pH value is full marks, otherwise points are deducted according to the degree of deviation. For dissolved oxygen, when it is not less than 5 mg / L, the evaluation function of dissolved oxygen is full marks, otherwise points are deducted according to the degree of below. For chemical oxygen demand and biochemical oxygen demand, points can be deducted according to whether it exceeds the specified limit, that is, when it does not exceed the specified limit, the evaluation function of chemical oxygen demand and the evaluation function of biochemical oxygen demand are full marks, otherwise points are deducted according to the degree of exceeding.

[0026] Step M3: Obtain a water supply stability scoring function. That is, water supply stability is to ensure continuous and stable water supply services to avoid water outages caused by water source depletion or facility failure. It is used to obtain the relationship between water supply pressure and flow rate change rate. In this embodiment, the formula for obtaining the water supply stability scoring function is specifically: ; in: is the water supply stability scoring function, is the evaluation function of water supply pressure, is the evaluation function of the flow rate change rate, is the weight of water supply pressure, is the weight of the flow rate change rate.

[0027] In the process of specific implementation, for the water supply pressure, the corresponding evaluation function size is determined according to the set interval. For the flow rate change rate, the corresponding evaluation function size is determined according to the fluctuation range. Specifically, the interval of the water supply pressure setting in this embodiment is 0.3MPa-0.5MPa, that is, when the obtained water supply pressure is in this interval, the evaluation function of the water supply pressure is full marks, otherwise it is deducted according to the degree of deviation.

[0028] Step M4: Obtain a comprehensive score. That is, according to the water efficiency score function obtained in step M1 , the water quality protection scoring function obtained in step M2 , the water supply stability scoring function obtained in step M3 , get the comprehensive score of the three, specifically: ; in: is the comprehensive evaluation score, is the water use efficiency scoring function, is the water quality protection scoring function, is the water supply stability scoring function, is the weight of the water use efficiency scoring function, is the weight of the water quality protection scoring function, is the weight of the water supply stability scoring function.

[0029] It is worth noting that the weights of the water efficiency scoring function, the water quality protection scoring function, and the water supply stability scoring function in this embodiment are specifically set through scenario simulation and optimization, as follows: Step N1: Set the weights for extreme weather conditions. That is, adjust the weights of each weight according to its initial weight. Specifically, when in drought, adjust the weights of the water efficiency scoring function Increase the weight of the water quality protection scoring function and the weight of the water supply stability scoring function When it is in the flood season, the weight of the water use efficiency scoring function is Lower the weight of the water quality protection scoring function and the weight of the water supply stability scoring function Adjust it higher.

[0030] In the process of specific implementation, the weight of the water efficiency scoring function The initial size of is set to 0.4, and the weight of the water quality protection scoring function The initial size of is set to 0.3, and the weight of the water supply stability scoring function The initial size of is set to 0.3. Furthermore, when in drought period, the weight of water use efficiency score function is The size is set to 0.5, and the weight of the water quality protection scoring function The size of is set to 0.25, the weight of the water supply stability scoring function The size of is set to 0.25. Furthermore, when in the flood period, the weight of the water use efficiency scoring function is The size is set to 0.2, and the weight of the water quality protection scoring function The size of is set to 0.4, the weight of the water supply stability scoring function The size is set to 0.4.

[0031] Step N2: Obtain the comprehensive evaluation score of each stage. That is, according to the size of each weight set in step N1, obtain the comprehensive evaluation score corresponding to each weight.

[0032] In the specific implementation process, the water efficiency scoring function is 10.1, the water quality protection scoring function is 0.96, and the water supply stability scoring function is 0.98. Therefore, the comprehensive evaluation score corresponding to the initial weight is 4.622. When it is in a drought period, the comprehensive evaluation score corresponding to its weight size is 5.535. When it is in a flood period, the comprehensive evaluation score corresponding to its weight size is 2.796.

[0033] Step N3: Determine the final weight. That is, repeat steps N1 and N2, continuously adjust the weight under extreme weather conditions, and obtain the comprehensive evaluation score corresponding to each adjusted weight. Further, according to all the comprehensive evaluation scores obtained, determine the corresponding time series graph, compare their changing trends, and determine the final weight.

[0034] In the process of specific implementation, the data list shown in Table 1 below is obtained, specifically: Table 1: Data list ; According to the data in Table 1 above and Figure 2 From the time series diagram in the figure, it can be seen that the comprehensive evaluation score corresponding to the initial weight is relatively stable overall, the comprehensive evaluation score corresponding to the drought period weight increases significantly between the 7th and 9th months, and the comprehensive evaluation score corresponding to the flood period weight decreases during the flood period (for example, the 11th month). In other words, by continuously adjusting the weight size, the optimal weight can be obtained. Therefore, the optimal weight in this embodiment is: the weight of the water efficiency scoring function The size is set to 0.25, and the weight of the water quality protection scoring function The size of is set to 0.4, the weight of the water supply stability scoring function The size is set to 0.35.

[0035] Step S2: Obtain a preliminary water-saving plan. That is, through the constructed water-saving model, the pre-processed sensor data obtained in step S1.1 is used as input, and the corresponding predicted water consumption is obtained as output. At the same time, according to the predicted water consumption and the comprehensive evaluation score obtained in step M4, a preliminary water-saving plan is determined. The details are as follows: Step S2.1: Obtain predicted water consumption. That is, a water-saving model is constructed through physical models and statistical models, and the preprocessed sensor data is used as the input of the water-saving model, and the corresponding predicted water consumption is obtained as output. The details are as follows: Step W1: Obtain the predicted water consumption corresponding to the first layer model. That is, set up three different basic models (such as long short-term memory network, gated recurrent unit and Transformer model), and use the preprocessed sensor data obtained in step S1.1 as the input of the three basic models, and output the predicted water consumption corresponding to the three basic models. Specifically: ; in: is the predicted water consumption corresponding to the long short-term memory network, is the predicted water consumption corresponding to the gated circulation unit, is the predicted water consumption corresponding to the Transformer model, is the model function of the long short-term memory network, is the model function of the gated recurrent unit, is the model function of the Transformer model, is the feature vector input at time t, is the feature vector input at time t-1, is the feature vector input at time tn.

[0036] Step W2: Obtain the predicted water consumption corresponding to the second-layer model. The three predicted water consumptions obtained in step W1 and the original features (i.e., the preprocessed sensor data obtained in step S1.1) are combined to form a feature matrix, and the constructed feature matrix is ​​used as input through the random forest model to obtain the final predicted water consumption as output, specifically: ; in: To predict the final water consumption, is the model function of the random forest model, is the feature matrix.

[0037] In the process of specific implementation, the characteristic data table of Table 2 below is obtained, specifically: Table 2: Feature data table ; According to the data in Table 2 above, the predicted water consumption data table corresponding to the first-level model in Table 3 below can be obtained, specifically: Table 3: Predicted water consumption data corresponding to the first-level model ; According to the data in Table 3 above, the data table of the feature matrix shown in Table 4 below can be obtained, specifically: Table 4: Data table of feature matrix ; According to the data in Table 4 above, the data table of the final predicted water consumption can be obtained as shown in Table 5 below, specifically: Table 5: Data table of final predicted water consumption ; Step S2.2: Obtain a preliminary water-saving plan. The comprehensive evaluation score obtained according to step M4 Step W2: The final predicted water consumption is obtained , as follows: Step E1: Set the water saving target. That is, the final predicted water consumption obtained in step W2 , get the target water consumption, specifically: ; in: The target water consumption is To predict the final water consumption.

[0038] In the specific implementation process, according to the data table of the final predicted water consumption in Table 5 above, the data table of the target water consumption as shown in Table 6 below can be obtained, specifically: Table 6: Data table of target water consumption ; Step E2: Divide the time. That is, the comprehensive evaluation score obtained according to step M4 , divide the time. Specifically, according to the comprehensive evaluation score , determine the minimum score function corresponding to each day. In other words, determine the time of low water use efficiency, poor water quality protection and unstable water supply.

[0039] Step E3: Formulate preliminary water-saving measures. That is, according to the target water consumption obtained in step E1 Set corresponding water-saving measures according to the time divided in step E2. Specifically: When water use efficiency is low, avoid concentrated time periods of high water consumption operations.

[0040] When water quality is poor, reduce the operating load of sewage treatment facilities.

[0041] When the water supply is unstable, inspect the water supply network and repair leaks.

[0042] Step S3: Obtain the final water-saving plan. That is, through the test module, adjust the preliminary water-saving measures formulated in step E3, obtain the risk assessment coefficient corresponding to each water-saving measure, and compare all risk assessment coefficients to determine the minimum risk assessment coefficient. The water-saving measure corresponding to the minimum risk assessment coefficient is the final water-saving plan. The details are as follows: Step S3.1: Obtain the risk assessment coefficient of the preliminary water-saving measures. The risk assessment coefficient of the preliminary water-saving measures is obtained through the risk assessment coefficient acquisition formula, which is specifically: ; in: is the risk assessment coefficient corresponding to the water-saving measures, is the severity of water conservation measures, is the occurrence frequency corresponding to the water-saving measures, The difficulty of detecting water-saving measures.

[0043] Furthermore, the severity of water-saving measures can be comprehensively determined by water loss, economic loss and environmental impact. Specifically, the average daily additional water consumption can be determined by obtaining the amount of water loss in the failure records of water-saving measures. Similarly, water fee losses (the product of additional water consumption and water price) and indirect costs (the product of shutdown market and unit time output value) can be obtained through water quality sensor data (such as COD and BOD exceeding the standard record) and the time point of the failure event. Similarly, the excess multiples of pollutant leakage can be obtained through water quality sensor data.

[0044] Furthermore, in the process of obtaining water loss, economic loss and environmental impact, it is possible to score the water loss, economic loss and environmental impact according to the division range of the preset threshold size (which can be specifically divided according to actual needs) to obtain their corresponding score sizes. Specifically, the score size of water loss, the score size of economic loss and the score size of environmental impact are compared to determine the maximum score size, which is the severity size corresponding to the water-saving measure.

[0045] Furthermore, the occurrence frequency of water-saving measures can be obtained through the implementation records of water-saving measures and failure events (such as valve failure, sensor false alarm, and operation error). Specifically, according to the implementation records of water-saving measures and failure events, the number of failures and the total number of observations in the failure records of water-saving measures are determined, so as to obtain the corresponding occurrence frequency, which is as follows: ; in: is the occurrence frequency corresponding to the water-saving measures, is the number of failures actually observed, is the prior success parameter, is the total number of operating days of the water-saving measures, is the a priori failure parameter.

[0046] Furthermore, the detection difficulty corresponding to the water-saving measures can be determined by the equipment coverage, data delay and manual inspection frequency. Specifically, by obtaining the size of the equipment coverage, data delay and manual inspection frequency, and according to the division range of the preset threshold (which can be specifically divided according to actual needs), the corresponding threshold size is determined, and the threshold of the equipment coverage, the threshold of the data delay and the threshold of the manual inspection frequency are compared to determine the maximum value threshold size, which is the size of the detection difficulty corresponding to the water-saving measures.

[0047] Specifically, the equipment coverage is determined by the proportion of sensors installed in key nodes (such as water inlets, water outlets, and high-water-consuming equipment). Data latency is determined by obtaining the timestamp of sensor data uploaded to the monitoring system and the timestamp of the system generating an alarm, that is, the time interval from data collection to triggering an alarm. The frequency of manual inspections is determined by the cycle of manual inspections (such as daily, weekly, and monthly).

[0048] Step S3.2: Obtain risk assessment coefficients for different water-saving measures. That is, based on the preliminary water-saving measures, the preliminary water-saving measures are adjusted accordingly through the set test module, that is, different combinations of water-saving measures are obtained. It is worth noting that in this embodiment, the preliminary water-saving measures are adjusted accordingly through the test module, which can be specifically adjusted and set according to the actual needs of the user, so it is not specifically explained in this embodiment. Furthermore, the risk assessment coefficient corresponding to the adjusted water-saving measure is obtained at the same time.

[0049] Step S3.3: Obtain the final water-saving plan. That is, compare the risk assessment coefficient of the preliminary water-saving measures in step S3.1 with the risk assessment coefficients of different water-saving measures in step S3.2, and determine the minimum risk assessment coefficient. The water-saving measure corresponding to the minimum risk assessment coefficient is the final water-saving plan.

[0050] In the process of specific implementation, the severity of the preliminary water-saving measure A is 7, the frequency of occurrence is 4, and the detection difficulty is 3, so its corresponding risk assessment coefficient is 84. Furthermore, the severity of the water-saving measure B for the water recycling system is 6, the frequency of occurrence is 3, and the detection difficulty is 2, so its corresponding risk assessment coefficient is 36. Furthermore, the severity of the water-saving measure C for optimizing the water use in the production process is 5, the frequency of occurrence is 4, and the detection difficulty is 2, so its corresponding risk assessment coefficient is 40. In other words, the risk assessment coefficient corresponding to water-saving measure B is the smallest, so water-saving measure B is the final water-saving solution, that is, the water recycling system.

[0051] This embodiment also provides an artificial intelligence-based industrial cluster water resource demand forecasting and water-saving regulation system, which uses an artificial intelligence-based industrial cluster water resource demand forecasting and water-saving regulation method in the above-mentioned embodiment.

[0052] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is limited by the attached embodiments and their equivalents.

Claims

1. A method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence, characterized in that: Included are: S1: Establishing an evaluation index model: acquiring sensor data and preprocessing it, and establishing an evaluation index model based on the preprocessed sensor data; S2: Obtaining a preliminary water-saving plan: Obtaining the predicted water consumption by using the preprocessed sensor data through the constructed water-saving model, and determining a preliminary water-saving plan based on the predicted water consumption and the comprehensive evaluation score; S3: Obtaining the final water-saving plan: adjusting the preliminary water-saving plan and obtaining the risk assessment coefficient corresponding to each water-saving plan. The final water-saving plan is determined by the risk assessment coefficient, including: S3.1: Obtaining the risk assessment coefficient of the preliminary water-saving measures: Obtain the risk assessment coefficient of the preliminary water-saving measures through the risk assessment coefficient acquisition formula, specifically: ; in: is the risk assessment coefficient corresponding to the water-saving measures, is the severity of water conservation measures, is the occurrence frequency corresponding to the water-saving measures, The difficulty of detecting water-saving measures; S3.2: Obtaining risk assessment coefficients for different water-saving measures: adjusting the preliminary water-saving measures through the test module to obtain different combinations of water-saving measures and corresponding risk assessment coefficients; S3.3: Obtain the final water-saving plan: compare the risk assessment coefficients of the preliminary water-saving measures and the risk assessment coefficients corresponding to different combinations of water-saving measures to determine the minimum risk assessment coefficient. The water-saving measure corresponding to the minimum risk assessment coefficient is the final water-saving plan.

2. According to the method of artificial intelligence-based water resource demand forecasting and water-saving regulation in industrial clusters according to claim 1, it is characterized in that: Establish an evaluation index model, including: M1: Obtain water use efficiency scoring function: Obtain water use efficiency scoring function through water volume and temperature, specifically: ; in: is the water use efficiency scoring function, is the evaluation function of water quantity, is the evaluation function of temperature, is the weight of water, is the weight of temperature; M2: Obtain water quality protection scoring function: Obtain water quality protection scoring function through pH value, dissolved oxygen, chemical oxygen demand and biochemical oxygen demand, specifically: ; in: is the water quality protection scoring function, is the evaluation function of pH value, is the evaluation function of dissolved oxygen, is the evaluation function of chemical oxygen demand, is the evaluation function of biochemical oxygen demand, is the pH weight, is the weight of dissolved oxygen, is the weight of chemical oxygen demand, is the weight of biochemical oxygen demand; M3: Obtain water supply stability scoring function: Obtain water supply stability scoring function through water supply pressure and flow rate change rate, specifically: ; in: is the water supply stability scoring function, is the evaluation function of water supply pressure, is the evaluation function of the flow rate change rate, is the weight of water supply pressure, is the weight of the flow rate change rate; M4: Obtaining a comprehensive score: Obtain a comprehensive evaluation score based on the water use efficiency scoring function, water quality protection scoring function and water supply stability scoring function, specifically: ; in: is the comprehensive evaluation score, is the water use efficiency scoring function, is the water quality protection scoring function, is the water supply stability scoring function, is the weight of the water use efficiency scoring function, is the weight of the water quality protection scoring function, is the weight of the water supply stability scoring function.

3. The method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence according to claim 2 is characterized in that: The process of obtaining the weights of the water efficiency scoring function, the water quality protection scoring function and the water supply stability scoring function includes: N1: Set the weights for extreme weather conditions: adjust the weights for drought and flood periods according to the initial weights of each weight; N2: Obtain the comprehensive evaluation scores of each stage: According to the adjusted weights, obtain the comprehensive evaluation scores corresponding to each adjusted weight; N3: Determine the final weight size: Repeat steps N1 and N2 to adjust the weight size, and obtain the comprehensive evaluation score corresponding to each adjusted weight, and draw a time series graph corresponding to the weight based on all the comprehensive evaluation scores, and determine the optimal weight through the time series graph.

4. The method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence according to claim 3 is characterized in that: The weights of drought and flood periods are adjusted as follows: When in a drought period, the weight of the water use efficiency scoring function is increased, and the weight of the water quality protection scoring function and the weight of the water supply stability scoring function are decreased; When it is in the flood season, the weight of the water use efficiency scoring function is lowered, and the weight of the water quality protection scoring function and the weight of the water supply stability scoring function are increased.

5. The method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence according to claim 1 is characterized in that: Get forecasted water usage, including: W1: Get the predicted water consumption corresponding to the first-layer model: Set multiple basic models and get the predicted water consumption corresponding to each basic model, specifically: ; in: is the predicted water consumption corresponding to the long short-term memory network, is the predicted water consumption corresponding to the gated circulation unit, is the predicted water consumption corresponding to the Transformer model, is the model function of the long short-term memory network, is the model function of the gated recurrent unit, is the model function of the Transformer model, is the feature vector input at time t, is the feature vector input at time t-1, is the feature vector input at time tn; W2: Obtain the predicted water consumption corresponding to the second-layer model: According to the predicted water consumption corresponding to each of the basic models and the preprocessed sensor data, a feature matrix is ​​constructed, and the feature matrix is ​​used as the input of the random forest model to obtain the final predicted water consumption, specifically: ; in: For the final predicted water consumption, is the model function of the random forest model, is the feature matrix.

6. The method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence according to claim 1 is characterized in that: Determine the preliminary water conservation plan, including: E1: Set water saving target: According to the final predicted water consumption, obtain the target water consumption, specifically: ; in: The target water consumption is To finally predict water consumption; E2: Divide time: Divide time according to the comprehensive evaluation score to determine the time of low water use efficiency, poor water quality protection and unstable water supply; E3: Develop preliminary water-saving measures: According to the divided time and the target water consumption, set water-saving measures, specifically: When the water use efficiency is low, avoid the concentrated time period of high water consumption operations; When the water quality protection is poor, reducing the operating load of the sewage treatment facilities; When the water supply is unstable, the water supply network is inspected and leaks are repaired.

7. An artificial intelligence-based water resource demand forecasting and water-saving control system for industrial clusters, characterized in that: A method for predicting water resource demand and water-saving regulation in industrial clusters based on artificial intelligence as described in any one of claims 1 to 6 is used.

Images