An industrial water quota bayesian network uncertainty modeling and correction method

Abstract

The application discloses an industrial water quota Bayesian network uncertainty modeling and correction method and relates to the technical field of industrial water. The method comprises the following steps: obtaining historical data and real-time monitoring data of an industrial water quota to obtain an industrial water data set; pre-processing the obtained industrial water data set; adopting a hybrid learning algorithm to construct a Bayesian network topology; adopting a Markov chain Monte Carlo method to train parameters in the Bayesian network to obtain trained network parameters; outputting a probability distribution of the water quota based on the Bayesian network; and updating the Bayesian network node state through real-time data assimilation to dynamically predict water demand. The application integrates historical data and real-time monitoring information by adopting the Bayesian network, quantifies the uncertainty of the water quota, dynamically corrects model parameters through probability reasoning, adapts to industrial process changes, reduces prediction errors, and improves the robustness and reliability of water-saving measures.

Description

Technical Field

[0001] This invention relates to the field of industrial water technology, and in particular to a method for modeling and correcting uncertainty in Bayesian networks for industrial water quotas. Background Technology

[0002] Industrial water use quotas refer to the limited water consumption per unit of water user within a certain period. Water use quotas are an essential tool for water conservation and water resource management, and are widely used in water-related planning, water resource assessment, water abstraction permits, planned water use management, water conservation evaluation, construction of water conservation facilities, water conservation management, progressive pricing for exceeding quotas, and water conservation supervision and inspection.

[0003] Industrial water consumption quotas, as a management tool, have played an important role in water conservation and water resource management. However, their practical application also has several limitations, mainly reflected in the following aspects:

[0004] (1) Quotas are usually based on historical data or specific technical conditions, and it is difficult to dynamically reflect the improvement in water efficiency brought about by technological progress, process innovation or changes in production scale.

[0005] (2) There are significant differences in production processes and water demand between different industries (such as steel, textiles, and food processing) and even between different enterprises in the same industry. A unified quota may not be able to accurately match the actual needs.

[0006] (3) The scientific nature of quota setting depends heavily on the accuracy and completeness of water monitoring data, but many regions or enterprises have problems such as insufficient metering facilities and inaccurate data reporting.

[0007] A Bayesian network is a graphical network based on probabilistic reasoning. Simply put, it consists of nodes and edges. Nodes represent random variables, which can be various factors that may affect safety, such as equipment status or personnel behavior; edges represent the conditional dependencies between variables. Through this network structure, we can intuitively present the interactions between various factors, and then use probability theory and graph theory methods for reasoning and analysis.

[0008] Traditional security risk assessment methods often struggle to fully consider the complex interactions between various factors. Bayesian network models, however, can easily integrate multiple security-influencing factors for analysis. They clearly demonstrate the relationships between these factors, accurately assessing the overall security risk level.

[0009] Traditional industrial water use quotas often employ static indicators or linear regression models, making it difficult to quantify uncertainties caused by missing data, process differences, and environmental changes. Quota adjustments rely on manual experience or periodic adjustments, lagging behind actual production changes and potentially leading to irrational water resource allocation or excessively high compliance costs for enterprises. Existing methods do not systematically integrate multi-source data (such as real-time monitoring, meteorological data, and equipment status), failing to dynamically respond to fluctuations in water use efficiency.

[0010] Therefore, proposing a Bayesian network uncertainty modeling and correction method for industrial water quotas to address the difficulties in existing technologies is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0011] In view of this, the present invention provides a Bayesian network uncertainty modeling and correction method for industrial water quotas. It uses a Bayesian network to integrate historical data and real-time monitoring information to quantify the uncertainty of water quotas. Through probabilistic reasoning, the model parameters are dynamically corrected to adapt to changes in industrial processes, reduce prediction errors, and improve the robustness and reliability of water-saving measures.

[0012] To achieve the above objectives, the present invention adopts the following technical solution:

[0013] A method for modeling and correcting uncertainties in Bayesian networks for industrial water quotas includes the following steps:

[0014] S1. Data Acquisition: Acquire historical and real-time monitoring data of industrial water consumption quotas to obtain an industrial water consumption dataset;

[0015] S2. Data Preprocessing: Preprocess the acquired industrial water dataset;

[0016] S3. Bayesian Network Structure Learning: Bayesian network topology is constructed using a hybrid learning algorithm;

[0017] S4. Network Training: The parameters in the Bayesian network are trained using the Markov chain Monte Carlo method to obtain the trained network parameters;

[0018] S5. Uncertainty Quantification: Based on Bayesian networks, output the probability distribution of water consumption quotas;

[0019] S6. Dynamic Correction: Update the state of Bayesian network nodes through real-time data assimilation to dynamically predict water demand.

[0020] Optionally, the industrial water dataset includes water intake data, circulating water rate data, product output data, temperature data, and humidity data.

[0021] Optionally, in S2, the acquired industrial water dataset is preprocessed by data cleaning, missing value imputation, data transformation, and outlier detection.

[0022] Optionally, S3 uses a domain knowledge-constrained PC algorithm and greedy search to construct a Bayesian network topology.

[0023] Optional, the specific details of the domain knowledge-constrained PC algorithm are as follows:

[0024] Input: Calculate conditional mutual information between variables; perform independence tests using an industrial significance level of α = 0.01; preserve strong relationships determined by domain knowledge during edge orientation;

[0025] Output: The initial network structure obtained by the domain knowledge-constrained PC algorithm;

[0026] The specific content of the greedy search is:

[0027] Input: The initial network structure obtained by the domain knowledge-constrained PC algorithm, the industrial water dataset, and the domain knowledge constraints, which include prohibiting reverse causality and enforcing key connections;

[0028] Output: The optimized Bayesian network structure and the weights and confidence scores of each edge.

[0029] Optionally, S4 also includes dividing the pre-processed industrial water dataset into training, testing, and validation sets in an 8:1:1 ratio.

[0030] Optionally, in S5, based on a Bayesian network, the specific content of the probability distribution for outputting water consumption quotas is as follows:

[0031]

[0032] Where Q is the target quota variable, E is the observed evidence variable, and Pa is the set of parent nodes.

[0033] Optionally, in S6, the Bayesian network node states are updated through real-time data assimilation to dynamically predict water demand, and a quota correction factor is proposed as follows:

[0034]

[0035] Among them, Q 实际 The actual water consumption is represented by θ, which is a network parameter. Quota revision is triggered when B > B threshold.

[0036] As can be seen from the above technical solution, compared with the prior art, the present invention provides a method for modeling and correcting the uncertainty of Bayesian networks for industrial water quotas, which has the following beneficial effects:

[0037] (1) This invention uses Bayesian networks to integrate historical data and real-time monitoring information, quantifies the uncertainty of water quotas, dynamically corrects model parameters through probabilistic reasoning, adapts to changes in industrial processes, reduces prediction errors, and improves the robustness and reliability of water-saving measures.

[0038] (2) This invention constructs a Bayesian network based on mixed uncertainty, improves physical interpretability, couples quota output with water-saving cost function, and supports enterprises in balancing compliance and economy. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0040] Figure 1 The flowchart of the Bayesian network uncertainty modeling and correction method for industrial water quota provided by the present invention is shown. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] Reference Figure 1 As shown, this invention discloses a method for modeling and correcting uncertainties in Bayesian networks for industrial water quotas, comprising the following steps:

[0043] S1. Data Acquisition: Acquire historical and real-time monitoring data of industrial water consumption quotas to obtain an industrial water consumption dataset;

[0044] S2. Data Preprocessing: Preprocess the acquired industrial water dataset;

[0045] S3. Bayesian Network Structure Learning: Bayesian network topology is constructed using a hybrid learning algorithm;

[0046] S4. Network Training: The parameters in the Bayesian network are trained using the Markov chain Monte Carlo method to obtain the trained network parameters;

[0047] S5. Uncertainty Quantification: Based on Bayesian networks, output the probability distribution of water consumption quotas;

[0048] S6. Dynamic Correction: Update the state of Bayesian network nodes through real-time data assimilation to dynamically predict water demand.

[0049] Furthermore, the industrial water dataset includes water intake data, circulating water rate data, product output data, temperature data, and humidity data.

[0050] Furthermore, in S2, the acquired industrial water dataset undergoes data cleaning, missing value imputation, data transformation, and outlier detection preprocessing.

[0051] Furthermore, S3 uses a domain-knowledge-constrained PC algorithm and greedy search to construct a Bayesian network topology.

[0052] Furthermore, the specific details of the domain knowledge-constrained PC algorithm are as follows:

[0053] Input: Calculate conditional mutual information between variables; perform independence tests using an industrial significance level of α = 0.01; preserve strong relationships determined by domain knowledge during edge orientation;

[0054] Output: The initial network structure obtained by the domain knowledge-constrained PC algorithm;

[0055] The specific content of the greedy search is:

[0056] Input: The initial network structure obtained by the domain knowledge-constrained PC algorithm, the industrial water dataset, and the domain knowledge constraints, which include prohibiting reverse causality and enforcing key connections;

[0057] Output: The optimized Bayesian network structure and the weights and confidence scores of each edge.

[0058] Furthermore, S4 also includes dividing the pre-processed industrial water dataset into training, testing, and validation sets in an 8:1:1 ratio.

[0059] Furthermore, the specific content of the probability distribution of water consumption quota output in S5 based on Bayesian networks is as follows:

[0060]

[0061] Where Q is the target quota variable, E is the observed evidence variable, and Pa is the set of parent nodes.

[0062] Furthermore, S6 updates the Bayesian network node states through real-time data assimilation to dynamically predict water demand, and proposes a quota correction factor as follows:

[0063]

[0064] Among them, Q 实际The actual water consumption is represented by θ, which is a network parameter. Quota revision is triggered when B > B threshold.

[0065] In one specific embodiment, the following is included:

[0066] Taking a chemical company as an example:

[0067] Data input: Collect water consumption data (water intake, circulating water rate, product output) for the past 3 years, meteorological data (temperature, humidity), and equipment operation logs;

[0068] Network training: After initializing the Bayesian network, it converged through 100,000 Markov chain Monte Carlo iterations. The sensitivity of cooling tower efficiency to water consumption was found to be 0.23; 0.23 (95% HDI [0.18, 0.28]), that is, for every 10% increase in efficiency, water consumption decreased by 2.3%.

[0069] Output prediction distribution: mean E 预测 =69.5 m³ / h, standard deviation σ = 5.2 m³ / h, 95% confidence interval [59.7, 79.1] m 3 / h;

[0070] When the actual water consumption Q 实际 When the score is 125, the Z-score is used to quantify the degree of anomaly.

[0071]

[0072] Typically, |z|>3 is considered abnormal, far exceeding the usual threshold;

[0073] Dynamic correction: Actual water consumption during the summer high-temperature period exceeds prediction, triggering correction factor α.

[0074]

[0075] α = 1.8 (> threshold 1.5). When α > 1.5, the system performs the following action:

[0076] Parameter reestimation: Assuming the temperature node distribution should be updated from N(28,2) to N(32,3), new parameters θ are obtained through retraining. 新 ;

[0077] Ratio of expected values ​​between old and new distributions:

[0078]

[0079] Therefore, the quota is increased by 15% (from 10.0 to 11.5m). 3 / ton);

[0080] Short-term measures include activating the backup water cooling system (increasing the heat exchange area by 20%); long-term measures include recommending the installation of a spray cooling device.

[0081] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0082] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An industrial water rationing Bayesian network uncertainty modeling and revision method, characterized in that, Includes the following steps: S1. Data Acquisition: Acquire historical and real-time monitoring data of industrial water consumption quotas to obtain an industrial water consumption dataset; S2. Data Preprocessing: Preprocess the acquired industrial water dataset; S3. Bayesian Network Structure Learning: Bayesian network topology is constructed using a hybrid learning algorithm; S4. Network Training: The parameters in the Bayesian network are trained using the Markov chain Monte Carlo method to obtain the trained network parameters; S5. Uncertainty Quantification: Based on Bayesian networks, output the probability distribution of water consumption quotas; S6. Dynamic Correction: Update the state of Bayesian network nodes through real-time data assimilation to dynamically predict water demand; S3 uses a domain knowledge-constrained PC algorithm and greedy search to construct a Bayesian network topology; The specific content of the domain knowledge-constrained PC algorithm is as follows: Input: Calculate conditional mutual information between variables; perform independence tests using an industrial significance level of α=0.01; preserve strong relationships determined by domain knowledge during edge orientation; Output: The initial network structure obtained by the domain knowledge-constrained PC algorithm; The specific content of the greedy search is: Input: The initial network structure obtained by the domain knowledge-constrained PC algorithm, the industrial water dataset, and the domain knowledge constraints, which include prohibiting reverse causality and enforcing key connections; Output: The optimized Bayesian network structure and the weights and confidence scores of each edge.

2. The method for modeling and correcting uncertainties in Bayesian networks for industrial water quotas according to claim 1, characterized in that, The industrial water dataset includes water intake data, circulating water rate data, product output data, temperature data, and humidity data.

3. The method for modeling and correcting uncertainty in Bayesian networks for industrial water quotas according to claim 1, characterized in that, In S2, the acquired industrial water dataset undergoes preprocessing including data cleaning, missing value imputation, data transformation, and outlier detection.

4. The method for modeling and correcting uncertainties in Bayesian networks for industrial water quotas according to claim 1, characterized in that, S4 also includes dividing the pretreated industrial water dataset into training, testing, and validation sets in an 8:1:1 ratio.

5. The method for modeling and correcting uncertainties in Bayesian networks for industrial water quotas according to claim 1, characterized in that, The specific content of the probability distribution of water consumption quota output based on Bayesian network in S5 is as follows: in, For the target quota variable, For observed evidence variables, The set of parent nodes.

6. The method for modeling and correcting uncertainties in Bayesian networks for industrial water quotas according to claim 1, characterized in that, S6 updates the state of Bayesian network nodes through real-time data assimilation to dynamically predict water demand, and proposes a quota correction factor as follows: in, To measure the actual water consumption, For network parameters, when > Quota revision is triggered when the threshold is reached.

Images