A method, device and medium for optimizing sensor arrangement in sewage treatment process

Abstract

The present invention discloses a method, device and medium for optimizing the arrangement of sensors in a sewage treatment process, wherein the method is as follows: establishing a system directed graph model for the sewage treatment process, defining a decision variable matrix for the sensor arrangement; establishing a sewage treatment process sensor optimization arrangement model based on the decision variable matrix and considering the importance of variables to be measured in different links of sewage treatment based on the structural observability and redundancy criteria; using a discrete multi-objective state transfer algorithm to solve the sewage treatment process sensor optimization arrangement model to obtain an optimal solution set; introducing an efficient non-dominated sorting strategy and a crowding distance diversity retention strategy to ensure the effectiveness and distribution of the optimal solution; based on the corrosion degree and maintenance cost evaluation strategy, selecting a unique solution suitable for application in the sewage treatment process from the optimal solution set. The technology of the present invention has guiding significance for the optimal arrangement of sensors in the sewage treatment process, and further provides a technical basis for optimizing the operation of sewage treatment plants.

Description

Technical Field

[0001] The invention belongs to the technical field of sewage treatment process optimization control, and in particular relates to a sewage treatment process sensor optimization arrangement method, equipment and medium. Background Art

[0002] Urban sewage treatment plants comprehensively treat urban domestic sewage, industrial wastewater and rainfall runoff, etc., which is a strategic measure to achieve comprehensive utilization of water resources and a virtuous cycle. In order to ensure the safety and reliability of sewage treatment plants, process parameter monitoring is crucial, which promotes the widespread application of sensors, thus providing a technical basis for the safe and reliable operation of sewage treatment plants.

[0003] In the process of urban sewage treatment, in order to achieve optimized control and ensure that various water quality parameters are stable and meet the standards, a large number of detection instruments and sensors need to be installed in the entire sewage treatment process to monitor the parameters comprehensively. However, on the one hand, the sewage treatment process is a distributed parameter system with infinite dimensional freedom. Due to the limitations of the detection environment, technology and cost, it is unrealistic to conduct full coverage detection of the entire plant; on the other hand, although sewage treatment plants are data-rich, due to improper installation or maintenance of sensors, the collected data is not converted into effective and actionable knowledge for system optimization, resulting in a waste of storage space and investment funds. Therefore, it is extremely important to optimize the installation location, type and number of sensors in sewage treatment plants. Summary of the invention

[0004] The present invention provides a method, device and medium for optimizing the arrangement of sensors in a sewage treatment process, which optimize the arrangement of sensors installed in the entire sewage treatment process and achieve multi-objective optimization of sensor cost, process parameter information richness and reliability.

[0005] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0006] A method for optimizing the arrangement of sensors in a sewage treatment process, comprising:

[0007] Step 1: Establish a system directed graph model for the sewage treatment process and define the decision variable matrix of sensor layout;

[0008] Step 2: Based on the structural observability and redundancy criteria, establish a sensor optimization layout model for the entire sewage treatment process based on the decision variable matrix;

[0009] Step 3: Using a discrete multi-objective state transition algorithm, the sensor optimization layout model of the entire sewage treatment process is solved to obtain the optimal solution set; an efficient non-dominated sorting strategy and a crowding distance diversity preservation strategy are introduced to ensure the effectiveness and distribution of the optimal solution;

[0010] Step 4: Based on the evaluation strategy of corrosion degree and maintenance cost, a unique solution suitable for industrial application is selected from the optimal solution set obtained in step 3.

[0011] Furthermore, the directed graph model of the system includes a number of nodes and directed edges, the nodes represent different unit processes, flow connection points and environmental nodes in the sewage treatment process, and the directed edges represent the flows between the two end nodes; the elements of the decision variable matrix indicate whether sensors are arranged for each flow and each variable to be measured.

[0012] Furthermore, the sensor optimization layout model for the entire sewage treatment process established in step 2 is a weighted multi-objective optimization model that takes into account the importance of the variables to be measured in different treatment links of the sewage treatment process, and is expressed as:

[0013]

[0014] Where x is the decision variable matrix describing the sensor layout, where the decision variable x ij Indicates whether the jth variable to be measured in the i-th stream in the sewage treatment process is equipped with a sensor, N is the total number of streams in the sewage treatment process, and M is the total number of variables to be measured in each stream; f C (x) is the sensor installation cost objective function, f O (x) is the variable observability objective function, f R (x) is the measurement redundancy objective function; δ O is the variable observability matrix, δ O,ij is the observability of the jth variable to be measured in the i-th stream; δ R is the measurement redundancy matrix, δ R,ij is the redundancy of the jth variable to be measured in the i-th stream; w C is the sensor cost matrix, w C,ij is the cost of placing sensors for the jth variable to be measured in the i-th stream; w O is the observable importance matrix of the variables to be measured, w O,ij is the observable importance of the jth variable to be measured in the i-th stream; w R is the redundant importance matrix of the variables to be measured, w R,ij is the redundant importance of the jth variable to be measured in the i-th stream; D ij is the decision variable x ij The value range of .

[0015] Furthermore, the observability refers to the fact that the value of the variable to be measured can be directly measured by a sensor or can be indirectly calculated from known variables; the redundancy refers to the fact that when the sensor measuring the variable is removed, the variable remains observable.

[0016] Furthermore, the triangular fuzzy analytic hierarchy process is used to determine the observable importance matrix w of the variables to be measured. O and the redundant importance matrix w R , specifically:

[0017] Firstly, the observable importance of the jth measured variable in the sewage treatment process is evaluated as the goal, volatility, reliability and risk are used as the importance evaluation attributes, and the measured variable of all flows in the sewage treatment process is used as the object to establish a three-layer hierarchical structure;

[0018] Then, according to the characteristics of the sewage treatment process, the triangular fuzzy multiplication preference relationship matrix of the attributes is constructed. The matrix elements are triangular fuzzy numbers Represents the preference ratio of the relative importance of any two attributes a and b; in addition, based on each attribute, the observability triangular fuzzy multiplication preference relationship matrix of the object is constructed in the same way;

[0019] Then, the contradiction and acceptability of each triangular fuzzy multiplication preference relationship matrix constructed are checked. If the conditions are not met, the matrix element values ​​are adjusted. If the conditions are met, the fuzzy preference programming method is used to derive the optimal fuzzy weight vector from the matrix; and a parameterized defuzzification method is used to obtain the deterministic optimal real number weight vector from the optimal fuzzy weight vector;

[0020] Then obtain the weight vectors of all attributes in the evaluation target and the weight vectors of objects based on each attribute in the evaluation target, and aggregate the local weights at different levels into the global weight of the variable j to be measured on each stream through weighted summation:

[0021]

[0022] Among them, A i is the global weight of the variable j to be measured on the i-th stream; w v is the weight of attribute v, a iv is the weight evaluation value of the variable j to be measured in the i-th stream based on the attribute v; A i The larger the value, the more important the i-th stream is to the measured variable j. The observable importance weight vector of the measured variable j on the N streams of the sewage treatment process is expressed as w O,j =[A1,A2,...,A n ];

[0023] Finally, the observability weight vectors of all variables to be measured in the sewage treatment process are obtained according to the above process, and the observable importance matrix w of the variables to be measured is obtained by aggregation O ;

[0024] In addition, according to the above observability matrix w for all variables to be measured OThe determination method is to obtain the redundant importance matrix w of all variables to be measured. R .

[0025] Furthermore, in step 3, the discrete multi-objective state transfer algorithm uses a state transformation operator to generate candidate solutions, and introduces an efficient non-dominated sorting strategy and a crowding distance diversity retention strategy to ensure the effectiveness and distribution of the optimal solution; the specific steps of solving the model to obtain the optimal solution set include:

[0026] Step 3.1: Set the search intensity SE and the maximum allowed number of iterations Maxiter of the discrete multi-objective state transfer algorithm according to the configuration of the sewage treatment plant; randomly generate the initial solution Best of the sensor layout plan k ;

[0027] Step 3.2: Best for the current optimal solution k Perform exchange transformation, move transformation, symmetry transformation and substitution transformation respectively to generate SE candidate solutions;

[0028] Step 3.3: Use an efficient non-dominated sorting strategy to compare each candidate solution with the candidate solutions that have been assigned a frontier level, quickly divide the frontier level of the candidate solution set, and the solution with a higher frontier level is better;

[0029] Step 3.4: Adopt the crowding distance diversity preservation strategy to calculate the crowding distance of each candidate solution. When two candidate solutions are at the same frontier level, the solution with a larger crowding distance is better; update the current optimal solution Best k ;

[0030] Step 3.5: Let k = k + 1. If the number of iterations is less than Maxiter, return to step 3.2 until the termination condition is met.

[0031] Furthermore, step 4 specifically includes:

[0032] Step 4.1: Establish the corrosion degree calculation function

[0033] A sensor corrosion condition assessment system is established, which takes into account multiple corrosion types, and each corrosion type corresponds to a score s that can quantify the degree of corrosion of the sensor. t , t is used to distinguish different corrosion types; the sensor corrosion condition evaluation system determines the calculation formula for the average corrosion degree of all sensors in the sensor optimization layout scheme:

[0034]

[0035] Among them, x p is any solution in the optimal solution set, E C (x p ) is the optimal arrangement of sensors xp is the average corrosion degree of all sensors in the optimal solution set; K is the total number of solutions in the optimal solution set; N is the total number of streams in the sewage treatment plant; M is the total number of variables to be measured in each stream; λ is the weight vector of all corrosion types; η is the corrosion effect compensation coefficient; T is the number of corrosion types;

[0036] Step 4.2, construct the maintenance cost calculation function, expressed as:

[0037]

[0038] Among them, E M (x p ) is the optimal arrangement of sensors x p The average maintenance cost of all sensors; C j is the basic maintenance cost of different types of sensors; ij and f ij is the maintenance difficulty coefficient and maintenance frequency of the sensor layout for the jth variable to be measured in the i-th stream;

[0039] Step 4.3: The above corrosion degree and maintenance cost calculation functions are integrated and normalized to obtain the overall evaluation function of the sensor layout scheme:

[0040]

[0041] Among them, ω is a weight coefficient based on the configuration and operating experience of the sewage treatment plant.

[0042] Step 4.4, select the sensor layout scheme with the smallest overall evaluation function value from the optimal solution set, which is the only solution suitable for sewage treatment process application.

[0043] Furthermore, the sensor corrosion condition assessment system includes six corrosion types, namely: acid corrosion, alkali corrosion, organic corrosion, inorganic corrosion, microbial corrosion and mechanical impact corrosion.

[0044] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor implements the method for optimizing the arrangement of sensors in a sewage treatment process as described in any one of the above technical solutions.

[0045] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for optimizing the arrangement of sensors in a sewage treatment process as described in any of the above technical solutions.

[0046] Beneficial Effects

[0047] The present invention provides a method for optimizing the arrangement of sensors in a sewage treatment process. The method is based on graph theory knowledge, structural observability and redundancy criteria, and a triangular fuzzy hierarchical analysis method. A sewage treatment full-process sensor optimization arrangement model that considers the importance of variables to be measured in different treatment links is established, so that the model is more in line with the characteristics of actual optimization problems, significantly reduces the number of solutions in a feasible domain, and reduces the complexity of solution. A discrete multi-objective state transfer algorithm is used to solve the optimal solution of the sensor optimization arrangement problem, and two update strategies, namely an efficient non-dominated sorting strategy and a crowded distance diversity maintenance strategy, are combined to ensure the effectiveness and distribution of the optimal solution. An evaluation strategy based on corrosion and maintenance cost is designed to select a solution with the smallest evaluation function value in the optimal solution set as a group of solutions most suitable for further application in the sewage treatment process. The sensor arrangement scheme achieves the goal of obtaining higher process information richness and reliability at a lower sensor cost, and ensures that the data quality, operating cost and reliability of the sewage treatment plant are at a comprehensive optimal level.

[0048] By utilizing the sewage treatment process sensor optimization layout method of the present invention, based on structural judgment criteria and structured goals, without the need for sensor performance information and the need to collect measurement data in advance, the optimal sensor layout plan in the sewage treatment plant can be quickly and effectively obtained to ensure that the data quality, operating cost and reliability of the sewage treatment plant are at a comprehensive optimal level.

[0049] The present invention has the characteristics of quick and convenient calculation, high information richness and reliability of the production process, etc. It is suitable for optimizing the layout of sensors in the whole process of sewage treatment process, and is of great significance to the stable operation of the sewage treatment process, ensuring the quality of effluent, and reducing costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A flow chart of a method for optimizing sensor arrangement in a sewage treatment process provided by the present invention;

[0051] Figure 2 A process flow chart of a sewage treatment process according to an embodiment of the present invention;

[0052] Figure 3 A directed graph model of a sewage treatment process system according to an embodiment of the present invention;

[0053] Figure 4 A hierarchical structure diagram constructed by using the triangular fuzzy hierarchical analysis method in an embodiment of the present invention;

[0054] Figure 5 The optimal solution distribution diagram for optimizing the arrangement of flow sensors in a sewage treatment process according to an embodiment of the present invention;

[0055] Figure 6 This is an application diagram of the optimal arrangement scheme of flow sensors in a sewage treatment process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following is a detailed description of an embodiment of the present invention. This embodiment is based on the technical solution of the present invention, and provides a detailed implementation method and a specific operation process to further explain the technical solution of the present invention.

[0057] like Figure 1 As shown, the present invention provides a method for optimizing the arrangement of sensors in a sewage treatment process. This embodiment takes the urban sewage treatment process using an activated sludge treatment process as an example. The sewage treatment process process flow chart is shown in FIG. Figure 2 As shown in the figure, the optimal arrangement of flow sensors in the urban sewage treatment process is studied. The method includes the following steps:

[0058] Step 1: Establish a system directed graph model for the sewage treatment process and define the decision variable matrix of sensor layout;

[0059] According to the process flow chart of sewage treatment process and based on graph theory knowledge, the system directed graph model of sewage treatment process is obtained, such as Figure 3 As shown in the figure, nodes a to h represent unit processes (reaction tanks, sedimentation tanks, etc.) and flow connection points in the flow chart, node i represents the environment node, which connects the system boundary point with the external environment to ensure the mass balance of the entire process, and edges 1 to 13 represent the flow between every two nodes in the flow chart.

[0060] The decision variable matrix of sensor layout is defined, and the matrix element value indicates whether a sensor is arranged for each variable to be measured in each stream during the sewage treatment process.

[0061] Step 2: Based on the structural observability and redundancy criteria, a sensor optimization layout model for the entire sewage treatment process is established based on the decision variable matrix and the importance of the variables to be measured in different stages of sewage treatment;

[0062] In this embodiment, the optimization objectives of the sensor optimization layout problem are sensor cost, variable observability and measurement redundancy, where sensor cost refers to the purchase and installation cost of the sensor, variable observability refers to the value of the measured variable being directly measured by the sensor or indirectly calculated from known variables, and variable redundancy refers to the measured variable remaining observable after the sensor is removed. Then, according to the number of sensors installed in the sensor layout scheme, the sensor cost objective function is obtained; according to the structural observability and redundancy criteria, a classification algorithm based on loop analysis is used to mark the observability of all process variables and the redundancy of the measured variables, and the variable observability and measurement redundancy objective functions are obtained.

[0063] For the variable observability objective, it is usually assumed that the importance of all process variables in the objective function is consistent, and the same is true for the measurement redundancy objective, which leads to great flexibility in the optimization process and abundant non-dominated optimal solutions. Moreover, considering the actual situation of the sewage treatment process, the importance of variable observability and measurement redundancy in different treatment links is not exactly the same. Therefore, this embodiment uses the triangular fuzzy hierarchical analysis method to design the weight factor of the flow in the variable observability and measurement redundancy objective functions. The design process is briefly described as follows:

[0064] First, with the goal of evaluating the observable importance of the jth variable to be measured in the sewage treatment process, volatility, reliability and risk as the importance assessment attributes, and the variable to be measured in all flows in the sewage treatment process as the object, a three-layer hierarchy is established; Figure 4 As shown;

[0065] Then, according to the characteristics of the sewage treatment process, the triangular fuzzy multiplication preference relationship matrix of the attributes is constructed. The matrix elements are triangular fuzzy numbers Represents the preference ratio of the relative importance of any two attributes a and b; in addition, based on each attribute, the observability triangular fuzzy multiplication preference relationship matrix of the object is constructed in the same way;

[0066] Then, the contradiction and acceptability of each triangular fuzzy multiplication preference relationship matrix constructed are checked. If the conditions are not met, the matrix element values ​​are adjusted. If the conditions are met, the fuzzy preference programming method is used to derive the optimal fuzzy weight vector from the matrix; and a parameterized defuzzification method is used to obtain the deterministic optimal real number weight vector from the optimal fuzzy weight vector;

[0067] Then obtain the weight vectors of all attributes in the evaluation target and the weight vectors of objects based on each attribute in the evaluation target, and aggregate the local weights at different levels into the global weight of the variable j to be measured on each stream through weighted summation:

[0068]

[0069] Among them, A i is the global weight of the variable j to be measured on the i-th stream; w v is the weight of attribute v, a iv is the weight evaluation value of the variable j to be measured in the i-th stream based on the attribute v; A i The larger the value, the more important the i-th stream is to the measured variable j. The observable importance weight vector of the measured variable j on the N streams of the sewage treatment process is expressed as w g =[A1,A2,...,A n ];

[0070] Finally, the observability weight vectors of all measured variables in the sewage treatment process are obtained according to the above process, and the observable importance matrix of the measured variables is obtained by aggregation.

[0071] The above process is a process of determining the observability matrix of all variables to be measured. The redundant importance matrix of all variables to be measured can be determined by the same method.

[0072] In summary, the weighted multi-objective optimization layout model of flow sensors in the whole process of sewage treatment has been established, which can be expressed as:

[0073]

[0074] Where x is the decision variable matrix describing the sensor layout, where the decision variable x ij Indicates whether the jth variable to be measured in the i-th stream in the sewage treatment process is equipped with a sensor, N is the total number of streams in the sewage treatment process, and M is the total number of variables to be measured in each stream; f C (x) is the sensor installation cost objective function, f O (x) is the variable observability objective function, f R (x) is the measurement redundancy objective function; δ O is the variable observability matrix, δ O,ij is the observability of the jth variable to be measured in the i-th stream; δ R is the measurement redundancy matrix, δ R,ij is the redundancy of the jth variable to be measured in the i-th stream; w C is the sensor cost matrix, w C,ij is the cost of placing sensors for the jth variable to be measured in the i-th stream; w O is the observable importance matrix of the variables to be measured, w O,ij is the observable importance of the jth variable to be measured in the i-th stream; w R is the redundant importance matrix of the variables to be measured, w R,ij is the redundant importance of the jth variable to be measured in the i-th stream; D ij is the decision variable x ij The value range of .

[0075] Step 3: Use a discrete multi-objective state transfer algorithm to solve the sensor optimization layout model of the entire sewage treatment process to obtain the optimal solution set; introduce two update strategies, namely, an efficient non-dominated sorting strategy and a crowding distance diversity maintenance strategy, to quickly and effectively obtain the optimal solution and ensure the distribution of the solution; specifically include:

[0076] Step 3.1: Set the search intensity SE and the maximum allowed number of iterations Maxiter of the discrete multi-objective state transfer algorithm according to the configuration of the sewage treatment plant; randomly generate the initial solution Best of the sensor layout plank ;

[0077] Step 3.2: Best for the current optimal solution k Perform exchange transformation, move transformation, symmetry transformation and substitution transformation respectively to generate SE candidate solutions;

[0078] Step 3.3: Use an efficient non-dominated sorting strategy to compare each candidate solution with the candidate solutions that have been assigned a frontier level, quickly divide the frontier level of the candidate solution set, and the solution with a higher frontier level is better;

[0079] In this strategy, the frontier of each candidate solution is determined one by one until the frontiers of all candidate solutions are determined. Its advantage is that only the solution to be assigned a frontier needs to be compared with the solution that has been assigned to a frontier, which greatly reduces the number of repeated comparisons and improves the computational efficiency.

[0080] Step 3.4: Adopt the crowding distance diversity preservation strategy to calculate the crowding distance of each candidate solution. When two candidate solutions are at the same frontier level, the solution with a larger crowding distance is better; update the current optimal solution Best k ;

[0081] In this embodiment, a crowding distance diversity preservation strategy is used to select solutions to the multi-objective optimization problem, filter out densely distributed candidate solutions, and ensure that the optimal solution to the optimization problem is evenly distributed on the Pareto frontier. The crowding distance of the candidate solution is defined as follows:

[0082]

[0083] Where Q is the number of objective functions; l is the serial number of candidate solutions on a certain frontier; and They are the objective functions f q The maximum and minimum values ​​of .

[0084] Step 3.5: Let k = k + 1. If the number of iterations is less than Maxiter, return to step 3.2 until the termination condition is met.

[0085] In summary, the optimal solution set for the optimal arrangement of flow sensors in the whole process of sewage treatment process in this embodiment is obtained, as follows: Figure 5 As shown in the figure, each point in the figure represents an optimal solution, and all optimal solutions are evenly distributed on the Pareto front.

[0086] Step 4: Based on the evaluation strategy of corrosion degree and maintenance cost, select the only solution suitable for sewage treatment process application from the optimal solution set obtained in step 3. Specifically include:

[0087] Step 4.1: The first part of the evaluation strategy is the average corrosion degree of all sensors in an optimal solution. First, a flow sensor corrosion condition evaluation system is established to evaluate the corrosion degree of flow sensors in different treatment links of sewage treatment plants; six types of corrosion are considered in the corrosion condition evaluation system, including acid corrosion, alkali corrosion, organic corrosion, inorganic corrosion, microbial corrosion and mechanical impact corrosion; in order to quantify the corrosion degree of the sensor, the impact of each corrosion type on the sensor is assigned a score s t ∈{0,1,2,3}(t=1,2,...,6), s t The larger the value, the greater the impact of the corrosion type on the sensor. In addition, according to the super-additive effect of corrosion, that is, the sum of the single corrosion effects of each corrosion factor is less than the combined effect of all factors, a corrosion effect compensation coefficient η is defined; the calculation formula for the average corrosion degree of all sensors in an optimal solution is as follows:

[0088]

[0089] Among them, x p is an optimal solution in the optimal solution set; K is the total number of solutions in the optimal solution set; N is the total number of streams in the sewage treatment plant; M is the total number of variables to be measured in each stream; λ is the weight vector of the six corrosion types.

[0090] In this embodiment, the total corrosion score vector of the corrosion condition evaluation system for the flow sensor in the whole process of sewage treatment is F corrosion =[14.04,15.6,15.6,14.04,14.04,10.92,10.92,4.68,10.92,10.92,10.92,14.04,9.36].

[0091] Step 4.2: The second part of the evaluation strategy is the average maintenance cost of all sensors in an optimal solution. In order to keep the sensors in good working condition and thus ensure the authenticity and sustainability of the measurement data, the sensors need to be maintained, mainly in the form of cleaning, calibration and repair; the average maintenance cost of all sensors in an optimal solution is calculated as follows:

[0092]

[0093] Among them, C j is the basic maintenance cost of different types of sensors; ξ is the maintenance difficulty coefficient of sensors in different processing links; f is the maintenance frequency of sensors.

[0094] In this embodiment, the maintenance cost vector of the flow sensor in the whole process of sewage treatment is M corrosion=[1800,1800,1800,1080,1080,960,960,720,1920,1920,1920,960,1800]

[0095] Step 4.3: Based on the above two local calculation functions, the normalized form of the overall evaluation function of the evaluation strategy is as follows:

[0096]

[0097] Wherein, ω is a weight coefficient based on the configuration and operation experience of the sewage treatment plant, and in this embodiment, ω=0.5.

[0098] For the optimal solution set of the optimal arrangement problem of flow sensors in the whole process of sewage treatment process obtained in step 2 in this embodiment, the evaluation function value of each optimal solution is calculated. The evaluation results are as follows: Figure 5 As shown, each point in the figure represents an optimal solution, and the area of ​​the point represents the size of the evaluation function value of the optimal solution. The optimal solution marked with a five-pointed star is the solution with the smallest evaluation function value in the optimal solution set.

[0099] Step 4.4: Select Figure 5 The solution with the smallest overall evaluation function value is taken as the most suitable solution for further sewage treatment process application. The flow sensor layout scheme corresponding to this solution is as follows: Figure 6 As shown, the objective function values ​​are f C =6, f O =10.24 and f R =8.18. The experimental results show that the scheme achieves the goal of obtaining higher process information richness and reliability at a lower sensor cost, ensuring that the data quality, operating cost and reliability of the sewage treatment plant are at the comprehensive optimal level.

[0100] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive. Therefore, the present invention is not limited to the embodiments described in the specific embodiments. Any other embodiments derived by those skilled in the art based on the technical solution of the present invention that do not depart from the purpose and scope of the present invention, whether modified or replaced, also belong to the protection scope of the present invention.

Claims

1. A method for optimizing the arrangement of sensors in a sewage treatment process, characterized in that: include: Step 1: Establish a system directed graph model for the sewage treatment process and define the decision variable matrix of sensor layout; Step 2: Based on the structural observability and redundancy criteria, establish a sensor optimization layout model for the entire sewage treatment process based on the decision variable matrix; The sensor optimization layout model for the whole process of sewage treatment established in step 2 is a weighted multi-objective optimization model that takes into account the importance of the variables to be measured in different treatment links of the sewage treatment process, and is expressed as: Where x is the decision variable matrix describing the sensor layout, where the decision variable x ij Indicates whether the jth variable to be measured in the i-th stream in the sewage treatment process is equipped with a sensor, N is the total number of streams in the sewage treatment process, and M is the total number of variables to be measured in each stream; f C (x) is the sensor installation cost objective function, f O (x) is the variable observability objective function, f R (x) is the measurement redundancy objective function; δ O is the variable observability matrix, δ O,ij is the observability of the jth variable to be measured in the i-th stream; δ R is the measurement redundancy matrix, δ R,ij is the redundancy of the jth variable to be measured in the i-th stream; w C is the sensor cost matrix, w C,ij is the cost of placing sensors for the jth variable to be measured in the i-th stream; w O is the observable importance matrix of the variables to be measured, w O,ij is the observable importance of the jth variable to be measured in the i-th stream; w R is the redundant importance matrix of the variables to be measured, w R,ij is the redundant importance of the jth variable to be measured in the i-th stream; D ij is the decision variable x ij The value range of Step 3, using a discrete multi-objective state transfer algorithm, solve the sensor optimization layout model for the entire sewage treatment process to obtain the optimal solution set; Step 4: Based on the evaluation strategy of corrosion degree and maintenance cost, a unique solution suitable for industrial application is selected from the optimal solution set obtained in step 3.

2. The method according to claim 1, characterized in that: The system directed graph model includes a number of nodes and directed edges, wherein the nodes represent different unit processes, flow connection points and environmental nodes in the sewage treatment process, and the directed edges represent the flows between the two end nodes; The elements of the decision variable matrix indicate whether a sensor is arranged for each variable to be measured in each flow.

3. The method according to claim 1, characterized in that The observability means that the value of the variable to be measured can be directly measured by a sensor or can be indirectly calculated from known variables; the redundancy means that when the sensor measuring the variable is removed, the variable still remains observable.

4. The method according to claim 1, characterized in that: The triangular fuzzy analytic hierarchy process is used to determine the observable importance matrix w of the variables to be measured. O and the redundant importance matrix w R , specifically: Firstly, the observable importance of the jth measured variable in the sewage treatment process is evaluated as the goal, volatility, reliability and risk are used as the importance evaluation attributes, and the measured variable of all flows in the sewage treatment process is used as the object to establish a three-layer hierarchical structure; Then, according to the characteristics of the sewage treatment process, the triangular fuzzy multiplication preference relationship matrix of the attributes is constructed. a=1,2,3;b=1,2,3,the matrix elements are triangular fuzzy numbers Represents the preference ratio of the relative importance of any two attributes a and b; in addition, based on each attribute, the observability triangular fuzzy multiplication preference relationship matrix of the object is constructed in the same way; Then, the contradiction and acceptability of each triangular fuzzy multiplication preference relationship matrix constructed are checked. If the conditions are not met, the matrix element values ​​are adjusted. If the conditions are met, the fuzzy preference programming method is used to derive the optimal fuzzy weight vector from the matrix; and a parameterized defuzzification method is used to obtain the deterministic optimal real number weight vector from the optimal fuzzy weight vector; Then obtain the weight vectors of all attributes in the evaluation target and the weight vectors of objects based on each attribute in the evaluation target, and aggregate the local weights at different levels into the global weight of the variable j to be measured on each stream through weighted summation: Among them, A i is the global weight of the variable j to be measured on the i-th stream; w v is the weight of attribute v, a iv is the weight evaluation value of the variable j to be measured in the i-th stream based on the attribute v; A i The larger the value, the more important the i-th stream is to the measured variable j. The observable importance weight vector of the measured variable j on the N streams of the sewage treatment process is expressed as w O,j =[A1,A2,...,A n ]; Finally, the observability weight vectors of all variables to be measured in the sewage treatment process are obtained according to the above process, and the observable importance matrix w of the variables to be measured is obtained by aggregation O ; In addition, according to the above observability matrix w for all variables to be measured O The determination method is to obtain the redundant importance matrix w of all variables to be measured. R .

5. The method according to claim 1, characterized in that In step 3, the discrete multi-objective state transfer algorithm uses a state transformation operator to generate candidate solutions, and introduces an efficient non-dominated sorting strategy and a crowding distance diversity preservation strategy to ensure the effectiveness and distribution of the optimal solution; The specific steps of solving the model to obtain the optimal solution set include: Step 3.1: Set the search intensity SE and the maximum allowed number of iterations Maxiter of the discrete multi-objective state transfer algorithm according to the configuration of the sewage treatment plant; randomly generate the initial solution Best of the sensor layout plan k ; Step 3.2: Best for the current optimal solution k Perform exchange transformation, move transformation, symmetry transformation and substitution transformation respectively to generate SE candidate solutions; Step 3.3: Use an efficient non-dominated sorting strategy to compare each candidate solution with the candidate solutions that have been assigned a frontier level, quickly divide the frontier level of the candidate solution set, and the solution with a higher frontier level is better; Step 3.4: Adopt the crowding distance diversity preservation strategy to calculate the crowding distance of each candidate solution. When two candidate solutions are at the same frontier level, the solution with a larger crowding distance is better; update the current optimal solution Best k ; Step 3.5: Let k = k + 1. If the number of iterations is less than Maxiter, return to step 3.2 until the termination condition is met.

6. The method according to claim 1, characterized in that Step 4 specifically includes: Step 4.1: Establish the corrosion degree calculation function A sensor corrosion condition assessment system is established, which takes into account multiple corrosion types, and each corrosion type corresponds to a score s that can quantify the degree of corrosion of the sensor. t , t is used to distinguish different corrosion types; the sensor corrosion condition evaluation system determines the calculation formula for the average corrosion degree of all sensors in the sensor optimization layout scheme: Among them, x p is any solution in the optimal solution set, E C (x p ) is the optimal arrangement of sensors x p is the average corrosion degree of all sensors in the optimal solution set; K is the total number of solutions in the optimal solution set; N is the total number of streams in the sewage treatment plant; M is the total number of variables to be measured in each stream; λ is the weight vector of all corrosion types; η is the corrosion effect compensation coefficient; T is the number of corrosion types; Step 4.2, construct the maintenance cost calculation function, expressed as: Among them, E M (x p ) is the optimal arrangement of sensors x p Sensor maintenance cost; C j is the basic maintenance cost of different types of sensors; ij and f ij is the maintenance difficulty coefficient and maintenance frequency of the sensor layout for the jth variable to be measured in the i-th stream; Step 4.3: The above corrosion degree and maintenance cost calculation functions are integrated and normalized to obtain the overall evaluation function of the sensor layout scheme: Among them, ω is a weight coefficient based on the configuration and operation experience of the sewage treatment plant; Step 4.4, select the sensor layout scheme with the smallest overall evaluation function value from the optimal solution set, which is the only solution suitable for sewage treatment process application.

7. The method according to claim 6, characterized in that The sensor corrosion condition assessment system includes six types of corrosion: acid corrosion, alkali corrosion, organic corrosion, inorganic corrosion, microbial corrosion and mechanical impact corrosion.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the computer program is executed by the processor, the processor is caused to implement the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images