Energy-saving control method and system for water pump operation based on multi-source data analysis

Abstract

The present invention discloses a water pump operation energy-saving control method and system based on multi-source data analysis, which relates to the technical field of water pump energy saving. By obtaining the measured state parameters and the corresponding water pump energy consumption value data at the same moment, abnormal data points in the sampled data are eliminated, and the neural network is trained using the sampled data until the expected value of the mean square error of its prediction result is less than the set expected threshold, and the mapping relationship between the state parameters and the water pump energy consumption value is output. Using the mapping relationship, the adjustable range of the active state parameters and the range of the water pump energy consumption value corresponding to the measured values of the passive state parameters are obtained at the current moment. Abnormal detection is performed on each water pump energy consumption value in the range with λ as the detection step length to obtain the reasonable range of the water pump energy consumption value at the current moment. The value of the active state parameter corresponding to the upper limit of the reasonable range of the water pump energy consumption value at the current moment is taken as the optimal state value of the state parameter, and the active state parameter is adjusted.

Description

Technical Field

[0001] The present invention relates to the technical field of water pump energy saving, and particularly relates to a water pump operation energy saving control method and system based on multi-source data analysis. Background Technique

[0002] As one of the indispensable devices in industrial and agricultural production, water pumps consume a large proportion of the total energy consumption in the entire production process. Traditional water pumps consume a large amount of energy during long-term operation, directly resulting in energy waste and high operation costs. With the continuous rise of energy prices, enterprises and institutions are increasingly feeling the increasing pressure of energy costs and are urgently in need of finding effective ways to reduce energy consumption. Water pump energy saving transformation can optimize the production process while reducing energy consumption, achieving effective cost control. In addition, the high energy consumption of traditional water pumps not only increases greenhouse gas emissions but also has a negative impact on the rational use of water resources. Through water pump energy saving transformation, energy consumption can be reduced, environmental pollution can be mitigated, and the green development goals of enterprises can be achieved.

[0003] During the operation of a water pump, the heat generated by its internal components is one of the important factors affecting energy consumption increase. Generally, a matching cooling fan is configured for the water pump to reduce its operating temperature and achieve the purpose of energy saving and consumption reduction. On the one hand, the higher the operating frequency of the cooling fan of the water pump motor, the better the cooling effect usually is, which is beneficial to reducing energy consumption; on the other hand, the higher the operating frequency of the cooling fan of the water pump motor, the more power support is required, resulting in an increase in overall energy consumption. In addition, the higher the operating frequency, the louder the noise will be, and too high a frequency will impose an additional burden on the motor and shorten its service life.

[0004] At present, for the control of the operating frequency of the cooling fan of the water pump motor, it is usually considered based on a single influencing factor of the impact of the cooling effect on energy consumption. It does not consider the influence between the operating frequency of the cooling fan and the comprehensive energy consumption, and the fineness and intelligence level in the control process are relatively low, resulting in poor control effects and often unable to determine the optimal operating frequency of the cooling fan during the operation of the water pump. For this reason, we propose a water pump operation energy saving control method and system based on multi-source data analysis. Summary of the Invention

[0005] The main object of the present invention is to provide a water pump operation energy saving control method and system based on multi-source data analysis, which can effectively solve the problems in the background technique.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A water pump operation energy saving control method based on multi-source data analysis includes:

[0008] Sample the state parameters during the operation of the water pump to obtain the measured $i$-th state parameter $x$ at the same moment i and the corresponding water pump energy consumption value $E_f$ i of the historical data, and perform anomaly detection on the obtained water pump energy consumption value $E_f$ i to eliminate the abnormal data points of the water pump energy consumption value $E_f$ i and the corresponding state parameter $x$ i where $i = 1, 2, \cdots, n$; $n$ is the type of state parameters;

[0009] Use the normalized state parameters as input parameters, and use the normalized water pump energy consumption value as the output parameter to construct a neural network model, and use the sampled data to train the neural network model until the expected value $E$ of the squared error of its prediction result p is less than the set expected threshold $\varepsilon$, and use the trained neural network model to output the mapping relationship $f$ between the state parameter $x$ i and the water pump energy consumption value $E_f$ i The expression is: $f: ((x_1, x_1, \cdots, x$ i $)$ T ) \to E_f$ i ,

[0010] Classify the state parameters into active state parameters and passive state parameters according to whether they are controlled objects, and obtain the adjustable range of the $u$-th active state parameter and the measured value of the $v$-th passive state parameter at the current $t$ moment where is the lower limit of the adjustable range of the $u$-th active state parameter at the current $t$ moment; is the upper limit of the adjustable range of the $u$-th active state parameter at the current $t$ moment, where $u, v \in n$, and $u + v = n$;

[0011] Use the mapping relationship $f$ to obtain the adjustable range of the active state parameter and the measured value of the passive state parameter at the current $t$ moment t corresponding to the initial interval of the water pump energy consumption value $E_f$ is the minimum value of the water pump energy consumption value at the current $t$ moment, is the maximum value of the water pump energy consumption value at the current $t$ moment;

[0012] Perform anomaly detection on each water pump energy consumption value in the initial interval with $\lambda$ as the detection step size, and obtain the water pump energy consumption value $E_f$ at the current $t$ moment according to the detection resultt Reasonable range Among them, is the pump energy consumption value Ef t Lower limit of the reasonable range; is the pump energy consumption value Ef t Upper limit of the reasonable range;

[0013] Take the active state parameter value corresponding to the upper limit of the reasonable range of the pump energy consumption value Ef at the current time t as the optimal state value of the state parameter, and adjust the u-th active state parameter at the current time t to the optimal state value. t Upper limit of the reasonable range Take the active state parameter value corresponding to the upper limit of the reasonable range of the pump energy consumption value Ef at the current time t as the optimal state value of the state parameter, and adjust the u-th active state parameter at the current time t to the optimal state value.

[0014] The state parameters include the operating frequency of the cooling fan of the pump motor. The state parameters also include at least one of the ambient temperature, ambient humidity, three-phase winding terminal temperature of the pump motor, front bearing temperature of the pump motor, and rear bearing temperature of the pump motor. Among them, the operating frequency of the cooling fan of the pump motor can only be used as an active state parameter, and there is at least one passive state parameter.

[0015] Expected value of mean squared error E p The calculation formula is:

[0016]

[0017] In the formula, is the ideal output value of the q-th prediction result in the neural network training sample; Y q is the actual output value of the q-th prediction result in the neural network training sample; N is the total amount of data in the neural network training sample.

[0018] State parameter The calculation formula is: In the formula, x ij is the j-th sampling value of the i-th state parameter; Q is the total amount of data sampling; Pump energy consumption value The calculation formula is: In the formula, Ef ij is the pump energy consumption value corresponding to the j-th sampling value of the i-th state parameter.

[0019] The process of anomaly detection specifically includes the following steps:

[0020] Step S11: Among the sampling values of the pump energy consumption value, set the k-th sampling point with the smallest distance from the r-th sampling point as r ∈ Q, Q is the total amount of data sampling; According to the formula: Calculate the k-nearest neighbor distance of the r-th sampling point ​

[0021] Step S12: Calculate the distances between the r-th sampling point and all the other sampling points respectively to calculate the set of sampling points with distance values less than the k-nearest neighbor distance as the k-distance neighborhood of the r-th sampling point

[0022] Step S13: Take the maximum value among the distances from the k-th sampling point closest to the sampling point in the k-distance neighborhood as the k-th reachable distance between the r-th sampling point and the k-th sampling point

[0023] Step S14: Calculate the local reachability density ρ i r of the r-th sampling point Ef k (Ef i r ), where the calculation formula for the local reachability density is:

[0024] Step S15: Calculate the local outlier factor of the r-th sampling point Ef i r according to the obtained local reachability density, and the calculation formula is: In the formula, ρ k (Ef i k ) is the local reachability density of the k-th sampling point Ef i k ;

[0025] Step S16: Use the obtained local outlier factor to determine the abnormality of the r-th sampling point , and the determination principle is:

[0026] When it means that the r-th sampling point is abnormal data;

[0027] When it means that the r-th sampling point is normal data.

[0028] The determination process of the reasonable interval of the pump energy consumption value Ef t includes the following steps:

[0029] Step S51: Use the abnormal detection process to detect the lower limit of the interval with the detection step ​​​Perform detection and determination whether it is normal data;

[0030] Step S52: When it is determined that the lower limit is normal data, increase a step size λ, and use the increased data as the data to be detected, and determine whether the current data to be detected is normal data;

[0031] Step S53: Repeat Step S52. If the data to be detected is determined to be abnormal data, use the interval where the data to be detected is determined to be normal data as the reasonable interval of the pump energy consumption value Ef t of.

[0032] The pump operation energy-saving control system based on multi-source data analysis includes a historical data acquisition module, a data processing module, a neural network construction module, a real-time data acquisition module, an energy consumption interval acquisition module, a reasonable adjustment interval acquisition module, and a state parameter adjustment module;

[0033] The historical data acquisition module is used to sample the state parameters during the operation of the pump to obtain the measured ith state parameter x i and the corresponding pump energy consumption value Ef i of historical data;

[0034] The data processing module is used to perform anomaly detection on the obtained pump energy consumption value Ef i and eliminate the abnormal data points of the pump energy consumption value Ef i and the corresponding state parameter x i ;

[0035] The neural network construction module is used to use the normalized state parameters as input parameters and the normalized pump energy consumption value as output parameters to construct a neural network model, and use the sampling data to train the neural network model until the expected value E of the squared error of its prediction result p is less than the set expected threshold ε, and use the trained neural network model to output the mapping relationship f i between the state parameter x i and the pump energy consumption value Ef;

[0036] The real-time data acquisition module is used to obtain the adjustable interval of the u-th active state parameter and the measured value

[0037] The energy consumption interval acquisition module is used to obtain the adjustable interval of the active state parameter at the current time t by using the mapping relationship f. and the measured value of the passive state parameter corresponding to the pump energy consumption value Ef t of the initial interval

[0038] The reasonable adjustment interval acquisition module is used to perform anomaly detection on each pump energy consumption value in the initial interval with λ as the detection step, and obtain the reasonable interval of the pump energy consumption value Ef t at the current time t according to the detection result.

[0039] The state parameter adjustment module is used to use the upper limit of the reasonable interval of the pump energy consumption value Ef t at the current time t corresponding active state parameter value as the optimal state value of the state parameter, and adjust the u-th active state parameter at the current time t to the optimal state value.

[0040] The system includes a memory, a processor, and a computer program stored on the memory and executable on the processor.

[0041] The present invention has the following beneficial effects.

[0042] Compared with the prior art, by sampling the state parameters during the operation of the pump, the measured i-th state parameter x i at the same moment and the corresponding pump energy consumption value Ef i historical data are obtained, anomaly detection is performed on the obtained pump energy consumption value Ef i , and the abnormal data points of the pump energy consumption value Ef i and the corresponding state parameter x i in the sampling data are eliminated. Using the normalized state parameter as the input parameter and the normalized pump energy consumption value as the output parameter to construct a neural network model, and using the sampling data to train the neural network model until the expected value E p of the squared error of its prediction result is less than the set expected threshold ε, using the trained neural network model to output the mapping relationship f between the state parameter x i and the pump energy consumption value Ef i , classifying the state parameters into active state parameters and passive state parameters according to whether they are controlled objects, obtaining the adjustable interval of the u-th active state parameter at the current time t Use the mapping relationship f to obtain the adjustable range of the active state parameter at the current time t and measured values ​​of passive state parameters The corresponding pump energy consumption value Ef t The initial interval Take λ as the detection step size for the initial interval The energy consumption value of each water pump in the test is detected abnormally, and the energy consumption value Ef of the water pump at the current time t is obtained according to the detection results. t Reasonable range Take the energy consumption value of the water pump Ef at the current time t t The upper limit of the reasonable range The corresponding active state parameter value is used as the optimal state value of the state parameter, and the u-th active state parameter at the current time t is adjusted to the optimal state value, thereby improving the precision and intelligence of the control process, keeping the cooling fan of the water pump in the optimal operating frequency range during operation, and improving the energy-saving effect of the water pump operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flow chart of the water pump operation energy-saving control method based on multi-source data analysis of the present invention;

[0044] Figure 2 It is a structural block diagram of the water pump operation energy-saving control system based on multi-source data analysis of the present invention;

[0045] Figure 3 This is a structural diagram of the neural network constructed in the solution of the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0047] The specific implementation process of the technical solution of the present invention includes the following steps:

[0048] Step 1: Sample the state parameters of the water pump during operation and obtain the i-th state parameter x measured at the same time i And the corresponding pump energy consumption value Ef i Historical data of, where i = 1, 2, ..., n; n is the type of state parameter;

[0049] Among them, the state parameters include the operating frequency of the cooling fan of the water pump motor, and the state parameters also include at least one of the ambient temperature, ambient humidity, three-phase winding terminal temperature of the water pump motor, front bearing temperature of the water pump motor, and rear bearing temperature of the water pump motor.

[0050] In this embodiment, taking the state parameters including the operating frequency of the cooling fan of the water pump motor, ambient temperature, ambient humidity, three-phase winding terminal temperature of the water pump motor, front bearing temperature of the water pump motor, and rear bearing temperature of the water pump motor as an example for illustration.

[0051] Step 2: Perform anomaly detection on the obtained water pump energy consumption value Ef i and eliminate the abnormal data points of the water pump energy consumption value Ef i and the corresponding state parameter x i ; the specific process is as follows:

[0052] Step S21: Among the sampled values of the water pump energy consumption value, the kth sampling point with the smallest distance from the rth sampling point is set as r ∈ Q, where Q is the total amount of data sampling; according to the formula: Calculate the k-nearest neighbor distance of the rth sampling point

[0053] Step S22: Calculate the distances between all the other sampling points and the rth sampling point respectively, and use the set of sampling points with the calculated distance values less than the k-nearest neighbor distance as the k-distance neighborhood of the rth sampling point

[0054] Step S23: Take the maximum value of the distances between the k-nearest neighbor distance and the k-distance neighborhood to the kth sampling point closest to the sampling point as the k-reachable distance between the rth sampling point and the kth sampling point

[0055] Step S24: Calculate the local reachability density ρ i r of the rth sampling point Ef k (Ef i r ), where the calculation formula of the local reachability density is:

[0056] Step S25: Calculate the rth sampling point Ef ir The local outlier factor, the calculation formula is: In the formula, ρ k (Ef i k ) is the local reachability density of the kth sampling point Ef i k ;

[0057] Step S26: Use the obtained local outlier factor to determine the abnormality of the rth sampling point , and the determination principle is:

[0058] When indicates that the rth sampling point is abnormal data;

[0059] When indicates that the rth sampling point is normal data.

[0060] Step 3: Use the normalized state parameters as input parameters, and use the normalized pump energy consumption value as output parameters to construct a neural network model. Among them, the number of neurons in the input layer is equal to the type of input state parameters, the number of output neurons is 1, and the number of neurons in the middle hidden layer can be determined by an empirical formula, specifically: α and β are the number of neurons in the input layer and the output layer respectively. In this embodiment, α = 7, β = 1, and a is an integer from 1 to 9. The structural form of the constructed neural network model is as Figure 3 shown;

[0061] Among them, the state parameter The calculation formula is: In the formula, x ij is the jth sampling value of the ith state parameter; Q is the total amount of data sampling; the pump energy consumption value The calculation formula is: In the formula, Ef ij is the pump energy consumption value corresponding to the jth sampling value of the ith state parameter.

[0062] Step 4: Use the sampling data to train the neural network model until the expected value E of the squared error of its prediction result p is less than the set expected threshold ε, and use the trained neural network model to output the mapping relationship f between the state parameter x i and the pump energy consumption value Ef i , and the expression is: f: ((x1, x1,..., x i ) T ) → Ef i; where the expected value E of the squared error p is calculated as follows:

[0063]

[0064] In the formula, is the ideal output value of the q-th predicted result in the neural network training samples; Y q is the actual output value of the q-th predicted result in the neural network training samples; N is the total amount of data in the neural network training samples.

[0065] Step 5: Classify the state parameters into active state parameters and passive state parameters according to whether they are controlled objects, and obtain the adjustable range of the u-th active state parameter and the measured value of the v-th passive state parameter at the current time t is the lower limit of the adjustable range of the u-th active state parameter at the current time t; is the upper limit of the adjustable range of the u-th active state parameter at the current time t, where u, v ∈ n and u + v = n;

[0066] Among them, the operating frequency of the cooling fan of the water pump motor can only be used as an active state parameter, and there is at least one passive state parameter.

[0067] In this embodiment, the passive state parameters include ambient temperature, ambient humidity, the temperature of the three-phase winding terminals of the water pump motor, the temperature of the front bearing of the water pump motor, and the temperature of the rear bearing of the water pump motor.

[0068] Step 6: Use the mapping relationship f to obtain the adjustable range of the active state parameter and the measured value of the passive state parameter at the current time t t corresponding to the initial interval of the water pump energy consumption value Ef In this embodiment, the adjustable range of the active state parameter is the adjustable range of the operating frequency of the cooling fan of the water pump motor, is the minimum value of the water pump energy consumption value at the current time t,

[0069] Step 7: Perform anomaly detection on each water pump energy consumption value in the initial interval with λ as the detection step size, and obtain the reasonable interval t of the water pump energy consumption value Ef where, is the lower limit of the reasonable interval of the water pump energy consumption value Ef t ; For the energy consumption value Ef of the water pump t The upper limit of the reasonable range;

[0070] The energy consumption value Ef of the water pump t The determination process of the reasonable range includes the following steps:

[0071] Step S71: Use the abnormal detection process to detect the lower limit of the interval for the detection step size To determine Whether it is normal data;

[0072] Step S72: When it is determined that the lower limit Is normal data, increase A step size λ, and use the increased data As the data to be detected, determine whether the current data to be detected Is normal data;

[0073] Step S73: Repeat step S52. If the data to be detected is determined to be abnormal data, use the interval where the data to be detected is determined to be normal data as the reasonable range of the energy consumption value Ef of the water pump t Of the reasonable range.

[0074] It should be noted that when all the energy consumption values of the water pumps in the interval Are normal data, then the reasonable range of the energy consumption value Ef of the water pump t Of the reasonable range Is the interval

[0075] Step 8: Take the value of the active state parameter corresponding to the upper limit of the reasonable range of the energy consumption value Ef of the water pump at the current t moment t As the optimal state value of the state parameter, and adjust the u-th active state parameter at the current t moment to the optimal state value.

[0076] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.​

Claims

1. A method for energy-saving control of pump operation based on multi-source data analysis, characterized in that, Including: Sample the state parameters during the operation of the water pump to obtain the measured $i$-th state parameter $x$ at the same moment i and the corresponding water pump energy consumption value $E_f$ i of the historical data. Perform anomaly detection on the obtained water pump energy consumption value $E_f$ i and remove the abnormal data points of the water pump energy consumption value $E_f$ i and the corresponding state parameter $x$ i from the sampled data, where $i = 1, 2,\cdots, n$; $n$ is the type of state parameters With the state parameters after normalization as input parameters, and the normalized pump energy consumption value Ef i as output parameters to construct a neural network model, and using the sampled data to train the neural network model until the expected value E p of the squared error of its prediction result is less than the set expected threshold ε, and using the trained neural network model to output the state parameter x i and the mapping relationship f between the pump energy consumption value Ef i is expressed as: f: ((x1, x1,..., x i )) T ) → Ef i , Classify the state parameters into active state parameters and passive state parameters according to whether they are controlled objects, and obtain the adjustable range of the u-th active state parameter at the current time t and the measured value of the v-th passive state parameter where is the lower limit of the adjustable range of the u-th active state parameter at the current time t; is the upper limit of the adjustable range of the u-th active state parameter at the current time t, where u, v ∈ n and u + v = n; Obtain the adjustable range of the active state parameter at the current moment t by using the mapping relationship f and the measured value of the passive state parameter corresponding to the pump energy consumption value Ef t of the initial range is the minimum value of the pump energy consumption value at the current moment t, is the maximum value of the pump energy consumption value at the current moment t; With λ as the detection step size for the initial interval Anomaly detection is performed on the energy consumption values of each water pump in it, and according to the detection results, the reasonable interval of the water pump energy consumption value Ef t is obtained at the current time t Among them, is the lower limit of the reasonable interval of the water pump energy consumption value Ef t ; is the upper limit of the reasonable interval of the water pump energy consumption value Ef t ; Obtain the pump energy consumption value Ef at the current moment t t The upper limit of the reasonable range The corresponding active state parameter value is used as the optimal state value of the state parameter, and the u-th active state parameter at the current moment t is adjusted to the optimal state value.

2. The energy-saving control method for pump operation based on multi-source data analysis according to claim 1, wherein The state parameters include the operating frequency of the cooling fan of the water pump motor, and the state parameters further include at least one of the ambient temperature, ambient humidity, three-phase winding terminal temperature of the water pump motor, front bearing temperature of the water pump motor, and rear bearing temperature of the water pump motor. Among them, the operating frequency of the cooling fan of the water pump motor can only be used as an active state parameter, and there is at least one passive state parameter.

3. The energy-saving control method for pump operation based on multi-source data analysis according to claim 1, characterized in that, Expected value E of the mean squared error p The calculation formula is as follows: In the formula, is the ideal output value of the q-th prediction result in the neural network training sample; Y q is the actual output value of the q-th prediction result in the neural network training sample; N is the total amount of data in the neural network training sample.

4. The energy-saving control method for pump operation based on multi-source data analysis according to claim 1, wherein, State parameter The calculation formula is as follows: In the formula, x ij is the j-th sampling value of the i-th state parameter; Q is the total amount of data sampling; the pump energy consumption value The calculation formula is as follows: In the formula, Ef ij is the pump energy consumption value corresponding to the j-th sampling value of the i-th state parameter.

5. The energy-saving control method for pump operation based on multi-source data analysis according to claim 1, wherein The process of anomaly detection specifically includes the following steps: Step S11: Among the sampled values of the pump energy consumption, the k-th sampling point Ef that is closest to the r-th sampling point Ef i r is Ef i k , where r ∈ Q and Q is the total amount of data sampling; According to the formula: calculate the k-nearest neighbor distance dk(Efir) of the r-th sampling point Ef i r. Step S12: Calculate the distances between the r-th sampling point Ef and all the other sampling points respectively, and use the set of sampling points whose calculated distance values are less than the k-nearest neighbor distance d i r (Ef k (Ef i r ) as the k-distance neighborhood N i r (Ef k (Ef i r ); Step S13: Obtain the k-nearest neighbor distance d k (Ef i r ) and the k-distance neighborhood N k (Ef i r ), take the maximum value of the distances from the sampling point Ef i k to the k-th nearest sampling point as the distance from the r-th sampling point Ef i r to the k-th sampling point Ef i k which is the k-reachable distance reach_dist k (Ef i k , Ef i r ); Step S14: Calculate the local reachability density ρ of the r-th sampling point Ef i r where the formula for calculating the local reachability density is: k (Ef i r ), Step S15: Calculate the local outlier factor of the r-th sampling point Ef based on the obtained local reachability density. The calculation formula is as follows: i r where ρ in the formula, ρ k (Ef i k ) is the local reachability density of the k-th sampling point Ef i k ; Step S16: Determine the abnormality of the r-th sampling point Ef using the obtained local outlier factor. i r The determination principle is as follows: When the LOF k (Ef i r ) > 1 indicates that the r-th sampling point Ef i r is abnormal data; When LOF k (Ef i r ) ≤ 1 indicates that the r-th sampling point Ef i r is normal data.

6. The energy-saving control method for pump operation based on multi-source data analysis according to claim 1, characterized in that Pump energy consumption value Ef t The determination process of the reasonable range includes the following steps: Step S51: Detect the lower limit of the interval of the detection step length according to the anomaly detection process and determine whether it is normal data; Step S52: When it is determined that the lower limit is normal data, increase a step size λ, and use the increased data +λ as the data to be detected, and determine whether the currently detected data +λ is normal data; Step S53: Repeat Step S52. If the detected data is determined to be abnormal data, use the interval in which the detected data is determined to be normal data as the pump energy consumption value Ef t reasonable interval.

7. A pump operation energy-saving control system based on multi-source data analysis, characterized in that, The system is used to implement the steps of the water pump operation energy-saving control method based on multi-source data analysis described in any one of claims 1-6, and includes a historical data acquisition module, a data processing module, a neural network construction module, a real-time data acquisition module, an energy consumption range acquisition module, a reasonable adjustment range acquisition module, and a state parameter adjustment module; The historical data acquisition module is used to sample the state parameters during the operation of the water pump, and obtain the historical data of the measured ith state parameter xi and the corresponding water pump energy consumption value Ef at the same moment; i and the corresponding water pump energy consumption value Ef i at the same moment; The data processing module is used to perform anomaly detection on the obtained pump energy consumption value Ef i and eliminate the abnormal data points of the pump energy consumption value Ef i and the corresponding state parameter x i in the sampled data; The neural network construction module is used to use the normalized state parameters as input parameters and the normalized pump energy consumption value as output parameters to construct a neural network model, and use the sampling data to train the neural network model until the expected value E of the squared error of its prediction result p is less than the set expected threshold ε, and use the trained neural network model to output the mapping relationship f i between the state parameter x i and the pump energy consumption value Ef; The real-time data acquisition module is used to obtain the adjustable range of the u-th active state parameter and the measured value of the v-th passive state parameter at the current moment t. and the measured value of the v-th passive state parameter The energy consumption interval obtaining module is used to obtain the adjustable interval of the active state parameter at the current moment t by using the mapping relationship f and the measured value of the passive state parameter corresponding to the pump energy consumption value Ef t of the initial interval The reasonable adjustment range obtaining module is used to perform anomaly detection on the energy consumption values of each water pump in the initial range with λ as the detection step size, and obtain the reasonable range of the water pump energy consumption value Ef at the current moment t according to the detection results t ​ The state parameter adjustment module is used to take the upper limit of the reasonable range of the pump energy consumption value Ef at the current time t t as the optimal state value of the state parameter, and adjust the u-th active state parameter at the current time t to the optimal state value. The corresponding active state parameter value is used as the optimal state value of the state parameter, and the u-th active state parameter at the current time t is adjusted to the optimal state value.

8. The energy-saving control system for pump operation based on multi-source data analysis according to claim 7, characterized in that The system includes a memory, a processor, and a computer program stored on the memory and executable on the processor. Among them, when the processor executes the program, it can implement the steps of the water pump operation energy-saving control method based on multi-source data analysis described in any one of claims 1-6.

Images