Automatic drainage method and system for cable trench of wind power plant

Abstract

The invention provides an automatic drainage method and system for a wind power plant cable trench, and relates to the technical field of cable trench drainage, and the method comprises the steps: arranging a liquid level sensor, a temperature sensor and an air pressure sensor in the cable trench, and enabling the liquid level sensor, the temperature sensor and the air pressure sensor to be used for detecting the water level, the temperature and the air pressure in the cable trench respectively; arranging a humidity sensor at the soil closest to the cable trench, wherein the humidity sensor is used for acquiring the humidity of surrounding soil; multi-dimensional data including the water level, the temperature and the air pressure in the cable trench, the surrounding soil humidity and the predicted total precipitation in future U hours are collected periodically; predicting whether the water level has an abnormal rising risk or not through a trained water level prediction model based on the multi-dimensional data of continuous N periods; and performing drainage operation according to whether the abnormal rising risk exists or not and the current cable trench water level. The method has the advantage that predictive and reliable drainage can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of cable trench drainage, and in particular to an automatic drainage method and system for a cable trench in a wind farm. Background Art

[0002] Cable trenches are a crucial component of power transmission lines, providing cable installation and protection. However, due to environmental factors, cable trenches can be susceptible to water accumulation, particularly during heavy rains, groundwater infiltration, or poor drainage. Water accumulation can cause moisture to form in cables, leading to short circuits and even damage, seriously impacting the stability of the power supply system.

[0003] Therefore, cable trench drainage management is a crucial component of power operations and maintenance. Currently, traditional cable trench drainage methods rely primarily on fixed-time drainage or simple level-triggered drainage. When the liquid level reaches a set threshold, the drainage pump starts draining and stops when the water level drops. However, this approach has its problems. Relying solely on level-triggered drainage fails to predict water level changes in advance, potentially leading to a sudden and rapid rise in water levels and an inability to respond in a timely manner. Sometimes, even if the liquid level hasn't reached the threshold, heavy rainfall or external environmental factors may cause the water level to rise rapidly, making it difficult to implement timely measures with traditional methods, making it easy for the cable trench to flood. Furthermore, drainage systems without predictive capabilities may frequently start and stop, resulting in energy waste and reduced system efficiency.

[0004] Therefore, there is an urgent need to improve the drainage scheme of the cable trench to achieve predictable and reliable drainage to avoid the hidden dangers caused by the rapid rise of liquid level and avoid waste of resources. Summary of the Invention

[0005] The object of the present invention is to provide a method and system for automatic drainage of a cable trench in a wind farm, which can achieve predictable and reliable drainage.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for automatically draining a cable trench in a wind farm comprises the following steps:

[0008] A liquid level sensor, a temperature sensor and an air pressure sensor are installed in the cable trench to detect the water level, temperature and air pressure in the cable trench respectively;

[0009] A humidity sensor is provided at the soil closest to the cable trench, and the humidity sensor is used to obtain the surrounding soil humidity;

[0010] Periodically collect multi-dimensional data, including water level, temperature, air pressure in the cable trench, surrounding soil moisture, and predicted total precipitation in the next U hours;

[0011] Based on the multi-dimensional data of N consecutive periods, the trained water level prediction model is used to predict whether there is a risk of abnormal water level rise;

[0012] Drainage operations are carried out based on whether there is a risk of abnormal water rise and the current cable trench water level.

[0013] Preferably, the method for periodically collecting multidimensional data is: collecting the multidimensional data at multiple fixed time points every day.

[0014] Preferably, the predicted total precipitation in the next multiple hours is the predicted total precipitation within 24 hours from the collection time point of the multi-dimensional data.

[0015] Preferably, the predicted total precipitation is obtained through meteorological forecast data.

[0016] Preferably, the method for predicting whether there is a risk of abnormal water level rise by using a trained water level prediction model based on multi-dimensional data of multiple consecutive periods is:

[0017] The multidimensional data is input into the input layer of the water level prediction model to form the input vector X at the current time t:

[0018] X=[X t ,X t-1 ,X t-2 ,…,X t-N-1 ] T ;

[0019] X t-i =[Te t-i ,P t-i ,H t-i ,R t-i ], i=0,1,2,…,N-1;

[0020] Among them, the input vector X at time ti t-i In Te t-i ,P t-i ,H t-i and R t-i They are the water level, temperature, air pressure, surrounding soil moisture in the cable trench collected at time ti and the predicted total precipitation in the next time period. T represents the transpose;

[0021] Perform feature extraction through the feature extraction layer of the water level prediction model to obtain a feature vector X″;

[0022] Based on the feature vector X″, the output layer of the water level prediction model predicts whether there is an abnormal rise risk.

[0023] Preferably, the method for performing feature extraction through the feature extraction layer of the water level prediction model is:

[0024] Get the weighted eigenvector X′:

[0025]

[0026] Among them, α i is the weight of time ti, e is a natural constant, τ is a time parameter, and the larger τ is, the greater the influence of data closer to time t;

[0027] Through nonlinear operation, the eigenvector X″ is obtained:

[0028] X″=tanh(β×X′+γ);

[0029] Among them, tanh is the hyperbolic tangent function, β and γ are the nonlinear weight and bias respectively.

[0030] Preferably, the method for obtaining the time parameter τ is:

[0031]

[0032] Among them, y i is the intermediate parameter, var represents the variance, and τ0 is the basic time parameter value and is an empirical value.

[0033] Preferably, the method for predicting whether there is an abnormal rise risk through the output layer of the water level prediction model is:

[0034]

[0035] Among them, ρ is the probability of abnormal rising risk, π k is the training weight, X″(k) is the kth element in X″, and sigmoid[.] represents the sigmoid function.

[0036] Preferably, the method of performing drainage operation according to whether there is an abnormal rising risk and the current cable trench water level is:

[0037] A first water level and a second water level are preset, and the first water level is greater than the second water level;

[0038] When the risk of abnormal water rise is predicted, if the current cable trench water level is lower than the second water level, no operation will be performed; otherwise, drainage will be performed;

[0039] When it is predicted that there is no risk of abnormal rise, if the current cable trench water level is lower than the first water level, no operation will be performed; otherwise, drainage will be carried out.

[0040] The present invention further provides a wind farm cable trench automatic drainage system, which is applied to the above-mentioned wind farm cable trench automatic drainage method, comprising:

[0041] Measuring device, a liquid level sensor, a temperature sensor and an air pressure sensor are set in the cable trench, and the liquid level sensor, the temperature sensor and the air pressure sensor are used to detect the water level, temperature and air pressure in the cable trench respectively;

[0042] A humidity sensor is provided at the soil closest to the cable trench, and the humidity sensor is used to obtain the surrounding soil humidity;

[0043] Data acquisition module, used to periodically collect multi-dimensional data, including water level, temperature, air pressure in the cable trench, surrounding soil moisture and predicted total precipitation in the next few hours;

[0044] The water level prediction module is used to predict whether there is a risk of abnormal water level rise based on multi-dimensional data of N consecutive periods using a trained water level prediction model;

[0045] The drainage module is used to perform drainage operations based on whether there is an abnormal rise risk and the current cable trench water level.

[0046] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0047] The present invention makes dynamic decisions by predicting water level trends, comprehensively considering current water levels and future risks, and ensuring that drainage operations are carried out at the appropriate time with foresight. This can not only avoid difficulties in timely drainage caused by sudden water level rises, but also avoid frequent starting and stopping of water pumps or unnecessary drainage operations, thereby improving system stability and energy efficiency.

[0048] The present invention uses liquid level sensors, temperature sensors, air pressure sensors, and soil moisture sensors to monitor the cable trench environment in real time. Combined with future precipitation, this method can build an accurate water level prediction model with higher prediction accuracy compared to traditional methods that rely on a single data source.

[0049] The present invention is applicable to cable trench drainage management in different types of areas and can be adaptively adjusted according to local rainfall patterns and environmental characteristics, with strong adaptability.

[0050] When the present invention predicts that there is no risk of abnormal water level rise in the future, a higher first water level is used as the threshold for drainage, further avoiding the waste of resources caused by frequent starting and stopping of drainage. When the risk of abnormal water level rise in the future is predicted, a lower second water level is used as the evaluation standard for drainage, thereby reducing the risk of rapid water accumulation in the future.

[0051] The present invention has a reasonable design and high cost-effectiveness. Through intelligent drainage management, it reduces the need for manual intervention and lowers long-term maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A schematic flow chart of a method for automatically draining a cable trench in a wind farm provided in Example 1 of the present invention;

[0053] Figure 2 This is a schematic diagram of the principle of the wind farm cable trench automatic drainage system provided by Example 2 of the present invention. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0055] Example 1

[0056] This embodiment provides a method for automatically draining cable trenches in wind farms. Figure 1 , including the following steps:

[0057] Install sensors: Install liquid level sensors, temperature sensors, and air pressure sensors in the cable trench to detect the water level, temperature, and air pressure in the cable trench, respectively.

[0058] A humidity sensor is provided at the soil closest to the cable trench, and the humidity sensor is used to obtain the surrounding soil humidity;

[0059] Multi-dimensional data collection: Periodically collect multi-dimensional data, including water level, temperature, air pressure, surrounding soil moisture, and total precipitation forecast for the next U hours in the cable trench;

[0060] Predict abnormal water level rise risk: Based on multi-dimensional data from N consecutive periods, use the trained water level prediction model to predict whether there is a risk of abnormal water level rise;

[0061] Conduct drainage operations: Conduct drainage operations based on whether there is a risk of abnormal water rise and the current cable trench water level.

[0062] Furthermore, the predicted total precipitation in the next multiple hours is the predicted total precipitation within 24 hours from the collection time point of the multi-dimensional data.

[0063] In addition, the predicted total precipitation is obtained through meteorological forecast data.

[0064] This embodiment makes intelligent dynamic decisions by predicting water level trends. It obtains drainage strategies based on a comprehensive consideration of the current water level status and the risk of abnormal water level rises in the future, thereby achieving predictive drainage management. Compared with the traditional passive drainage method of detecting water levels and comparing them with thresholds to determine whether to drain, this embodiment can start drainage in advance before the water level is about to rise abnormally, effectively avoiding the problem of untimely drainage caused by sudden water level surges, and avoiding safety hazards such as equipment damage and degradation of cable insulation performance caused by water accumulation in cable trenches. By using liquid level sensors, temperature sensors, air pressure sensors, and soil moisture sensors, and combining them with future precipitation forecast data, the internal environment and external meteorological conditions of the cable trench are monitored in real time in multiple dimensions. This multi-source data analysis method can provide more comprehensive and accurate water level trend forecasts, reduce misjudgments and delayed response problems, and improve the intelligence level and decision-making reliability of the drainage system.

[0065] In this embodiment, the method for periodically collecting multi-dimensional data is: collecting the multi-dimensional data at multiple fixed time points every day.

[0066] As a preferred solution of this embodiment, the method for predicting whether there is a risk of abnormal water level rise by using a trained water level prediction model based on multi-dimensional data of multiple consecutive periods is:

[0067] The multidimensional data is input into the input layer of the water level prediction model to form the input vector X at the current time t:

[0068] X=[X t ,X t-1 ,X t-2 ,…,X t-N-1 ] T ;

[0069] X t-i =[Te t-i ,P t-i ,H t-i ,R t-i ], i=0,1,2,…,N-1;

[0070] Among them, the input vector X at time ti t-i In Te t-i ,P t-i ,H t-i and R t-i They are the water level, temperature, air pressure, surrounding soil moisture in the cable trench collected at time ti and the predicted total precipitation in the next time period. T represents the transpose;

[0071] Perform feature extraction through the feature extraction layer of the water level prediction model to obtain a feature vector X″;

[0072] Based on the feature vector X″, the output layer of the water level prediction model predicts whether there is an abnormal rise risk.

[0073] Specifically, the method for performing feature extraction through the feature extraction layer of the water level prediction model is:

[0074] Get the weighted eigenvector X′:

[0075]

[0076] Among them, α i is the weight of time ti, e is a natural constant, τ is a time parameter, and the larger τ is, the greater the influence of data closer to time t;

[0077] Through nonlinear operation, the eigenvector X″ is obtained:

[0078] X″=tanh(β×X′+γ);

[0079] Among them, tanh is the hyperbolic tangent function, β and γ are the nonlinear weight and bias respectively.

[0080] On this basis, the method for obtaining the time parameter τ is:

[0081]

[0082] Among them, y i is the intermediate parameter, var represents the variance, and τ0 is the basic time parameter value and is an empirical value.

[0083] Next, the method for predicting whether there is an abnormal rise risk through the output layer of the water level prediction model is as follows:

[0084]

[0085] Among them, ρ is the probability of abnormal rising risk, π k is the training weight, X″(k) is the kth element in X″, and sigmoid[.] represents the sigmoid function.

[0086] This embodiment uses multi-dimensional data input based on multiple consecutive cycles to predict water levels, and uses a trained water level prediction model to perform feature extraction and risk prediction. It can combine historical trend information to construct time series features, thereby more comprehensively reflecting the changes in parameters that affect water level changes. When performing feature extraction, different weights are assigned to input vectors at different time points, specifically, the closer to the current moment, the heavier the weight. Because the change in the water level in the cable trench is dynamically affected by factors such as rainfall, infiltration, and evaporation, recent data often has a more direct impact on the current water level status. This embodiment gives higher weights to the most recent moments, allowing the model to more keenly capture the latest changing trends in the water level, while not ignoring the impact of historical data from more distant times, thereby improving the accuracy and real-time nature of the risk prediction of abnormal water level rises.

[0087] In addition, since the larger the time parameter τ in the weight is, the greater the influence of the data closer to the time t is, this embodiment is set so that the larger the variance of the input vector between different time points, the larger the value of τ. When the variance of the input vector is large, it means that the changes in environmental factors such as water level, temperature, air pressure, and soil moisture are more drastic, and the system is in a relatively unstable state. At this time, a higher weight is given to the current time point, so that the prediction pays more attention to the changing trend at the current moment and reduces the interference of historical data.

[0088] Finally, the method for performing drainage operations based on whether there is an abnormal rise risk and the current cable trench water level is as follows:

[0089] A first water level and a second water level are preset, and the first water level is greater than the second water level;

[0090] When the risk of abnormal water rise is predicted, if the current cable trench water level is lower than the second water level, no operation will be performed; otherwise, drainage will be performed;

[0091] When it is predicted that there is no risk of abnormal rise, if the current cable trench water level is lower than the first water level, no operation will be performed; otherwise, drainage will be carried out.

[0092] This embodiment adopts a dynamic water level threshold control strategy, adaptively adjusting the drainage triggering criteria based on the predicted results of future water level changes. Specifically, when it is predicted that there will be no risk of abnormal water accumulation in the future, the system uses a higher first water level threshold for drainage to avoid ineffective energy consumption and frequent starting and stopping of water pumps caused by premature drainage, thereby improving energy utilization. When it is predicted that there may be a risk of abnormal water level rise in the future, the system actively lowers the drainage triggering threshold and uses a lower second water level threshold for early drainage, thereby reserving sufficient drainage space before the arrival of rainfall or water seepage peak, effectively reducing the risk of a sharp rise in water levels in a short period of time, and improving the initiative and safety of overall drainage management.

[0093] Example 2

[0094] This embodiment provides a wind farm cable trench automatic drainage system, see Figure 2 , a wind farm cable trench automatic drainage method applied to the above embodiment includes:

[0095] Measuring device, a liquid level sensor, a temperature sensor and an air pressure sensor are set in the cable trench, and the liquid level sensor, the temperature sensor and the air pressure sensor are used to detect the water level, temperature and air pressure in the cable trench respectively;

[0096] A humidity sensor is provided at the soil closest to the cable trench, and the humidity sensor is used to obtain the surrounding soil humidity;

[0097] Data acquisition module, used to periodically collect multi-dimensional data, including water level, temperature, air pressure in the cable trench, surrounding soil moisture and predicted total precipitation in the next few hours;

[0098] The water level prediction module is used to predict whether there is a risk of abnormal water level rise based on multi-dimensional data of N consecutive periods using a trained water level prediction model;

[0099] The drainage module is used to perform drainage operations based on whether there is an abnormal rise risk and the current cable trench water level.

[0100] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for automatic drainage of cable trenches in wind farms, characterized in that: The following steps are involved: A liquid level sensor, a temperature sensor and an air pressure sensor are installed in the cable trench to detect the water level, temperature and air pressure in the cable trench respectively; A humidity sensor is provided at the soil closest to the cable trench, and the humidity sensor is used to obtain the surrounding soil humidity; Periodically collect multi-dimensional data, including water level, temperature, air pressure in the cable trench, surrounding soil moisture, and predicted total precipitation in the next U hours; Based on the multi-dimensional data of N consecutive periods, the trained water level prediction model is used to predict whether there is a risk of abnormal water level rise; Drainage operations are carried out based on whether there is a risk of abnormal water rise and the current cable trench water level.

2. The automatic drainage method for a wind farm cable trench according to claim 1, characterized in that: The method for periodically collecting multidimensional data is: collecting the multidimensional data at multiple fixed time points every day.

3. The automatic drainage method for cable trenches in wind farms according to claim 2, characterized in that: The predicted total precipitation in the next multiple hours is the predicted total precipitation within 24 hours from the collection time point of the multi-dimensional data.

4. The automatic drainage method for cable trenches in wind farms according to claim 3, characterized in that: The predicted total precipitation is obtained through weather forecast data.

5. The automatic drainage method for cable trenches in wind farms according to claim 1, characterized in that: The method for predicting whether there is a risk of abnormal water level rise by using a trained water level prediction model based on multi-dimensional data of multiple consecutive periods is: The multidimensional data is input into the input layer of the water level prediction model to form the input vector X at the current time t: X=[X t ,X t-1 ,X t-2 ,…,X t-N-1 ] T ; X t-i =[The t-i ,P t-i ,H t-i ,R t-i ], i=0,1,2,…,N-1; Among them, the input vector X at time ti t-i In Te t-i ,P t-i ,H t-i and R t-i They are the water level, temperature, air pressure, surrounding soil moisture in the cable trench collected at time ti and the predicted total precipitation in the next time period. T represents the transpose; Perform feature extraction through the feature extraction layer of the water level prediction model to obtain a feature vector X″; Based on the feature vector X″, the output layer of the water level prediction model predicts whether there is an abnormal rise risk.

6. The method for automatic drainage of a cable trench in a wind farm according to claim 5, characterized in that: The method for performing feature extraction through the feature extraction layer of the water level prediction model is as follows: Get the weighted eigenvector X′: Among them, α i is the weight of time ti, e is a natural constant, τ is a time parameter, and the larger τ is, the greater the influence of data closer to time t; Through nonlinear operation, the eigenvector X″ is obtained: X" = tanh(β × X' + γ); Among them, tanh is the hyperbolic tangent function, β and γ are the nonlinear weight and bias respectively.

7. The method for automatic drainage of a cable trench in a wind farm according to claim 6, characterized in that: The method for obtaining the time parameter τ is: Among them, y i is the intermediate parameter, var represents the variance, and τ0 is the basic time parameter value and is an empirical value.

8. The method for automatic drainage of a cable trench in a wind farm according to claim 7, characterized in that: The method for predicting whether there is an abnormal rise risk through the output layer of the water level prediction model is: Among them, ρ is the probability of abnormal rising risk, π k is the training weight, X "" (k) is X "′ The kth element in , sigmoid[.] represents the sigmoid function.

9. The method for automatic drainage of a cable trench in a wind farm according to claim 1, characterized in that: The method for performing drainage operations based on whether there is an abnormal rise risk and the current cable trench water level is as follows: A first water level and a second water level are preset, and the first water level is greater than the second water level; When the risk of abnormal water rise is predicted, if the current cable trench water level is lower than the second water level, no operation will be performed; otherwise, drainage will be performed; When it is predicted that there is no risk of abnormal rise, if the current cable trench water level is lower than the first water level, no operation will be performed; otherwise, drainage will be carried out.

10. An automatic drainage system for cable trenches in wind farms, applied to an automatic drainage method for cable trenches in wind farms according to any one of claims 1 to 9, characterized in that: include: Measuring device, a liquid level sensor, a temperature sensor and an air pressure sensor are set in the cable trench, and the liquid level sensor, the temperature sensor and the air pressure sensor are used to detect the water level, temperature and air pressure in the cable trench respectively; A humidity sensor is provided at the soil closest to the cable trench, and the humidity sensor is used to obtain the surrounding soil humidity; Data acquisition module, used to periodically collect multi-dimensional data, including water level, temperature, air pressure in the cable trench, surrounding soil moisture and predicted total precipitation in the next few hours; The water level prediction module is used to predict whether there is a risk of abnormal water level rise based on multi-dimensional data of N consecutive periods using a trained water level prediction model; The drainage module is used to perform drainage operations based on whether there is an abnormal rise risk and the current cable trench water level.

Images