A water-cooled high-temperature resistant oil cylinder and its use method
Through the design of water-cooled high-temperature resistant oil cylinder, combined with built-in and external water-cooling components and intelligent algorithms, the problems of low cooling efficiency and complex structure of hydraulic systems in high-temperature environments are solved, and efficient and stable temperature control and energy consumption optimization are achieved. Improves the reliability and life of the system.
Patent Information
- Application Number
- CN202410722990.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-05
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-06-05
AI Technical Summary
The hydraulic system cannot effectively reduce the temperature in a high-temperature environment, resulting in problems such as lower oil viscosity, accelerated aging of seals, leakage of internal hydraulic system, low cooling efficiency, complex structure, difficult maintenance and high cost.
Design a water-cooled high-temperature resistant oil cylinder, which uses built-in and external water-cooling components to achieve efficient cooling through a two-way pump and cooling water tank, and dynamically adjusts the cooling water distribution and bidirectional pump working frequency using temperature sensors and intelligent algorithms, combining LSTM and GRU models Perform temperature prediction and energy consumption optimization.
Effectively reduce the temperature of the hydraulic cylinder, maintain oil viscosity and lubrication performance, improve system reliability and stability, simplify the structure, reduce energy consumption, extend service life, and achieve efficient and stable temperature control.
Smart Images

Figure CN118499307B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of hydraulic equipment, and in particular to a water-cooled high-temperature resistant oil cylinder and a use method thereof. Background Art
[0002] In modern industrial applications, hydraulic systems are widely used in various machinery and equipment due to their high transmission efficiency, flexible configuration and precise control. However, when the hydraulic system is running in a high temperature environment or under continuous high load conditions, the heat generated by the system will increase significantly, causing the oil temperature to rise, thus affecting the stability and working efficiency of the system, and may even cause equipment damage.
[0003] Traditional hydraulic cylinders usually rely on natural heat dissipation or simple heat sinks for cooling. These methods may be sufficient under mild working conditions, but they often cannot effectively reduce the temperature in high temperature environments, resulting in reduced oil viscosity, accelerated aging of seals, and internal leakage of the hydraulic system, which in turn affects the reliability and safety of the entire system. In order to solve these problems, hydraulic cylinders need to adopt more efficient cooling technology. At present, some advanced hydraulic cylinders use special designs, but there are still many limitations, such as low cooling efficiency, complex structure, difficult maintenance, and high cost.
[0004] To sum up, in order to ensure the safe, stable and efficient operation of the hydraulic cylinder, an innovative cooling system design is needed to improve the performance and life of the hydraulic cylinder under high temperature working conditions. Therefore, the present application provides a water-cooled high-temperature resistant cylinder and its use method to meet the needs. Summary of the invention
[0005] The technical problem to be solved by the present invention is to provide a water-cooled high-temperature resistant oil cylinder and a method of using the same to solve the problems raised in the above-mentioned background technology that the temperature cannot be effectively reduced when facing a high-temperature environment, resulting in a decrease in oil viscosity, accelerated aging of seals, internal leakage of the hydraulic system, low cooling efficiency, complex structure, difficult maintenance, and high cost.
[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:
[0007] A water-cooled high-temperature resistant oil cylinder and a method for using the same, including: an oil cylinder, a cooling water tank, the oil cylinder including a piston oil cylinder, a piston cylinder cover, and a piston, the piston cylinder cover is fixedly installed on the side of the piston oil cylinder, the piston is slidably installed in the inner cavity of the piston oil cylinder, a water-cooling cavity is provided inside the piston oil cylinder, and a built-in water-cooling component is fixedly installed in the water-cooling cavity; the built-in water-cooling component includes a connecting water pipe and a two-way pump, one side of the connecting water pipe is fixedly connected to the cooling water tank, the connecting water pipe and the inner cavity of the cooling water tank are mutually communicated, and the two ends of the two-way pump are respectively fixedly connected with a connecting water pipe and an inner cooling pipe, and the inner cooling pipe is evenly distributed by a water pipe distribution plate; an external water-cooling component, the external water-cooling component is fixedly installed on the outside of the piston oil cylinder, the external water-cooling component includes a connecting water pipe and a two-way pump, one side of the connecting water pipe is fixedly connected to the cooling water tank, the connecting water pipe and the inner cavity of the cooling water tank are mutually communicated, and the two ends of the two-way pump are respectively fixedly connected with a connecting water pipe and an external cooling pipe.
[0008] Preferably, an oil outlet is provided on the outer side of the piston cylinder, and the oil outlet is communicated with the inner cavity. The water cooling cavity is located between the outer side and the inner cavity of the piston cylinder, and the water cooling cavity is not communicated with the oil outlet and the inner cavity. Heat dissipation ribs are evenly distributed on the outer side of the piston cylinder, and mounting holes are provided on the sides of the heat dissipation ribs, and an external water cooling component is fixedly mounted in the mounting holes.
[0009] Preferably, an installation step and an installation ring groove are provided on one end surface of the piston cylinder cover, the installation step is fixedly sleeved in the water-cooling cavity, the installation ring groove is fixedly sleeved in the inner cavity side wall, the piston cylinder cover is provided with a water pipe sleeve hole and a piston hole, a cooling water pipe is fixedly sleeved in the water pipe sleeve hole, and a piston rod is slidably installed in the piston hole.
[0010] Preferably, a water cavity and a water cavity are respectively provided inside the cooling water tank, and the water cavity and the water cavity can respectively accommodate multiple water cooling cycle usage of the built-in water cooling component and the external water cooling component. The water volume inside the water cavity is inversely proportional to the water volume when the water cavity is in use.
[0011] Preferably, the inner cooling pipe is composed of a cooling straight pipe, a cooling straight bent pipe, and a cooling bent pipe. The inner cooling pipe is fixedly mounted in the water pipe sleeve hole as a cooling straight pipe. The cooling straight bent pipe and the two ends of the cooling bent pipe are fixedly connected in sequence. One end face of the cooling straight pipe is fixedly connected to the cooling straight bent pipes at both ends, and the other end face is fixedly connected to a bidirectional pump.
[0012] Preferably, a temperature sensor is fixedly mounted on the outside of the cooling straight pipe, and the two-way pump is controlled and driven by a built-in controller. When the temperature sensor detects that the temperature inside the cooling pipe is too high, the water temperature in the water cavity and the water cavity is detected and compared at the same time, and the two-way pump is started to pump water into the water cavity with lower temperature, and discharge the high-temperature water into the high-temperature water cavity for cooling.
[0013] Preferably, the water pipe distribution plate is fixedly mounted in the water cooling cavity, an open groove is provided on the side of the water pipe distribution plate, and when the water pipe distribution plate is installed, the open groove is tangent to the outer wall of the oil outlet for easy installation, and the end surface of the water pipe distribution plate is evenly distributed with distribution holes, and a cooling straight and bent pipe is fixedly mounted in the distribution hole.
[0014] Preferably, the external cooling pipe is composed of a cooling spiral pipe, a cooling short bend pipe, and a cooling long bend pipe. The cooling spiral pipe is fixedly mounted in the mounting hole. Both ends of the cooling spiral pipe are respectively fixedly connected to one end face of the cooling short bend pipe and the cooling long bend pipe. The other end faces of the cooling short bend pipe and the cooling long bend pipe are fixedly connected to the bidirectional pump.
[0015] Preferably, temperature sensors are fixedly mounted on the outer sides of the short cooling bend pipe and the long cooling bend pipe respectively, and the two-way pump is controlled and driven by a built-in controller. When the temperature sensor detects that the temperature inside the outer cooling pipe is too high, the water temperature in the water cavity and the water cavity is detected and compared at the same time, and the two-way pump is started to pump water into the water cavity with lower temperature, and discharge the high-temperature water into the high-temperature water cavity for cooling.
[0016] A method for using a water-cooled high-temperature resistant oil cylinder comprises the following steps:
[0017] S1. Start the temperature sensor to detect the temperature of the cooling water in the built-in cooling component and the external cooling component during the operation of the oil cylinder;
[0018] S2. When the detected temperature exceeds the set value, the temperature sensor immediately detects the cooling water temperature of the double water chambers of the cooling water tank and compares them;
[0019] S3. After the comparison is completed, the cooling water with lower temperature is selected, such as the water chamber, and the bidirectional pump and the bidirectional pump are started;
[0020] S4. The two-way pump and the two-way pump connected to the water cavity draw cooling water into the inner cooling pipe and the outer cooling pipe, and the two-way pump and the two-way pump connected to the water cavity discharge the cooling water in the inner cooling pipe and the outer cooling pipe to the water cavity for cooling, and the cooling water circulates to the allowable range set by the temperature detection.
[0021] According to S4, based on real-time temperature data and historical temperature data, the intelligent algorithm can adaptively adjust the cooling water distribution of the dual water chambers in the cooling water tank (2). According to the temperature changes of the inner cooling tube (303) and the outer cooling tube (403), the cooling water exchange frequency and flow rate of the water chamber (201) and the water chamber (202) are dynamically adjusted; and then the energy consumption is reduced by optimizing the operating frequency of the bidirectional pump (302) and the bidirectional pump (402). Under the premise of ensuring the cooling effect, the intelligent algorithm can select the lowest energy consumption operation mode to reduce the overall energy consumption of the system. The specific process is as follows:
[0022] Step 1: Data collection and preprocessing: Install temperature sensors (307) and (407) in the inner cooling pipe (303) and the outer cooling pipe (403) as well as the water chambers (201) and (202) of the cooling water tank (2) to collect real-time temperature data and historical temperature data of the cooling system under different working conditions; then clean and normalize the data to remove noise and outliers, and normalize the temperature data to a standard range for input into the intelligent algorithm model.
[0023] Installation of temperature sensors: A temperature sensor (307) is installed on the inner cooling pipe (303) to detect the temperature of the cooling water in the inner cooling pipe in real time; a temperature sensor (407) is installed on the outer cooling pipe (403) to detect the temperature of the cooling water in the outer cooling pipe in real time. Temperature sensors are installed in the water chamber (201) and the water chamber (202) of the cooling water tank (2) respectively to detect the temperature of the cooling water in the double water chambers.
[0024] Sampling frequency: Collect temperature data once per second and record temperature changes.
[0025] Among them, the inner cooling tube temperature T inner (t), external cooling pipe temperature T outer (t), water chamber (201) temperature T 201 (t), water chamber (202) temperature T 202 (t) Real-time temperature data is stored in the database and marked with timestamps to form time series data. Historical temperature data is backed up regularly for subsequent model training.
[0026] Remove noise and handle outliers:
[0027] Use a sliding average filter to smooth the data and reduce high-frequency noise;
[0028]
[0029] Where n is the sliding window size, T smooth (t) is the temperature after smoothing filtering, and T(ti) is the original temperature value at time ti. The box plot method is used to detect and process outliers. If a data point exceeds 1.5 times the upper and lower quartile range, it is considered an outlier and replaced by the median.
[0030] Data normalization:
[0031]
[0032] Among them, T min and T max They are the minimum and maximum values in the historical data respectively. The normalized data are:
[0033] Step 2: Model establishment and training: Select a time series prediction model, a mixture of long short-term memory network LSTM and gated recurrent unit GRU. Use historical temperature data as training data and set the input and output of the model. Use past temperature data as input to predict temperature changes in a certain period of time in the future. Divide the training set and validation set, perform model training and validation, and adjust model parameters to optimize prediction accuracy.
[0034] 1. Long short-term memory network LSTM model for time series prediction
[0035] LSTM contains three gates: forget gate, input gate and output gate.
[0036] Forget Gate:
[0037] f t =σ(W f ·[h t-1 , x t ]+b f )
[0038] Among them, f t is the output of the forget gate, W f and b f are weights and biases, σ is the activation function, [h t-1 , x t ]The hidden state h of the previous moment t-1 and the current input x t The stitching vector.
[0039] Input Gate:
[0040] i t =σ(W i ·[h t-1 , x t ]+b i )
[0041]
[0042] Among them, i t is the output of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate, [h t-1 , x t ] is the hidden state h of the previous moment t-1 and the current input x t The concatenation vector of is the candidate memory cell state, tanh is the hyperbolic tangent activation function, W C The weight matrix of candidate cell states, b C is the bias vector of the candidate cell state.
[0043] Memory Cell Renewal:
[0044]
[0045] Among them, C t is the cell state at the current moment, C t-1 The cell state at the previous moment, f is the activation value of the forget gate, i t is the activation value of the input gate.
[0046] Output Gate:
[0047] o t =σ(W o ·[h t-1 , x t ]+b o )
[0048] h t =o t tanh(C t )
[0049] Among them, t is the activation value of the output gate, W o is the weight matrix of the output gate, b o is the bias vector of the output gate, [h t-1 , x t ] is the hidden state h at the previous moment t-1 and the current input x t The concatenation vector of h. t is the hidden state at the current moment, tanh(C t ) is the value of the cell state at the current moment after being processed by the tanh activation function.
[0050] To meet the temperature prediction requirements in patent applications, we improved the LSTM model: the temperature of the inner cooling tube, the temperature of the outer cooling tube, and the temperature of the water chambers 201 and 202 in the past 30 minutes are used as input. The temperature of the inner cooling tube and the temperature of the outer cooling tube in the next 5 minutes are used as output. The attention mechanism is introduced to give different weights to the input data of different time steps, thereby improving the sensitivity of the model to key time steps. The energy consumption optimization target is added to the LSTM loss function, so that the model can not only predict the temperature, but also optimize the energy consumption while ensuring the cooling effect.
[0051] The improved LSTM model has multiple inputs and multiple outputs:
[0052] Input vector: X t =[T inner (t), T outer (t), T 201 (t), T202 (t)]
[0053] Prediction vector:
[0054] Among them, T inner (t) is the temperature data of the inner cooling pipe. outer (t) is the temperature data of the external cooling pipe. 201 (t) is the real-time temperature data of the water chamber (201). 202 (t) is the real-time temperature data of the water chamber (202). and Predict the temperature of internal and external cooling components at a certain point in the future.
[0055] Attention Mechanism: Weighted Inputs after Calculating Attention Weights
[0056] e t =tanh(W e ·h t-1 +U e ·X t +b e )
[0057]
[0058] X′ t =α t ·X t
[0059] Among them, e t is the attention score, W e is the weight matrix acting on the hidden state of the previous moment, h t-1 is the hidden state of the previous moment, h t-1 is the weight matrix acting on the current input, X t is the input data at the current moment, α t is the attention weight, X′ t is the weighted input data.
[0060] Energy consumption optimization goal: In order to optimize the energy consumption of the cooling system, the energy consumption function is introduced into the loss function to achieve a balance between energy consumption and temperature control.
[0061] Define the energy consumption function E:
[0062]
[0063] Among them, E is the total energy consumption, P 302 (t) and P 402 (t) are the operating frequencies of the bidirectional pump (302) and the bidirectional pump (402) respectively.
[0064] Add the energy consumption target to the loss function:
[0065] Loss = MSE + λE
[0066] Among them, Loss is the total loss function, MSE is the mean square error, and λ is the energy consumption weight coefficient.
[0067]
[0068] Where N is the number of samples, For the time t i The predicted temperature, T(t i ) is the time at t i The actual temperature.
[0069] After improving the model, the training model is trained: the temperature data of the past 30 minutes, with a sampling frequency of 1Hz, a total of 1800 data points as input data, and the temperature data of the next 5 minutes, a total of 300 data points as output data. The training set is 80%, and the validation set is 20%. Initialize the LSTM model parameters, use historical temperature data for training, optimize the loss function, and use the back propagation algorithm and gradient descent to optimize the model parameters. Use the validation set to evaluate the model performance, and adjust the hyperparameters such as the learning rate and the number of LSTM layers to improve the prediction accuracy.
[0070] 2. Improved Gated Recurrent Unit (GRU) model for time series prediction
[0071] The gated recurrent unit GRU contains two gates: update gate and reset gate
[0072] Update Gate:
[0073] z t =σ(W z ·[h t-1 , x t ]+b z ).
[0074] Among them, z t is the output of the update gate, W z and b z is the weight and bias of the update gate, σ is the activation function, [h t-1 , x t ] is the hidden state h of the previous moment t-1 and the current input x t The stitching vector.
[0075] Reset the gate:
[0076] r t =σ(W r ·[h t-1 , xt ]+b r )
[0077] Among them, r t is the output of the reset gate, W r is the weight matrix of the reset gate, b r is the bias vector for the reset gate.
[0078] Candidate activation status:
[0079]
[0080] in, is the candidate activation state, tanh is the hyperbolic tangent activation function, W h is the weight matrix of the candidate activation state, represents element-wise multiplication, represents the hidden state at the previous moment after reset, b h Bias vector for candidate activation states.
[0081] Hide status updates:
[0082]
[0083] Among them, h t is the hidden state at the current moment, h t-1 is the hidden state at the previous moment.
[0084] 1-z t It is the reverse output of the update gate, indicating the proportion of the hidden state at the previous moment to be retained.
[0085] It is the product of the output of the update gate and the candidate activation state, indicating the proportion of updating the hidden state at the current moment.
[0086] To meet the temperature prediction requirements in patent applications, we improved the GRU model in the same way as the LSTM algorithm model, improving the multi-input and output architecture, attention mechanism, and energy optimization target. Through the energy optimization target predicted by the model, we adjusted the operating frequency of the bidirectional pump and reduced the system energy consumption.
[0087] 3. Hybrid LSTM and GRU model to control oil cylinder temperature
[0088] The LSTM and GRU algorithms are improved according to the above methods. The historical temperature data is used to train the LSTM and GRU models and adjust the model parameters, such as the learning rate and the number of layers, to optimize the prediction accuracy. The hidden layer outputs of the LSTM and GRU modules are then weighted and fused, that is, a hybrid model of LSTM and GRU is constructed to obtain the final prediction results. The performance of the hybrid model is evaluated using the validation set, and the weights and structure of the hybrid model are adjusted to further improve the prediction accuracy.
[0089] The final realization is: Use the hybrid model to predict the temperature changes of the cooling water in the built-in cooling component and the external cooling component in a certain period of time in the future. According to the predicted temperature changes, select the cooling water with lower temperature, and start the corresponding two-way pump to extract and discharge the cooling water. Combined with the predicted temperature changes and the system working status, dynamically adjust the working frequency of the two-way pump to improve cooling efficiency and reduce energy consumption.
[0090] Step 3: Model deployment and real-time application. The trained model is deployed to the control unit of the cooling system. Based on the real-time collected temperature data, the model is used to predict the future temperature change trend. The predicted temperature change is combined with the actual temperature data to adjust the cooling water distribution of the dual water chambers in the cooling water tank (2) and the operating frequency of the bidirectional pumps 302 and 402.
[0091] Model deployment: Save the trained hybrid model as a model file that can be used for real-time inference in TensorFlow SavedModel or ONNX format. Deploy the model file to the control unit of the cooling system to ensure that the control unit has sufficient computing resources and real-time inference capabilities. Establish an interface in the control unit to receive real-time collected temperature data and feed back the prediction results to the system controller.
[0092] Real-time applications:
[0093] First, the cooling system is adjusted, and the predicted temperature change trend is combined with the actual temperature data to determine the current working state and effect of the cooling system. According to the predicted results, the cooling water distribution of the double water chambers in the cooling water tank (2) and the working frequency of the two-way pump (302) and the two-way pump (402) are adjusted. Then feedback and optimization are performed, and the adjusted temperature change is monitored and fed back to the hybrid model in real time. According to the real-time feedback, the model parameters are optimized or the model structure is adjusted to improve the prediction accuracy and system stability.
[0094] Step 4: Adaptive control and energy consumption optimization. According to the model's prediction results and real-time temperature data, dynamically adjust the cooling water exchange frequency and flow rate of water chambers (201) and (202). If the predicted temperature rise trend is obvious, increase the cooling water exchange frequency and flow rate. If the predicted temperature tends to be stable, appropriately reduce the cooling water exchange frequency and flow rate. Monitor the temperature changes after adjustment and feed back to the hybrid model in real time for model parameter optimization and structural adjustment to ensure that the system strikes a balance between optimal cooling effect and minimum energy consumption.
[0095] Compared with the prior art, the present invention has at least the following beneficial effects:
[0096] 1. In the above scheme, by setting up a built-in water cooling component and utilizing the high thermal conductivity and high specific heat capacity of water, heat can be more effectively extracted from the hydraulic oil, thereby keeping the oil within the ideal operating temperature range, providing important technical guarantee for the stable operation of hydraulic equipment in high temperature environment.
[0097] 2. In the above scheme, an external water cooling component is provided to circulate cooling water to effectively transfer the heat in the hydraulic oil to the external environment, thereby keeping the oil within a suitable operating temperature range, maintaining the viscosity and lubrication properties of the oil, and ensuring the stable operation of the system.
[0098] 3. In the above scheme, the performance of the hydraulic cylinder in a high temperature environment is improved by setting up a dual cooling component of an internal water cooling component and an external water cooling component, ensuring its reliability and long-term stability. In addition, the internal water cooling component and the external water cooling group are designed to be more compact and integrated, which helps to simplify the overall structure and reduce system complexity.
[0099] 4. In the above scheme, by setting a temperature sensor to monitor the working temperature of the cylinder in real time, and by adjusting the flow and temperature of the cooling water in the built-in water cooling component and the external water cooling component, the heat generated by long-term work can be effectively dissipated, the hydraulic cylinder can be kept running stably and the service life can be extended.
[0100] 5. In the above scheme, by setting up a two-way pump and a double water chamber design of the cooling water tank, the temperature in the double water chamber of the cooling water tank can be detected by the temperature sensor, the cooling water in the water chamber suitable for the current work can be selected and the two-way pump can be started to control the flow direction of the cooling water for cooling.
[0101] 6. By improving the LSTM model and the improved GRU model and combining the attention mechanism and energy consumption optimization goals, the patented system can realize intelligent adjustment of the cooling system through real-time collected temperature data and model prediction results. According to the current state and predicted trend, the distribution of cooling water and the working frequency of the bidirectional pump can be optimized to improve cooling efficiency, reduce energy consumption, and achieve efficient and stable temperature control. In addition, by mixing the LSTM and GRU models, it can better adapt to the complexity of temperature changes in practical applications and provide more effective solutions for system optimization. Through adaptive regulation and energy consumption optimization, the cooling system can achieve the lowest energy consumption operating state while ensuring the cooling effect, thereby improving the energy efficiency and environmental friendliness of the system. Both are suitable for time series data processing, but have different characteristics and advantages. By mixing them, the advantages of the two models can be fully utilized to improve prediction accuracy and system performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0102] Figure 1 It is a schematic diagram of the three-dimensional structure of the present invention;
[0103] Figure 2 It is a cross-sectional structural schematic diagram of the present invention;
[0104] Figure 3 It is a schematic diagram of the exploded structure of the oil cylinder of the present invention;
[0105] Figure 4 It is a schematic diagram of the cross-sectional structure of the oil cylinder of the present invention;
[0106] Figure 5 It is a schematic diagram of the structure of the built-in water cooling component of the present invention;
[0107] Figure 6 It is a schematic diagram of the cooling water pipe structure in the built-in water cooling assembly of the present invention;
[0108] Figure 7 It is a schematic diagram of the structure of the external water cooling assembly of the present invention;
[0109] Figure 8 It is a schematic diagram of the internal structure of the water tank of the present invention.
[0110] Reference numerals
[0111] 1. Oil cylinder; 101. Piston oil cylinder; 102. Piston cylinder cover; 103. Piston; 104. Inner cavity; 105. Water cooling cavity; 106. Oil outlet; 107. Heat dissipation rib; 108. Mounting hole; 109. Water pipe sleeve hole; 110. Mounting ring groove; 111. Piston hole; 112. Mounting step; 113. Piston rod; 2. Cooling water tank; 201. Water cavity; 202. Water cavity; 3. Built-in water cooling assembly; 301. Connecting water pipe; 302, two-way pump; 303, inner cooling pipe; 304, water pipe distribution plate; 305, distribution hole; 306, cooling elbow; 307, temperature sensor; 308, cooling straight pipe; 309, cooling straight elbow; 310, open slot; 4, external water cooling assembly; 401, connecting water pipe; 402, two-way pump; 403, outer cooling pipe; 404, cooling spiral pipe; 405, cooling short elbow; 406, cooling long elbow; 407, temperature sensor.
[0112] As shown in the figure, in order to clearly implement the structure of the embodiment of the present invention, specific structures and devices are marked in the figure, but this is only for illustrative purposes and is not intended to limit the present invention to the specific structure, device and environment. According to specific needs, ordinary technicians in this field can adjust or modify these devices and environments, and the adjustments or modifications made are still included in the scope of the attached claims. DETAILED DESCRIPTION
[0113] The following is a detailed description of a water-cooled high-temperature resistant oil cylinder and its use method provided by the present invention in combination with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternative methods to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.
[0114] like Figure 1-8As shown, an embodiment of the present invention provides a water-cooled high-temperature resistant oil cylinder, including: an oil cylinder 1, a cooling water tank 2, the oil cylinder 1 includes a piston oil cylinder 101, a piston cylinder cover 102, and a piston 103, the piston cylinder cover 102 is fixedly installed on the side of the piston oil cylinder 101, the piston 103 is slidably installed in the inner cavity 104 of the piston oil cylinder 101, the piston oil cylinder 101 is provided with a water-cooling cavity 105, and a built-in water-cooling component 3 is fixedly installed in the water-cooling cavity 105; the built-in water-cooling component 3 includes a connecting water pipe 301 and a two-way pump 302, one side of the connecting water pipe 301 is fixedly connected to the cooling water tank 2, and the connecting water pipe 301 is connected to the cooling water tank 2. The water connection pipe 301 is interconnected with the inner cavity of the cooling water tank 2, and the two ends of the two-way pump 302 are respectively fixedly connected with the connection water pipe 301 and the inner cooling pipe 303, and the inner cooling pipe 303 is evenly distributed by the water pipe distribution plate 304; the external water cooling component 4 is fixedly installed on the outside of the piston cylinder 101, and the external water cooling component 4 includes a connection water pipe 401 and a two-way pump 402. One side of the connection water pipe 401 is fixedly connected with the cooling water tank 2, and the connection water pipe 401 is interconnected with the inner cavity of the cooling water tank 2, and the two ends of the two-way pump 402 are respectively fixedly connected with the connection water pipe 401 and the outer cooling pipe 403. The wiring diagram in the present invention belongs to the common knowledge in the art, and its working principle is a well-known technology. The model is selected according to the actual use. Therefore, the control method and wiring arrangement of the internal water cooling component 3 and the external water cooling component 4 are no longer explained in detail.
[0115] like Figure 3 and 4 As shown, an oil outlet 106 is provided on the outer side of the piston cylinder 101, and the oil outlet 106 is communicated with the inner cavity 104, and a water-cooling cavity 105 is located between the outer side of the piston cylinder 101 and the inner cavity 104, and the water-cooling cavity 105 is not communicated with the oil outlet 106 and the inner cavity 104, and heat dissipation ribs 107 are evenly distributed on the outer side of the piston cylinder 101, and mounting holes 108 are provided on the sides of the heat dissipation ribs 107, and an external water cooling component 4 is fixedly mounted in the mounting holes 108; cooling water is circulated through the external water cooling component 4, and the heat in the hydraulic oil is effectively transferred to the external environment, so as to keep the oil within a suitable working temperature range, maintain the viscosity and lubrication performance of the oil, and ensure the stable operation of the system.
[0116] like Figure 4 As shown, one side end face of the piston cylinder cover 102 is provided with an installation step 112 and an installation groove 110. The installation step 112 is fixedly mounted in the water cooling cavity 105, and the installation groove 110 is fixedly mounted in the side wall of the inner cavity 104. The piston cylinder cover 102 is provided with a water pipe sleeve hole 109 and a piston hole 111. The water pipe sleeve hole 109 is fixedly mounted with an inner cooling pipe 303, and the piston rod 113 is slidably mounted in the piston hole 111.
[0117] like Figure 8As shown, a water cavity 201 and a water cavity 202 are respectively provided inside the cooling water tank 2. The water volumes in the water cavities 201 and 202 are inversely proportional when in use. The water cavities 201 and 202 can respectively accommodate the water volumes required for multiple water cooling cycles of the built-in water cooling component 3 and the external water cooling component 4.
[0118] like Figure 5 and 6 As shown, the inner cooling pipe 303 is composed of a cooling straight pipe 308, a cooling straight curved pipe 309, and a cooling curved pipe 306. The inner cooling pipe 303 is fixedly mounted in the water pipe sleeve hole 109 as the cooling straight pipe 308. The cooling straight curved pipe 309 and the two ends of the cooling curved pipe 306 are fixedly connected in sequence. One end face of the cooling straight pipe 308 is fixedly connected to the cooling straight curved pipes 309 at both ends, and the other end face is fixedly connected to the two-way pump 302.
[0119] like Figure 2 and 5 As shown, a temperature sensor 307 is fixedly mounted on the outside of the cooling straight pipe 308, and the two-way pump 302 is controlled and driven by a built-in controller; when the temperature sensor 307 detects that the temperature inside the inner cooling pipe 303 is too high, the water temperatures in the water chamber 201 and the water chamber 202 are detected and compared at the same time, and the built-in controller starts the two-way pump 302 to pump water into the water chamber with lower temperature, and discharges the high-temperature water to the high-temperature water chamber for cooling.
[0120] like Figure 2 and 5 As shown, the water pipe distribution plate 304 is fixedly mounted in the water cooling cavity 105, and an opening groove 310 is provided on the side of the water pipe distribution plate 304. When the water pipe distribution plate 304 is installed, the opening groove 310 is tangent to the outer wall of the oil outlet 106 for easy installation. The end surface of the water pipe distribution plate 304 is evenly distributed with distribution holes 305, and a cooling straight bend pipe 309 is fixedly mounted in the distribution hole 305.
[0121] like Figure 7 As shown, the external cooling pipe 403 is composed of a cooling spiral pipe 404, a cooling short bend pipe 405, and a cooling long bend pipe 406. The cooling spiral pipe 404 is fixedly mounted in the mounting hole 108. Both ends of the cooling spiral pipe 404 are fixedly connected to one end surface of the cooling short bend pipe 405 and the cooling long bend pipe 406 respectively. The other end surfaces of the cooling short bend pipe 405 and the cooling long bend pipe 406 are fixedly connected to the bidirectional pump 402.
[0122] like Figure 1 and 7As shown, temperature sensors 407 are fixedly mounted on the outside of the cooling short bend pipe 405 and the cooling long bend pipe 406, respectively, and the two-way pump 402 is controlled and driven by a built-in controller; when the temperature sensor 407 detects that the temperature inside the outer cooling pipe 403 is too high, the water temperatures in the water cavity 201 and the water cavity 202 are detected and compared at the same time, and the built-in controller starts the two-way pump 402 to pump water into the water cavity with lower temperature, and discharges the high-temperature water to the high-temperature water cavity for cooling.
[0123] Example 1
[0124] A method for using a water-cooled high-temperature resistant oil cylinder comprises the following steps:
[0125] S1, start the temperature sensor 307 and the temperature sensor 407 to detect the temperature of the cooling water in the internal cooling component and the external cooling component during the operation of the oil cylinder 1;
[0126] S2. When the detected temperature exceeds the set value, the temperature sensor 307 and the temperature sensor 407 immediately detect the cooling water temperature of the double water chambers of the cooling water tank 2 and compare them;
[0127] S3, after the comparison is completed, select cooling water with a lower temperature such as water chamber 201, and start the two-way pump 302 and the two-way pump 402;
[0128] S4, the two-way pump 302 and the two-way pump 402 connected to the water cavity 201 draw cooling water into the inner cooling pipe 303 and the outer cooling pipe 403, the two-way pump 302 and the two-way pump 402 connected to the water cavity 202 discharge the cooling water in the inner cooling pipe 303 and the outer cooling pipe 403 to the water cavity 202 for cooling, and the cooling water circulates to the allowable range set by the temperature sensor.
[0129] According to the technical solution provided by the present invention, during operation, the temperature sensor 307 and the temperature sensor 407 detect the temperature of the cooling water in the built-in cooling component and the external cooling component. When the temperature exceeds the set value, the temperature sensor 307 and the temperature sensor 407 detect the temperature of the cooling water in the double water chambers of the cooling water tank, and select cooling water with a lower temperature such as the water chamber 201, then start the two-way pump 302 and the two-way pump 402, the two-way pump 302 and the two-way pump 402 connected to the water chamber 201 extract cooling water, and the two-way pump 302 and the two-way pump 402 not connected to the water chamber 201 discharge the cooling water in the inner cooling pipe 303 and the outer cooling pipe 403 to the water chamber 202, and the cooling water circulates to within the allowable range set by the temperature detection.
[0130] According to S4, based on real-time temperature data and historical temperature data, the intelligent algorithm can adaptively adjust the cooling water distribution of the dual water chambers in the cooling water tank (2). According to the temperature changes of the inner cooling tube (303) and the outer cooling tube (403), the cooling water exchange frequency and flow rate of the water chamber (201) and the water chamber (202) are dynamically adjusted; and then the energy consumption is reduced by optimizing the operating frequency of the bidirectional pump (302) and the bidirectional pump (402). Under the premise of ensuring the cooling effect, the intelligent algorithm can select the lowest energy consumption operation mode to reduce the overall energy consumption of the system. The specific process is as follows:
[0131] Step 1: Data collection and preprocessing: Install temperature sensors 307 and 407 in the inner and outer cooling pipes 303 and 403 and the water chambers (201) and (202) of the cooling water tank (2) to collect real-time temperature data and historical temperature data of the cooling system under different working conditions; then clean and normalize the data to remove noise and abnormal values, and normalize the temperature data to a standard range for input into the intelligent algorithm model. Installation of temperature sensors: Install a temperature sensor (307) on the inner cooling pipe (303) to detect the cooling water temperature in the inner cooling pipe in real time; install a temperature sensor (407) on the outer cooling pipe (403) to detect the cooling water temperature in the outer cooling pipe in real time. Install temperature sensors in the water chambers (201) and (202) of the cooling water tank (2) to detect the cooling water temperature in the double water chambers.
[0132] Sampling frequency: Collect temperature data once per second and record temperature changes.
[0133] Among them, the inner cooling tube temperature T inner (t), external cooling pipe temperature T outer (t), water chamber (201) temperature T 201 (t), water chamber (202) temperature T 202 (t) Real-time temperature data is stored in the database and marked with timestamps to form time series data. Historical temperature data is backed up regularly for subsequent model training.
[0134] Remove noise and handle outliers:
[0135] Use a sliding average filter to smooth the data and reduce high-frequency noise;
[0136]
[0137] Where n is the sliding window size, T smooth (t) is the temperature after smoothing filtering, and T(ti) is the original temperature value at time ti. The box plot method is used to detect and process outliers. If a data point exceeds 1.5 times the upper and lower quartile range, it is considered an outlier and replaced by the median.
[0138] Data normalization:
[0139]
[0140] Among them, T min and T max They are the minimum and maximum values in the historical data respectively. The normalized data are: Step 2: Model establishment and training: Select a time series prediction model, a mixture of long short-term memory network LSTM and gated recurrent unit GRU. Use historical temperature data as training data and set the input and output of the model. Use past temperature data as input to predict temperature changes in a certain period of time in the future. Divide the training set and validation set, perform model training and validation, and adjust model parameters to optimize prediction accuracy.
[0141] 1. Long short-term memory network LSTM model for time series prediction
[0142] LSTM contains three gates: forget gate, input gate and output gate.
[0143] Forget Gate:
[0144] f t =σ(W f .[h t-1 , x t ]+b f )
[0145] Among them, f t is the output of the forget gate, W f and b f are weights and biases, σ is the activation function, [h t-1 , x t ]The hidden state h of the previous moment t-1 and the current input x t The stitching vector.
[0146] Input Gate:
[0147] i t =σ(W i ·[h t-1 , x t ]+b i )
[0148]
[0149] Among them, i t is the output of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate, [h t-1 , x t] is the hidden state h of the previous moment t-1 and the current input x t The concatenation vector of is the candidate memory cell state, tanh is the hyperbolic tangent activation function, W C The weight matrix of candidate cell states, b C is the bias vector of the candidate cell state.
[0150] Memory Cell Renewal:
[0151]
[0152] Among them, C t is the cell state at the current moment, C t-1 The cell state at the previous moment, f is the activation value of the forget gate, i t is the activation value of the input gate.
[0153] Output Gate:
[0154] o t =σ(W o ·[h t-1 , x t ]+b o )
[0155] h t =o t tanh(C t )
[0156] Among them, t is the activation value of the output gate, W o is the weight matrix of the output gate, b o is the bias vector of the output gate, [h t-1 , x t ] is the hidden state h at the previous moment t-1 and the current input x t The concatenation vector of h. t is the hidden state at the current moment, tanh(C t ) is the value of the cell state at the current moment after being processed by the tanh activation function. In order to meet the temperature prediction requirements in the patent application, we improved the LSTM model: the temperature of the inner cooling tube, the temperature of the outer cooling tube, and the temperature of the water chambers 201 and 202 in the past 30 minutes are used as input. The temperature of the inner cooling tube and the temperature of the outer cooling tube in the next 5 minutes are used as output. The attention mechanism is introduced to give different weights to the input data of different time steps, thereby improving the sensitivity of the model to the key time steps. The energy consumption optimization target is added to the LSTM loss function, so that the model can not only predict the temperature, but also optimize the energy consumption while ensuring the cooling effect.
[0157] The improved LSTM model has multiple inputs and multiple outputs:
[0158] Input vector: X t =[T inner (t), T outer (t), T 201 (t), T 202 (t)]
[0159] Prediction vector:
[0160] Among them, T inner (t) is the temperature data of the inner cooling pipe. outer (t) is the temperature data of the external cooling pipe. 201 (t) is the real-time temperature data of the water chamber (201). 202 (t) is the real-time temperature data of the water chamber (202). and Predict the temperature of internal and external cooling components at a certain point in the future.
[0161] Attention Mechanism: Weighted Inputs after Calculating Attention Weights
[0162] e t =tanh(W e ·h t-1 +U e ·X t +b e )
[0163]
[0164] Among them, e t is the attention score, W e is the weight matrix acting on the hidden state of the previous moment, h t-1 is the hidden state of the previous moment, h t-1 is the weight matrix acting on the current input, X t is the input data at the current moment, α t is the attention weight, X′ t is the weighted input data.
[0165] Energy consumption optimization goal: In order to optimize the energy consumption of the cooling system, the energy consumption function is introduced into the loss function to achieve a balance between energy consumption and temperature control.
[0166] Define the energy consumption function E:
[0167] Among them, E is the total energy consumption, P 302 (t) and P 402(t) are the operating frequencies of the bidirectional pump (302) and the bidirectional pump (402) respectively.
[0168] Add the energy consumption target to the loss function:
[0169] Loss = MSE + λE
[0170] Among them, Loss is the total loss function, MSE is the mean square error, and λ is the energy consumption weight coefficient.
[0171]
[0172] Where N is the number of samples, For the time t i The predicted temperature, T(t i ) is the time at t i The actual temperature.
[0173] After improving the model, the training model is trained: the temperature data of the past 30 minutes, with a sampling frequency of 1Hz, a total of 1800 data points as input data, and the temperature data of the next 5 minutes, a total of 300 data points as output data. The training set is 80%, and the validation set is 20%. Initialize the LSTM model parameters, use historical temperature data for training, optimize the loss function, and use the back propagation algorithm and gradient descent to optimize the model parameters. Use the validation set to evaluate the model performance, and adjust the hyperparameters such as the learning rate and the number of LSTM layers to improve the prediction accuracy.
[0174] 2. Improved Gated Recurrent Unit GRU model for time series prediction The Gated Recurrent Unit GRU contains two gates: update gate and reset gate
[0175] Update Gate:
[0176] z t =σ(W z ·[h t-1 , x t ]+b z ).
[0177] Among them, z t is the output of the update gate, W z and b z is the weight and bias of the update gate, σ is the activation function, [h t-1 , x t ] is the hidden state h of the previous moment t-1 and the current input x t The stitching vector.
[0178] Reset the gate:
[0179] r t =σ(W r ·[ht-1 , x t ]+b r )
[0180] Among them, r t is the output of the reset gate, W r is the weight matrix of the reset gate, b r is the bias vector for the reset gate. Candidate activation states:
[0181]
[0182] in, is the candidate activation state, tanh is the hyperbolic tangent activation function, W h is the weight matrix of the candidate activation state, represents element-wise multiplication, represents the hidden state at the previous moment after reset, b h Bias vector for candidate activation states.
[0183] Hide status updates:
[0184]
[0185] Among them, h t is the hidden state at the current moment, h t-1 is the hidden state at the previous moment.
[0186] 1-z t It is the reverse output of the update gate, indicating the proportion of the hidden state at the previous moment to be retained.
[0187] It is the product of the output of the update gate and the candidate activation state, indicating the proportion of updating the hidden state at the current moment.
[0188] To meet the temperature prediction requirements in patent applications, we improved the GRU model in the same way as the LSTM algorithm model, improving the multi-input and output architecture, attention mechanism, and energy optimization target. Through the energy optimization target predicted by the model, we adjusted the operating frequency of the bidirectional pump and reduced the system energy consumption.
[0189] 3. Hybrid LSTM and GRU model to control oil cylinder temperature. Improve the LSTM and GRU algorithms according to the above methods, use historical temperature data to train the LSTM and GRU models and adjust model parameters, such as learning rate and number of layers, to optimize prediction accuracy. Then perform weighted fusion on the hidden layer outputs of the LSTM and GRU modules, that is, build a hybrid model of LSTM and GRU to obtain the final prediction result. Use the validation set to evaluate the performance of the hybrid model, and adjust the weight and structure of the hybrid model to further improve the prediction accuracy.
[0190] The final realization is: Use the hybrid model to predict the temperature changes of the cooling water in the built-in cooling component and the external cooling component in a certain period of time in the future. According to the predicted temperature changes, select the cooling water with lower temperature, and start the corresponding two-way pump to extract and discharge the cooling water. Combined with the predicted temperature changes and the system working status, dynamically adjust the working frequency of the two-way pump to improve cooling efficiency and reduce energy consumption.
[0191] Step 3: Model deployment and real-time application. The trained model is deployed to the control unit of the cooling system. Based on the real-time collected temperature data, the model is used to predict the future temperature change trend. The predicted temperature change is combined with the actual temperature data to adjust the cooling water distribution of the dual water chambers in the cooling water tank (2) and the operating frequency of the bidirectional pumps 302 and 402.
[0192] Model deployment: Save the trained hybrid model as a model file that can be used for real-time inference in TensorFlow SavedModel or ONNX format. Deploy the model file to the control unit of the cooling system to ensure that the control unit has sufficient computing resources and real-time inference capabilities. Establish an interface in the control unit to receive real-time collected temperature data and feed back the prediction results to the system controller.
[0193] Real-time applications:
[0194] First, the cooling system is adjusted, and the predicted temperature change trend is combined with the actual temperature data to determine the current working state and effect of the cooling system. According to the predicted results, the cooling water distribution of the double water chambers in the cooling water tank (2) and the working frequency of the two-way pump (302) and the two-way pump (402) are adjusted. Then feedback and optimization are performed, and the adjusted temperature change is monitored and fed back to the hybrid model in real time. According to the real-time feedback, the model parameters are optimized or the model structure is adjusted to improve the prediction accuracy and system stability.
[0195] Step 4: Adaptive control and energy consumption optimization. According to the model's prediction results and real-time temperature data, dynamically adjust the cooling water exchange frequency and flow rate of water chambers (201) and (202). If the predicted temperature rise trend is obvious, increase the cooling water exchange frequency and flow rate. If the predicted temperature tends to be stable, appropriately reduce the cooling water exchange frequency and flow rate. Monitor the adjusted temperature changes and feed back to the hybrid model in real time for model parameter optimization and structural adjustment to ensure that the system strikes a balance between optimal cooling effect and minimum energy consumption. Numerical case analysis:
[0196] The hybrid model uses a parallel structure, where LSTM and GRU process the same input data simultaneously and their outputs are fused. Assume that the input sequence is X = {x 1 , x 2 , ..., x T}, in the parallel structure, LSTM and GRU process the same input data X, and their outputs are and The final output is the weighted sum of the two:
[0197]
[0198] Among them, α is the mixing weight parameter.
[0199] Suppose we have a set of temperature data X, containing the temperatures of the past 10 time steps:
[0200] X={30.5, 31.0, 30.8, 31.2, 31.4, 31.6, 31.8, 32.0, 32.2, 32.5}
[0201] Data preparation: Standardize the data and divide it into training set and test set.
[0202] Model architecture: Design a parallel LSTM and GRU model.
[0203] Training model: Use the training set data to train the model and optimize the loss function.
[0204] Model predictions
[0205] Assume that after training, the model predicts the temperature for the next time step and obtains the following results:
[0206] LSTM model prediction:
[0207] GRU model prediction:
[0208] Assuming the mixing weight α = 0.6, the prediction result of the mixing model is:
[0209]
[0210] According to the predicted temperature Adjust the operating frequency and flow rate of the bidirectional pump to optimize the cooling effect.
[0211] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for using a water-cooled high-temperature resistant oil cylinder, the high-temperature resistant oil cylinder comprising: The oil cylinder (1) and the cooling water tank (2) are characterized in that the oil cylinder (1) comprises a piston oil cylinder (101), a piston cylinder cover (102), and a piston (103); the piston cylinder cover (102) is fixedly mounted on the side of the piston oil cylinder (101); the piston (103) is slidably mounted in an inner cavity (104) provided in the piston oil cylinder (101); a water cooling cavity (105) is provided inside the piston oil cylinder (101); and a built-in water cooling component (3) is fixedly mounted in the water cooling cavity (105); The built-in water cooling component (3) comprises a connecting water pipe (301) and a first bidirectional pump (302); one side of the connecting water pipe (301) is fixedly connected to the cooling water tank (2); the connecting water pipe (301) and the inner cavity of the cooling water tank (2) are in communication with each other; the two ends of the first bidirectional pump (302) are respectively fixedly connected to the connecting water pipe (301) and the inner cooling pipe (303); the inner cooling pipe (303) is evenly distributed by a water pipe distribution plate (304); An external water cooling component (4), the external water cooling component (4) being fixedly mounted on the outside of the piston cylinder (101), the external water cooling component (4) comprising a connecting water pipe (401) and a second bidirectional pump (402), one side of the connecting water pipe (401) being fixedly connected to the cooling water tank (2), the connecting water pipe (401) and the inner cavity of the cooling water tank (2) being interconnected, and both ends of the second bidirectional pump (402) being fixedly connected to the connecting water pipe (401) and the external cooling pipe (403); The outer side surface of the piston cylinder (101) is provided with an oil outlet (106), the oil outlet (106) and the inner cavity (104) are in communication with each other, the water cooling cavity (105) is located between the outer side of the piston cylinder (101) and the inner cavity (104), the water cooling cavity (105) is not in communication with the oil outlet (106) and the inner cavity (104), the outer side surface of the piston cylinder (101) is evenly distributed with heat dissipation ribs (107), the side surface of the heat dissipation ribs (107) is provided with mounting holes (108), and an external water cooling component (4) is fixedly sleeved inside the mounting hole (108); The cooling water box (2) is provided with a first water chamber (201) and a second water chamber (202) respectively, and the water volumes in the first water chamber (201) and the second water chamber (202) are inversely proportional when in use; The inner cooling pipe (303) is composed of a cooling straight pipe (308), a cooling straight curved pipe (309), and a cooling curved pipe (306); the piston cylinder cover (102) is provided with a water pipe sleeve hole (109); the inner cooling pipe (303) is fixedly sleeved in the water pipe sleeve hole (109) as a cooling straight pipe (308); the cooling straight curved pipe (309) and the cooling curved pipe (306) are respectively and sequentially fixedly connected at both ends; one end face of the cooling straight pipe (308) is fixedly connected to the cooling straight curved pipes (309) at both ends; and the other end face is fixedly connected to the first bidirectional pump (302); A first temperature sensor (307) is fixedly mounted on the outside of the cooling straight pipe (308), and the first bidirectional pump (302) is controlled and driven by a built-in controller; The external cooling pipe (403) is composed of a cooling spiral pipe (404), a cooling short curved pipe (405), and a cooling long curved pipe (406); the cooling spiral pipe (404) is fixedly sleeved in the mounting hole (108); two ends of the cooling spiral pipe (404) are respectively fixedly connected to one end surface of the cooling short curved pipe (405) and the cooling long curved pipe (406); the other end surfaces of the cooling short curved pipe (405) and the cooling long curved pipe (406) are fixedly connected to the second bidirectional pump (402); The second temperature sensors (407) are respectively fixedly mounted on the outside of the cooling short curved pipe (405) and the cooling long curved pipe (406), and the second bidirectional pump (402) is controlled and driven by a built-in controller; The method of use comprises the following steps: S1, during the operation of the oil cylinder (1), starting the first temperature sensor (307) and the second temperature sensor (407) to detect the temperature of cooling water in the internal water cooling component (3) and the external water cooling component (4); S2. When the detected temperature exceeds the set value, the first temperature sensor (307) and the second temperature sensor (407) immediately detect the cooling water temperature of the double water chambers of the cooling water tank (2) and compare them; S3, after the comparison is completed, cooling water with a lower temperature is selected to enter the first water chamber (201), and the first bidirectional pump (302) and the second bidirectional pump (402) are started; S4, the first two-way pump (302) and the second two-way pump (402) connected to the first water chamber (201) extract cooling water into the inner cooling pipe (303) and the outer cooling pipe (403), the first two-way pump (302) and the second two-way pump (402) connected to the second water chamber (202) discharge the cooling water in the inner cooling pipe (303) and the outer cooling pipe (403) into the second water chamber (202) for cooling, and the cooling water circulates until it is within the allowable range set by the temperature detection; According to S4, based on the real-time temperature data and the historical temperature data, the cooling water distribution of the double water chambers in the cooling water tank (2) is adaptively adjusted; according to the temperature changes of the inner cooling pipe (303) and the outer cooling pipe (403), the cooling water exchange frequency and flow rate of the first water chamber (201) and the second water chamber (202) are dynamically adjusted; and the energy consumption is reduced by optimizing the operating frequency of the first bidirectional pump (302) and the second bidirectional pump (402); under the premise of ensuring the cooling effect, the intelligent algorithm selects the operation mode with the lowest energy consumption to reduce the overall energy consumption of the system. The specific process is as follows: Step 1: Data collection and preprocessing: Install temperature sensors in the inner cooling pipe (303), the outer cooling pipe (403) and the water cavity of the cooling water tank (2) to collect real-time temperature data and historical temperature data of the cooling system under different working conditions; then clean and normalize the data to remove noise and outliers, and normalize the temperature data to a standard range for input into the intelligent algorithm model; Sampling frequency: collect temperature data once per second and record temperature changes; Among them, the inner cooling tube temperature , external cooling pipe temperature , the temperature of the first water chamber (201) , the temperature of the second water chamber (202) Real-time temperature data is stored in the database and marked with timestamps to form time series data; historical temperature data is backed up regularly for subsequent model training; Remove noise and handle outliers: Use a sliding average filter to smooth the data and reduce high-frequency noise; in, is the sliding window size, is the temperature after smoothing filtering, for The original temperature value at the time; the box plot method is used to detect and process outliers; if the data point exceeds 1.5 times the upper and lower quartile range, it is regarded as an outlier and replaced by the median; Data normalization: in, and They are the minimum and maximum values in the historical data respectively. The normalized data are: ; Step 2: Model establishment and training: Select a time series prediction model that is a mixture of a long short-term memory network (LSTM) and a gated recurrent unit (GRU); use historical temperature data as training data and set the model input and output; use past temperature data as input to predict temperature changes in a certain period of time in the future; divide the training set and validation set, perform model training and validation, and adjust model parameters to optimize prediction accuracy; 2.1 Long Short-Term Memory Network LSTM Model for Time Series Prediction LSTM contains three gates: forget gate, input gate and output gate; Forget Gate: in, is the output of the forget gate, are weights and biases, is the activation function, The hidden state at the previous moment and the current input The concatenation vector of Input Gate: in, is the output of the input gate, is the weight matrix of the input gate, is the bias vector of the input gate, is the hidden state at the previous moment and the current input The concatenation vector of is the candidate memory cell state, tanh is the hyperbolic tangent activation function, The weight matrix of candidate cell states, is the bias vector of the candidate cell state; Memory Cell Renewal: in, is the cell state at the current moment, The cell state at the previous moment, is the activation value of the forget gate, is the activation value of the input gate; Output Gate: in, is the activation value of the output gate, is the weight matrix of the output gate, is the bias vector of the output gate, is the hidden state at the previous moment and the current input The concatenation vector of is the hidden state at the current moment, is the value of the cell state at the current moment after being processed by the tanh activation function; Improvements were made to the LSTM model: the temperature of the inner cooling tube, the temperature of the outer cooling tube, and the temperature of the water cavity in the past 30 minutes were used as input; the temperature of the inner cooling tube and the temperature of the outer cooling tube in the next 5 minutes were used as output; an attention mechanism was introduced to assign different weights to input data at different time steps to improve the model's sensitivity to key time steps; an energy consumption optimization objective was added to the LSTM loss function, so that the model could not only predict the temperature, but also optimize energy consumption while ensuring the cooling effect; The improved LSTM model has multiple inputs and multiple outputs: Input vector: Prediction vector: in, is the temperature data of the inner cooling pipe; is the temperature data of the external cooling pipe; is the real-time temperature data of the first water chamber (201); is the real-time temperature data of the second water chamber (202); and Predict the temperature of internal and external water cooling components at a certain point in the future; Attention Mechanism: Weighted Inputs after Calculating Attention Weights in, Score for attention, is the weight matrix acting on the hidden state of the previous moment, is the hidden state at the previous moment, The weight matrix of is the input data at the current moment, is the attention weight, is the weighted input data; Energy consumption optimization goal: In order to optimize the energy consumption of the cooling system, the energy consumption function is introduced into the loss function to achieve a balance between energy consumption and temperature control; Define the energy consumption function : in, E is the total energy consumption, and are respectively the operating frequencies of the first bidirectional pump (302) and the second bidirectional pump (402); Add the energy consumption target to the loss function: ; Among them, Loss is the total loss function, MSE is the mean square error, is the energy consumption weight coefficient; Where N is the number of samples, For at the moment The predicted temperature, For at the moment The actual temperature; After the model is improved, the training model is trained: the temperature data of the past 30 minutes, with a sampling frequency of 1Hz, a total of 1800 data points as input data, and the temperature data of the next 5 minutes, a total of 300 data points as output data; 80% of the training set and 20% of the validation set; initialize the LSTM model parameters, use historical temperature data for training, optimize the loss function, and use the back propagation algorithm and gradient descent to optimize the model parameters; use the validation set to evaluate the model performance, and adjust the learning rate and LSTM layer number hyperparameters to improve the prediction accuracy; 2.2 Improved Gated Recurrent Unit (GRU) model for time series prediction The gated recurrent unit GRU contains two gates: update gate and reset gate Update Gate: ; in, is the output of the update gate, To update the gate weights and biases, is the activation function, is the hidden state at the previous moment and the current input The concatenation vector of Reset the gate: in, is the output of the reset gate, To reset the gate weight matrix, is the bias vector for resetting the gate; Candidate activation status: in, is the candidate activation state, tanh is the hyperbolic tangent activation function, is the weight matrix of the candidate activation state, represents element-wise multiplication, Indicates the hidden state at the previous moment after reset. The bias vector of the candidate activation state; Hide status updates: in, is the hidden state at the current moment, is the hidden state at the previous moment; is the reverse output of the update gate, indicating the proportion of the hidden state at the previous moment to be retained; It is the product of the output of the update gate and the candidate activation state, indicating the proportion of updating the hidden state at the current moment; Improve the GRU model in the same way as the LSTM algorithm model, including improving the multi-input and output architecture, attention mechanism, and energy consumption optimization target; adjust the working frequency of the bidirectional pump through the energy consumption optimization target predicted by the model to reduce the system energy consumption; 2.3 Hybrid LSTM and GRU model to control oil cylinder temperature The LSTM and GRU algorithms are improved according to the above methods respectively, and the historical temperature data is used to train the LSTM and GRU models and adjust the model parameters; then the hidden layer outputs of the LSTM and GRU modules are weighted and fused, that is, a hybrid model of LSTM and GRU is constructed to obtain the final prediction results; the performance of the hybrid model is evaluated using the validation set, and the weight and structure of the hybrid model are adjusted to further improve the prediction accuracy; The final result is: using the hybrid model to predict the temperature change of the cooling water in the built-in water cooling component (3) and the external water cooling component (4) in a certain period of time in the future, selecting cooling water with a lower temperature based on the predicted temperature change, and starting the corresponding bidirectional pump to extract and discharge the cooling water; then combining the predicted temperature change and the system working state, dynamically adjusting the working frequency of the bidirectional pump to improve cooling efficiency and reduce energy consumption; Step 3: Deploy the trained model to the control unit of the cooling system, and use the model to predict future temperature change trends based on the real-time collected temperature data; combine the predicted temperature change with the actual temperature data to adjust the cooling water distribution of the double water chambers in the cooling water tank (2) and the operating frequency of the bidirectional pump; Model deployment: Save the trained hybrid model as a model file that can be used for real-time inference in TensorFlowSavedModel or ONNX format; deploy the model file to the control unit of the cooling system to ensure that the control unit has sufficient computing resources and real-time inference capabilities; establish an interface in the control unit to receive real-time collected temperature data and feed back the prediction results to the system controller; Real-time application: firstly, the cooling system is adjusted, the predicted temperature change trend is combined with the actual temperature data, the current working state and effect of the cooling system is judged, and according to the predicted result, the cooling water distribution of the double water chambers in the cooling water tank (2) and the working frequency of the first bidirectional pump (302) and the second bidirectional pump (402) are adjusted; then feedback and optimization are performed, the temperature change after adjustment is monitored, and real-time feedback is given to the hybrid model; according to the real-time feedback, the model parameters are optimized or the model structure is adjusted to improve the prediction accuracy and system stability; Step 4: Adaptive control and energy consumption optimization; According to the model's prediction results and real-time temperature data, dynamically adjust the cooling water exchange frequency and flow rate of the water chamber. If the predicted temperature rise trend is obvious, increase the cooling water exchange frequency and flow rate; if the predicted temperature tends to be stable, appropriately reduce the cooling water exchange frequency and flow rate; monitor the temperature changes after adjustment, and feed back to the hybrid model in real time for model parameter optimization and structural adjustment to ensure that the system strikes a balance between optimal cooling effect and minimum energy consumption.
2. The method according to claim 1, characterized in that: The piston cylinder cover (102) is provided with a mounting step (112) and a mounting ring groove (110) on one side end surface; the mounting step (112) is fixedly sleeved in the water-cooling cavity (105); the mounting ring groove (110) is fixedly sleeved in the side wall of the inner cavity (104); the piston cylinder cover (102) is provided with a water pipe sleeve hole (109) and a piston hole (111); an inner cooling pipe (303) is fixedly sleeved in the water pipe sleeve hole (109); and a piston rod (113) is slidably installed in the piston hole (111).
3. The method according to claim 1, characterized in that: The water pipe matching plate (304) is fixedly sleeved in the water cooling cavity (105); an opening groove (310) is provided on the side of the water pipe matching plate (304); when the water pipe matching plate (304) is installed, the opening groove (310) is tangent to the outer wall of the oil outlet (106) for easy installation; matching holes (305) are evenly distributed on the end surface of the water pipe matching plate (304); and a cooling straight bend pipe (309) is fixedly sleeved in the matching hole (305).
Citation Information
Patent Citations
Self-circulation cooling type oil cylinder
CN114215820A
Automatic cooling system of new energy battery case forming machine
CN117644621A
High-efficient heat dissipation hydraulic pressure jar
CN204512064U
A hydraulic cylinder with a cooling device
CN215171184U
Internal and external double-cooling adjustable air compressor cylinder
CN217602877U
