Condensate water recycling system and method

Abstract

The present application relates to the field of industrial water treatment, and discloses a condensate water recycling system and a recycling method, which comprises the following steps: collecting the conductivity original signal of condensate water and the liquid level signal of a downstream water tank; obtaining an effective conductivity estimation value; dynamically adjusting the anti-shake holding time and the hysteresis half-width; comparing the effective conductivity estimation value with upper and lower threshold values; performing trend prediction based on the effective conductivity estimation value; performing fluctuation mode identification on the effective conductivity estimation value; recording the cumulative switching times of each device; and switching the condensate water between the high-value recycling path and the low-value recycling path according to the fusion result. By performing sliding window filtering and Kalman filtering on the conductivity original signal to obtain an effective estimation value, noise interference is filtered out, and the active optimization switching strategy is adjusted according to the remaining life of the device, thereby realizing efficient, stable and long-life operation of the condensate water recycling system.

Description

Technical Field

[0001] This invention relates to the field of industrial water treatment technology, specifically to a method for timely recovery of condensate in a condensate recovery and utilization system. Background Technology

[0002] In industries such as chemical, pharmaceutical, printing and dyeing, and food processing, double-effect evaporator technology is widely used for the treatment of high-salt wastewater. During the evaporation process, the steam condensate formed by the direct condensation of live steam and the evaporation condensate produced by the evaporation of high-salt wastewater are usually characterized by high temperature and low ion content. They have the potential to replace softened water in the boiler soft water system and reuse the condensate in high-value processes, which can effectively reduce the consumption of tap water and softening salt, and achieve energy saving and consumption reduction.

[0003] Currently, existing condensate recovery systems mostly rely on a preset conductivity threshold. When the monitored value is below the threshold, the water quality is deemed acceptable, and the control valve sends the condensate to a high-value reuse path such as a soft water tank or circulating water tank. When the monitored value is above the threshold, the water quality is deemed unacceptable, and the condensate is switched to a low-value reuse path such as fire water tank makeup or circulating alkaline washing tank. However, when the actual conductivity value is just near the preset threshold, signal noise can cause the sampled value to frequently cross the threshold. The control system then repeatedly switches valves and pumps, resulting in a ping-pong switching phenomenon. This can easily cause electric valves to endure a large number of invalid actions in a short period of time, shortening their mechanical life and increasing motor wear. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a timely recovery method for condensate recovery systems, which solves the problem that existing timely recovery methods for condensate recovery systems easily lead to electric valves being subjected to a large number of ineffective actions in a short period of time, shortening their mechanical life and increasing motor wear.

[0005] To achieve the above objectives, the present invention provides a method for timely recovery of condensate in a condensate recovery and utilization system, comprising the following steps:

[0006] The original conductivity signal of the condensate and the liquid level signal of the downstream water tank are collected, and the liquid level change rate is obtained based on the liquid level signal.

[0007] The original conductivity signal is subjected to sliding window filtering and Kalman filtering to obtain an effective conductivity estimate.

[0008] Based on the fluctuation of the liquid level change rate and the estimated effective conductivity, the anti-shake holding time and hysteresis half-width are dynamically adjusted, and an upper threshold and a lower threshold are generated based on the adjusted hysteresis half-width.

[0009] The effective conductivity estimate is compared with the upper and lower thresholds by hysteresis, and the stabilization is confirmed based on the adjusted stabilization hold time to generate the first handover decision.

[0010] Based on the effective conductivity estimate, a trend prediction is made, and when the predicted value is not lower than the upper threshold and the trend is continuously rising, a second switching decision is generated.

[0011] Fluctuation patterns are identified in the effective conductivity estimate, and the filtering or anti-jitter parameters are adjusted according to the identified fluctuation patterns.

[0012] Record the cumulative number of switching times for each device, adjust the hysteresis half-width or the anti-shake hold time according to the remaining lifetime ratio, and switch to the standby device when the remaining lifetime ratio is lower than the threshold.

[0013] The first switching decision and the second switching decision are logically ORed and merged. Based on the fusion result, the condensate is switched between a high-value reuse path and a low-value reuse path.

[0014] By adopting the above technical solution, effective estimates are obtained by performing sliding window filtering and Kalman filtering on the original conductivity signal, thus filtering out noise interference. The anti-shake holding time and hysteresis half-width are dynamically adjusted according to the liquid level change rate and the fluctuation of the estimated value to avoid ping-pong switching near the threshold. At the same time, trend prediction is introduced to generate pre-switching decisions, and the filtering parameters are adaptively adjusted by combining fluctuation pattern recognition. The switching strategy is actively optimized according to the remaining life of the equipment. This reduces the number of invalid actions of valves and pumps, extends the mechanical life of the actuators, and reduces the additional losses of motors caused by frequent start-stop. It achieves efficient, stable, and long-life operation of the condensate recycling system and solves the problem that the existing condensate recycling system's timely recovery method easily leads to electric valves being subjected to a large number of invalid actions in a short period of time, shortening their mechanical life and increasing motor losses.

[0015] Preferably, obtaining the effective conductivity estimate includes the following steps:

[0016] Store the original conductivity sample values, calculate the median and absolute deviation of the median in the queue, and determine the robust standard deviation;

[0017] Replace the sampled values ​​in the queue that deviate from the median by more than a preset number of robust standard deviations with the median to obtain the adjusted queue;

[0018] The filtered value is obtained by weighting each sampled value in the adjusted queue according to the time decay weighting factor.

[0019] The filtered value is used as the observation value and input into the Kalman filter for time update and measurement update recursion. The output Kalman estimate is used as the effective conductivity estimate.

[0020] Preferably, the dynamic adjustment of the image stabilization hold time and hysteresis half-width includes the following steps:

[0021] The liquid level change rate is calculated based on the liquid level in the downstream water tank. The absolute value of the liquid level change rate is compared with the preset maximum change rate to obtain the demand normalization factor.

[0022] The adjusted image stabilization hold time is determined based on the baseline image stabilization hold time and the required normalization factor.

[0023] Obtain the standard deviation of the current effective conductivity estimate, and determine the fluctuation hysteresis term based on the baseline hysteresis half-width, the standard deviation, and the adaptive gain coefficient;

[0024] Based on the fluctuation hysteresis term and the demand normalization factor, the adjusted hysteresis half-width is determined and limited to between a preset minimum and maximum value.

[0025] Preferably, the step of generating the upper and lower thresholds based on the adjusted hysteresis half-width includes the following steps:

[0026] Obtain the preset nominal switching threshold;

[0027] The difference between the nominal switching threshold and the adjusted hysteresis half-width is used as the lower threshold;

[0028] The upper threshold is the sum of the nominal switching threshold and the adjusted hysteresis half-width.

[0029] Preferably, generating the first handover decision includes the following steps:

[0030] Obtain the path state at the previous moment, and determine the candidate state based on the path state at the previous moment, the current effective conductivity estimate, the upper threshold, and the lower threshold;

[0031] When the candidate state is different from the path state at the previous time step, record the time during which the candidate state is continuously maintained.

[0032] When the continuous holding time is not less than the adjusted anti-shake holding time, the first switching decision is set to true, and the candidate state is used as the new path state.

[0033] Preferably, determining the candidate state includes the following steps:

[0034] When the previous state was a high-value reuse path and the current effective conductivity estimate is not lower than the upper threshold, the candidate state is set to a low-value reuse path state.

[0035] When the previous state was a low-value reuse path and the current effective conductivity estimate is not higher than the lower threshold, the candidate state is set as a high-value reuse path state.

[0036] If neither of the above two conditions is met, the candidate state remains the same as the path state at the previous time step.

[0037] Preferably, generating the second switching decision includes the following steps:

[0038] Take the estimated effective conductivity values ​​and their time points at a preset number of sampling times, fit a linear trend, and obtain the slope and intercept;

[0039] The predicted conductivity value is determined based on the slope, the intercept, the current time, and the preset future prediction time.

[0040] When the predicted conductivity value is not lower than the upper threshold at the current time and the slope is greater than zero, the second switching decision is set to true and the locking state is set to valid. During the period when the locking state is valid, the generation of the second switching decision is prohibited.

[0041] Preferably, adjusting the filtering parameters or anti-shake parameters according to the identified fluctuation pattern includes the following steps:

[0042] Based on a sliding window, the variance of the effective conductivity estimate, the frequency of crossing the nominal switching threshold, and the pulse peak ratio are calculated within the window.

[0043] The fluctuation pattern is determined by comparing the variance, crossover frequency, and pulse peak ratio with their respective preset thresholds.

[0044] Adjust the filtering or anti-shake parameters according to the fluctuation pattern.

[0045] Preferably, the step of switching to a standby device when the remaining lifetime ratio is below a threshold includes the following steps:

[0046] After each switching operation is performed, the cumulative number of switching operations for the corresponding device is increased.

[0047] The remaining lifespan ratio is determined based on the cumulative number of switching operations and the design lifespan of the execution device;

[0048] Based on the remaining lifetime ratio, adjust the hysteresis half-width and limit its upper limit, as well as the image stabilization hold time;

[0049] When the remaining lifetime ratio is lower than the preset lifetime threshold, the execution device is deactivated and switched to the standby device.

[0050] Preferably, the step of switching condensate between high-value and low-value reuse paths based on the fusion result includes the following steps:

[0051] Perform a logical OR operation on the first switching decision and the second switching decision to obtain a fusion decision;

[0052] When the fusion decision is true, a switching command is sent, and the condensate is switched from the current reuse path to the target reuse path according to the switching command, and the path status record is updated.

[0053] This invention provides a method for timely recovery of condensate in a condensate recovery and utilization system. It has the following beneficial effects:

[0054] 1. This invention obtains an effective estimate of the conductivity by performing sliding window filtering and Kalman filtering on the original conductivity signal, filtering out noise interference. It dynamically adjusts the anti-shake holding time and hysteresis half-width according to the liquid level change rate and the fluctuation of the estimated value to avoid ping-pong switching near the threshold. At the same time, it introduces trend prediction to generate pre-switching decisions, combines fluctuation pattern recognition to adaptively adjust filtering parameters, and actively optimizes the switching strategy according to the remaining life of the equipment. This reduces the number of invalid actions of valves and pumps, extends the mechanical life of the actuators, reduces the additional losses of motors caused by frequent start-stop, and realizes efficient, stable and long-life operation of the condensate recycling system.

[0055] 2. This invention introduces a pre-switching decision mechanism based on linear trend prediction. By fitting the recent trend of conductivity estimates, it predicts the water quality status in the future. When the predicted value is not lower than the upper threshold and the trend continues to rise, a switching command is issued before the actual exceedance. This prevents unqualified condensate from entering the high-value reuse path during the anti-shake delay period, ensuring the safe operation of downstream soft water systems and boiler equipment.

[0056] 3. This invention dynamically adjusts the anti-shake holding time and hysteresis half-width according to the liquid level change rate of the downstream water tank, which can respond to the urgency of water demand. When the liquid level drops rapidly, the anti-shake time is automatically shortened, the switching sensitivity is improved, and the emergency water replenishment is guaranteed to be timely. When the liquid level is stable, the anti-shake time returns to the normal value to maintain switching stability. At the same time, the hysteresis region width is adaptively adjusted according to the fluctuation of the conductivity estimate. The dead zone is automatically widened when the water quality fluctuates drastically and automatically narrowed when the water quality is stable, realizing real-time matching between control parameters and operating conditions.

[0057] 4. This invention identifies the fluctuation patterns of the effective conductivity estimate, classifies the interference types, and adopts differentiated filtering or anti-jitter parameter adjustment strategies for different patterns. It can automatically replace abnormal values ​​without triggering malfunctions when the signal is subjected to short-term pulse interference, and respond quickly when real water quality deterioration occurs, thereby improving the robustness and adaptability of the system in complex industrial environments. Attached Figure Description

[0058] Figure 1 This is a flowchart of a timely recovery method for a condensate recovery and utilization system proposed in this invention;

[0059] Figure 2 This is an architectural diagram of a condensate recovery and utilization system proposed in an embodiment of the present invention.

[0060] Figure 3 This is a schematic diagram of the signal processing flow of sliding window filtering and Kalman filtering proposed in the embodiments of the present invention;

[0061] Figure 4 This is a schematic diagram of the logical state flow of hysteresis comparison and anti-jitter confirmation proposed in an embodiment of the present invention;

[0062] Figure 5 This is a schematic diagram of the fluctuation pattern identification and parameter adjustment decision-making process proposed in an embodiment of the present invention. Detailed Implementation

[0063] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0064] Example 1:

[0065] In a first embodiment of the present invention, the present invention provides a method for timely recovery of condensate in a condensate recovery and utilization system, such as... Figure 1 , Figures 3-5 As shown, it includes the following steps:

[0066] The original conductivity signal of the condensate and the liquid level signal of the downstream water tank were collected, and the liquid level change rate was obtained based on the liquid level signal.

[0067] Specifically, after acquiring the source language speech signal, it is first subjected to multi-layer preprocessing to extract stable initial semantic features. The original speech signal is mixed with environmental noise, equipment interference and vocal tract response differences. Direct feature extraction will lead to distortion of subsequent semantic representation. First, the source language speech signal is denoised to eliminate background noise and circuit thermal noise. Then, a pre-emphasis operation is performed to compensate for the energy attenuation of high-frequency components in the speech signal. The pre-emphasis coefficient is usually 0.97. Next, the speech signal is divided into several short time frames. The frame length is generally set to 25 milliseconds and the frame shift is 10 milliseconds. A Hamming window is applied to each frame to reduce spectral leakage.

[0068] Perform a Fast Fourier Transform on each frame of the windowed signal to obtain the power spectrum. , where x nLet N be the nth sampling point, and N be the frame length. Based on the power spectrum, a spectral mapping is performed using a Mel filter bank. The Mel filter output is... H m (k) is the frequency response function of the m-th Mel filter. For S... m After taking the logarithm, a discrete cosine transform is performed to obtain the Mel frequency cepstral coefficients. Where M is the number of filters, typically 26. The cepstral coefficients of each frame are concatenated in time sequence to form the acoustic feature matrix. The acoustic feature matrix is ​​input into a deep neural acoustic model. Local time-frequency patterns are extracted through convolutional layers, inter-frame temporal dependencies are modeled through a bidirectional long short-term memory network, and global context association is performed through a self-attention encoding layer, outputting an initial semantic feature vector. In this embodiment, self-attention encoding uses the formula... Where Q, K, and V are obtained from s through linear transformation, and d k The key vector dimension is used; through the above steps, initial semantic features with rich contextual semantic information can be obtained, providing high-fidelity input for subsequent quantum adaptive compression.

[0069] Please see the appendix Figure 3 The original conductivity signal is subjected to sliding window filtering and Kalman filtering to obtain an effective conductivity estimate.

[0070] Furthermore, obtaining an estimate of the effective conductivity includes the following steps:

[0071] Store the original conductivity samples, calculate the median and absolute deviation of the median in the queue, and determine the robust standard deviation;

[0072] Replace the sampled values ​​in the queue that deviate from the median by more than a preset number of robust standard deviations with the median to obtain the adjusted queue;

[0073] The filtered value is obtained by weighting each sampled value in the adjusted queue according to the time decay weighting factor.

[0074] The filtered value is used as the observation value and input into the Kalman filter for time update and measurement update recursion. The output Kalman estimate is used as the effective conductivity estimate.

[0075] Specifically, to ensure that the conductivity measurement value truly reflects the water quality of the condensate and to suppress spikes and random noise interference during the sampling process, the original conductivity signal is subjected to dual processing of sliding window filtering and Kalman filtering. High-frequency noise and bubble entrainment generated during evaporator operation will cause instantaneous abnormal fluctuations in the original sampled value, which will lead to misjudgment if used directly for decision-making. First, median sliding window filtering is used to remove outliers, then time decay weighted averaging is used to obtain a stable filtered value, and finally, the optimal estimate is achieved through Kalman recursion.

[0076] Establish a data queue of length N ,in The value is the most recent i-th original conductivity sample; in this embodiment, N is preferably 10. The median of the queue is calculated. and median absolute deviation Therefore, robust standard deviation is determined. For the queue that satisfies The sampled value is replaced with m to obtain the adjusted queue. The values ​​in the adjusted queue are weighted by a time decay factor. Perform a weighted average, where As the forgetting factor, in this embodiment The preferred value is 0.93, resulting in the first filtered value: ,in To adjust the queue of the first There are 1 element, where W is the weight normalization factor.

[0077] Will The observations are input into the Kalman filter to establish the state equation. and observation equations ,in This is an estimate of the true conductivity. For observations , For process noise, For measuring noise. Kalman recursion includes time updates and measurement updates: predicting prior estimates. Prior covariance ; Calculate Kalman gain Update posterior estimates Posterior covariance In this embodiment, the process noise covariance Q is preferably 0.01, and the measurement noise covariance R is preferably 0.04. The final output Kalman estimate... This is the estimated effective conductivity. When Exceeding the preset divergence threshold When this happens, the Kalman filter automatically fails and falls back to the initial state. The aforementioned dual filtering process suppresses noise interference, providing a reliable data foundation for subsequent hysteresis comparisons and trend predictions.

[0078] Based on the fluctuation of the liquid level change rate and the estimated effective conductivity, the anti-shake holding time and hysteresis half-width are dynamically adjusted, and the upper and lower thresholds are generated based on the adjusted hysteresis half-width.

[0079] Furthermore, dynamically adjusting the image stabilization hold time and hysteresis half-width includes the following steps:

[0080] The liquid level change rate is calculated based on the liquid level in the downstream water tank. The absolute value of the liquid level change rate is compared with the preset maximum change rate to obtain the demand normalization factor.

[0081] The adjusted image stabilization hold time is determined based on the baseline image stabilization hold time and the required normalization factor.

[0082] Obtain the standard deviation of the current effective conductivity estimate, and determine the fluctuation hysteresis term based on the baseline hysteresis half-width, standard deviation, and adaptive gain coefficient;

[0083] Based on the fluctuation hysteresis term and the demand normalization factor, the adjusted hysteresis half-width is determined and limited to between the preset minimum and maximum values.

[0084] Furthermore, the upper and lower thresholds are generated based on the adjusted hysteresis half-width, including the following steps:

[0085] Obtain the preset nominal switching threshold;

[0086] The difference between the nominal switching threshold and the adjusted hysteresis half-width is used as the lower threshold;

[0087] The upper threshold is the sum of the nominal switching threshold and the adjusted hysteresis half-width.

[0088] Specifically, in this embodiment, in order to balance the response sensitivity to downstream water demand and the ability to suppress water quality fluctuations, the anti-shake holding time and hysteresis half-width are dynamically adjusted according to the water urgency represented by the liquid level change rate and the real-time fluctuation range of the effective conductivity estimate, thereby generating upper and lower thresholds for water quality judgment. When the liquid level in the downstream water tank drops rapidly, the anti-shake time needs to be shortened to improve the response speed, while when the conductivity fluctuates drastically, the hysteresis region needs to be widened to avoid frequent switching.

[0089] First, calculate the rate of change of water level based on the water level signal from the downstream water tank. Where L3 is the current water level in the downstream tank. T represents the liquid level at the previous moment. s In this embodiment, T is the sampling period. s Take 5 seconds. Compare the absolute value of the liquid level change rate with the preset maximum change rate. By comparison, the demand normalization factor is obtained. ,in The preferred value is 0.01 m / s. The closer this factor is to 1, the more urgent the water demand.

[0090] Adjusted image stabilization hold time Calculate using the following formula: ,in As a baseline for anti-shake hold time, this embodiment uses 30 seconds for live steam condensate and 60 seconds for evaporated condensate. When demand is urgent... Shorten, but not less than 0.2. .

[0091] At the same time, obtain the standard deviation of the current effective conductivity estimate. This standard deviation is based on the most recent The calculation is performed at each sampling point to reflect the current degree of water quality fluctuation. First, the fluctuation hysteresis term is calculated: ,in As the baseline hysteresis half-width, the live steam condensate is taken as 10 microsiemens per centimeter, and the evaporated condensate as 30 microsiemens per centimeter; λ is the adaptive gain coefficient, preferably 0.6. Then, demand correction is introduced to obtain the adjusted hysteresis half-width: ,in and These are the preset maximum and minimum hysteresis half-widths, respectively. In this embodiment, the live steam condensate... =50、 =5, Evaporation of condensate =200、 =10.

[0092] Based on the adjusted hysteresis half-width Combined with the preset nominal switching threshold Generate upper and lower thresholds: In this embodiment, the condensate from the live steam 200 microsiemens per centimeter, evaporation condensate The value is 1000 micro Siemens per centimeter. Through the above dynamic adjustment, it can respond quickly in case of emergency water demand and remain stable when water quality fluctuates, avoiding unnecessary switching.

[0093] The effective conductivity estimate is compared with the upper and lower thresholds by hysteresis, and the stabilization is confirmed based on the adjusted stabilization hold time to generate the first handover decision.

[0094] Furthermore, generating the first handover decision includes the following steps:

[0095] Obtain the path state of the previous time step, and determine the candidate state based on the path state of the previous time step, the current effective conductivity estimate, the upper threshold and the lower threshold.

[0096] When the candidate state is different from the path state at the previous time step, record the time during which the candidate state is continuously maintained.

[0097] When the duration of continuous hold is not less than the adjusted anti-shake hold time, the first switching decision is set to true, and the candidate state is used as the new path state.

[0098] Furthermore, determining the candidate states includes the following steps:

[0099] When the previous state was a high-value reuse path and the current effective conductivity estimate is not lower than the upper threshold, the candidate state is set to a low-value reuse path state.

[0100] When the previous state was a low-value reuse path and the current effective conductivity estimate is not higher than the lower threshold, the candidate state is set as a high-value reuse path state.

[0101] If neither of the above two conditions is met, the candidate state remains the same as the path state at the previous time step.

[0102] Specifically, in this embodiment, to avoid frequent path switching caused by small fluctuations in the conductivity estimate near the threshold, the effective conductivity estimate is compared with the dynamically generated upper and lower thresholds using hysteresis, and a first switching decision is generated in conjunction with the anti-jitter confirmation mechanism. If a single threshold is used for direct comparison, a large number of invalid switching commands will be generated when the estimate oscillates near the threshold. Hysteresis comparison introduces state memory characteristics, and only when the estimate continuously deviates from outside the hysteresis region is the candidate state allowed to change, thereby effectively suppressing the jitter in the critical region.

[0103] First, obtain the path state from the previous time step. Its value is the high-value reuse path status. or low-value reuse path status Based on the current estimated effective conductivity... Upper threshold and lower threshold Determine candidate state H k The specific judgment rules are as follows: If and ,but ;like and ,but ;otherwise .

[0104] When candidate state H k Path state compared to the previous time step At different times, start the anti-shake timer and record H. k Continuous holding time The timer is updated in each sampling period: if H k If the candidate state is the same as that in the previous sampling period, then Accumulated sampling period T s If changes occur, then Reset to zero. When Not less than the adjusted image stabilization hold time When the first switching decision D1 is true, the path status is updated to H. k .

[0105] In this embodiment, the above-mentioned decision logic is equivalent to a state machine with hysteresis characteristics, specifically: When D1=1, path switching is performed. Through the synergistic effect of hysteresis comparison and anti-shake confirmation, the transient out-of-bounds of conductivity signals near the threshold is effectively filtered out, ensuring that the switching action only occurs when water quality undergoes continuous changes. The anti-shake hold time can be dynamically adjusted according to the urgency of downstream water use, ensuring both timely response and avoiding unnecessary actions.

[0106] Please see the appendix Figure 4 Based on the effective conductivity estimate, trend prediction is performed. When the predicted value is not lower than the upper threshold and the trend is continuously rising, a second switching decision is generated.

[0107] Furthermore, generating a second handover decision includes the following steps:

[0108] Take the estimated effective conductivity values ​​and their time points at a preset number of sampling times, fit a linear trend, and obtain the slope and intercept;

[0109] The predicted conductivity value is determined based on the slope, intercept, current time, and preset future prediction time.

[0110] When the predicted conductivity value is not lower than the upper threshold at the current moment and the slope is greater than zero, the second switching decision is set to true and the lockout state is set to valid. During the period when the lockout state is valid, the generation of the second switching decision is prohibited.

[0111] Specifically, in order to predict the upward trend of conductivity in advance and complete the path switching before the actual exceedance, a linear trend prediction is made based on the effective conductivity estimate. When the predicted value is not lower than the upper threshold and the trend continues to rise, a second switching decision is generated. Simply relying on the comparison between the current value and the threshold has a lag. When the water quality deteriorates rapidly, some unqualified condensate may enter the high-value path during the anti-vibration confirmation period. By fitting the recent conductivity change trend, the future state can be predicted one time window in advance, thus achieving predictive protection.

[0112] Take the estimated effective conductivity values ​​and their corresponding time points from the most recent M sampling times, where M is preferably 10. Let the sampling time be... The corresponding estimated effective conductivity is The least squares method is used to fit the linear trend, and the slope a and intercept b are calculated: ,in and These are the arithmetic mean of the time points and the estimated conductivity values, respectively. The slope 'a' reflects the rate of change of conductivity, with a positive value indicating an upward trend.

[0113] Based on the current time t k and the preset future prediction time Calculate the predicted conductivity value: In this embodiment, the future prediction time For live steam condensate, 60 seconds is preferred, and for evaporative condensate, 120 seconds is preferred.

[0114] when When a > 0, set the second switching decision D2 to true. Simultaneously, set the lock status flag... Set to active; during the active lock state, a second handover decision is prohibited from being generated again. Lock time. The preferred duration is 120 seconds. This mechanism prevents multiple switching commands from being triggered repeatedly during a single upward trend.

[0115] When the lockout is active, the subsequent second switching decision is ignored, but the normal execution of the first switching decision is not affected. Through trend prediction and pre-switching control, switching can be initiated before the conductivity actually exceeds the standard, preventing substandard water from entering the high-value reuse path. At the same time, the lockout mechanism prevents repeated actions in a short period of time.

[0116] Please see the appendix Figure 5 The effective conductivity estimate is subjected to fluctuation pattern identification, and the filtering parameters or anti-jitter parameters are adjusted according to the identified fluctuation pattern.

[0117] Furthermore, the filter parameters or anti-shake parameters are adjusted based on the identified fluctuation pattern, including the following steps:

[0118] Based on a sliding window, the variance of the effective conductivity estimate, the frequency of crossing the nominal switching threshold, and the pulse peak ratio are calculated within the window.

[0119] The fluctuation pattern is determined by comparing the variance, crossover frequency, and pulse peak ratio with their respective preset thresholds.

[0120] Adjust the filter parameters or anti-shake parameters according to the fluctuation pattern.

[0121] Specifically, to differentiate the sources of fluctuations in conductivity estimates and implement targeted suppression measures, three statistical characteristics—variance, crossover frequency, and peak pulse ratio—are calculated based on a sliding window. These characteristics are then used to identify fluctuation patterns and adaptively adjust filtering or anti-jitter parameters. Generally, default parameters are maintained when water quality is stable; enhanced filtering and smoothing are needed during critical oscillations; outliers need to be replaced during spikes; and rapid response is required when actual deterioration occurs.

[0122] Specifically, based on a sliding window of length W, with a window value of W=20, the estimated sequence of effective conductivity within the window is calculated. variance ,in This is the arithmetic mean of the estimates within the window. Crossing frequency. Defined as the estimated value crossing the nominal switching threshold within the window. The number of times divided by the window duration. Pulse peak ratio ε = 0.1 microSiemens per centimeter is a small constant to prevent division by zero.

[0123] Compare the variance with the first variance threshold. Second variance threshold Comparison of crossing frequency and threshold Comparison per minute, pulse peak ratio and threshold The comparison is as follows: When the variance is less than the first threshold and the crossover frequency is zero, it is determined to be in a stable mode, and the current parameters are maintained. When the variance is not less than the first threshold and less than the second threshold, and the crossover frequency is not lower than the threshold, it is determined to be in a critical oscillation mode, and the sliding window filter window length is increased to 2N, and the image stabilization hold time is extended to 1.5 times. When the pulse peak ratio is greater than the threshold and the pulse duration is less than three sampling periods, it is determined to be in a spike pulse mode, and the current sampled value is replaced with the median within the window. When the variance is not less than the second threshold and the linear trend slope is greater than 0.5, it is determined to be in a true degradation mode, and the image stabilization hold time is shortened to 5 seconds. Through the above adaptive modes, stability can be maintained under different interference environments.

[0124] Record the cumulative number of switching times for each device, adjust the hysteresis half-width or anti-shake hold time according to the remaining lifetime ratio, and switch to the standby device when the remaining lifetime ratio is lower than the threshold.

[0125] Furthermore, switching to standby equipment when the remaining lifetime ratio falls below a threshold includes the following steps:

[0126] After each switching operation is performed, the cumulative number of switching operations for the corresponding device is increased.

[0127] The remaining lifespan ratio is determined based on the cumulative number of switching operations and the design lifespan of the execution equipment.

[0128] Adjust the hysteresis half-width and limit its upper limit based on the remaining lifetime ratio range, as well as the image stabilization hold time;

[0129] When the remaining lifetime ratio is lower than the preset lifetime threshold, the execution device is deactivated and switched to the standby device.

[0130] Specifically, in this embodiment, to extend the service life of the actuators and avoid premature failure due to excessive switching, the cumulative number of switching operations for each device is recorded. The hysteresis half-width or anti-jitter hold time is dynamically adjusted based on the remaining lifespan ratio. When the remaining lifespan ratio falls below a threshold, the device is actively switched to a backup device. Since actuators such as solenoid valves have limited mechanical lifespans, frequent switching accelerates their wear. After each switching operation, the cumulative number of switching operations for the corresponding actuator is recorded. Increase by 1. Based on the cumulative number of switching operations and the equipment's design life. Calculate the remaining lifetime ratio: In this embodiment, the design life is... The optimal number of times is 100,000.

[0131] Based on remaining lifespan ratio The interval in which it is located will be adjusted accordingly: when When the hysteresis width is greater than 0.3, maintain the current hysteresis half-width. and image stabilization duration Unchanged; when 0.1 < When ≤0.3, increase the hysteresis half-width to And limit it to no more than twice the original value, extending the image stabilization hold time to ×1.5; when When the value is ≤0.1, the executing device is shut down and automatically switched to the backup device. If no backup device is available, the water source is forcibly switched to a low-value reuse path and an alarm signal is issued. Through the above life management mechanism, the switching frequency of endangered equipment is actively reduced, thereby extending the overall maintenance cycle.

[0132] The first switching decision and the second switching decision are logically ORed and merged. Based on the fusion result, the condensate is switched between a high-value reuse path and a low-value reuse path.

[0133] Furthermore, based on the fusion results, the condensate is switched between high-value and low-value reuse paths, including the following steps:

[0134] Perform a logical OR operation on the first and second switching decisions to obtain the fusion decision;

[0135] When the fusion decision is true, a switching command is sent, and the condensate is switched from the current reuse path to the target reuse path according to the switching command, and the path status record is updated.

[0136] Specifically, in this embodiment, in order to integrate the first switching decision from hysteresis anti-shake and the second switching decision from trend prediction, the two are logically ORed and fused, and the switching of condensate between high-value reuse path and low-value reuse path is controlled according to the fusion result. The first switching decision focuses on the stable judgment of the current water quality status, and the second switching decision focuses on the early warning of future deterioration trends. The two complement each other, and when either decision is true, it is considered necessary to perform the switching operation.

[0137] Let the first switching decision be D1, and the second switching decision be D2, both of which are Boolean variables, taking values ​​of true or false. The fusion decision D... fuse Determined by logical OR operation: ,in This represents the logical OR operator. D is true when at least one of D1 or D2 is true. fuse It is true.

[0138] When D fuse When true, the controller generates a switching command, which includes the direction signal of the three-way electric valve and the start / stop signal of the booster pump. Based on the difference between the current path state and the target path state, the following actions are performed: if the current path is a high-value reuse path and a switch to a low-value reuse path is required, the high-value side valve is closed and the low-value side valve is opened; otherwise, the switch is performed. After the switch is completed, the current path state is updated to the target path state, and the switch event is recorded for subsequent equipment life counting.

[0139] For example, when the conductivity continues to rise and triggers the first switching decision, or when the trend prediction triggers the second switching decision in advance, the fusion decision will take effect to ensure that the condensate water switches to the high-value path in a timely manner. Through this dual protection mechanism, neither the switching required due to actual water quality exceeding the standard will be missed, nor the pre-switching required due to the deterioration of the trend will be missed, thus fully ensuring the safety of reuse.

[0140] Example 2:

[0141] In a second embodiment of the present invention, the present invention provides a condensate recovery and utilization system, such as a timely recovery system. Figure 2 As shown, it includes the following modules:

[0142] Acquisition module: used to acquire the raw conductivity signal of condensate and the liquid level signal of the downstream water tank, and obtain the liquid level change rate based on the liquid level signal;

[0143] Filtering module: used to perform sliding window filtering and Kalman filtering on the raw conductivity signal to obtain an effective conductivity estimate;

[0144] Adjustment module: used to dynamically adjust the anti-shake holding time and hysteresis half-width based on the fluctuation of the liquid level change rate and the effective conductivity estimate, and generate upper and lower thresholds based on the adjusted hysteresis half-width;

[0145] Comparison module: Used to perform hysteresis comparison between the estimated effective conductivity value and the upper and lower thresholds, and to confirm the anti-shake based on the adjusted anti-shake hold time, generating the first switching decision;

[0146] Prediction module: Used to predict trends based on the estimated effective conductivity. When the predicted value is not lower than the upper threshold and the trend is continuously rising, a second switching decision is generated.

[0147] Identification module: used to identify fluctuation patterns in the effective conductivity estimate and adjust filtering or anti-jitter parameters based on the identified fluctuation patterns;

[0148] Switching module: Used to record the cumulative number of switching times for each device, adjust the hysteresis half-width or anti-shake hold time according to the remaining lifetime ratio, and switch to the standby device when the remaining lifetime ratio is lower than the threshold;

[0149] Decision module: Used to logically OR and merge the first switching decision and the second switching decision, and switch the condensate between the high-value reuse path and the low-value reuse path according to the fusion result.

[0150] In the daily operation of industrial wastewater treatment plants, a large amount of condensate produced by double-effect evaporators is directly discharged into fire-fighting water tanks or circulating alkaline washing tanks, resulting in inefficient use of high-quality water resources. If the condensate is reused in a soft water system, the frequent switching problem caused by conductivity signal noise needs to be addressed. Frequent valve operations not only shorten equipment lifespan but also increase maintenance costs and system instability. To solve these problems, a condensate recovery and utilization system provided by this invention is adopted, the architecture of which is as follows: Figure 2 As shown. The specific implementation process of this system is as follows:

[0151] First, the acquisition module collects the raw conductivity signal of the condensate and the liquid level signal of the downstream water tank in real time, and obtains the liquid level change rate based on the liquid level signal.

[0152] The filtering module performs sliding window filtering and Kalman filtering on the raw conductivity signal to obtain an effective conductivity estimate.

[0153] The adjustment module dynamically adjusts the anti-shake holding time and hysteresis half-width based on the fluctuation of the liquid level change rate and the estimated effective conductivity, and generates upper and lower thresholds based on the adjusted hysteresis half-width.

[0154] The comparison module performs a hysteresis comparison between the estimated effective conductivity and the upper and lower thresholds, and confirms the anti-shake based on the adjusted anti-shake hold time, generating the first handover decision;

[0155] Meanwhile, the prediction module makes trend predictions based on the effective conductivity estimate. When the predicted value is not lower than the upper threshold and the trend is continuously rising, a second switching decision is generated.

[0156] The identification module identifies fluctuation patterns in the effective conductivity estimate and adjusts the filtering or anti-jitter parameters based on the identified fluctuation patterns.

[0157] The switching module records the cumulative number of switching times for each device, adjusts the hysteresis half-width or anti-shake hold time based on the remaining lifetime ratio, and switches to the standby device when the remaining lifetime ratio is lower than the threshold.

[0158] Finally, the decision-making module performs a logical OR fusion of the first switching decision and the second switching decision, and switches the condensate between the high-value reuse path and the low-value reuse path based on the fusion result. Through the collaborative work of the above modules, the system can adaptively suppress noise interference, significantly reduce the number of invalid switching, extend the service life of the equipment, and achieve efficient and stable reuse of condensate.

[0159] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for timely recovery of condensate in a condensate recovery and utilization system, characterized in that, Includes the following steps: The original conductivity signal of the condensate and the liquid level signal of the downstream water tank are collected, and the liquid level change rate is obtained based on the liquid level signal. The original conductivity signal is subjected to sliding window filtering and Kalman filtering to obtain an effective conductivity estimate. Based on the fluctuation of the liquid level change rate and the estimated effective conductivity, the anti-shake holding time and hysteresis half-width are dynamically adjusted, and an upper threshold and a lower threshold are generated based on the adjusted hysteresis half-width. The effective conductivity estimate is compared with the upper and lower thresholds by hysteresis, and the stabilization is confirmed based on the adjusted stabilization hold time to generate the first handover decision. Based on the effective conductivity estimate, a trend prediction is made, and when the predicted value is not lower than the upper threshold and the trend is continuously rising, a second switching decision is generated. Fluctuation patterns are identified in the effective conductivity estimate, and the filtering or anti-jitter parameters are adjusted according to the identified fluctuation patterns. Record the cumulative number of switching times for each device, adjust the hysteresis half-width or the anti-shake hold time according to the remaining lifetime ratio, and switch to the standby device when the remaining lifetime ratio is lower than the threshold. The first switching decision and the second switching decision are logically ORed and merged. Based on the fusion result, the condensate is switched between a high-value reuse path and a low-value reuse path.

2. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: Obtaining the estimated effective conductivity includes the following steps: Store the original conductivity sample values, calculate the median and absolute deviation of the median in the queue, and determine the robust standard deviation; Replace the sampled values ​​in the queue that deviate from the median by more than a preset number of robust standard deviations with the median to obtain the adjusted queue; The filtered value is obtained by weighting each sampled value in the adjusted queue according to the time decay weighting factor. The filtered value is used as the observation value and input into the Kalman filter for time update and measurement update recursion. The output Kalman estimate is used as the effective conductivity estimate.

3. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The dynamic adjustment of the image stabilization hold time and hysteresis half-width includes the following steps: The liquid level change rate is calculated based on the liquid level in the downstream water tank. The absolute value of the liquid level change rate is compared with the preset maximum change rate to obtain the demand normalization factor. The adjusted image stabilization hold time is determined based on the baseline image stabilization hold time and the required normalization factor. Obtain the standard deviation of the current effective conductivity estimate, and determine the fluctuation hysteresis term based on the baseline hysteresis half-width, the standard deviation, and the adaptive gain coefficient; Based on the fluctuation hysteresis term and the demand normalization factor, the adjusted hysteresis half-width is determined and limited to between a preset minimum and maximum value.

4. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The process of generating the upper and lower thresholds based on the adjusted hysteresis half-width includes the following steps: Obtain the preset nominal switching threshold; The difference between the nominal switching threshold and the adjusted hysteresis half-width is used as the lower threshold; The upper threshold is the sum of the nominal switching threshold and the adjusted hysteresis half-width.

5. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The generation of the first handover decision includes the following steps: Obtain the path state at the previous moment, and determine the candidate state based on the path state at the previous moment, the current effective conductivity estimate, the upper threshold, and the lower threshold; When the candidate state is different from the path state at the previous time step, record the time during which the candidate state is continuously maintained. When the continuous holding time is not less than the adjusted anti-shake holding time, the first switching decision is set to true, and the candidate state is used as the new path state.

6. The timely recovery method of a condensate recovery and utilization system according to claim 5, characterized in that: Determining the candidate state includes the following steps: When the previous state was a high-value reuse path and the current effective conductivity estimate is not lower than the upper threshold, the candidate state is set to a low-value reuse path state. When the previous state was a low-value reuse path and the current effective conductivity estimate is not higher than the lower threshold, the candidate state is set as a high-value reuse path state. If neither of the above two conditions is met, the candidate state remains the same as the path state at the previous time step.

7. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The generation of the second handover decision includes the following steps: Take the estimated effective conductivity values ​​and their time points at a preset number of sampling times, fit a linear trend, and obtain the slope and intercept; The predicted conductivity value is determined based on the slope, the intercept, the current time, and the preset future prediction time. When the predicted conductivity value is not lower than the upper threshold at the current time and the slope is greater than zero, the second switching decision is set to true and the locking state is set to valid. During the period when the locking state is valid, the generation of the second switching decision is prohibited.

8. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The step of adjusting the filtering parameters or anti-shake parameters according to the identified fluctuation pattern includes the following steps: Based on a sliding window, the variance of the effective conductivity estimate, the frequency of crossing the nominal switching threshold, and the pulse peak ratio are calculated within the window. The fluctuation pattern is determined by comparing the variance, crossover frequency, and pulse peak ratio with their respective preset thresholds. Adjust the filtering parameters or anti-shake parameters according to the fluctuation pattern.

9. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The switching to standby equipment when the remaining lifetime ratio is below a threshold includes the following steps: After each switching operation is performed, the cumulative number of switching operations for the corresponding device is increased. The remaining lifespan ratio is determined based on the cumulative number of switching operations and the design lifespan of the execution device; Based on the remaining lifetime ratio, adjust the hysteresis half-width and limit its upper limit, as well as the image stabilization hold time; When the remaining lifetime ratio is lower than the preset lifetime threshold, the execution device is deactivated and switched to the standby device.

10. The timely recovery method of a condensate recovery and utilization system according to claim 1, characterized in that: The process of switching condensate between high-value and low-value reuse paths based on the fusion results includes the following steps: Perform a logical OR operation on the first switching decision and the second switching decision to obtain a fusion decision; When the fusion decision is true, a switching command is sent, and the condensate is switched from the current reuse path to the target reuse path according to the switching command, and the path status record is updated.

Images