Mastering Glasshouse Control: Unleashing the Power of True Digital Control

In the modern world of agriculture, especially within glasshouse systems, controlling environmental factors such as temperature and nutrient levels is critical for optimizing plant growth. Traditional control systems often fail to meet the complex demands of today’s fast-paced agricultural environment. But what if we told you there’s a better way to manage these systems, ensuring precision and efficiency through technology? Welcome to the world of True Digital Control (TDC).

This approach, built on decades of research, allows farmers and glasshouse operators to take full advantage of modern computing technology to automate processes with unparalleled accuracy. In this article, we’ll walk you through the ins and outs of TDC, explain how it differs from conventional methods, and show you how it can revolutionize the way you manage your crops.

The Shortcomings of Conventional Control Systems

Before diving into TDC, let’s talk about the standard methods used in most glasshouses. Most of the control systems you’ll encounter rely on Proportional-Integral-Derivative (PID) or Proportional-Integral (PI) controllers. These are designed in continuous-time and then digitized to work with digital hardware—a practice known as Direct Digital Control (DDC).

Here’s the problem: these systems aren’t designed with modern digital capabilities in mind. They are based on trial-and-error tuning, meaning they often need manual readjustments if anything in the system changes. They can also become unstable when dealing with complex factors like time delays—think of how long it takes for your glasshouse to cool down after the vents are opened. This can lead to less-than-optimal growth conditions for your crops.

What Makes True Digital Control (TDC) Different?

TDC flips this approach on its head. Instead of retrofitting analog systems for digital use, TDC starts with a fully digital mindset. The idea is to design the control system using digital data from the ground up. It’s all about precision and flexibility, offering control that adapts to the changing dynamics of your glasshouse without needing manual intervention.

The 3 Key Steps in TDC Design:

1. Data-Driven Model Creation
   TDC begins with collecting data—either through planned experiments or using high-order simulations. This data is then used to create precise digital models of your system, allowing for more accurate control.
2. Iterative System Design
   With the model in place, the TDC design goes through an iterative process. This means the system is tested, adjusted, and optimized in a controlled environment (using Monte Carlo simulations) to ensure it meets your goals without any guesswork.
3. Implementation and Adaptation
   Once the design is finalized, it is tested in both simulations and real-world conditions. If necessary, TDC systems can adapt on the fly, automatically adjusting the control parameters as your glasshouse conditions change.

Why Use TDC for Glasshouse Management?

Now that we’ve introduced TDC, you might wonder: “Why should I make the switch?” The answer lies in its many advantages over traditional systems, especially when managing complex environments like glasshouses.

1. Model-Based Precision

Unlike conventional PID controllers that rely on manual tuning, TDC systems are based on mathematical models derived from real-time data. These models allow the system to make precise adjustments, ensuring that your crops always receive the optimal conditions.

2. State Variable Feedback (SVF)

TDC uses State Variable Feedback (SVF), a more powerful and flexible control method compared to the basic PID controllers. SVF can handle more complex objectives like optimal temperature control or precise nutrient distribution, helping you get the most out of your setup.

3. No More Integral Wind-Up

One of the common issues with traditional controllers is integral wind-up, which happens when the controller overshoots its target and takes too long to stabilize. TDC avoids this, offering a more stable control environment.

4. Robust to Time Delays

Glasshouses often have delayed responses (e.g., the time it takes for temperature to adjust after changing the vent settings). TDC’s unique design minimizes the impact of these delays, keeping the system responsive and accurate.

PIP Controller: A Game Changer in Horticulture

A standout feature of TDC is its use of the Proportional-Integral-Plus (PIP) controller. The PIP controller builds on the traditional PI controller but takes advantage of the digital model to achieve much better results. Since it was developed with glasshouses in mind, PIP is particularly suited for managing the variables involved in agricultural environments, like controlling temperature or nutrient levels.

TDC in Action: Glasshouse and Nutrient Control

TDC has already proven its effectiveness in several real-world applications:

• Glasshouse Temperature Control
  TDC systems have been used to maintain consistent temperatures in glasshouses, ensuring plants grow in ideal conditions without fluctuations that can harm productivity.
• Nutrient Film Technique (NFT)
  When managing nutrient levels in systems like NFT, where plants receive nutrients via a thin film of water, TDC has shown superior control, ensuring that plants get just the right amount of nutrients at the right time.

Wrapping It Up: Why TDC Matters for You

True Digital Control offers a step-change improvement over traditional methods, bringing model-based precision, flexibility, and automatic adjustments to your glasshouse. It’s especially helpful for growers who want to reduce manual interventions and ensure their crops receive the best care possible.

Key Takeaways for Social Media and Infographics

• Engaging Headline: “Take Control of Your Glasshouse with True Digital Control!”
• What is TDC?: A digital-first approach to managing glasshouse systems, providing precision and adaptability.
• TDC vs Traditional Control: TDC is model-based, while traditional systems like PID are manually tuned.
• Three Steps of TDC: Data-driven modeling, iterative design, and adaptive implementation.
• PIP Controller: Designed specifically for horticulture, offering superior control over temperature and nutrients.
• Benefits: Model-based, stable, avoids integral wind-up, robust to time delays.
• Call to Action: “Switch to True Digital Control and optimize your glasshouse operations today!”

This summary should help you create engaging reels and infographics that captivate and inform your audience. Happy growing! 🌱

This excerpt describes the iterative process of the Stochastic Recursive Instrumental Variable (SRIV) algorithm and its application in transfer function (TF) modeling. Here’s a breakdown of the key concepts:

1. Prefiltering Process:
   • Variables like y(k)y(k)y(k), u(k)u(k)u(k), and the instrumental variable x(k)x(k)x(k) are passed through an adaptive prefilter Vh(z−1)V_h(z^{-1})Vh​(z−1) to obtain the prefiltered variables y∗(k)y^*(k)y∗(k), u∗(k)u^*(k)u∗(k), and x∗(k)x^*(k)x∗(k), respectively.
   • The adaptive prefilter is updated based on the parameter estimates from the previous iteration.
2. SRIV Algorithm:
   • This method uses an instrumental variable approach, where the goal is to iteratively estimate the model parameters (a^(k)\hat{a}(k)a^(k)) using recursive equations similar to Recursive Least Squares (RLS).
   • The algorithm’s main recursive form is as follows: a^(k)=a^(k−1)+g(k)[y∗(k)−z∗(k)Ta^(k−1)]\hat{a}(k) = \hat{a}(k-1) + g(k) \left[ y^*(k) – z^*(k)^T \hat{a}(k-1) \right]a^(k)=a^(k−1)+g(k)[y∗(k)−z∗(k)Ta^(k−1)] where g(k)g(k)g(k) is the gain vector, and P(k)P(k)P(k) is the covariance matrix updated at each step.
3. Auxiliary Model:
   • An “auxiliary model” generates the instrumental variable x∗(k)x^*(k)x∗(k), and this is updated iteratively as well.
   • The auxiliary model consists of polynomials A(z−1)A(z^{-1})A(z−1) and B(z−1)B(z^{-1})B(z−1), with prefilter parameters also updated based on earlier estimates.
4. Model Order Identification:
   • This is critical to avoid over-parameterization, ensuring the model is well-behaved for control system design.
   • The YIC (Young Identification Criterion) helps in selecting the best model order, balancing model fit and parameter precision: YIC=log⁡e(σ2σy2)+log⁡e(NEVN)YIC = \log_e \left( \frac{\sigma^2}{\sigma_y^2} \right) + \log_e (NEVN)YIC=loge​(σy2​σ2​)+loge​(NEVN)
     • The first term measures how well the model explains the data, while the second measures the efficiency of parameter estimates.
5. State Representation for Control:
   • The Non-Minimal State-Space (NMSS) representation is introduced to handle the transfer function model. The state vector includes the current and past outputs y(k)y(k)y(k), inputs u(k)u(k)u(k), and an error integral z(k)z(k)z(k), which ensures type-1 servomechanism performance.
   • In control terms, this structure simplifies the state-variable feedback (SVF) law, avoiding the need for state observers, as all relevant signals are measurable or storable.
6. Controllability Conditions:
   • The NMSS model is controllable if:
     • The polynomials A(z−1)A(z^{-1})A(z−1) and B(z−1)B(z^{-1})B(z−1) are coprime (i.e., no common factors).
     • The sum of the coefficients of the input polynomial B(z−1)B(z^{-1})B(z−1) is non-zero.

In summary, this approach combines adaptive prefiltering, recursive estimation, and state-space methods to model dynamic systems efficiently. The SRIV algorithm is particularly suited for systems where noise and parameter estimation are key concerns, and its iterative nature helps to refine model parameters while ensuring the control system remains robust

The document you’ve shared focuses on the design and implementation of Proportional-Integral-Plus (PIP) control systems within a state-variable feedback (SVF) framework, using both z-domain and delta-domain operators. Here’s a breakdown of the key points:

1. Coprimeness and System Conditions: The coprimeness condition ensures no pole-zero cancellations in the system’s transfer function. This is crucial for maintaining the integrity of the control design, especially in systems employing integral action. If integral action is not used, certain zero-pole constraints (e.g., avoiding a zero at unity) are not needed.
2. SVF Control Law: The main control objective is to design an SVF control law defined by a gain vector k=[f0,f1,…,fn−1,g1,…,gm−1,kI]k = [f_0, f_1, …, f_{n-1}, g_1, …, g_{m-1}, k_I]k=[f0​,f1​,…,fn−1​,g1​,…,gm−1​,kI​], where the control signal u(k)u(k)u(k) depends on the current and past system outputs and inputs as well as the integral of error z(k)z(k)z(k). This formulation provides flexibility in determining the desired closed-loop pole locations or optimizing the system’s performance using LQ or LQG methods.
3. Closed-Loop System: In the case of pole assignment, the closed-loop system takes the PIP form, and the characteristic polynomial of the system is designed to ensure that poles are placed at desired positions in the complex z-plane. The equations governing the closed-loop system are derived from equating the system’s characteristic polynomial to the desired closed-loop characteristic polynomial, resulting in a set of linear algebraic equations.
4. Linear Quadratic (LQ) Optimization: In the LQ approach, the control gains are computed by minimizing a quadratic cost function JJJ, which includes weightings on the system output, control effort, and integral of error. These weightings Wy,Wu,WzW_y, W_u, W_zWy​,Wu​,Wz​ allow the designer to balance performance metrics, such as output accuracy and control energy, to achieve desired closed-loop behavior.
5. Delta Operator Model: The document extends the PIP control design to the delta operator framework. This formulation can be advantageous in situations where discretization of the system is critical, particularly in fast sampling systems. The control law in the delta operator form is similar to the z-domain but operates on discrete differentials rather than past samples.
6. Computer-Aided Control System Design (CACSD) with TDC: The package, referred to as TDC, is integrated with Matlab/Simulab, providing a user-friendly interface for control design. It shields users from the Matlab command line through a graphical user interface (GUI) while allowing access to Matlab commands when needed. The TDC system facilitates various stages of control design, including pole assignment, LQ optimization, and Monte Carlo sensitivity analysis. The GUI enables interactive specification of design parameters, real-time evaluation of system behavior, and control system optimization.
7. Control Design in Practice: The TDC package supports a range of control system operations, from basic time and frequency response analysis to more advanced tasks like model uncertainty analysis and Monte Carlo sensitivity studies. The user can select between different design methods (e.g., pole assignment, LQ optimization) and specify the weighting terms that influence the control strategy.

The document is a comprehensive guide for implementing PIP control systems in both theoretical and practical contexts. It emphasizes flexibility and usability through the CACSD package and the integration of different control methodologies within the PIP framework.

Control of a Nutrient Film Technique (NFT) System

The Nutrient Film Technique (NFT) is a hydroponic system widely used in greenhouse environments, especially for growing crops like tomatoes. It involves a continuous flow of nutrient solution over plant roots to provide essential water, nutrients, and oxygen. This system comprises sloping channels where plants are placed, allowing a shallow stream of the nutrient solution to flow through, while the excess solution is collected in a tank and recirculated by a pump.

Key Components of the NFT System:

1. Nutrient Solution: A mixture of water and essential nutrients required for plant growth. It can contain up to 12 nutrients, although only a few (e.g., potassium and nitrogen) are absorbed in large quantities.
2. Channels: Plants are placed in channels with a slight slope, enabling the nutrient solution to flow over the plant roots.
3. Settling Trench: A tank that collects the solution before it is pumped back into the channels, completing the feedback loop.
4. Pump: Recirculates the nutrient solution back to the channels. The flow rate can be controlled to optimize nutrient delivery.
5. Sensors: Devices for monitoring critical parameters such as temperature, ion concentration, and pH levels.

Dynamics and Mathematical Modelling

The NFT system’s nutrient delivery and flow dynamics are modeled using a high-order simulation model, incorporating both advective time delay (the time it takes for the nutrient to travel) and dispersive time constant (related to nutrient mixing in the solution). A simplified model known as the Aggregated Dead Zone (ADZ) is used, which has two components:

1. Continuous Stirred Tank Reactor (CSTR): Simulates the mixing and dispersion of nutrients.
2. Pipeline: Simulates the delay due to the movement of the solution through the system.

The time constant (T) of the CSTR is given by the equation:T=VQT = \frac{V}{Q}T=QV​

where:

• VVV is the active mixing volume,
• QQQ is the flow rate.

A physical flow model for experimentation involves long plastic pipes (e.g., 70 meters for channels and 40 meters for trenches) to simulate these time delays. The system’s output is simulated by a black dye concentration, which is measured by photoelectric sensors. The system input, in turn, is controlled by a peristaltic pump.

Control System Design

Control of the NFT system involves managing the flow of nutrients based on real-time feedback. The Proportional-Integral-Plus (PIP) control system is implemented, relying on self-tuning and self-adaptive mechanisms. These approaches account for changes in system dynamics due to plant growth and varying nutrient uptake over time.

Mathematical Representation:

The system’s dynamics are modeled using a 5th order transfer function (TF) to represent the relationship between nutrient concentration and system inputs, such as flow rate. The general form of the transfer function is:y(k)=b0+b1z−1+b2z−2+b3z−31+a1z−1+a2z−2+a3z−3+a4z−4+a5z−5u(k)y(k) = \frac{b_0 + b_1 z^{-1} + b_2 z^{-2} + b_3 z^{-3}}{1 + a_1 z^{-1} + a_2 z^{-2} + a_3 z^{-3} + a_4 z^{-4} + a_5 z^{-5}} u(k)y(k)=1+a1​z−1+a2​z−2+a3​z−3+a4​z−4+a5​z−5b0​+b1​z−1+b2​z−2+b3​z−3​u(k)

Where:

• y(k)y(k)y(k) is the output concentration,
• u(k)u(k)u(k) is the input signal from the peristaltic pump,
• z−1z^{-1}z−1 represents the time delay operator.

The PIP control system adapts to changes in the plant growth environment by adjusting control parameters based on real-time data. Two control strategies are commonly used:

1. Self-Tuning Control (STC): Periodically updates the model parameters and control gains to maintain system performance.
2. Self-Adaptive Control (SAC): Continuously adjusts control parameters to compensate for dynamic changes, providing more precise nutrient delivery.

Experimental Results

Self-tuning PIP controllers have been tested on pilot-scale NFT systems. The model was tuned to all 8 parameters of the 5th order transfer function. Initial fluctuations in the input signal were used to provide richer information for the recursive estimation algorithm. The experimental results demonstrated that self-tuning control systems can effectively handle the changing dynamics of NFT systems, maintaining desired nutrient levels despite varying plant uptake.

Conclusion

Controlling an NFT system requires sophisticated feedback mechanisms to manage nutrient delivery effectively. The PIP control system, especially when using self-tuning and self-adaptive techniques, allows for optimal plant growth by dynamically adjusting to the changing conditions in the greenhouse environment. This control system offers a practical solution to maintaining consistent nutrient concentrations, improving the overall performance of the NFT setup.

The text you shared is an excerpt from a technical discussion on control systems, particularly focusing on Proportional-Integral-Plus (PIP) control design based on the delta (δ) operator model. Here’s a breakdown of key points:

1. Simplified Adaptive Gain Control:
   • Fig. 15 shows results from a simpler adaptive gain control system, where only the gain GGG is updated recursively. This system effectively manages a sinusoidal disturbance (simulating nutrient uptake by plants) despite using a simpler method than the full self-adaptive system. The full system, however, performs slightly better in compensating for these losses.
2. Delta Operator Model Identification:
   • A third-order delta operator SRIV (Stochastic Recursive Instrumental Variable) model with either zero or unity time delay is identified. The model captures the system dynamics, including non-minimum phase (nmp) characteristics, which substitute for an explicit time delay, offering a realistic simulation of the system.
3. PIP Control Design:
   • The PIP control system, designed using the delta operator model, is based on optimal Linear Quadratic (LQ) design with specific weightings (e.g., qy=1.0q_y = 1.0qy​=1.0, qu=0.5q_u = 0.5qu​=0.5, qz=0.1q_z = 0.1qz​=0.1). The initial design results, presented in Fig. 16, show:
     • (a) Closed-loop step and impulse responses.
     • (b) Closed-loop pole locations.
     • (c) Sensitivity analysis to parameter uncertainty using Monte Carlo simulations.
     • (d) Stochastic root loci, showing potential sensitivity to oscillations in the system, especially in uncertain scenarios.
4. Design Sensitivity and Revisions:
   • Although the nominal design performs well, Monte Carlo analysis reveals sensitivity to uncertainty, indicating potential instability due to oscillations in certain realizations. This suggests the need to refine the design by adjusting the performance index weightings to improve stability margins.
5. Comparison with z−1z^{-1}z−1 Operator Models:
   • Although both delta and z−1z^{-1}z−1 operator models are capable of controlling the system, the delta operator design, though robust and adaptable to high-frequency sampling, does not show clear advantages over the z−1z^{-1}z−1 operator in this scenario.
6. Conclusion:
   • The PIP control system, part of the broader True Digital Control (TDC) philosophy, offers a sophisticated alternative to traditional PID controllers. It can be extended to multivariable and stochastic systems and, when paired with recursive techniques, serves as a foundation for self-tuning and adaptive control systems.

Would you like more specific details or clarification on any part of this explanation?

7. True Digital Control (TDC) and PIP Design:

• The True Digital Control (TDC) philosophy is presented as a modern approach to automatic control system design, offering a fresh perspective compared to traditional PI (Proportional-Integral) and PID (Proportional-Integral-Derivative) controllers.
• The Proportional-Integral-Plus (PIP) controller, which is central to the TDC approach, stands out due to its model-based design, which ensures the complexity of the controller matches the complexity of the system. This results in more powerful and robust control designs, tailored to the dynamics of the controlled system.
• Unlike conventional PI controllers, which rely on fixed parameters, the PIP controller incorporates advanced techniques, like the Non-Minimum State Space (NMSS) formulation. This provides a deeper understanding of the system’s state and allows for more accurate control adjustments, especially in dynamically complex environments.

8. Single-Input Single-Output (SISO) and Multivariable Extensions:

• While the chapter mainly discusses Single-Input Single-Output (SISO) systems, the PIP control strategy can be extended to Multivariable and Stochastic Systems. This is important for more complex control problems where multiple inputs and outputs interact, as is common in industrial systems or complex environments like glasshouses.
• Extensions to multivariable systems allow PIP control to handle larger, interconnected systems with multiple interacting variables, making it suitable for more sophisticated real-world applications.

9. Self-Tuning and Self-Adaptive Control:

• One of the key advantages of the PIP control system within the TDC framework is its ability to function as a self-tuning or self-adaptive control system. Self-tuning systems can automatically adjust their parameters in real time, based on changing conditions in the system.
• The combination of the PIP control system with on-line recursive techniques, such as the SRIV (Stochastic Recursive Instrumental Variable) algorithms, enables continuous updating of system parameters. This allows the control system to adapt to slow changes in the environment or the system dynamics, making it more robust and responsive to long-term variability.

10. Applications in Glasshouse Systems:

• The case study discussed in the paper relates specifically to glasshouse horticulture, where precise control over variables such as temperature, humidity, and nutrient levels is critical for plant growth and productivity.
• In this context, the PIP controller shows its superiority by managing external disturbances (like the simulated nutrient uptake) effectively, ensuring stable control even in the face of uncertainty. The sinusoidal leak introduced in the study simulates nutrient uptake by plants, and the control system’s ability to maintain stability despite this disturbance demonstrates its robustness.
• This application shows the potential of PIP control for real-time control in agricultural environments, where external disturbances and environmental variability are common challenges.

11. Advantages of PIP over Traditional PI/PID Controllers:

• Model-based Design: Unlike PI/PID controllers, which use fixed control parameters, the PIP controller adapts based on a model of the system, offering more precise and flexible control, especially in complex, dynamic environments.
• Handling Uncertainty: The PIP system is better equipped to handle system uncertainty and external disturbances, as illustrated by the sensitivity analysis (Monte Carlo simulations) in the example. By accounting for stochastic variability in the system model, the PIP controller can maintain stability where traditional controllers might fail.
• Flexibility in System Complexity: The PIP system can scale its complexity to match the system being controlled, ensuring that it isn’t overly simplistic or unnecessarily complex, which is a limitation of traditional PI/PID controllers.
• Future-Proof: The PIP controller’s ability to adapt to multivariable and stochastic systems makes it a future-proof option for complex, interconnected control systems, such as those found in industries, agriculture, or advanced manufacturing.

12. Delta Operator vs z−1z^{-1}z−1 Operator:

• The delta (δ) operator, used in the example, offers advantages in certain scenarios, particularly when higher frequency sampling is required. However, the study concludes that, for the coarsely sampled system in this glasshouse application, the z−1z^{-1}z−1 operator (backward shift) model is adequate, and no significant advantages of using the delta operator were observed in this case.
• The delta operator is often preferred when dealing with high-speed systems because it can handle continuous-time signals more effectively when sampling rates are high. In contrast, the z−1z^{-1}z−1 operator works well for systems with slower dynamics or coarser sampling intervals, as seen in this example.

13. Future Outlook:

• The paper suggests that the PIP design, combined with self-tuning and self-adaptive capabilities, can serve as a foundation for more advanced control systems. These systems will be able to handle more complex and dynamic environments, making them suitable for applications beyond horticulture, such as industrial automation, climate control, energy systems, and advanced robotics.
• The NMSS formulation and recursive identification techniques like SRIV make PIP controllers versatile and adaptable, ensuring they can meet the evolving demands of modern control systems where flexibility and robustness are paramount.

In conclusion, the PIP controller, as part of the TDC framework, presents a modern and robust alternative to conventional PI/PID controllers. Its adaptability to system dynamics, ability to handle uncertainty, and potential for self-tuning make it ideal for complex control problems, such as those found in glasshouse horticulture and beyond. The delta operator model adds flexibility, particularly for high-speed systems, although in this specific case, the z−1z^{-1}z−1 operator proved sufficient. The PIP system’s strengths lie in its ability to match the complexity of the control system to the dynamics of the environment, leading to more efficient and stable control solutions.

If you’d like to dive deeper into specific aspects or have further questions, feel free to ask!

The article “Physical Modelling of Greenhouse Climate” by Gerard P.A. Bot delves into the dynamics of greenhouse climate management through physical modeling. It outlines how greenhouses function as bioreactors where crops are grown under controlled environmental conditions. Here’s a summary of key points from the article:

I. Introduction

Greenhouses have been used historically as structures to protect crops and enhance productivity, particularly in unfavorable climates. Advances in technology have enabled the construction of greenhouses with high light transmission and automated environmental controls, making modern greenhouses highly efficient.

II. General Aspects

Greenhouse climate is defined as the set of environmental factors that influence crop growth. The primary differences between the internal climate of a greenhouse and the external weather are due to:

1. Enclosure of air: This leads to reduced air exchange between the inside and outside, affecting the energy and mass balance.
2. Radiation processes: Glass coverings allow shortwave radiation (from the sun) to enter, but trap longwave radiation (emitted from inside the greenhouse), leading to the so-called “greenhouse effect.” However, this trapping mechanism has less effect on temperature than commonly thought.

III. Physical Processes

The article breaks down the physical processes governing the greenhouse climate into:

1. Radiative Exchange:
   • Greenhouse crop production relies heavily on solar radiation, which is partly converted into photosynthesis.
   • About half of the incoming solar energy is in the Photosynthetically Active Radiation (PAR) range (400-700 nm), critical for plant growth.
   • The interaction of solar radiation with greenhouse materials (such as glass) plays a role in determining how much light reaches the crop. Optical properties of materials and the angle of incoming radiation affect this.
   • In terms of thermal radiation, surfaces inside the greenhouse exchange heat, which is modeled using the Stefan-Boltzmann law.
2. Ventilation Exchange:
   • Greenhouse ventilation affects energy and mass transfer and is driven by wind and temperature differences.
   • Ventilation helps regulate internal climate by replacing air inside the greenhouse with outside air. It can be described as a volumetric flow that depends on factors such as wind speed and window openings.
   • Bot (1983) described the factors influencing ventilation, including wind and temperature effects, which were later validated through experiments.

The physical model developed in this paper incorporates these processes to provide a dynamic simulation of the greenhouse climate, useful for design and control purposes. Key components of the model include radiative, convective, and conductive heat transfer, along with crop-specific factors like transpiration and photosynthesis.

The article concludes that through physical modeling and in-situ measurements, researchers can gain better control over the greenhouse environment, optimizing crop growth and minimizing energy costs.

Crop Transpiration

In greenhouses, plants transform a dry, hot climate into a warm, humid environment through physical processes, particularly crop transpiration. This process involves the exchange of water vapor between the plant’s leaf surface and the surrounding air. Water evaporates from the internal cavities of leaves, moving through stomatal openings, and into the greenhouse air. The rate of this process is influenced by environmental factors and plant physiology. Stomata, the small pores on leaves, regulate both the release of water vapor and the absorption of carbon dioxide, essential for photosynthesis. However, a detailed physiological model of stomatal behavior is beyond the scope of physical greenhouse models.

The crop resistance, a key factor in the model, accounts for the plant’s physiological reactions to the environment. This resistance, alongside the boundary layer resistance influenced by local air movement and temperature differences, determines the rate of transpiration. The transpiration rate is described by the equation:

mtr=(es−ea)rtotm_{\text{tr}} = \frac{(e_s – e_a)}{r_{\text{tot}}}mtr​=rtot​(es​−ea​)​

Where:

• ese_ses​ is the saturated vapor concentration at the leaf surface,
• eae_aea​ is the vapor concentration in the air,
• rtotr_{\text{tot}}rtot​ is the total resistance (crop and boundary layer).

Additionally, the latent energy required for water evaporation impacts the crop’s energy budget, defined as:

E=H×mtrE = H \times m_{\text{tr}}E=H×mtr​

Where HHH is the heat of evaporation.

III.5 Exchange between the Cover and the Air

Energy and water vapor exchanges between the greenhouse air and the inner surface of the cover follow the principle of convection. At the inside of the greenhouse, low air velocities result in natural convection due to temperature differences. On the outside, forced convection occurs due to external wind. The heat transfer coefficient (α\alphaα) relates the heat flux to the temperature difference between the air and cover surface. This can be expressed as:

Qconv=α(Ta−Tg)Q_{\text{conv}} = \alpha (T_a – T_g)Qconv​=α(Ta​−Tg​)

Where TaT_aTa​ is the air temperature and TgT_gTg​ the surface temperature.

For mass transfer (condensation), an analogous relation to heat transfer is used:

mg=kg(eg−ea)m_g = k_{\text{g}} (e_{\text{g}} – e_{\text{a}})mg​=kg​(eg​−ea​)

Where kgk_{\text{g}}kg​ is the mass transfer coefficient, and ege_{\text{g}}eg​ and eae_{\text{a}}ea​ are the vapor concentrations at the cover and in the air, respectively. The heat and mass transfer coefficients are related via the Lewis number.

III.6 Exchange between the Heating System and the Air

Greenhouses often use heating pipes to warm the air. The heat exchange between the heating pipes and air follows natural convection, as indicated by low Reynolds numbers but high Grashof numbers. Experiments show that natural convection dominates heat transfer from the pipes to the air, introducing nonlinearities due to temperature differences.

III.7 Exchange with and Transport in the Soil

Although heat exchange with the soil plays a minor role in the overall energy budget, it impacts daily temperature dynamics. Heat transfer in the soil occurs through conduction, driven by temperature gradients, while the soil surface also exchanges radiation with the greenhouse air. The soil’s thermal conductivity, dependent on moisture content, governs the rate of heat transport. In well-watered systems, typical in greenhouse cultivation, soil properties remain stable.

IV. Dynamic Model

The dynamic model of the greenhouse climate integrates the energy and mass balances for key compartments: the greenhouse cover, air, crop, and soil. The model tracks temperature and water vapor concentrations over time, considering their interactions. Energy balances for each compartment are expressed as:

VjCpdTjdt=∑(Qj,n)−Ej+SjV_j C_p \frac{dT_j}{dt} = \sum(Q_{j,n}) – E_j + S_jVj​Cp​dtdTj​​=∑(Qj,n​)−Ej​+Sj​

Where:

• VjV_jVj​ is the compartment volume,
• CpC_pCp​ is the specific heat capacity,
• Qj,nQ_{j,n}Qj,n​ are energy fluxes to neighboring compartments,
• EjE_jEj​ is the energy used for transpiration,
• SjS_jSj​ is the solar radiation absorbed.

The mass balance for water vapor concentration is similarly expressed, allowing the model to simulate both short-term and long-term greenhouse conditions. Numerical integration techniques, such as Euler or Runge-Kutta methods, solve the resulting differential equations. The model’s predictions, validated by experimental data, are realistic and accurate within a 10% margin. It is also adaptable for control purposes, helping optimize greenhouse operations, like heating and ventilation, based on climate set points.

V. Conclusions

The dynamic model offers a reliable representation of the greenhouse climate by quantifying the interactions between environmental factors and plant processes. It can be applied in both practical and theoretical studies to optimize greenhouse production and design. Simplified models can be used for specific control tasks, while the full model serves as a validation tool, reducing the need for experimental testing

Introduction

Biotechnical processes, from traditional ones (like brewing beer) to modern bio-manufacturing, increasingly rely on automation and control technologies. Factors like increasing product demands, legal requirements, and cost pressures drive this shift. Additionally, the agricultural sector, with its environmental and economic constraints, can benefit from similar technological advances. The paper discusses measurement techniques, process modeling, and control methods relevant to biotechnical and agricultural applications.

II. Measurement Techniques for Biotechnical Systems

Measurement in biotechnical systems is complex due to the nature of the data:

• Quasi-continuous measurements (e.g., pressure, temperature) are fast and accurate.
• Discontinuous measurements (e.g., gas analysis) have a short delay.
• Long-lag discontinuous measurements (e.g., biomass composition) are mostly used post-process.

Effective data management is essential due to the varying data sources. Direct measurements like temperature, pH, gas flow, and dissolved gases are common, but many current sensors still have limitations like cross-sensitivities or application restrictions. Analytical techniques, such as using external devices like gas chromatography (GC) and high-performance liquid chromatography (HPLC), are useful but expensive and require skilled personnel.

III. Model-Based Predictive Control

For biotechnical processes, model-based predictive control (like Open-Loop Feedback Optimal Control, OLFO) is a promising strategy. This approach uses process models to predict future states and optimize control actions. However, challenges arise due to incomplete models and the non-linear nature of bioprocesses.

IV. Information Content of Measurements

The effectiveness of predictive control depends on the quality of measurement data. The precision of parameter estimates is influenced by the experimental conditions and the measurement system. Several optimization criteria (A-criterion, D-criterion, G-criterion, and E-criterion) evaluate the information content of measurements, aiming to minimize variances and maximize the accuracy of predictions.

In summary, the paper emphasizes the importance of precise measurements and accurate models for the control and optimization of biotechnical processes. Advanced control methods, like OLFO, can enhance productivity and efficiency, provided the measurement data is reliable and models are well-constructed.

technical paper discussing parameter estimation in biotechnical processes, specifically focusing on the dynamics of a tower reactor used for biomass cultivation. It provides a detailed analysis of the impact of sensor placement, types of measurements, and the identifiability of parameters within the cultivation process. Here’s a breakdown of the key points and concepts presented:

Key Concepts

1. Modeling of Biotechnical Processes:
   • The system is described using equations that model the concentrations of dissolved oxygen, oxygen molar fractions in the gas phase, and biomass concentration.
   • The equations account for the dynamics of the reactor and the interactions between different variables.
2. Random Noise and Parameter Estimation:
   • The presence of noise in measurements introduces randomness in the optimization problem for parameter estimation.
   • The expected values of the estimated parameters can be biased due to this noise, but under certain conditions, unbiased estimates may be possible.
3. Cramér-Rao Inequality:
   • This is used to provide a lower bound on the covariance of parameter estimates, emphasizing the importance of the Fisher information matrix FFF in determining the precision of estimates.
   • The covariance SSS is derived from the Fisher information matrix, indicating how accurately parameters can be estimated.
4. Sensor Allocation and Measurement Types:
   • The paper investigates optimal sensor placements to improve the accuracy of parameter estimates. For example, it discusses the placement of dissolved oxygen sensors in the reactor.
   • Various configurations of sensors (DO, biomass concentration, off-gas analysis) were evaluated for their contribution to improving parameter estimation.
5. Identifiability of Parameters:
   • The identifiability of parameters can change over time during the cultivation process, which may impact how well certain parameters can be estimated at different stages.
   • The analysis uses eigenvector components of the information matrix to evaluate parameter identifiability.

Results Summary

1. Sensor Placement:
   • The most informative sensor placement for dissolved oxygen measurements is near the bottom and top of the reactor, depending on the additional types of measurements taken.
   • Adding cell concentration and off-gas measurements significantly improves the accuracy of parameter estimates.
2. Parameter Identifiability:
   • Certain parameters, such as the maximum specific growth rate and yield coefficient, were found to be easily identifiable during certain stages of cultivation.
   • In conditions of oxygen limitation or during early growth stages, some parameters became less identifiable.
3. Conclusions and Implications:
   • The findings indicate the need for careful consideration of sensor placement and the types of measurements used in biotechnical processes.
   • The research emphasizes the need for adaptive control systems that can modify input functions based on real-time evaluation of parameter identifiability.

Application and Future Research

• The findings from the study not only apply to biotechnical processes but also have potential implications for agricultural systems, such as greenhouses and animal husbandry.
• Future work may focus on real-time evaluation of identifiability and adaptive modifications to ensure reliable parameter estimation throughout the cultivation process.

Overall, this paper presents a comprehensive approach to understanding and optimizing parameter estimation in biotechnical processes, emphasizing the importance of model-based strategies and sensor technologies.

ENVIRONMENTAL CONTROL IN PLANT TISSUE CULTURE AND ITS APPLICATION FOR MICROPROPAGATION

---

I. INTRODUCTION

Plant tissue culture refers to the aseptic cultivation in vitro of various excised plant parts for purposes such as propagation, breeding, biomass production, and preservation of genetic resources. Micropropagation is a specialized aspect of plant tissue culture focused on the aseptic propagation of plants in vitro. The micropropagation process consists of several stages:

1. Stage 0: Stock plant selection and preparation
2. Stage 1: Initiation and establishment of an aseptic culture
3. Stage 2: Multiplication
4. Stage 3: In vitro rooting and conditioning
5. Stage 4: Acclimatization to greenhouse conditions

Despite its advantages over traditional methods like cutting and grafting, the commercial use of micropropagation is limited by high production costs, primarily due to labor intensity, slow multiplication rates, and low survival rates during acclimatization.

The ultimate goal of micropropagation is to produce numerous genetically identical and physiologically uniform plantlets that can survive harsh greenhouse conditions, achieving this efficiently and cost-effectively. The development of automated environmental control systems and improved culture systems is crucial for reducing production costs. Recent research has focused on automating processes such as medium preparation, image recognition, and transplanting.

However, limited research has been done on the effects and control of environmental factors in vitro, which affects plant productivity. This oversight is partly due to the small, airtight nature of conventional culture vessels, which makes controlling and measuring environmental factors challenging.

This section discusses the in vitro microenvironment and its impact on photosynthesis and plantlet growth, drawing on previous reviews by Kozai and colleagues.

---

II. SIGNIFICANCE OF ENVIRONMENTAL CONTROL IN PLANT TISSUE CULTURE

The environmental dynamics in a tissue culture vessel resemble those in a greenhouse. Both systems rely on the interrelationship between plants and their surrounding environment. A culture vessel functions similarly to a miniature greenhouse, where the explant can be viewed as a small cutting in conventional propagation.

Research on the ecological and environmental aspects of micropropagation is essential, yet it has often been neglected. Traditionally, reliance on exogenous plant growth regulators has overshadowed studies on the dynamic growth modeling of plantlets in vitro, despite extensive research on crops in greenhouses and fields. Developing a growth model for in vitro plantlets is likely simpler because microenvironmental conditions, like air temperature and light intensity, can be kept constant.

The conventional in vitro environment differs significantly from greenhouse conditions, often leading to physiological and pathological issues. These conditions typically include:

1. High relative humidity (RH)
2. Constant temperature
3. Low photosynthetic photon flux density (PPFD)
4. Fluctuating CO2 concentrations
5. High concentrations of sugars, salts, and growth regulators
6. Accumulation of toxic substances

Such conditions can negatively impact plantlet growth, resulting in reduced transpiration and photosynthesis, nutrient uptake, and increased dark respiration rates.

Historically, plantlets have been cultured under heterotrophic or photomixotrophic conditions, relying heavily on sugars for energy. However, recent studies indicate that chlorophyllous explants can photosynthesize effectively, growing more rapidly under photoautotrophic conditions if the physical and chemical environments are adequately managed.

Controlling the microenvironment in vitro is essential for practical reasons, especially in reducing production costs. Effective environmental control can:

1. Enhance plant growth and development
2. Improve rooting and branching
3. Mitigate morphological and physiological disorders
4. Decrease losses due to biological contamination
5. Encourage uniform plant growth and reduce reliance on exogenous growth regulators
6. Facilitate rapid growth during acclimatization

Improving environmental conditions for increased photosynthetic and transpirational activity is necessary for developing a more efficient micropropagation system.

---

III. CONSIDERATIONS FOR MEASUREMENT AND CONTROL OF THE IN VITRO ENVIRONMENT

A. Light

Different light sources emit varying spectral distributions, influencing plant growth. Fluorescent lamps are commonly used in micropropagation because their spectrum aligns well with the needs of in vitro cultures. However, the PPFD distribution within culture vessels can be uneven, heavily influenced by the vessel type and arrangement.

Studies have shown that sideward lighting can promote better growth than traditional downward lighting. Future micropropagation systems should consider lateral lighting to enhance light availability and reduce energy consumption. Additionally, light-emitting diodes (LEDs) may be a practical and cost-effective lighting option.

B. Gas Exchange Characteristics

The type of vessel closure impacts the gaseous composition and plant growth, with loose closures generally preferred to reduce issues like vitrification. The air exchange rate can be expressed as the number of air changes per hour (E), which can be enhanced with gas-permeable materials.

C. Relative Humidity and Medium Water Status

Water exchange among the gaseous, liquid, and plant environments in the vessel is crucial for plant development. Higher initial RH has been associated with better growth, with significant effects on shoot length and leaf area. Sensors can accurately measure RH, and the water potential of the culture medium influences plant growth dynamics.

D. Temperature

Temperature distribution within culture vessels can vary. Studies have shown that manipulating the temperature difference between light and dark periods (DIF) can effectively control plantlet height while minimizing energy costs.

E. CO2, O2, and Ethylene

Monitoring gas concentrations, particularly CO2, O2, and ethylene, is vital for optimizing plant growth. Various methods can be employed to modify the gaseous environment within the vessel, including using gas-permeable closures and ventilation systems.

F. CO2 Concentration During the Photoperiod

Studies have shown significant decreases in CO2 concentration in culture vessels during the photoperiod. This limitation on CO2 availability can restrict photosynthesis and shift plant growth dynamics toward heterotrophy. Providing adequate CO2 levels can facilitate faster growth and development.

G. Net Photosynthetic Rate

Estimating the net photosynthetic rate in vitro requires specialized equipment to assess gas exchange effectively. Understanding and controlling the photosynthetic environment within the culture vessels can enhance plant growth and productivity.

---

This overview highlights the crucial role of environmental control in plant tissue culture and micropropagation, emphasizing the need for ongoing research and innovation in this field to optimize growth conditions and improve efficiency

Photoautotrophic Growth and Development of the Plantlet in Vitro

In photoautotrophic micropropagation, the growth and development of plantlets in vitro are significantly influenced by various physical environmental factors, including light source and intensity, CO₂ and O₂ concentrations, humidity, airflow rate, and temperature. The effects of these environmental factors on net photosynthetic rate (NPR), growth, and development have been extensively reviewed by Kozai (1991a, 1991b, 1991c). This section briefly discusses the influence of these factors on the NPR and overall plantlet development.

A. Photosynthetic Response of Plantlets In Vitro

Kozai et al. (1990a) investigated the photosynthetic response of Cymbidium in relation to CO₂, photosynthetic photon flux density (PPFD), and temperature levels. They found that the response curves of in vitro Cymbidium plantlets resembled those of plants grown in shaded greenhouse conditions. When CO₂ concentration was maintained at approximately 200 ppm, NPR in Primula malacoides (a C3 plant) increased significantly, showing rates approximately 3 times higher at 1% O₂ and 1.5 times higher at 10% O₂ compared to the 21% O₂ environment, due to reduced photorespiration at lower O₂ concentrations. Similar results were noted in rose and potato plantlets, where lowering sucrose concentration in the culture medium enhanced NPR. Furthermore, increased leaf starch content was associated with higher sucrose concentrations, which in turn reduced NPR.

The relative humidity (RH) also significantly impacted NPR. In strawberry plantlets, NPR was higher in vessels with forced ventilation compared to those with natural ventilation, likely because CO₂ diffusion into the stomata was restricted under stagnant air conditions (Kozai et al., 1989). Under saturated PPFD, 340 ppm CO₂, and a leaf temperature of 20°C, NPR of in vitro plantlets was consistent, indicating that while CO₂ concentration affected NPR, the other factors were also significant (Pospisilova et al., 1987). However, the importance of the physical environment on the photosynthetic response of in vitro plantlets has only recently gained recognition, highlighting the need for further investigation.

B. CO₂ Enrichment Under High PPFD

CO₂ enrichment under high PPFD (100-200 μmol·m⁻²·s⁻¹) effectively promotes shoot and plantlet growth in potato (Kozai et al., 1988) and tobacco (Mosseau, 1986) regardless of sugar content in the medium. Enhanced growth of carnation plantlets was observed under CO₂ concentrations of 1000-1500 vpm and PPFD of 150 μmol·m⁻²·s⁻¹, attributed mainly to CO₂ enrichment (Kozai and Iwanami, 1988).

To enhance NPR and thereby promote plantlet growth in vitro, several practical methods can increase CO₂ concentrations:

1. Use of CO₂ Permeable Film: Research indicates that gas-permeable films, used as closures under high PPFD, positively impact NPR and plantlet growth (Kozai, 1991d). Plantlets from leafy single-node cuttings exhibit faster growth when cultured in a vessel closed with gas-permeable film compared to traditional heterotrophic methods in airtight vessels. Such passive CO₂ enrichment, requiring minimal adjustments in existing micropropagation facilities, significantly enhances in vitro plantlet growth while also reducing vitrification due to lower RH and improved gas exchange.
2. CO₂ Enrichment in the Culture Room: Enriching CO₂ levels (100-200 μmol·m⁻²·s⁻¹) in the culture room promotes growth in chlorophyllous tobacco (Mosseau, 1986), Cymbidium (Kozai et al., 1987), carnation (Kozai and Iwanami, 1988), and potato (Kozai et al., 1988) plantlets, irrespective of the sugar content in the medium.
3. Large Culture Vessel with CO₂ Supply System: Strawberry plantlets cultured on sugar-free liquid medium in a large vessel with forced ventilation exhibited greater dry weight and NPR under a PPFD of 96 μmol·m⁻²·s⁻¹ compared to conventional methods (Fujiwara et al., 1988). However, forced ventilation with atmospheric air or a N₂-O₂-CO₂ mixture may negatively impact propagule weight and shoot number in specific cases (Walker et al., 1988).

While the systems described above not only adjust CO₂ concentration but also influence RH, ethylene concentration, and gas diffusion within the vessel, the observed changes in plantlet growth are primarily attributed to CO₂ enrichment. More comprehensive studies on the effects of varied gaseous environments and forced ventilation on photoautotrophic growth and development are essential.

---

V. Advantages of Photoautotrophic Micropropagation Over Conventional Methods

Heterotrophic and photomixotrophic micropropagation methods present several challenges:

1. Increased Contamination Risk: The addition of sugar as a carbon source raises the likelihood of biological contamination, necessitating the use of airtight, small vessels, which complicates automation and robotization of micropropagation systems.
2. Saturated Conditions: The air within vessels tends to become saturated with water vapor, leading to abnormal concentrations of CO₂ and ethylene, diminishing the efficacy of high PPFD for promoting plantlet growth.
3. Reliance on Growth Regulators: These conditions often necessitate the use of growth regulators for effective plant regeneration.
4. Physiological Disorders: Unfavorable environmental conditions can induce physiological and morphological disorders, growth retardation, and mutations.
5. Unstable Production Cycles: Such factors contribute to an unstable production cycle, non-uniform plantlet growth, and elevated mortality rates during acclimatization, all of which raise production costs.

Conversely, photoautotrophic micropropagation offers numerous advantages over conventional methods, as summarized in Table 1. These benefits may significantly enhance plant productivity and quality while reducing production costs. While some advantages are experimentally validated, further investigation into the effects of diverse environmental factors on photosynthetic growth in vitro remains necessary. The potential for photoautotrophic micropropagation utilizing chlorophyllous shoots or nodal cuttings is promising, whether automated or not.

Table 1: Advantages of Photoautotrophic Micropropagation

Stable production cycles lead to lower production costs.

Improved environmental conditions promote plantlet growth and development.

Minimization of growth regulators and organic matter application.

Utilization of larger vessels with environmental control reduces contamination risks.

Reduced losses from biological contamination and simplified rooting and acclimatization processes.

Decreased physiological, morphological, and genetic disorders enhance plantlet quality.

Easier control of plantlet growth and development through environmental regulation.

Feasibility of automation, robotization, and computerization in the micropropagation system.

Photoautotrophic Growth and Development of the Plantlet in Vitro

In photoautotrophic micropropagation, the photosynthesis, growth, and development of the in vitro plantlet are significantly influenced by physical environmental factors, such as light source and intensity, CO2 and O2 concentrations, humidity, airflow rate, and temperature. The literature on the effect of environmental factors on photoautotrophic micropropagation has been extensively reviewed by Kozai (1991a, 1991b, 1991c). This section will briefly review the environmental effects on the net photosynthetic rate (NPR), growth, and development of plantlets in vitro.

A. Photosynthetic Response of Plantlets in Vitro

Kozai et al. (1990a) studied the photosynthetic response of Cymbidium as affected by CO2, photosynthetic photon flux density (PPFD), and temperature levels. They found that the response curves of the in vitro Cymbidium plantlet in situ were similar to those of plants grown in the shade in the greenhouse.

When the CO2 concentration was fixed at approximately 200 ppm, the NPR in Primula malacoides (C3 plant) plantlets in vitro in 1% and 10% O2 was approximately 3 and 1.5 times higher, respectively, than that in 21% O2. This resulted from reduced photorespiration at lower O2 concentrations. The NPR of the in vitro rose plantlet increased when cultured on medium with a lowered sucrose concentration. A similar result was observed in potato plantlets in vitro. The leaf starch content of the plantlet increased when cultured on medium with a raised sucrose concentration, while increased leaf starch content was associated with a lowered NPR

The NPR was also affected by relative humidity (RH). In strawberries, the NPR was higher when plantlets were cultured in a vessel with forced ventilation than with natural ventilation (Kozai et al., 1989). Under natural ventilation, CO2 diffusion into the stomata was probably restricted since the air was almost stagnant, and air movement in the vessel was caused only by natural convection.

The NPR of the plantlet and seedling in vitro under saturated PPFD, 340 ppm CO2, and a leaf temperature of 20 °C were similar, regardless of the slight differences in NPR affected by CO2 concentration. The significance of the photosynthetic response of the in vitro plantlet in situ, as affected by the in vitro physical environment, has only recently been recognized, and there are many aspects that need further investigation.

B. CO2 Enrichment Under High PPFD

Carbon dioxide enrichment under high PPFD (100-200 μmol m⁻² s⁻¹) effectively promotes shoot and plantlet growth of potato and tobacco when cultured on media with and without sugar. Kozai and Iwanami (1988) observed enhanced carnation plantlet growth under conditions with CO2 concentrations of 1000-1500 vpm and PPFD of 150 μmol m⁻² s⁻¹, largely due to CO2 enrichment. This growth promotion was observed in treatments with and without sugar addition in the medium.

Based on the data presented above, an increase in NPR and, hence, growth and development of the plantlet in vitro can be expected if the CO2 concentration in the vessel is raised during the photoperiod. There are a few practical methods for raising vessel CO2 concentration:

1. Use of CO2 Permeable Film in the Closure
   Several reports indicate positive effects of gas-permeable film as closure under high PPFD on NPR and growth of the plantlet in vitro. Plantlets of some species derived from leafy single-node cuttings grew faster when cultured photoautotrophically in vessels closed with gas-permeable film than when cultured heterotrophically in relatively airtight vessels. Under high PPFD, even passive CO2 enrichment, requiring only a minor change in the existing micropropagation facility, can significantly enhance plantlet growth in vitro. The percentage of vitrification also decreased with the use of the gas-permeable film, likely resulting from lowered RH and increased gas exchange and dehydration of the medium.
2. CO2 Enrichment in the Culture Room
   CO2 enrichment under high PPFD (100-200 μmol m⁻² s⁻¹) effectively promotes the growth of chlorophyllous tobacco, Cymbidium, carnation, and potato plantlets, regardless of the medium sugar content.
3. A Large Culture Vessel with a CO2 Supply System
   Dry weight and NPR of strawberry plantlets cultured on sugar-free liquid medium were greater when cultured in a large vessel with a forced ventilation system under a PPFD of 96 μmol m⁻² s⁻¹ compared to those of plants cultured using conventional methods. However, forced ventilation with atmospheric air or a N2-O2-CO2 mixture reduced propagule weight and shoot number in stage 2 Rhododendron cultured in vessels with a 400 ml air volume and under a PPFD of 39 μmol m⁻² s⁻¹.

In the systems described above, not only CO2 concentration but also RH, ethylene concentration, and gas diffusion in the vessel are modified. Therefore, changes in the growth of plantlets in vitro resulting from the use of these systems cannot be attributed entirely to CO2 enrichment. However, in most cases, the changes are probably caused primarily by CO2 enrichment. More studies on the effect of different gaseous environments and forced ventilation on the photoautrophic growth and development of the plantlet in vitro are needed.

V. Advantages of Photoautotrophic Micropropagation Over Conventional Micropropagation Methods

Several disadvantages and problems associated with hetero- and photomixotrophic micropropagation include:

1. The addition of sugar as a carbon source in the medium increases the incidence of biological contamination, and airtight, small vessels are commonly used to reduce these contaminations. Thus, automation, robotization, and computerization of the micropropagation system are practically difficult.
2. The air inside the vessel is nearly saturated with water vapor, and vessel CO2 and ethylene concentrations become abnormal. Therefore, high PPFD becomes ineffective in promoting plantlet growth.
3. Growth regulators are often necessary for plant regeneration.
4. These undesirable environmental conditions induce physiological and morphological disorders, growth retardation, and mutation.
5. Ultimately, an unstable production cycle, non-uniform plantlet growth, and a high plantlet death rate during the acclimatization stage raise production costs.

On the other hand, photoautotrophic micropropagation has several advantages over conventional micropropagation methods (Table 1) and may enhance plant productivity and plantlet quality, thereby considerably reducing production costs. Some of the advantages shown in Table 1 have been experimentally proven. However, many of those points still need further investigation, especially concerning the effect of different environmental factors in the air and medium on the photosynthetic growth of the plantlet in vitro. The potential benefits of photoautotrophic micropropagation, using chlorophyllous shoots or nodal cuttings as explants, whether automated or not, seem significant.

Table 1: Some Advantages of Photoautotrophic Micropropagation

The production cycle is stable, and costs are lowered.

Growth and development of plantlets are promoted resulting from improved environmental conditions for normal growth and development.

Application of growth regulators and other organic matter is minimized.

A larger vessel with environmental control and monitoring systems can be used, decreasing the incidence of biological contaminations.

The loss of plantlets due to biological contaminations is reduced, and procedures for rooting and acclimatization are simplified.

Physiological, morphological, and genetic disorders are reduced, thereby improving plantlet quality.

The control of plantlet growth and development through environmental control is relatively easier.

Automation, robotization, and computerization can be practically achieved.