Plant Combined Simulation for Compost Production starting from Sewage Sludge
From Proceeding of Summer Computer Simulation Conference,SCSC95, Ottawa, July 24-26 1994
Pietro Giribone, Agostino G.Bruzzone, P. Cereda, R.F. Dell'Acqua, Domenico Rivarolo

---

THE SIMULATION SYSTEM

The plant simulation was developed based on Object-Oriented logic. Given the nature of the process, the discrete process component had to be integrated with the continuous process component and therefore create a hybrid simulation.
The basic object of the entire system is represented by the material unit (RMO). Each RMO is defined by the significant characteristics of the process:

Mass
Density
Type
Humidity
Temperature
Size
Maturation level
Nitrogen Carbon ratio
Acidity

For material stored or handled in a discontinuous manner (e.g. handling operations in yards with bulldozers and trucks), the RMO characteristics are updated by methods which include the following main operating phases:

Addition to the pile
Pick up from the pile
Crushing
Chipping
Mixing
Forced maturation
Maturation in the maturing area: Phase one
Maturation in the maturing area: Phase two
Screening

The RMOs are internal components of almost all the other objects. They are classified into:

TO Treatment objects:
Transport and Handling Equipment
Machinery

SO Storage objects:
Storage areas
Composting Aisles

AO Auxiliary objects:
Personnel

Generally, the TO objects perform discrete and/or continuous operations on RMOs so that they can be applied to process modelling with an initial, linear transient and final state. In this case, the system equations are integrated by referring to the final conditions measured and, by means of continuity and mass balance equations, the material is made to flow while continuously updating the final values.
Instead, the SO are characterised by complex non-linear processes in which the RMOs must be classified to correctly model the process.
For example, the forced maturation lines are divided into 22 sections and each one corresponds to a specific RMO. Thus, during the process (which lasts 21 days), it is possible to evaluate the thermal dispersions, due to the non-homogeneous level of material, and the non-uniform maturation of the compost. Since the material advances progressively along the aisles, each section corresponds to the advancement of the material during the day while moving at normal speed.
Instead, each storage area is divided into 20 sections and each one corresponds to a different layer in the pile. This makes it possible to vary the maturation process in relation to the position in the pile and the external weather conditions. On the other hand, the AOs refer to personnel. These are characterised by a level of specialisation which can activate or disable the single unit to operate with each machine in the plant.
This is particularly important since some operations require skilled personnel (operating the scale, performing maintenance on some components).
The logic structure was developed using the C++ language which offers good operating efficiency, excellent portability and high execution speed.
Each type of object was included in a corresponding list of simulation objects which is used to manage the discrete (event) and continuous (flow) operations.
The equations which regulate the continuous process are integrated for each event; thus, it is possible to monitor the transition conditions in each significant point. With this approach, the numerical calculation occurs normally at the intervals characterised by specific discontinuities at the ends of the interval. This reduces internal discontinuities to a minimum and therefore speeds up the convergence of the algorithm and the execution speed.
Thus, as a time characterisation, it was also necessary to add the end times of each process operated by the TOs. This time is calculated on the basis of the relationship between the quantity to be treated and the rated capacity of the machinery. Process management is controlled by an automatic decision system that makes choices based on the boundary conditions.
This system was derived from a decision support tool already developed during a study which focused on a plant similar to the one being studied through the integrated use of the simulation and AI techniques based on neural networks.
The exogenous variables relative to the orders are modelled in this case as discrete and stochastic External Objects (EO). The raw material orders are generated at stochastic intervals according to probability distributions with the Montecarlo technique.
Loads of material entering on trucks are generated in correspondence with the similar main intervals of each order. Each trucks transports RMOs with characteristics extracted statistically from the characteristics of the product of the corresponding order.
The sales orders are similar. They are also characterised by the quality of the RMO, quantities, pick-up period, frequency and size of the loads.
Therefore, there are objects relative to operating activities and practices which are carried out by personnel. These activities are discrete and may be characterised as follows:

AO Activity Objects
Internal Maintenance Practices
Sample measurement

The decision system performs internal maintenance (when possible) based on the available manpower. Similarly, in correspondence with the arrival of material, controls and measurements are performed on a random basis on the incoming product to evaluate the quality and to obtain information about the acceptance of the next loads.
Process quality controls are performed during the entire production cycle (in maturation aisles, maturation areas, yards with the finished product). This provides a series of data similar to the operating information on which to base the technical choices to guarantee the quality of the product (ventilation flow, number of times the material is turned over in the yard, etc.).
In reality, the plant also has an automatic monitoring system, however the large temperature gradients still require the use of manual measurements and thus personnel, which will take them away from other operating activities.